Home

Cooper-S-Dissertatio.. - Department of Computer Science

1. 3 8 1 Main priorities 3 82 Secondary priorities 3 9 PROJECESCOPEL A AS Athe 3 10 REQUIREMENTS SPECIFICATION eene terne nnn tat tette ttt ta tote tonta tata tota soto oo so so sooo so so so so sooo so 46 jrn 51 4 1 DESIGN RATIONALE tace a edet ates eet tee de ede Gabe a eec ee 4 2 LOW FIDELITY PROTOTYPING scssssssssssssesssssssessessssecsesscsucsnsacsucsusasescancasescsnsesesesnsesesesnsesesesnsesesesasssescsassacassassacansassatansacsaeanes 421 NayvigationalstFuctuFe uper t eut desea ELM Door LOL poet 42 2 Task 42 3 TaskSupport 4 3 CARD USING STORYBOARDING 31 Results iain RD RH E RT e e ER ERE RE ER ERR RE HUE ad 4 4 ENVISIONED HT aR 4 5 HIGH FIDELITY PROTOTYPING ssssssssssssssesssssssessessssecsesecsucsnsecsucsnsasesesnsasesesnsesescsnsesesesnsesesesnsesesesusssesesassesansassacansassaeansaceseanes 4 6 CONCLUSION vant tenia tite ties IMPLEMENTATION AND TEST ING eerere eere en errare nnnm anna ta tnam
2. XUXNAS cre xref lot 2e voot Scores Chooses F cereus thag Wen NOD aem dem Powe u o X voxwas Lo bottom cC 2 Arethere any features you would like to see retained from the tool Or possibly there is a Gyre process or understanding you had from completing the task manually that you would like to be able to see do in a new system vo X eo x Qoa 2 O be Gb wc et TEN eSak Eronda dosao oe veng manno goss how mokrica uw Oaa arnoak Carve Oye 2 60 Od c cA MOAXA NGC So lon Se CLK 3 Do you find the current tool easy to use 229 A Qood MmarecSe ex uen Of gcecisely gt ro Anke reeds to eroVided XC dora sie an terms cC Formac Sutter anman Ge mot acc Pak Ur If so what makes it easy to use If not can you describe why it s difficult to use Ned a Wigh techn CON eun Fo Gin o ves RL orourd ke coo Mor upa Scu code CaL hep osse e d Weedk Vo ork er SO war Ge gt Aw 9 eos UI NT ee Grom ve CO e st eos SUA Coo epe d f cS perore yor Moda baira abh ee oe ee coc otc enc Coo ba enougr sec Ver O be cle to o Ome co QoS ISVS cs o eoret r haoa Sve pq AA 0 OS A Laue Ad doa Ut Rs qd eo
3. 2 5 1 Why user centred d sign a ta a V t i ke uh e t t deme 14 2 5 2 Task Analysis 2 5 3 Participatory design 2 6 USABILITY EVALUATION 2 61 Analytical Evaluation 2 6 2 Empirical Evaluation gb a e ta n e cie 2 7 CONCLUSION I uUnn 23 3 1 DATA GATH ERI HE tee 23 3 2 OVERVIEW OF THE TARGETING PROCESS sssssssssssesssssssessssecsscsssscsscsusessscsnsesescsnsssesesusssesesnsssesesassensesessscansassacansassseansaceaeanes 24 3 3 INTERVIEW WITH STAKEHOLDERS 3 2 1 Interview Process ius ae aan RE ERE ania ala EARS PEE nce c M M 3 3 3 Background Of interviewees tetra aha 26 3 32 4 nete atr i date Ht ces HE E d qe P oed Hr e E rd tei E er ieia 3 3 5 Summary of findings 3 4 HEURISTIC EVALUATION OF THE EXISTING TOOL eee eerennn terrena 30 3 5 ASK ANALYSIS 32 25 Observations seu eie edita egli te egeta lade cet 32 3 5 2 Observation summ ry attt Et etd ret Ea Ea td Hee t 39 3 6 CARD SORTING SESSION 3 61 Results and analysis et e te e ota teo p e ns 40 3 7 HIERARCHADI TASKANALYSIS2 e boda SR 42 3 8 SETTING PRIORITIES
4. umop uayosq ueg mee eng w n Kay 10 K2uenbai4 4980 BIEIBUAD ZZE prey uonisod uo oS zt TWNOILdO uode1 mum Jeu x joo e edeug Z sag 0 sioe qo ssaursnq Byep mau uodui z z TS m uodai 5129 SSBUISNG amp ejeJeuac Figure 17 HTA for the Access solution 44 Note that step 7 2 1 in figure 14 can be divided and these steps are shown in step 3 of figure 13 This is a large number of steps and after discussing these further with the users it became apparent that these steps could also be automated Therefore the new system should treat the business plan data creation just as the objectives and call frequency whereby the user will complete checks rather than manually enter each field s value This is a fundamental shift in user to system functional allocation It is also useful to note that the final stage of both of these solutions was completed externally to the application the rest was completed within This was discussed with the users and it was found that they consider this step to be time consuming and it is also another opportunity for an error to occur when transferring data It would improve the overall efficiency and user experience if this step could be part of the new system this again is another functional shift to the s
5. 55 NIELSON J 2007 Breadcrumb navigation increasingly useful online available from http www useit com alertbox breadcrumbs html Accessed 12 04 08 116 Personal Communications 40 N Lindsay amp N Langley 01 04 08 11 30 12 00 Telephone communications discussing design options 57 C Goddard 28 03 08 14 30 15 00 Telephone communications following up from evaluation session 58 N Lindsay 31 03 08 09 00 09 30 Telephone communications following up from evaluation session 117 Appendix A Requirements Supporting Documentation Section 1 Interview Transcripts and signed copy of email sent to participants Email sent to all interview participants The first part of requirements gathering will be informal interviews to help understand your opinions on the current system and what differences you would like to see in the new design For your information here is an overview of the questions that will be asked this is not a strict list and could vary depending on your responses 1 Are there any features that you like about the current tool 2 Are there any features you would like to see retained from the tool or possibly there is a process or understanding you had from completing the task manually that you would like to be able to see do in a new system 4 Do you find the current tool easy to use If so what makes it easy to use If not can you describe why its difficult to use 5 What is bad about the c
6. db CrossReferences DeleteAllOnSubmit CrossReference db CrossRefere nces When the user clicks save the first step will be to save the selected mappings or if required delete and then save to the CrossReferences table and then finally use this table containing the selected mappings to update the relevant tables The session variables described earlier are 82 again used but this time to determine whether a particular query is required to be run depending on whether the user selected that particular data type at the overview screen The code listings for each of these queries can be found in Appendix E Lastly the user is redirected to the data screen this is where all generated data can be viewed and edited Figure 38 shows example print screens of the prototypes 2 Overview Import X Ref Data gt Epot Lit Overview Import L Lb 5 View and update generated data Export this to Excel CUERO M And ify You are logged in as You are logged in as Please select the data you wish to verify Songer Business Plans BSPL aus 20083821 elete MANUAL ciam Objectives OBJT Admin pren 20083821 MANUAL Call Frequency Objectives OBCF Help 20085522 Help Sales Quota Objectives OBSQ MANUAL 125526 Restat MANUAL 20085522 dit Delete y 125526 Restart Log Out TER 20085522 Log Out ner 125526 Figure 38 Print screen of viewing and editing
7. 7 3 2 Research Moving Forward ttt aaa aa hah BIBLIOGRAPHY issa aaa eem an APPENDIX A REQUIREMENTS SUPPORTING DOCUMENTATION mereri eaten tetra tetas 118 iii APPENDIX B DESIGN SUPPORTING DOCUMENTATION eeseseneneneneneneninin intenta tete tet tenant ntn cata saa 136 APPENDIX C IMPLEMENTATION SUPPORTING 144 APPENDIX D EVALUATION SUPPORTING DOCUMENTATION 151 APPENDIX E CODE MAIN CLASSES ONLY eere 156 APPENDIX F ETHICS CHECKLIST sesa eae 167 iv List of Figures FIGURE 1 NORMAN S EXECUTION EVALUATION 6 FIGURE 2 AN EXAMPLE OF BREADCRUMBS aaa a ab ebb a ana e 9 FIGURE 3 THE USER CENTRED DESIGN PROCESS sssssssssssssssessessssessesscsecsnsscsucsnsacsuesusassscsnsesescensesesesnsesesesnsssesesassassesassatansassarsneaceatanes 13 FIGURE 4 GRAPH SHOWING PROPORTION OF USABILITY PROBLEMS FOUND BY EVALUATORS eren entren tnter tn 19 FIGURE 5 FORMATION OF THE REQUIREMENTS DOCUMENT ennt tenente tenente trente tte ta tette to tete to tetto to soto FIGURE 6 STRUCTURE OF THE TARGETING DATA FIGURE 7 AN EXAMPLE OF THE DATA REQUIRED FOR ONE SALES REPRESENTATIVE FIGURE 8 AN EXAMPLE OF A MESSAGE PROMPT ssssssssessessssesssssssecsssscsscsnsscsucsusacsucsusesescsnsesescsnsese
8. She also commented that because of experience of creating objectives data she usually estimates the number of records expected to be in the OBJT table providing a useful guideline for ensuring the data is correct During the observation there were no problems with the generated data so the user explained why she completes the checks and that although problems rarely occur the checks are still completed for her peace of mind The inexperienced user was unable to locate the correct query to run as the name did not match the one given in her guide Upon receiving assistance to locate the query she successfully verified the table A new system must allow users to complete these checks if they so wish it is anticipated that a similar method as described for the previous step could be used STEP 9 This step is to export both the OBJT and the OBCF tables and prepare the final file for loading into OneLilly The user explained that she uses the Export as xls such that she can process the data further in Excel A print screen of the Excel template used to combine all data types is shown in Appendix A Section 7 It consists of a worksheet for each data type and a further worksheet containing command buttons for completing verification checks on the data It also contains a worksheet for sales quota data OBSQ which is generated in the required format externally to the targeting process It is verified and exported along with the rest
9. Please upload the sales quota data Restart P 3 Upload an Excel file Browse Import Save amp Proceed Figure 35 Print screen of the import screen Error Prevention Upload Control box visibility As decided in section 4 2 3 the upload control will only be visible if the user selected the tick box to generate this data on the overview page This has been implemented using session variables 77 when the user clicks Save amp Proceed from the overview page each of the required variables are saved to a session variable An example Session OBSQ required BSPL Checked Session variables can be accessed from any other ASP page and they can be remembered from one page to another Here is an example of the code used to assess the variable from another page Boolean OBSQ required Boolean Session OBSQ required So for this particular page the control box is only visible if the OBSQ required variable was set to true Uploading data from Excel to Relational Database Inserting data from a local Excel file into the SelectedRatings table of the OneLilly database was done using ADO NET ActiveX Data Objects ADO is a library contained within the NET framework and is used to access data and data services The Excel file was opened through the OLE DB Object Linking and Embedding Database provider of ADO NET this is an API again designed by Microsoft for accessing different types of data stored
10. SUPERVISOR IF APPLICABLE Hilary Johnson SECOND READER IF APPLICABLE PROJECT TITLE Examining the effect of navigational redesign on end user experiences DATE 169
11. 0930 235959 break case Q4 objt Start Date year 1001 000000 objt End Date year 1231 235959 objt LLY Local Start Date year 1001 000000 objt LLY Local End Date year 1231 235959 break no default required as it can only be a value from drop down box objt Business Plan Name quarter year Territory position objt Status Planned objt Locked Flag Y objt Primary Position Flag Y objt Position Name position to work out position division String positionLetter position Substring 10 1 Switch positionLetter case B objt Position Division UK LPG BP break case N objt Position Division UK LPG MOOD break case objt Position Division UK LPG SCHIZ break case R objt Position Division UK LPG RSP break default as division is required make sure there is something if no match objt Position Division UNKNOWN break objt Report Flag Y db OBJTs InsertOnSubmit 0bjt db SubmitChanges select all distinct business plan names from OBJT List string q5 from o in db OBJTs Select o Business Plan Name Bistinct 0 to create BSPL if selected to do so on overview page if BSPL required foreach string bPlan in q5 BSPL bspl new BSPL work out position string position bPlan ToString Substring bPlan ToString Length 11 31 bspl Af
12. 137 ADp gt RCfF BSPL gt OBIT 2 o8 5Q gt o8 CE gt MAT Ad Another Cle A Or have menu Khe Wis ADD gt XREF DATA gt EXPORT gt SEND BSPL_ gt OBST gt ORGE gt OBCE visible when on data tak 55 Expert th code Record Tyee 138 VERVE L txeoec Sep see Your iS highlighted Bac D WSE oes AD L Rows Position Rating lt 19 bo j Automati tay Comoleked 139 140 Section 2 Copy of email communications for the CARD activity Original Email Sent On 26 March 2008 at 09 18 Samantha Cooper wrote Hi Nicola Carol Following on from our phone call here are the instructions for the CARD activity Thanks again for agreeing to complete this activity in such short notice Attached are two files 1 A JPEG containing a picture of 11 cards which you will need to print and cut out 2 An example data set for one sales representative although slightly shortened to make it more appropriate to complete this task What you need to do Imagine the cards are now the system and you need to generate OBCF and the corresponding data only for the 192R sales representative Simply arrange the cards in the order you would expect like to s Screen whilst you complete the targeting data you may use a card more than o
13. A position of authority or influence over any participant must not be allowed to pressurise participants to take part in or remain in any experiment No 10 Will the participants be informed that they could withdraw at any time All participants have the right to withdraw at any time during the investigation They should be told this in the introductory script Yes this is written in the initial briefing email 11 Will the participants be informed of your contact details All participants must be able to contact the investigator after the investigation They should be given the details of the Unit Lecturer or Supervisor as part of the debriefing Yes all participants are able to contact my supervisor or myself at any time throughout the project 12 Will participants be de briefed The student must provide the participants with sufficient information in the debriefing to enable them to understand the nature of the investigation Yes all participants will be informed of what the next steps will be and how I plan to use the data I have collected Here they will also have the opportunity to ask any other questions 13 Will the data collected from the participants be stored in an anonymous form All participant data hard copy and soft copy should be stored securely and in anonymous form As stated in point 4 names will be anonymous No data will be passed on and will be used for relevant purpose only 168 NAME Samantha Cooper
14. Integration ID Default Value Record Type Source System Id Operation Operation Date Name Type Description Start Date YYYYMMDD HHMISS End Date YYYYMMDD HHMISS Status Target Business Plan Name Summary Locked flag Geography Name 9 iL 5 2 2 Primary Position Flag Position Name Position Division Report Flag LLY Local Start Date YYYYMMDD HHMISS LLY Local End Date YYYYMMDD HHMISS Data Objective Time Management Record Type OBTM Field Name Affiliate Code Max Length Required Format Integration ID 0 Default Value Ref to DD o NE GE n a Record Type 127 Input Data Fields Input Data Fields n a Source System Id 30 Y 0 E n a Operation 1 n a Operation Date 15 Type 30 Y 0 Days 22 7 Hu Target Calls Day 22 7 0 T 0 Ratio 22 7 0 0 Comments 255 0 T E Parent Objective Name 100 Y Data Objective Sales Quota Record Type OBSQ Ref to DD Field Name Affiliate Code Max Length Required Format Integration ID Default Value Record Type Source System Id Operation Operation Date Parent Objective Name Product Id Primary Product Fl
15. M i we a ec T CannotOpenFile Excenton BSPL Figure 34 High Level Class Structure 76 LINQ TargetingDataContext Export GridViewExport 5 4 1 Navigational Structure Master Page To provide consistent layout throughout the application a master page has been used This is where the site navigation structures are contained Following earlier designs See Appendix B Section 1 this page will contain three content panes 1 Left Bar This is the left menu and will contain the Lilly logo at the top and buttons for logging out and accessing admin and help 2 A breadcrumbs placeholder This will be placed at the top of the page and implements the breadcrumb items as described in section 4 3 1 allowing a user to navigate to any sub task w Content placeholder This is where the custom content for each individual page will be added As shown in figure 34 this master page also requires a code behind file this is where the functions for each button click are contained 5 4 2 Importing Data This page is again built based on design decisions made in section 4 2 3 this is shown in figure 35 A Lb Overview Import X Ref gt Data Export gt Import You are logged in as sam cooper Call Frequency OBCF Please upload the Business Objects report for sales representatives selected ratings Admin Upload an Excel file Browse Import Sales Quota OBSQ
16. New application 3 I found it easy to locate the call frequency data Access solution 1 New application t 4 didn t need to rely a user manual or help functions Access solution New application 5 I find the system easy to navigate through Access solution New application 5 6 I found it easy to update the cross reference table and subsequently the data set Access solution New application 5 7 1 found it easy to edit the data produced Access solution 4 New application 5 Please provide all your general comments on the task or new interface 153 OneLilly Targeting Questionnaire Thank you for taking the time to participate in the OneLilly targeting application comparison experiment if you could please reflect upon the tasks you were asked to do by completing this questionnaire On a scale of 1 5 with 1 being strongly disagree and 5 being strongly agree please rank the following for both the existing and the new system 1 I found it easy to complete the first task Access solution New application 5 2 Ifound it easy to complete the second task Access Solution 1 New application 5 3 I found it easy to locate the call frequency data Access solution 18 New application 4 I didn t need to rely on a user manual or help functions Access solution 1 New application 5 5 I find the system easy to navigate through Access solution New application 5 6 I found it easy to update th
17. items in section 4 3 navigational structure it can be seen that this activity has confirmed breadcrumb items need to be in the following order Overview gt Add data gt Cross reference Review data gt Export Specify data types required The results also show that an additional card was required to view feedback from the import this is from a decision made in the previous section As this activity has discovered the design of this feedback is yet to be considered consequently further questions needs to be considered Earlier the question how data should be validated Q7 was answered by choosing the option to only allow valid data to be added as a result of this design decision further questions needs to be considered Q7a How should feedback be returned to the user Q7b What feedback is required The options for these questions were considered via a telephone call 40 with two of the users These can be summarised as follows 01 Displaying the results in a table below the import box 02 Displaying a message box 03 Simply display the invalid data so that the user is aware of exactly what was not loaded 04 Above plus row number 05 Above plus reason for rejection From the observation it was seen that users tend to ignore message boxes so the second option is more appropriate to ensure users take note of the feedback For part b all users would prefer to have the final option but the first option is required at a
18. made of where the user experienced issues Due to the complexity of this tool this would only be useful if an experienced user firstly gave a demonstration of how they complete the task The task will be to generate objectives data for one sales representative from the LPG Lilly osteoporosis group sales force Typically the tool would be used to generate data for an entire sales force and therefore approximately 100 sales representatives so it is important to note some areas of the tool will be simplified for example the cross reference The evaluation session will be recorded using a Dictaphone supplemented by hand written notes 32 Background of participants The experienced user has been using the tool in question since it was developed and will assist and demonstrate the tool to the inexperienced user She has acquired the required technical skills however they are not a requirement for her job role The inexperienced user has never used the tool She has a basic understanding of the objectives data as she has assisted with the task before the adoption of the existing tool She does however use Microsoft Access for other tasks and consequently some of the concepts are familiar Results and analysis The experienced user took around 20 minutes to complete the task in contrast to the inexperienced user who took approximately one hour It is reasonable based on past experience to assume that these figures would be doubled with
19. participatory design ethnography This is an observational technique it seeks to understand settings as they naturally occur rather than in artificial or experimental conditions Although this technique has the most significant benefit of fully understanding the end users it also has a number of issues It usually involves quite lengthy periods of time at the study site and once completed the presentation of the results is frequently not represented in a way that is useful to designers 31 There is a vast amount of stakeholders that could potentially be involved in participatory design these include end users developers human factory workers technical writers systems analysts trainers and marketing staff All can make a difference if they are part of the design and all will have their work lives influenced by the design one way or another 28 Once a design has been generated it will need to be evaluated not only against the requirements specification but also against previous designs again it will also require direct involvement of stakeholders 2 6 Usability Evaluation Usability evaluation is a means of collecting information about the users experience when interacting with a prototype As stated throughout this chapter a user centred approach is being taken for this to work effectively evaluation techniques will need to be used throughout the design and development life cycle to provide feedback this means that the results of evalua
20. system The reason for this could be the fact that she knows the existing system in greater depth 94 meaning that there has been a negative transfer of existing knowledge that other participants wouldn t have had to the new system This meant that she had to unlearn the existing system as well as learn the new to a much greater extent than the novice users During the interview it became apparent that she was trying to map how she completed certain tasks within the old system to the new one whereas other participants do not have the existing knowledge to do this meaning they were able to complete the task in a quicker time Figures 40 and 41 show the difference between experienced and novice users Mean Time Taken in Existing Solution B Experienced o 5 E X is E B Novice Task Number Figure 41 Bar Chart displaying differences for experienced and novice users existing system Mean Time Taken in New Solution 5 4 3 2 1 B Experienced 0 Time Taken Minutes B Novice 456 Task Number Figure 42 Bar Chart displaying differences for experienced and novice users new system When comparing the results between novice and expert users it can be seen from figure 41 that in general novice users took a greatest amount of time to complete each task within the existing system If this is compared to figure 42 this is actually turned around such that in the experienced user
21. van Cres Ad ABER Cose xo ao Oc OS GQ wo ce c C nef ua s A nace Ouar aA 119 wa Aan a wom QV uwa yanes Gre OA ready Coco the Cabla Extra uor ores by Sar Che Ue Vos a 7 we rer Cos ace Ye COSA Cursors C s it ons ans vo how WSer co moe Wor mead SF PONY Cor AK Sic Avand Sag tSkion oC c onu My o ed hove eda utora aser lecce Seer SF a oSe Says LS ma kawur Uan the Oa e mosod US Of Ex cor S ereadsroaks CAN con we varg Comegucoked SQL code Sivie 47 ASER Kos undar Stand o See ecm messages Con he ALR LA FO Wrohar Sound names cx dune gt teo COM IO Omm amp o ce moe khad gje ASS Seonw KO oder Syr anrd under Ws error messages be more UnhormMakivue 5021 Suggevceck b More XO 99i Uno S o qorar Luh ee System T o or hoan 5 og We Ga CADO D i monarx POS reas CeSoas Suna Acc e Ve M iain Cos monr See cae e Ce c 9 Lowe Co poligokicn ext amis Row fanon aways Reading to Stck 5 r g Aona O O Cx Orc ouf Ute aO O cede v v wore Q y von 0850 Geers v faoned Sra hadnt von uar e
22. 0 5 Last Revision Date 28 Jul 2006 Confidential to Eli Lilly and Company 134 Section 7 Print Screen of template used to combine and export data EI A C mm m cmd 1 For any issues regarding this spreadsheet please refer to OBJT template documentation 2 For any issues regarding sending files to OneLilly please refer to sending objectives troubleshooting guide EIS Run All Validation Export E 15 Send to OneLilly 16 a gt N Run All Validation Clicking this will run all the validation checks This includes Required fields date formats valid dates and ensuring that parent objectives and business plans exist Only run this validation if you have completed all tabs Export Clicking the export will publish the OBJT input file the file name will contain the date amp time that this button was clicked and the output file will be stored at Ybsvr09 uk onelilly_data Objectives exported input files Send to OneLilly Clicking this button will place a UK OBJV dat and a UK OBJV rdy file in the following location Ybsvr09 uk onelilly_data Objectives to send From here the file will be automatically routed to OneLilly n Individual Sheet Validation 22 If you are NOT loading all record types for objectives then use these buttons to run separate validation PLEASE NOTE You can edit the list of values but If you change the number of valid values in
23. 51 of the usability problems they evaluated Figure 2 8 shows how increasing the number of evaluators just by 2 or 3 there is significant improvement 100 75 50 Problems Found 25 Proportion of Usability 0 0 5 10 15 Number of Evaluators Figure 4 Graph Showing Proportion of Usability Problems Found By Evaluators The study showed that heuristic evaluation is difficult and that you should not rely on the results of having a single person look at an interface They will be much better if you have several people conducting the evaluation and should do some independently of each other As figure 2 8 shows the number of usability problems found grow rapidly in the interval from one to five evaluators but reaches a point around 10 evaluators Nielsen and Molich 5 recommend that heuristic evaluation is done with between 3 and 5 evaluators and that any additional resources are spent on alternative methods To aid the evaluators in discovering the usability problems a set of heuristics which were previously discussed in section 2 2 2 are provided to evaluate whether the interface elements conform to the principals these are listed and discussed in section 2 2 Each evaluator assesses the system and notes violations of any of the heuristics that would cause potential usability issues The severity is then assessed based on four factors 1 How common is the problem 2 How easy is it for the user to overcome 3 Wil
24. Challenges for educational multimedia and web designers SIGCSE Vol 33 Number 4 37 R Macfarland 1995 Ten design points for the human interface 38 V Bellotti A Maclean amp T Moran Oct 1991 What Makes A Good Design Question ACM SIGCHI Bulletin 39 A MacLean R Young V Bellotti amp T Moran Sept 1991 Questions options and criteria elements of design space analysis HCI volume 6 pg 201 250 45 PIROLLI P CARD S K 1999 Information foraging Psychological Review 106 4 pp 643 675 115 47 KATZ M A BYRNE M D 2003 Effects of Scent and Breadth on Use of Site specific search on E Commerce Web Sites ACM Transactions on Computer Human Interaction 10 3 pp 198 220 48 DUMAIS S T 1988 Using Latent Semantic Analysis to improve access to textual information Proceedings of the Conference on Human Factors in Computing Systems CHI 88 50 B Lida amp S Hull 2003 Breadcrumb Navigation An exploratory study of usage 53 MALDONADO C A PESNICK M L 2002 Do common user design patterns improve navigation Proceedings of the human factors and ergonomics 46 annual meeting Baltimore September 2002 pp 1315 1319 54 HUDSON W 2004 Breadcrumb navigation There s more to hansel and grettel than meets the eye Interactions 11 5 September October 2004 pp 79 80 New York ACM Press Web based 6 J Nielson 2005 Severity ratings for usability problems online Available at http
25. Developing predictions and choosing a means to test this From the hypothesis a small number of predictions can be developed Firstly the time to complete the task will be quicker than that of the existing system This will primarily be due to the structural guidance through the task that the new interface provides The use of breadcrumbs provides the user with a mental model of where they are within the system and how far they are through the task The new application will be easier for inexperienced users to use due to this as well as the fact that a user does not need a technical understanding since everything is completed through the interface and no knowledge of SQL is required Furthermore data quality and integrity will be improved within the new application since it completes verification on behalf of the user both during the import phase as well as any edits that are completed by a user This will lead to increased productivity and user preference for the new application Finally as the new interface was developed using a participatory design approach users have already approved the new interface and so the prediction is that the new interface will be superior to the existing with the most significant factor being to due to easier task completion in reduced time There is a prediction that the number of clicks required for a user to reach their goal successfully complete targeting data will be reduced due to the navigational redesign As stat
26. For example a web page on a particular type of car might have a breadcrumb style trail of Ford gt hatchback gt under 1000 gt used gt blue The difference to location trail is that this does not necessarily reflect the actual hierarchy of the web site This paper will focus exclusively on the use of location breadcrumb trails Neilson advocates the use of breadcrumbs as e They improve the user s understanding of where they are in the site e They are very easy to use and have never confused users in experiments e They take up very little space on the page and never have negative effects 9 Previous research on breadcrumbs as a navigation tool has been limited by using simple site structures and completed within controlled experimental conditions 50 Bernard 51 suggested that the use of breadcrumbs can aid in improving the user s mental model of the site s structure Opponents of the breadcrumb argue that breadcrumbs are not used and they have no effect on overall navigational efficiency Studies by Rogers and Chaparro 52 found evidence that Bernard s theory was well founded and that users mental model of the system does seem improved when breadcrumbs are used Maldonado and Pesnick 53 concurred that most users do not use breadcrumbs nor does it improve navigational efficiency for those who do but Rogers and Chaparro 52 backed this but went on to conclude it can and does have a positive effect when users are trained to use br
27. However it is clear that the experimental mode still requires comparing and decision making but this is just achieved automatically without us always realising For instance in driving a car your actions will include things like deciding on when to change gear and comparing other drivers speed with your own According to Preece both of these will require different technological support 2 There has been research into the evolution of cognitive modelling for understanding and predicting user behaviour in a computer system One example of a cognitive model is to consider the theory of an action this is how the user interacts in terms of their goals and what they need to do to achieve them Norman s execution evaluation cycle is the most influential in HCI The cycle can be divided into two major phases execution and evaluation These are subdivided and form seven stages Establishing the goal Forming the intention Specifying the action sequence Executing the action Perceiving the system state Interpreting the system state DON cm Geo ps Evaluating the system state with respect to the goals and intentions The theory proposes that the stages take place sequentially but in reality activity would not proceed in such an orderly sequential manner some stages could be missed repeated or completed in a different order As a model it is an approximation of what might happen and could aid designers to think how best to design
28. It is important to note that the empirical evaluation is actually an experimental manipulation because there is an independent variable the system being varied with the affects on the dependant variable being measured However it will not be undertaken using a large number of users or in a laboratory setting This experiment is a small scale pilot study due to there being limited participants The application has only four main users each of which participate in the experiment and it is not possible to recruit any more suitable users due to the fact that the experiment is within industry requiring users to participate during office hours Section 6 2 6 discusses participants in greater detail 6 2 Experimental evaluation The experiment is the primary means by which it is possible to establish cause effect relationships between certain variables and the events that occur as a result of this Johnson 42 states that designing HCI experiments involves the following steps Formulating a hypothesis Developing predictions from the hypothesis Choosing a means to test the prediction Identifying all variables that might affect the results of the experiment ge 00 See p Deciding which are the independent variables and levels of the independent variable dependent variables and which variables need to be controlled by some means 6 Designing the experimental task and method 87 7 Subjectselection 8 Deciding the experimental desi
29. String Format Provider Microsoft Jet OLEDB 4 0 Data Source 0 Extended Properties Excel 8 0 HDR YES fileName DbProviderFactory factory DbProviderFactories GetFactory System Data OleDb List EISelectedRating okToImport new List lt EISelectedRating gt List lt KeyValuePair lt int EISelectedRating gt gt problems new List lt KeyValuePair lt int EISelectedRating int currentRow 1 using DbConnection connection factory CreateConnection connection ConnectionString excelConnectionString using DbCommand command connection CreateCommand Main comes from the name of the worksheet command CommandText SELECT Division Position Con Integration ID My Selected Rating FROM Main try connection Open catch Exception This is if the file cannot be opened throw new CannotOpenExcelFileException try when you got to the end of reading close connection using DbDataReader dr command ExecuteReader CommandBehavior CloseConnection 156 while dr Read currentRow EISelectedRating r new EISelectedRating r Division dr Division ToString r Position dr Position ToString r ConIntegrationID dr Con Integration ID ToString if dr My Selected Rating ToString Equals Non Target r SelectedRating N else r SelectedRating dr My Selected Rating ToString string sl dr Position
30. The web provides access to an unparalleled volume of information and this will continue to expand with an increasing amount of web based information technologies being introduced The ways in which people produce find and use information are rapidly evolving which means that the field of information retrieval is far from being a solved problem Navigation directly 112 impacts information retrieval and developing interfaces to effectively support this is likely to be a significant research area moving forward With the trend towards increasing ubiquitous computing and the emergence of mobile devices this affords the opportunity for investigating navigation design from a new perspective Much existing research has considered the usage of navigational aids based on the size and positioning on the page but what are the implications for mobile devices with small screens Is there enough room for navigational aids If they are hidden will users be aware of their existence Increasingly standard websites are becoming available on smaller screens but is it possible to navigate in the same way These questions are likely to be those that are considered in future research Finally how navigation supports searching is a crucial research theme Today s retrieval systems support common functions for everyone regardless of who where or what the individual is or looking for Such systems are being used by an increasingly diverse population for increasingl
31. ToString int length sl Length take the last letter of position r PositionLetter sl length 1 if SelectedRatingsImporter RatingOKForImport r okToImport Add r 1 problems Add new KeyValuePair lt int EISelectedRating currentRow r catch Exception This is if reading the data fails throw new ExcelFileInWrongFormatException results NumberOfRatingsSuccessfullyValidated okToImport Count results RatingsWithValidationErrors problems int importCount 0 ExcellmportersDataContext db new ExcellmportersDataContext foreach EISelectedRating r in okToImport db EISelectedRatings InsertOnSubmit r db SubmitChanges results NumberOfRatingsSuccessfullyImported okToImport Count return results private static bool RatingOKForImport EISelectedRating r if string IsNullOrEmpty r Division r Division Length gt 15 return false if string IsNullOrEmpty r Position r Position Length 15 return false if string IsNullOrEmpty r ConIntegrationID r ConIntegrationID Length gt 15 return false if string IsNullOrEmpty r SelectedRating return false Switch r SelectedRating case A 157 return true case return true case return true case return true case C return true case Cr return true case N return true default return false SelectedRatingsImportResults c
32. about record types There is no need to sort the record types as the OneLilly Interface will do it in order to process the record types in the correct order first the Business Plans and Objective Headers the Objective facts The file should be generated with UTF8 unicode character set The file shouldn t contain any header 3 character header common in UTF 8 files Each file per affiliate Input Format will be transported from Affiliate server to One Lilly Informatica Windows Server through WebSphere MQ with specific routing ID one per affiliate and Input Format e g one for Japan Product 5 2 Allowed Operations The process will handle Insert and Update it will not handle physical deletion The name of the objective cannot be updated The update operation may perform Logical deletion changing the objective or business plan status from Current to Expired 5 3 List of Record Types The Objective file contains the following record types e OBJT Used for most of the Objective Headers Sales Quota Key Account Key Contact Expenses Call Frequency Coaching e OBTM Used for Time Management objective facts e OBSQ Used for Sales Quota objective facts e OBCF Used for Call Frequency objective facts e BSPL Used for Business plan header BPSW Used for Business plan facts notes 5 4 Technical Data Rules Title OneLilly R1 Entity Specification Subject Objectives Page 9 of 9 Documentum ID Status Released Version
33. actions 8 Reduce short term memory load by keeping display simple consolidating multiple page displays and providing time for learning action sequences These rules are guidelines only and are not applicable to every eventuality They will need the designers interpretation for each new situation However they are useful and their application will help most design project It is now important to gain a background understanding of how these guidelines and rules can be considered through to design 2 4 Design considerations From preliminary research within this project it became apparent that the main issue with the existing solution was that inexperienced users and even experienced experience issues with understanding the sequence of actions they are required to perform in order to achieve their goal The interface is preventing them from completed the task effectively this follows Norman s gulf of execution model This section will discuss how the heuristics from above can be interpreted to consider design options it will focus on considerations for navigational design designing to suit both novice and experiences users and information presentation 2 4 1 Navigation design Much of interaction design involves goal seeking behaviour Users will have some idea of what they are looking for and a partial model of the system When users are navigating through a system the important thing is not that they take the most efficient route but at
34. and assessed 4 Evaluating what is being built throughout the process and the user experience it offers Evaluating what has been done is the main focus of interaction design The main focus is to ensure that the product is usable and this is addressed through a user centred approach which considered users throughout As described in the introduction chapter the project aim is to redevelop the OneLilly targeting application to more effectively support the user With an aim to provide greater navigational support it will be hoped productivity can also be increased With this aim in mind this chapter will be used to gain a background understanding of the key factors principles and methodologies in user interface design and development using a user centred approach The chapter will begin with a summary and discussion of human computer interaction HCI principles and the design considerations needed to support these principles It will then move onto discuss navigation user support and information presentation The chapter will conclude by discussing user centred design tasks and usability evaluation This chapter will also highlight some of the current research impacting navigation design decisions This will include the field of information foraging theory and its related concepts such as information scent and latent semantic analysis 2 1 Introduction to HCI With the vast increase in use of PC s in the early 1980s new usability challenges
35. and ensuring the user knows what they have done which is a heuristic listed at the start of section 2 5 2 4 2 Designing for both novice and expert users Different users have different needs and skills meaning that help within the system might be essential to some users whereas for more regular users of a system user support may not always be required and could potentially become irritating Aberg 23 discusses there is a strong need to provide support to a whole range of users J Wu 25 also makes the same comment by stating that novice and expert users of a system differ greatly and accommodating both in one interface is a challenge as well as a necessity for most user interfaces Shneiderman 24 comments that expert and frequent users have special needs and as his golden rule states it is a good idea to enable frequent users to use shortcuts Nielson s seventh heuristic also suggest this they are known as accelerators and are elements that allow a user to perform a frequent task quickly even though the same tasks can be performed in a more general or slower way These can include things like abbreviations function keys or an entire command in a single key press 25 Another shortcut could be allowing users to jump directly to desired locations within the system 3 When using an interface expert users can form goals and sequences of actions to achieve their goal much quicker than a novice user They would prefer the number of inte
36. at each iteration The process of user centred design has shaped the project and as a result of giving users the opportunity to engage in hands on activities they have happily and immediately adopted and made use of the new system Therefore on significant success of the project is that the software is now being used on a daily basis in the commercial setting for which it was designed In general using participatory design techniques ran smoothly throughout the project However on occasion there were a limited number of comprehension and communication issues both between the author and between the end users themselves These resulted from the fact that not all the stakeholders and the project author could physically meet to discuss every small change in the implemented design and electronic means was utilised instead for such feedback needs Therefore as a recommendation if such a user centred project were to be completed again it would be sensible to set up a shared collaboration area where discussions and information could be held in one place rather than using email communication 7 2 4 Practical The practical project output is the successful redevelopment and deployment of the OneLilly targeting application This is a fully functional prototype which is now in use at Eli Lilly and during company testing was successfully used to generate targeting data for this quarter The software has provided the company with significant business benefits wit
37. be unable to fulfil its purpose Discussions during step 5 ofthe observation highlighted how the system should follow a user s way of thinking meaning that objectives are added to ratings Source Step 5 of the observation section 3 5 1 and the context section of the interviews section 3 3 4 The system shall now allow the cross reference to contain duplicate rows 47 10 11 12 13 14 15 Rationale As discussed in the interview the existing system maintains old values and places new below this causes user confusion and frustration and can be prevented through the use of this requirement Source Following Nielson s guideline of error prevention literature review section 2 3 and user comments during interviews section 3 3 4 The system shall be capable of producing the following data types BSPL OBJT OBCF and OBSQ following the OneLilly formal input specification see Appendix A Section 2 Rationale This is basic functionality of the system and is required to fulfil its purpose of generating all targeting data Source The context section of the interviews and the observation sections 3 3 4 and 3 5 1 The system shall allow a user to view the result for each data type Rationale This is a basic functionality of the system it is essential for a user to view the resultant data Source Interviews and observations section 3 5 1 steps 7 and 8 The system shall allow a user to add edit or delete any entr
38. currently the user is guiding the system but it should be the other way around the system should work in the same way that the user would think Several comments were made about how comfortable the users felt using the system These included e extremely worried that if I do the slightest thing wrong I will have to start again e Sometimes the system does things without telling me e Error messages can be very difficult to work out without a technical understanding of how Access works e Any invalid data is simply removed without informing the users Although such removal is good for removing undesirable incorrect data giving feedback to understand why this is being done and potentially allowing updating of the data would be better e There are numerous places where there is not enough feedback regarding what the tool is doing All of the SQL is visible to users instilling fear in non technical users Displaying SQL in this manner is not appropriate since it is very easy to edit tables or queries which may result in system failure However being able to access the code easily does allow for flexibility and makes support easier This is because in the current state support can be given immediately by a business integrator whereas a formally developed system would be supported by the main technical group which follows stringent change control procedures However the balance between usability and flexibility in this case is
39. design Proceedings of the SIGCHI conference on human factors in computing systems CHI 92 28 M J Muller M Wildman amp A White June 1993 Equal Opportunity PD using PICTIVE Communications of the AC volume 3 pg 64 66 29 M Muller T Dayton amp L Gayle 1993 A C A R D game for participatory task analysis and redesign Conference in human factors in computing 93 30 S Bodker amp O Iverson Oct 2002 Staging a professional participatory design practice Moving PD beyond initial fascination of user involvement NordiCHI conference on human computer interaction 2002 31 J Hughes V King T Rodden amp H Anderson Apr 1995 The role of ethnography in interactive system design Interactions Volume 2 Issue 2 ACM Press 32 H Lichter S Hufschmidt amp H Zullighoven May 1993 Prototyping in Industrial Software Projects bridging the gap between theory and practice Proceedings of the 15 International conference on Software Engineering ICSE 93 33 C Floyd 1984 A Systematic look at prototyping 34 J Rudd K Stern amp S Isensee Jan 1996 Low Vs High fidelity prototyping debate Interactions volume 3 Issue 1 ACM Press 35 L Gayle Tudor M Muller amp T Dayton Apr 1993 A C A R D game for participatory task analysis and redesign Macroscopic complement to PICTIVE INTERACT 93 and CHI 93 conference companion on human factors in computing systems 36 C D Ryan Dec 2001 The Human Computer Interface
40. differing opinions which again will reflect different options as follows 01 Queries will be hidden from the user such that they will be automatically ran when the proceed button within the cross reference section is clicked 02 As above but rather than the proceed button individual queries will run as each data type is clicked to be viewed 03 On the Data section include buttons which when clicked will run the relevant query The final two options are not appropriate since as previously discussed in the observation queries have to be ran in a particular order e g call frequency data has to exist before their objectives are generated so the first option is the only one where errors can be prevented DATA VALIDATION Two of the three users sketched designs for how data would be viewed and validated by the users Both used a standard table design displaying all columns as seen in figure 27 example of 0 3 ek sabor ee vor e came Sotti ey new Co e 3 rows You t0 ko Aciete m PTS 5 DELETE 25 Mesas mer a b ADD J e Co EI see cos ruin to ei Figure 27 Designs The only difference between these two designs is the presence of the add edit and delete controls When this was raised during the discussion the designer of figure 27 right hand side stated that although it was missed off her sketch there is a need for the ability to edit
41. each point in the interaction they make some assessment of whether they are getting closer to their often partially formed goal To enable the user to do this goal seeking each screen of the system needs to give the user enough knowledge of what to do to get closer to their goal 1 7 To do this goal seeking each state of the system or each screen needs to give the user enough knowledge of what to do to get closer to their goal Users will typically forage for information by navigating from page to page along links their patterns or actions are guided by their information needs 13 This is known as Information scenting and is a term used to describe how people evaluate the options they have and refers to the extent at which users can predict what they will find if they pursue a certain path through a website Pirolli and Card describe information scent as the imperfect perception of the value cost or access path of information sources obtained from proximal cues such as bibliographic citations www links or icons representing the sources 45 Users will use proximal cues such as snippets and graphics to assess the distal content 13 The cues such as text labels are used by users to make information seeking decisions and to guide them to their information goal If the user feel that they are on the right track to the information they require they will continue as long as they sense that they are getting warmer i e that the scent
42. either of the lists then you will need to re define the lists This can be done by using 23 S BSPL OBJT OB OBCF 26 27 List validation 29 30 Status OBTJ Type 31 Planned DSM Dual days 32 Ongoing Coaching Field Days Achieved z 33 Completed Contact Call Frequency Meetings with customers Not Achieved Direct Gales Nunta J the Insert menu gt Name gt Define ealact tha list uan chanaad and than 135 Appendix B Design Supporting Documentation Section 1 User Interface Design sketches completed by end users Welcome to the Las Tedesco Aq phe ion Loc aa TTA Ug name E zm Qe word B 1 yvur i PIDE mmo ij MAC 25 piens 2 ose re gt re eee xa ren khau Stage iS com qere OVERVIEW are a eed See ce ke deka e ore queo A i Aerecaxe Mere ree ve 4 peni wenke put dma es RL Buh rec e EE Y 1 oe onver oft F andlor pth t S otsa if Sele ck e Re beer o65Q guere Sele ex Quacker UN amp teoceen 136 Loco You are logged n as vinane te iz import esee lon Login Success OCA We loome Username qoe please Select Sales Farce eS wi lec oc ecd axe 2 Validate validate foes ogsa Z Ge load coke
43. emerged and became more salient Users consisted of cognitive scientists psychologists sociologists and philosophers they were interested in how people solve problems and learn new things This new area of shared interest between computer science and cognitive science is called Human Computer Interaction HCI 3 Human computer interaction is the study of the relationships which exist between human users and the computer systems they use in the performance of their various tasks 10 Along term goal of HCl is to minimize the barrier between the human s cognitive model of what they want to accomplish and the computers support for the user s task This means that HCI is concerned with providing an understanding of both the user and the computer system in an effort to make the interaction between the two easier and more satisfying However the emphasis should always be on the user The interest is in how the human user uses the computer as a tool to perform their desired task and in order to accomplish this task the user has to communicate with the computer To enable a successful interaction the product must be usable The term usability has many different definitions the Oxford English dictionary defines it as The effectiveness efficiency and satisfaction with which users can achieve tasks in a particular environment of a product High usability means a system is easy to learn and remember efficient visually pleasing and fun to use a
44. fle 5 PR dmg B groase ema Sant Ges SESE eae _ MED A quet v j ec Cony OAT NOE Moe tee es eb Ce Eni i sree craic we Ce on Qrew ous More a out Excel a US lexecec cers ave Riewe Sales e Qwose Selec khe objecrive X Foo nq ok n laxctac S Weoc o eon Ve rece R Wet ed 2 Obreckive NOT NULL oe of Coloured wa Edis eA geye o cannok BAVA Nk cada Line ko wes and edit xe daka Xet WAS een produced es YL OBIT Oct oe65Q ALL NALI DAT E QIN Att Same oJ excer VALAARTION c R sar es GET Nm CN CX AP Tandem y Servecte A M eAxX of COAT Xe AD N eec o EPO 238 ee Add ew uive L wit oM Cned un Wher tre Seve c cod ME adrcatie i repuwed eMecnue has been 85 OBST sovrte fae Example of ACE nakon ND iL CR COME Sec ks Ve coc Ac e ec rows yes vex Xo Acieke xe Monn PELETE enue ere Top D nes fow eow Seve co 5 wih Xo eo FRRORT aes INEO Gite uev Eces oe tare Wwe fre e Views voe QNMexed dora tortet Cocmrax XK Were Coe cComelered SNR Sena Onc VN
45. generating data for all sales representatives in a sales force because additional actions and verification checks would be required Both users relied on the step by step guide which can be found in Appendix A Section 3 although it was observed that this guide is not an accurate reflection of how the task is completed as it does not document workarounds for functional problems Figure 10 shows a summary of the task with steps labelled on the diagram and then referenced within the text Step 1 Step 2 Step 3 Step 4 gt Build E Finds unique selected ratings amp inserts into Ratings2Objectives Delete old amp Import SelectedRatings new raci required data splitting up position and letter Business objects report selected ratings Key Joins Ratings Q and Green QUERY Ratings2Objectives to generate OBCF in the correct input format currently this still contains invalid Step 6 selected ratings Step 5 Make Ratings2Objectives OBCF ONELILLY OBCF This acts as a X Ref Removes invalid ratings Step 8 Step 9 Create OBJT Uses OBCF Parent Select distinct amp OBJT ONELILLY OBC OBJT field to make table final F generate OBJT Figure 10 Overview of the Targeting Process Step 7 OBCF_ONELILLY final A In the following descriptions italic text describes how the inexperienced user approached each step 33 STEP 1 Generate the business objects report taking
46. if participants are likely to object or show unease when debriefed No 4 How will participants voluntarily give consent If the results of the evaluation are likely to be used beyond the term ofthe project for example the software is to be deployed or the data is to be published then signed consent is necessary A separate consent form should be signed by each participant Each participant will sign to state that they have read and understood how their data will be collected used and stored All names will be stated with initials only for reference and kept anonymously 5 Will the participants be exposed to any risks greater than those encountered in their normal work life Investigators have a responsibility to protect participants from physical and mental harm during the investigation The risk of harm must be no greater than in ordinary life No 6 Are you offering any incentive to the participants 167 The payment of participants must not be used to induce them to riskharm beyond that which they risk without payment in their normal lifestyle No 7 Are of your participants under the age of 16 Parental consent is required for participants under the age of 16 No 8 Do any of your participants have an impairment that will limit their understanding or communication Additional consent is required for participants with impairments No 9 Are you in a position of authority or influence over any of your participants
47. include those that can edit generated data and those that can only view This would be useful when sharing data to the sale force before it is complete It was therefore decided that the administration functionality was useful and should remain in place Finally discussions arose around the screens displaying generated data comments included the need to scroll far to the right hand side to see all columns Users had different opinions on this some stated some of the static columns could be removed and only displayed when the data is exported to Excel Other participants liked being able to see everything it enabled the user to verify everything within one screen Considering both of these comments an additional enhancement might be to have a button to allow a switch between two different types of views This discussion also led onto how the data is divided such that only 10 records are displayed per screen participants commented that this is useful but had comments on how this could be 103 improved by following the standard of their global CRM system Participants stated that it would improve the usability since they would immediately know how many records and pages there are Figure 46 and 47 show the difference 12345678910 Figure 46 Current paging structure Results 1 30 of 2000 Previous Page Figure 47 Requested global system paging structure Figure 46 shows page number only and this could become very large with sign
48. is used on a quarterly basis when data for each of the eight sales forces are generated There are other scenarios when data must be updated outside of this schedule for example if new sales representatives join the company or roles of existing sales people change The stakeholders refer to the data as objectives of which there are three types e Sales Quota the target amount in GBP of a given product that sales representatives must aim to sell in a specified time period e Time Management how a sales representative s time should be split for a specified time period This indicates how many days each sales representative should be visiting customers completing training or performing administrative tasks e Call frequency the frequency with which a sales representative must visit a targeted customer in the specified time period Each individual objective belongs to an objective and each objective is part of a business plan Figure 3 2 gives a pictorial view of how the data is structured Businss Plan BSPL Parent objective OBSQ OBTM OBCF Time Call management frequency Objective Sales quota Figure 6 Structure of the Targeting Data Time management does not form an inherent part of this project since this data type is currently unstable This decision was taken by the company and it is possible that when stability is achieved this maybe a future enhancement to the system 24 As an example figure 7
49. is getting stronger or the user will give up 46 With perfect information scent cues a user will make no navigation errors and will proceed directly to the desired information 45 If there is no or uninformative information scent i e there are no proximal cues for the user to assess Pirolli and Card found that the user will perform random choices in the search space as there is nothing to guide them A site with weak or no information scent could affect users decision to search or browse in a navigational space This could result in the user anticipating that browsing is a more costly strategy than searching In a study by Katz and Bryne on the effects of scent and breadth of use of site specific search they found that high information scent for menu options plays a key role in whether a menu will be used With high information scent on menu options their participant used menus even when there was a prominent search function 47 Nielson 14 states that the most obvious design lesson from information scent is to insure that links and category descriptions explicitly describe what the user will find at the destination e Do not make up words or slogans as navigation options Remember to think about the words in the users mind and therefore what they would be looking for e Each page should clearly indicate that they are still heading in the right direction It should provide feedback about the current location and how it relates to the users t
50. like about the current tool Ane we faos ee yoa can See SQL wan osagaie c re Qo C aseba For xa wog f amp CeesS S a muo quers Wan Matus _ Ww hr ungue ronas Xo vandoje a fram he Moines Obie ns egor 2 Arethere any features you would like to see retained from the tool Or possibly there is a process or understanding you had from completing the task manually that you would like to be able to see do in a new system N CX 3 Do you find the current tool easy to use ere gt 55 o cS Uo C ores cf Sof If so what makes it easy to use If not can you describe why it s difficult to use Qv ZAS xc Ov Bre una 125 4 What is bad about the current tool What frustrates you about it What features would you like to see removed a paw for 2a ch DN some Coreen bes eg Oe Coe ee eRe ies otsaga BCE cw s 3 m Ju nem Ov eues xxu ce o chor M Y you ov 7055 o0 5 What new features would you like to see Emro 9 Ure Sr Onn Car Wer on uw the oro cass take Resear Wan Loading Aowa ge CKO hom ax S S Mem G coda lt 5 ar nS D Oe thor ren an ve an i Ror cor A WeSoNEeSS sey oors Rx 3 0 Qvi eiivet Gre
51. minimum Obviously the designer also needs to be aware of the requirements that a data set needs to satisfy in order for it to be valid for the sales quota these are set out within the OneLilly formal input specification see Appendix A Section 2 For the selected ratings the experienced user provided an email see Appendix B Section 3 listing the requirements It can be seen from the results that the only difference in workflow was at card number 6 where each user chose a different data type to verify first This shows that requirement number 22 providing flexibility has been satisfied with this prototype This CARD session has provided an understanding of the high level actions required to complete the task it is now a good idea to consider each of these actions in greater details This 63 can be done using a hierarchal task analysis as previously seen in chapter 3 this involves breaking tasks down into sub task and then into sub subtasks and so on 4 4 Envisioned HTA An envisioned HTA will be generated based on findings from the previously conduced HTA section 3 7 and the design decisions that have been made so far Upon completion it will be discussed with stakeholders to make sure they have a good understanding of how the system will support each sub task allow them to suggest any restructuring Figure 30 shows the envisioned HTA 64 eijgissod Buuepio gare Kay BLOND sajes 9 s yoegpss m podu MOU
52. note that the ordering of steps was different each user undertook exactly the same steps After the initial discussion the first part of the session involved producing a post it note of each step and allowing each participant to arrange them in their preferred order Notes and pictures were taken from each of the user s choice Figure 15 shows the first user s result Figure 15 User s Results from Card Sorting Session The users described how they used an Excel template with column headers set out as the formal specification shows the specification can be found in Appendix A Section 2 To a certain extent itis then the user s choice on the ordering of which fields they decide to complete 40 The first user arranged her cards as follows In a separate spreadsheet import the sales representative data from OneLilly Divide the number and letter from each position using a formula Within the template For the number of records required enter the 5 System Fields For each record generate the Type based on the position Ov UT wm w doce For each record use a concatenation formula to complete the BSPL Name based on the date range type and position from the initial spreadsheet p Generate the Description a concatenation of name field and Business plan Enter the Start Date and End Date in the correct format 9 For each record enter Y for the Locked flag field 10 For each record enter Y for the Position flag field 11
53. now been added to the bottom these records are the distinct selected ratings d Now what you need to check is that there aren t any additional selectedRatings if there are then make sure you add INVALID all columns e Then remove any rows that already exist above f You should now have an up to date X ref 5 In the tables section delete any old copies of OBCF_ONELILLY and OBJT ONELILLY 6 Inthe query section run the Make OBCF ONELILLY query by double clicking it 7 click yes at the two prompts you will see 8 You will now have an OBCF_ONELILLY table in the table section it is a good idea to go and have a look in here to make sure you are happy with it and do some brief checks 9 Once you are happy you can now export this table right click export saving the type as Excel 10 Now you need to run the Make OBJT_ONELILLY query do this by double clicking it 11 There should now be a table OBJT ONELILLY in the tables section check and export this table as before Please note Not all databases have exactly the same naming conventions for the tables query but they will be very similar so you will be able to work it out didn t have time to go and make them all consistent AMA vM 4 130 Section 4 Notes taken during expert evaluation of existing system Following through each step Jason encountered numerous usability issues as well as some parts of the system where it displayed a good example of usabili
54. of each of these can be found on the participant instruction sheet in Appendix D Section 1 6 2 6 Participants Johnson 42 discusses how it is the participants that provide a sensitive test of the research hypothesis Sensitivity is the ability to detect differences when they are present It is possible to increase sensitivity when the number of participants is increased Generally the minimum number of participants in a controlled experiment should be at least six As previously stated during requirements section 3 3 the OneLilly targeting application only has 2 main users and 3 additional users that could potentially be asked to use the application to support the main users Four of these users are available to complete the experiment but this is not an issue since Dix et al 1 points out that the sample size should be large enough to be considered to be representative of the user population therefore the number of participants in this case is not an issue and cannot be changed Finally it is important to note that only generalisations within this small population will be able to be made from the results since this is only a small scale pilot study The participants for this experiment shown in table 6 2 Role Relation to OneLilly targeting Experience Gender application U1 OneLilly operational Primary user Experience of all solutions F data steward U2 Pharma IT Business Primary support of the application
55. of this project it was a keen discussion area after the initial sketches Throughout this discussion it was noted that the criteria needing to be satisfied are requirements 22 Flexibility 27 Easy to navigate Interestingly all three users designed slightly different navigational structures figure 20 shows one of which was particularly intriguing since it did not follow any standards 53 Figure 20 Navigational structure designed by an inexperienced user Further discussions with the user saw that she had drawn how the task maps out in her head It is possible to see how this compares to the comments raised during step 1 of the observation section 3 5 1 where the verification of data is set out in a hierarchical structure with a top down approach This design structure was discussed amongst the group and it was decided that this would not be appropriate Firstly this is because the design considers how one single user views and understands the process but as the requirements chapter highlighted users have differing preferences on the flexible tasks Secondly the other two users commented on how this was not a standard navigational structure and so may not be clear that the items are clickable it also takes up additional space on the screen and may lead to confusion for some users The designer agreed with these comments but also stated how she would like to keep the idea of highlighting the page or step the user currently at Fi
56. of tus Log Out oked on each pu s Showd Wee Bes VW teg Grad Video File s Vse Qo 145 You are logged in as sam cooper Please upload the Business Objects report for sales representatives selected ratings Upload an Excel file Help Validation Errors Number of records imported successfully 27 Number of records with validation errors 1 Restart VERN USEFUL How Sud Word be bekor NC Mas AS USA ap dr SWA Mee issue oC Column Name Sereo d s vas gt buswmess 7 ERR ChE e Vio NS yess ob eX ve of wr WO ose 3 146 La M we You are logged in as Ped creme a osition Selecte sam cooper Objective Frequency Cym Dep UK CYMBALTA Depression Cymbalta Help To edit the Objectives click here Restart Saeg Proceed pace rasto bukon ag oun kems Ae sod Por tess Vid wa rae God be ve Strother 7 Jo A cs 7 Auf Jure Los Nea Karaee roker Men Mae Bl Whaat y MERE user 5 SR used to 5 b 147 You are logged in as sam cooper Business Plans BSPL Objectives OBJT May be Shen odes only Admin Call Frequency Objectives OBCF br 0 uk uw e Help Sales Quota Objectives OBSQ das Coker Wer Log Out Lilt Overview Import Export this to Excel You are logged in as M sam cooper Affiliate ecord Type S
57. oso so TABLE 12 QUESTIONNAIRE RESPONSES FOR EXISTING SYSTEM E TABLE 13 QUESTIONNAIRE RESPONSES FOR NEW SYSTEM asessscsessssssssessteseesseesesneeseeneeseeseesnessesneeseeneeseeaeeaneesesneeseeneeaeeaeeaneeseaneeseeneeas vi Acknowledgements I would like to thank my supervisor Dr Hilary Johnson who has been an invaluable source of support and advice over the past year I would also like to thank both Jason Cooper and Jon Bolt for contributing their human computer interaction knowledge to the evaluations I would also like to say a huge thank you to all of the UK OneLilly team at Eli Lilly amp Company for their endless patience and constant enthusiasm throughout the project vii Chapter 1 Introduction The motivation for this project originated from an industrial placement year within a CRM project team at Eli Lilly amp Company Lilly is a global pharmaceutical company their products treat cancer diabetes depression schizophrenia and many other conditions During the development of a global Customer Relationship Management CRM solution the design of the tools to support the user tasks were not considered beforehand As a consequence they were implemented in the quickest way possible without any regards for user needs One particular system written in Microsoft Access was very poorly designed leading to a vast number of usability problems The system failed to support the task in the same way as the initial so
58. particular error e Number of users completing a task successfully This will provide empirical evidence to support a particular claim or hypothesis It is important to manipulate the context by moving users out of their usual working environment in order to uncover problems or observe less used procedures Any experiment has the same form and the evaluator will choose a hypothesis these are often based on theory or previous research findings to test A number of experimental conditions are considered which differ only in the values of certain controlled variables There are a number of factors that are important to the overall reliability of the experiment which must be considered 20 during the design This includes participants chosen variables tested and manipulated and the hypothesis tested According to Preece at al 2 it is considered that 5 12 users is an acceptable number to test in a usability study although sometimes it is possible to use fewer where there are budget and schedule constraints For example quick feedback about a design idea can be obtained from 2 or 3 users An effective technique during usability testing is to invite users to think aloud about what they are doing This can yield clues to assist the evaluator for example user comments like This text is too small or I can t find this menu for Afterwards the participants can be invited to make general comments or suggestions this informal atmosphere often leads t
59. position column on the cross reference table was sort able e Having statistics on the export page to inform the user of how many of each data type they will be exporting and sending this was also raised during the requirements phase but was secondary priority 104 Some suggestions for enhancements e Currently one of the major issues in any system that is used is if the selected ratings spreadsheet is changed Within the old system each column and title had to be precise in order for it to load the system has a slight improvement such that columns can now be in any order but the column titles still need to be exact A suggested further improvement is to extent the implementation to directly query and obtain the data from the secondary system rather than running a report and loading e Extending the system to include National sales managers as well as sales representatives If these interview comments were compared to those raised during the requirements phase it is possible to see a change in user attitude For example the following quotes can be compared am extremely worried that if I do the slightest thing wrong I will have to start again requirements interviews 30 11 07 It is fantastic that I don t need to worry about errors moving through each stage Post evaluation interview 28 03 08 This example along with other quotes has shown a positive attitude towards the new system with comments focused on navigational guidan
60. raised and discussed during the user comments in the interaction of section 6 3 If these links were to become blue then it would also be a good idea to change the breadcrumb colour design and Jon commented that if possible the most sensible choice would be make it consistent with the corporate standards Button design and colour matching was also mentioned the current design has green buttons on the left hand side menu but Jon commented that firstly buttons are not usually green and secondly it would be more appropriate to use a standard design for this type of functionality this would be to use hyperlinks rather than buttons Although this comment in valid for consistency with standard systems it would not be comment taken forward for this particular system This is because the design decisions were made solely by users and therefore the interface is customised to their preferences The discussion then moved onto navigation Jon suggested that to make the fact that users are able to navigate back more explicit there could be an additional Back button next to the Save amp Proceed button on each screen He also pointed out that when you are within each separate screen for data the breadcrumbs are broken since the Data tab is not in bold thus potentially confusing users Another navigational suggestion that Jon made was in relation to a user accessing the help videos to reach them users need to navigate away from their current locatio
61. rows This is a concept that is not familiar with non technical users it should match the real world in terms of viewing selecting and deleting data through an interface The system also displays a number of message prompts which are meaningless to most users Figure 8 shows one particular example lS A Security Warning Unsafe expressions are not blocked Do you want to block unsafe expressions w J e Figure 8 An Example of a Message Prompt The help button provided no further assistance on which option should be chosen Nielson s heuristic states the system should use words and phrases with which users are familiar Adopting this principle will lead to a reduction in errors encountered Neilson s tenth heuristic states that whilst it may be better that a system can be used without online documentation if it must be provided it should be simple to search task focused and easily navigable Currently users need to follow written documentation precisely to ensure tasks are completed successfully However the documentation makes assumptions about what experience of Microsoft Access a user may have The documentation of the existing tool fails to provide a complete guide to the steps that must be taken to achieve a desired goal with large sections missing Finally an interesting conversation arose during the interview regarding system icons Access is standard software thus the main icons may be familiar to some use
62. severity of usability problems the quality of the severity rating increases rapidly and using the mean of a set of ratings from three evaluators is satisfactory for many practical purposes 2 6 2 Empirical Evaluation Empirical methods involves studies of actual users they can be relatively informal for example observing people while they explore a prototype or they can be formal and systematic such as in a controlled laboratory to study performance times and errors Regardless of the care with which the data is collected the interpretation of the empirical results depends on having a good understanding of the system being evaluated 3 Preece et al 2 further split empirical evaluation out into two different approaches usability testing and field studies Usability testing is conducted in a laboratory or laboratory like conditions isolating the user from normal day to day interruptions and noise Taking users out of their working environment makes the commitment to usability clear to users 4 This method is important particularly at the later stages of design Preece at al 2 states usability testing involves measuring typical user performance on typical tasks and will include things like e Time to complete a task e Time to complete a task after a specified time away from the product e Number and type of errors per task e Number of errors per unit of time e Numbers of navigations to online help or manuals e Number of users making a
63. should do or how it should perform It is important to make each requirement as specific and unambiguous as possible The overall purpose of collecting data in this phase is to collect sufficient relevant and appropriate data to enable such requirements to be formed Data gathering is used to find out about the tasks users currently perform and their associated goals the context in which the tasks are performed and the rationale for the current situation The requirements will be formed from a number of different sources as shown in Figure 5 Understanding the underlying process 3 2 amp Evaluating exisiting 3 3 solution 3 4 Literature Participatory review chp 2 design chp 4 Requirements Figure 5 Formation of The Requirements Document 23 As stated in the previous chapter there is a strong focus on the necessity of involving stakeholders and the detailed process of participatory design achieves this objective It is also useful to gain an understanding of the purpose of the system in question the knowledge gained here will be applied throughout the project 3 2 Overview of the targeting process It is useful at this point to introduce to the targeting process and how the current system plays a vital role in the preparation of data This overview is based on personal knowledge gained whilst on placement at Lilly The targeting process generates data for loading into a global CRM system named OneLilly It
64. shown in figure 26 Both designs use a drop down list allowing the user to select the objective Two of the users agreed that both of these should be implemented since this would fix those issues described in the interview and observation of incorrect values being entered The third user discussed the potential difficulties of having set objectives the possibility of only an administrator to edit the contents of the drop down box This is obviously a company process decision whereby the current roles and responsibilities are being questioned and this decision will affect the system design This therefore leads to the following options for the objective field of the cross reference design 58 01 Drop down box to select objective contents only editable by an administrator 02 Drop down box to select objectives contents editable by any user 03 Free text entry The final option was immediately dismissed since this will cause the same errors currently being experienced with the existing system Therefore whichever decision it has been decided that a drop down box will be used GENERATING DATA The existing solution required users to manually run separate queries to generate each data type as already discussed the new system will prevent users from having direct access to such queries therefore this needs to be considered for the UI design None of the user sketches made any reference to how this could be done further discussions showed users had
65. shows the data required for one sales representative targeting 100 customers with two products One business plan for every sales representative There is an objective every month for sales 3forSQ quota and time management whereas for call 3for TM frequency there is only one objective for the 2 for CF whole 3 months but this is set for every product 8 100 customers x 2 products 3 months x 2 1 for each products 6 month 3 200 Figure 7 An Example of the Data required for one sales representative As shown 218 records are required for one sales representative This gives an impression of how time consuming completing this task manually would be and goes a way to demonstrating how important it is to have an effective support system in place The targeting solution has recently undergone a redevelopment during the new release of a CRM system Prior to the current Microsoft Access solution users completed the task manually albeit with use of a spreadsheet Previous to the introduction of this system users would take up to six weeks to compile targeting data the current solution reduced this significantly but as previously described the solution was not developed with any regards to users meaning a lot of their understanding was lost making the tool difficult to use The requirements gathering process will begin with interviews with stakeholders 3 3 Interview with stakeholders Beginning the requirements gatherin
66. simply redirects the user to the cross reference page 5 4 3 Cross Reference This cross reference table needs to be generated based on the data just added It needs to contain unique values of the SelectedRating and position letter fields from the EISelectedRating table and automatically place these into the cross reference table these values are non editable to the user Additionally there are two columns which will need to be updated using a drop down box by the user Objectives with a related Product Indication and Parent Objective and a Call Frequency An example generated table is shown in figure 37 80 Lilt Qverview Import X Ref gt Data Export J Cross Reference You are logged in as E EE x IL osition Selecte Ue a Cym Dep UK CYMBALTA Depression Cymbalta 1 3 Admin Cym Dep Depression Cymbalta 1 gt e IN Cym Dep UK CYMBALTA Depression Cymbalta 1 3 No Cym Dep UK CYMBALTA Depression Cymbalta 1 3 Help To edit the Objectives click here Save amp Proceed Restart Figure 37 Print Screen of an example cross reference table This table needs to be generated when the page loads The first step is to access the database to get the data required The following code shows this TargetingDataContext db new TargetingDataContext unique position amp letter items needed in x ref var ratingList from sr in d
67. tailor frequent actions 8 Aesthetic and minimalist design Dialogues should not contain information which is irrelevant or rarely needed Every extra unit of information in a dialogue competes with the relevant units of information and diminishes their relative visibility 9 Help users recognize diagnose and recover from errors Error messages should be expressed in plain language no codes precisely indicate the problem and constructively suggest a solution 10 Help and documentation Even though it is better if the system can be used without documentation it may be necessary to provide help and documentation Any such information should be easy to search focused on the user s task list concrete steps to be carried out and not be too large Similarly Shneiderman has developed eight golden rules 1 Consistency in action sequences layout terminology command use and so on 2 Enable frequent users to use shortcuts to perform regular familiar actions more quickly Offer information feedback for every user action 4 Design dialogs to yield closure so that the user knows when they have completed a task 5 Offer error prevention and simple error handling so that users are prevented from making mistakes and if they do are offered clear and informative instructions to enable them to recover 6 Permit easy reversal of actions 7 Support internal locus of control so the user is in control of the system which responds to his
68. that they are in an appropriate order to successfully support the user Figure 29 shows the cards that were designed for this session Note that cards can represent screens user s goals or their intentions User Mental Operation Decide which data types are required Select Items Select Quarter 1 L 1 Select Data Types to generate L OBSQ Select Year BSPL OBJT L OBCF Upload Excel File for OBCF Je Upload Excel File for OBSQ Select Objective e Em assess Add objective User Mental Operation Decide which objective each selected rating should have User Mental Operation Decide which data type to view 61 View Chosen Data Type User Mental Operation Decide if editing is required ma _ 4 Edit An Entry View Resultant File SERE E 2 24 You are about to BSPL Example view OBJV dat Text This is8 Export Resultant File In the correct OneLilly format View File Export File lt Send To OneLilly Figure 29 Cards used for CARD session Note Although itis not shown on this insert cards were numbered 1 11 working left to right Two users participated in this activity but due to travel requirements this CARD session had to be conducted electronically Each user was sent a copy of the cards along with a small subset of example data for one sales representative The foll
69. the functionality Participatory design is about establishing meaningful and productive interactions among those users who are directly impacted by the change in technology Participatory design represents a traditional break with traditional approaches to system development Computer applications need to be better suited to the skills and working practices of the people using the system Barriers between designers and users need to be broken down in order to build effective communication throughout the design process Bodker 30 states that users and designers have different backgrounds and belong to different communities of practice The practice of users is the starting point for design but at the same time users need to see and experience new ideas in order to transcend their own practice Participatory design stresses the importance of active participation of end users throughout the project from initial research through to design and evaluation they should take part in decisions that affect the system and the way in which it is designed and used Because technology is not developed in isolation participation in decisions about technology also involves decisions about work content and job design 7 Bridging the gap between designers and users is not an easy task as Kyne 22 discusses answering questionnaires and discussing requirements specifications are not suitable since users don t have the professional knowledge and skills that form th
70. this is by taking a user centred approach to development this means the users concerns direct the development rather than technical concerns To understand users it is necessary to understand processes capabilities and predictions that users might bring to the tasks they perform This will involve an understanding and knowledge of cognition The interface is the intermediate component between the user and the computer system it reflects the system model to its users and translates their intentions into appropriate system activity A user develops an understanding of how the system works and forms a model known as the mental model This model is developed during the interaction and will then form the basis of future interactions with the system the more someone learns about a system and how it functions the more their mental model develops The term cognitive model is proffered since itimplies cognition which is more representative of the process involved 15 2 2 Whatis cognition The term cognition refers to a facility for the human like processing of information it involves cognitive processes like thinking learning remembering and decision making Norman distinguishes between two general modes Experimental is a state of mind of which we perceive act or react to events around us effectively for example driving a car Reflective involves thinking comparing and decision making this is what leads onto new ideas and creativity
71. this reason two lists are required one which will store those records to be imported and the other to store records with errors ijst EISelectedRating okToImport new List EISelectedRating List lt KeyValuePair lt int EISelectedRating gt gt problems new List lt KeyValuePair lt int EISelectedRating gt gt EISelectedRating is the object which models the SelectedRating table within the database The list of problems uses a KeyValuePair to enable it to store the row number of where the validation failed providing the user with more information The list containing records that passed validation checks these now need to be added to the SelectedRating table via LINQ As discussed in section 5 1 2 LINQ allows the database to be modeled using a class this class is named ExcelImportersDataContext and data is added using the following code ExcelImportersDataContext db new ExcelImportersDataContext foreach EISelectedRating r in okToImport db EISelectedRatings InsertOnSubmit db SubmitChanges results NumberOfRatingsSuccessfullyImported okToImport Count return results EISelectedRating is the name of the table that data will be loaded into and okToImport is the list containing valid records to be inserted The code is firstly creating an instance of the database model and then looping through each item in okToImport and adding its values to the table Changes are then comm
72. this was then agreed by a selection of the target population 30 The system shall be capable of allowing a user to specify a call frequency for each selected rating Rationale The call frequency is decided by business partners based on the selected rating therefore this cannot be pre set meaning the user is required to input this value Source A user discovered this was missing when working through a scenario during the CARD session section 4 4 31 The system shall allow a user to terminate the transaction at any time Rationale If there are import errors then the user may wish to restart alternatively the user may wish to restart or terminate if new or different data set has been provided by business partners Source User designs section 4 3 within the task sequencing section Additional user requirements 32 The system shall allow a user to return and edit the cross reference during a single transaction Rationale Required to support the scenario of a user making a mistake or a business decision changing e g a call frequency of 2 is selected instead of 3 This requirement will prevent the need for the task to be restarted and so increases productivity Source Nielson s 9t heuristic of allowing users to easily recover from errors This was highlighted during user experimentations with higher fidelity prototypes section 4 6 it was seen that there is sometimes a need to make changes to the cross references This chapter has seen h
73. times shown within this results set This was also complemented with user comments in the follow up interviews one participant commented since it is not possible to edit a query it means the generated data will always follow the standard process If I were to use this system over a greater period of time I would become familiar with those fields I no longer need to check e g date format Looking at the standard deviations it is possible to see that they are much larger for the existing system This would have been assumed and expected due to the fact that the timings will differ significantly between the experienced and novice users Experienced users have found workarounds and become familiar with tasks that would have taken novice users longer to complete Number of errors made The aim of this measurement is to understand where and what type of common errors users experienced between the two systems Any action that causes the user to remove something e g delete an incorrectly inputted value or re iterate over a particular task e g re run a query will be noted as an error The number and type of errors made throughout all experiments were noted these are shown in table 8 and the number of errors made are summarised within tables 9 and 10 Within Existing Within New U1 T2 Accidentally let Access chose a primary T3 Didn t realise you need to select from T3 Used a keyboard shortcut to run but clicked drop down bo
74. to continue to consider new users and therefore it is advisable that screens should be simple it is easy to overload a new user with too much information as this can be frustrating 36 When presenting information it is also important to consider colour when used incorrectly it can seriously hamper communication Macfarland 37 recommends perceptual areas such as using a soft non intrusive background colour such as soft greys to prevent eye fatigue A consistent colour scheme should be used throughout and appropriate colours for each function for example red for stop It is also worth noting that colour is not completely reliable considering that 1096 of males show some indication of colour blindness 36 Preece et al 2 discusses the structure of information and show good examples where information has been ordered into meaningful categories which appropriate blank spacing between enabling easier reading for the user There is an advantage when presenting information in an interactive system in that it is easy to allow the user to choose among several representations so making it possible to achieve different goals Now that a good background understanding of the design principals and how these can be interpreted to consider design options has been achieved the next step is to consider the process by which such designs can be implemented 2 5 User centred design Now that I background research into HCI understood the heuristics and cons
75. to update the cross reference table and subsequently the data set 7 lfound it easy to edit the data produced The questionnaire scores can be seen in tables 12 and 13 Responses Table 12 Questionnaire responses for existing system 101 Responses Table 13 Questionnaire responses for new system Figure 45 is a bar chart displaying the mean scores for both systems Mean Questionnaire Responses i B pg B BN _ p B Existing New Question Number Figure 45 Bar chart displaying likert scale scores Figure 45 clearly shows a change in user preference with responses to questions 5 and 7 all at strongly agree An important fact to note is that even the most experienced user U1 still rated the new system higher than the existing Having analysed the discussions raised it is possible to categorise them into two sections Interaction and context Interaction Every participant discussed how to navigate around the new system and progress towards task completion Comments such as I like knowing how far got through the task and the bar across the top made locating the cross reference so much easier when I needed to make a change were raised 57 These provide good supporting evidence that the impact of navigational redesign through the use of breadcrumbs was successful Although whilst referencing t
76. tool section 3 4 25 The system shall provide a user with information about the current system state Rationale Keeping the updated on the current system state will prevent them from attempting to complete another task and potentially causing an error Source Heuristic evaluation of existing tool section 3 4 26 The system shall prevent errors where possible Rationale As the observation shows when a user encounters an error it confuses them further and could potentially lead to them giving up on the task Avoiding errors increases task efficiency Source It was seen that errors occurred throughout the observation This is also a guideline specified from Nielson literature review 2 3 27 The system shall be easy to navigate and links should be unambiguous Rationale The existing system contains no guidance through the task allowing a user to navigate anywhere completing any action resulted in an error The aim of this requirement is to allow the user to understand where they are and the next steps they can pursue Source HCI expert evaluation section 3 4 highlighted the lack of clearly defined start and end points having a negative effect on the navigational usability This was also discussed in the literature review section 2 4 28 The system should follow standard design principals as stated in the literature review Rationale To avoid user frustration and improve user experiences Source Literature review This chapter has
77. using Storyboarding section 4 4 e Envisioned HTA section 4 5 e Iterative higher fidelity prototyping section 4 5 The chapter will then conclude with a summary of the design rationales Each stage throughout the design will require small meetings of a selected number of stakeholders from OneLilly team at Eli Lilly It can be seen that the design is very much an iterative process meaning that implementation will begin as part of the design process to allow higher fidelity prototyping to take place As stated in the literature review it is often said that users are unable to know exactly what they want until they are able to see something iterations therefore occur to allow them to understand what is possible and explore different designs Therefore it should be noted that the requirements specification is still open to changes 4 1 Design Rationale In designing any system decisions are made as the product from a set of possible vague customer requirements to a deliverable entity Often it can be difficult to recreate the reasons or rationale behind each design decision Design rationale is the information about the decisions 51 made during a design process and the reasons why those decisions were made Benefits of completing a design rationale include e Encourages range of possible solutions to be explored rather than a narrow exploration of a single solution e The effort required forces the designer to carefully consider des
78. very unsatisfactory and misguided for the users of the application resulting in system failure or incorrect and un validated data being produced Context In terms of functionality the users like the current tool The following points were highlighted as being particularly important e The checking of input files to ensure data is valid important since the global OneLilly application allows the sales representative to enter any value e Updating the cross reference table automatically with any necessary additional values is useful e Removing invalid data from the output file automatically although note the point about feedback made in the above section e The ability to edit the generated data there will inevitably be some anomalies that require adjustment e The output file produced follows the OneLilly formal input specification This specification can be found in Appendix A Section 2 28 However there are pieces of functionality that confuse users and have the potential for improvement The cross reference table easily confuses users This is because old data values remain within the table and new values are placed below this rather than ignoring if the record is already present this problem must then be resolved manually Old data should be removed before the tool can be used with new data This has the following implications e corollary of the above problem is that it is not possible to see data from prev
79. woned out unless ener ess Cowie enver otc F andlor over Cem Grameen Pes See 9 E B Af wicked evs C o0S grote omy MI Qs eae Sele Quacker Um 6 eb out More ARRON _ aout guer eae wea CUR Sg Figure 19 Example user sketch Once initial sketches were complete all users were brought together and a telephone meeting took place to compare and contrast some of these initial ideas The interface design dicussions fell into three main areas Navigational structure task sequencing and task support It was decided that these three areas could shape the high level questions needing to be answered as part of the QOC notation Navigation Q1 How should the user navigate through the system Task Sequencing Q2 How should the user progress through each sub task Q3 How should the user help be provided Q4 How should errors be prevented Task Support Q5 How should data be added to the system Q6 How should mappings be made between ratings and objectives Q7 How should data be generated Q8 How should this data be validated Q9 How should the resultant file be exported Through analysis and discussions of these sketches it is now possible to have an initial attempt at answering these questions a summary of this and the conclusions made will now be given 4 2 1 Navigational structure Obviously with this being the most significant aspect
80. 4 it is possible to see a significant change in user opinion Finally another common discussed aspect was the change in user confidence this has been increased due to a number of reasons e Verification during import e Since there is no technical knowledge required participants commented I don t need to worry about whether I made a mistake when editing the query e The validation that is completed whilst editing produced data The new system restricts some functionality to administration users only there were mixed comments about this since it was stated that the lack of flexibility affects the usability since they are unable to change the objective details themselves The main experienced user stated that this was required whereas supporting users stated that this could waste time if the main user was away and data needed to be completed urgently The main discussion point was adding additional objectives to the drop down box but as the main user pointed out this is not going to be a frequent occurrence and if this was the case they would be notified well in advance This discussion is focused around a change in the roles and responsibilities of the job rather than the software It is useful that the software allowed this discussion to go ahead assisting the company to make business process decisions A further enhancement to the administration functionality was also suggested Adding an additional role rather than just admin and user to
81. 6 CURRENT PAGING STRUCTURE icut ta cactetac acude de onu dn dao du Cani a cac da sa da cR d eT abc cc bn FIGURE 47 REQUESTED GLOBAL SYSTEM PAGING STRUCTURE FIGURE 48 STANDARD BREADCRUMB DESIGN sesssssssssssssessssussessssusesesesusssssessssucsesassucsusassueansaseucsnsasescsnsaseacsnseseaesnssseacsnsesesesnssesseenseoess FIGURE 49 BREADCRUMB DESIGN cem ccc d rera Saa Ba Gd DN a d Gi RD Ea e E List of Tables TABLE 1 USER CENTRED DESIGN METHODS aS SSS SSSA TABLE 2 COMMON ERRORS AND POSSIBLE DESIGN SOLUTIONS TABLE EXPERIMENT VARIABLES scsssssssssssssssssesscsscsnsscsscsnsessacsnsssesesusssssesussscsesnsssssesassscsusassucausaseseansasesesneaseaesusaseacensesesesusseseenssesseeaseoess TABLE 4 BACKGROUND OF PARTICIPANTS io TABLES PARTICIPANT ALLOCATION TABLE 6 TIME TAKEN USING EXISTING SYSTEM TABLE 7 TIME TAKEN USING NEW SYSTEM uscsssssssssssssssssnssssscsnssessssnssescsusssssesnssscsusassscsusassucansaseseansaseseansasesesnseseacensesesesnessesesssessseeassoess MOREM BOO EDI OUBLIER TABLE 9 NUMBER OF ERRORS MADE USING EXISTING SYSTEM TABLE 10 NUMBER OF ERRORS MADE USING NEW SYSTEM ausssssssssssessessssessesussecsesassucsusassucsusacsucanseseuesusaseaesnsasesesusssesesussseseenssessesneseess TABLE 11 COMPARING THE NUMBER OF CLICKS REQUIRED entere tette tete toten t ta tate tita aso soto sesso
82. 70 5 1 AN INTRODUCTION TO THE NET FRAMEWORRK ee rennen rtnntn retenta tetas tata satt atta tates 70 Dd dee eee eie debebat aciei ead tite eiut L2 LO uo e aW ac e RU E n Ee o ae 5 2 DATABASE STRUCTURE 5 3 SYSTEM ARCHITECTURE AALS 5 4 SYSTEM DEVELOPMENT ste cuvette 5 41 Navigational Structure Master Page 5 42 Importing Data 5 43 Cross Reference 6 1 EVALUATION ACTIVITIES setae nets etn tne cd 6 2 EXPERIMENTAL EVALUATION 62 1 Hypothesis 6 2 2 Measurements sternere ottico 6 2 3 Developing predictions and choosing a means to test this tnn 89 6 2 4 Identifying variables see ttt e t tet EASES aa 6 2 5 Experimental task and method 6 2 6 Participants hacia dik 6 2 7 Experimental design and data collection method 6 28 Experimentalres lts oderit nr Hen oi ER n Hi rir eR 6 3 SUBJECTIVE USER EVALUATION 6 4 EXPERT EVALUATION MT 6 5 GONGEUSION E CONCLUSIONS asks scars aa 108 7 1 PROJECT OVERVIEW ASA SSA SSA SEAR RA SS SSS 108 7 2 PROJECT OUTPUTS 72 1 Theoretical n 7Z22 Empirical s zx 72 38 Methodological understanding ttt ttt ttt tnt ea 111 A e O O OLE KS a bas e te e e a en e tiet me ba 7 3 FUTURE WORK edn 7 3 1 System Further Enhancements
83. A good question might be How should menus be accessed therefore only dealing with one of the issues within the former question 38 To assist in ensuring that design decisions satisfy every requirements stated in the previous chapter it has been decided that this chapter will use the QOC notation to keep track of design rationales 4 2 Low fidelity prototyping A prototype does not have to be a complex piece of software it can be as simple as paper based outline of a screen an electronic picture or a video simulation of the task It allows stakeholders to interact with an envisioned product to gain some experience of using it allowing them to easily contribute to the next iterative design Having collected information what the users do and not like about the existing solution as well as observing the task being completed within chapter 3 it is now possible to try out ideas by 52 iteratively building prototypes This will begin by asking a selection of users were asked to independently sketch their interpretation of what they believe the new system should look like An example of such a sketch can be found in figure 19 all sketches can be found in appendix B Section 1 SORTED eire Ean te e ES xa Comer hd geen Linen Kaur Stage iS MOLT D N N EN ece x dara ore tae ETE Aerecaxe d excet Ale mF anere wht 04 ZU 11 8 Bars eecc ge
84. Examining the Effect of Navigational Redesign on End User Experiences Samantha Cooper Bachelor of Science in Computer Science with Honours The University of Bath May 2008 This dissertation may be made available for consultation within the University Library and may be photocopied or lent to other libraries for the purposes of consultation Signed II Examining the Effect of Navigational Redesign on End User Experiences Submitted by Samantha Cooper COPYRIGHT Attention is drawn to the fact that copyright of this dissertation rests with its author The Intellectual Property Rights of the products produced as part of the project belong to the University of Bath see http www bath ac uk ordinances intelprop This copy of the dissertation has been supplied on condition that anyone who consults it is understood to recognise that its copyright rests with its author and that no quotation from the dissertation and no information derived from it may be published without the prior written consent of the author Declaration This dissertation is submitted to the University of Bath in accordance with the requirements of the degree of Bachelor of Science in the Department of Computer Science No portion of the work in this dissertation has been submitted in support of an application for any other degree or qualification of this or any other university or institution of learning Except where specifically acknowledged it is the work of t
85. FileToDatabaseConnection filename ConfigurationManager ConnectionStrings LillyTargetingConnectionString ConnectionString lblCallFrequencyNumberOfValidationErrors Text results RatingsWithValidationErrors Count ToString lblCallFrequencyNumberImportedSuccessfully Text results NumberOfRatingsSuccessfullyImported ToString foreach KeyValuePair lt int EISelectedRating rowKV results RatingsWithValidationErrors int rn rowKV Key EISelectedRating rowKV Value TableRow row new TableRow TableCell celll new TableCell 111 rn ToString TableCell 112 new TableCell cell2 Text r Division ToString TableCell cell3 new TableCell cell3 Text r Position ToString TableCell cell4 new TableCell cell4 Text r ConIntegrationID ToString TableCell cell5 new TableCell cell5 Text r SelectedRating ToString 160 row Cells Add celll row Cells Add cell2 row Cells Add cell3 row Cells Add cell4 row Cells Add cell5 tblValidationErrors Rows Add row if results RatingsWithValidationErrors Count 0 lblValidationErrors Visible true tblValidationErrors Visible true else tblValidationErrors Visible false lblValidationErrors Visible false panelCallFrequencyResults Visible true cmdUploadCallFrequencyFile Enabled false fulCallFrequency Enabled false else lblFileError Text You must b
86. GNED BY AN INEXPERIENCED USER FIGURE 21 NAVIGATIONAL STRUCTURE DESIGNED BY A USER WITH MINIMAL EXPERIENCE eren rennen terere tenen tn ttnnun 54 FIGURE 22 NAVIGATIONAL STRUCTURE DESIGNED BY AN EXPERIENCED USER recen retener trente tnter teta tet teta 54 FIGURE 23 USER SKETCH SHOWING PROCEED BUTTON recente tnter ntn ttnt terres tatto stata ttai sh sosta 55 FIGURE 24 USER SKETCHES FOR UPLOAD FUNCTIONALITY ccsssssssssssscsessssscsecsssscsscsnsecsscsnsessscsnsesescsnsesesesusssesesassassesassacansassacaeassaeanes 57 FIGURE 25 CROSS REFERENCE DESIGN 1 FIGURE 26 CROSS REFERENCE DESIGN 2 FIGURE 27 ep ata oA FIGURE 28 EXPORT DESIGNS FIGURE 29 CARDS USED FOR CARD SESSION eres Saas 62 FIGURE 30 ENVISIONS HTA etta eite cr ee dre eR nde ue Cede e nere a 65 FIGURE 31 OVERVIEW OF THE NET COMMON LANGUAGE RUNTIME 71 FIGURE 32 DATABASE DIAGRAM a ve BERE CEN aae Bee A Gene RE RR END Re Enc fa e Bre t EC e DEED E Peces 73 FIGURE 33 OVERVIEW OF SYSTEM IMPLEMENTATION sssssssssssscsecsesscsecsesscsscsssacsucsnsacsscsnseseacsnsesescsnsesesesnsssesesnssassnsassatansassataneacsseanes 74 FIGURE 34 HIGH LEVEL CLASS STRUCTURE AAAS 76 FIGURE 35 PRINT SCREEN OF THE IMP
87. Good experience of M Integrator Needs to be fully aware of its daily existing solution use U3 OneLilly business data Secondary user Minimal experience ofthe F steward existing solution U4 RADS Technical Supports the application from a Minimal experience Only M support technical perspective e g where it limited understanding of is hosted amp code developments daily use of the system Table 4 Background of participants Regional Applications Development and Support The participants cover a range of roles and experiences it is a very good representation of the users OneLilly targeting application currently has It is possible to divide the four participants into two types Experienced and Novice This is important to be aware of when analysing the experiment results 91 6 2 7 Experimental design and data collection method The next step is to decide the experimental method to be used There are a number of different experimental design methods that could be used these include within between and mixed groups designs which are all one off experiments There are also longitudinal and cross sectional studies which span over a greater period of time but due to time constraints the latter two will not be considered although longer periods of study would give better learning rate data The two main design methods to consider are between groups and within groups it is important to consider a number of facto
88. ISISIS YYYYMMDD HHMISS Section 3 User Instructions for the Existing solution How to re use a targeting tool databases can be found in the following location Ybsvr10 uk onelilly Work In Progress Samantha Cooper Q3 2007 Objectives 1 Data steward to extract data 2 Open required database and delete all old data a b C Go the query section and select new Click ok where it says design view then click close when it prompts you to add a table On the top left hand side click SQL and then type the following into the box DELETE SelectedRatings FROM SelectedRatings d e Click run the red explanation mark on the top menu Click yes when it asks you if you are sure 3 Import your new data a File gt Get external data gt Import b Locate and open your new file c Follow through the wizard making sure you insert into existing table ilc SelectedRatings There is no need for a primary key WH v first tory onte 4 You now need to make sure you are happy with the X ref CC UMP noc a Bearing in mind that the MySelectedRatings field is free text reps are therefore able to select anything they want So you need to make sure your X ref is updated to check which reps have selected invalid Therefore the best way to do this is re run the Build Rating2Objective query b You will see another prompt just click yes c Now open the table Rating2Objective You will see that additional records have
89. ListItem String Format 0 1 2 Product Trim m Indication Trim entry pk ToString Trim list Items Add 1 The last column in the table contains the drop down list for the call frequency this is done in the same way as above firstly by creating the list item items needed are the numbers 1 4 and then adding each item to the drop down control At this stage it is important to note that the additional requirement found during high fidelity prototyping section 4 5 has been met this was to allow the user to return and update the cross reference table during the same transaction This was met by adding in an extra step which checks whether a mapping has already been saved and if so updates each variable CrossReference current null try current db CrossReferences SingleOrDefault p gt p PositionLetter posLetter amp amp p UniqueSelectedRating selRating catch Exception current null If there is not already a mapping specified then the current variable is set to null This is then used when updating the objectives drop down list by checking whether current is not null and if so selects the already chosen value The other factor to note is that when the table is saved for a second time previous data will need to be removed before new is added this is done as follows bool crossReferenceAlreadyHasData db CrossReferences Count gt 0 if crossReferenceAlreadyHasData
90. OBSQ required Boolean Session OBSQ required to create OBCF foreach SelectedRating a in ql OBCF obcf new OBCF obcf Affiliate Code UK obcf Record OBCE obcf Source System Id MANUAL obcf Operation I obcf Operation Date DateTime Now ToString yyyymmdd hhmmss obcf Contact ID a ConIntegrationID obcf Parent Objective Name quarter year a Position a CrossReference ObjectivesMapping ObjectiveParent obcf My Selected Rating a CrossReference ObjectivesMapping Objective obcf My Call Frequency a CrossReference CallFrequency obcf Product Id a CrossReference ObjectivesMapping Product Trim obcf Indication a CrossReference ObjectivesMapping Indication Trim db OBCFs InsertOnSubmit obcf db SubmitChanges distinct values of parent objt from obsq and obcf List string q3 from obcf in db OBCFs select obcf Parent Objective Name Union from obsq in db OBSQs select obsq ParentObjectiveName Distinct ToList 163 distinct values of just obcf List string q4 from obcf in db OBCFs select obcf Parent Objective Name Distinct ToList to create OBJT if selected to do so on overview page different query needed depending on whether OBSQ is included var queryNeeded q4 if OBSQ required queryNeeded q3 if OBJT required foreach string parent objt in queryNeeded needed to work out whether th
91. ORT SCREEN eerte treten tenen nnne ttti tetto tatto sosta Dose so 77 FIGURE 36 PRINT SCREEN OF THE IMPORT SCREEN AFTER FILE UPLOAD rennen rrnnnn rn rtnnnn tn 80 FIGURE 37 PRINT SCREEN OF AN EXAMPLE CROSS REFERENCE TABLE sssssssssssscsssssssescsnsscsscsnssssscsnsssesesnsssesesnssassnsassacansassatacsacsseanes 81 FIGURE 38 PRINT SCREEN OF VIEWING AND EDITING DATA ussssssssessssscsecsssscsscsnsacsucsnsecescsnssssscsnsesescsnsesesesnsssesesassessesessasansassaransassatanes 83 FIGURE 39 AN EXAMPLE OF USER FEEDBACK DURING IMPLEMENTATION scssssssssssssescsssessssnsscssesnsseesesusssesesassesscsassacansassarassaeeseanes 84 FIGURE 40 BAR CHART DISPLAYING MEAN AVERAGE TIME sssssssssssecsssscsessssscscsssecsscsnseesscsnsessscsnsesesesnsssesesassasscsassaeansassacassaeeaeanes 94 FIGURE 41 BAR CHART DISPLAYING DIFFERENCES FOR EXPERIENCED AND NOVICE USERS EXISTING SYSTEM 95 FIGURE 42 BAR CHART DISPLAYING DIFFERENCES FOR EXPERIENCED AND NOVICE USERS NEW 95 FIGURE 43 BAR CHART DISPLAYING THE MEAN NUMBER OF ERRORS FIGURE 44 BAR CHART DISPLAYING THE MEAN NUMBER OF CLICKS FIGURE 45 BAR CHART DISPLAYING LIKERT SCALE SCORES recenter entren terne teni FIGURE 4
92. Options and Criteria notation to record all design decisions that were made Chapter 5 Implementation described the technical details of how the findings from previous chapters were used to develop the new OneLilly targeting application It began by giving an introduction to the NET framework and discussing why ASP NET was chosen as the development environment It then moved onto an exploration of a recently added extension to the framework language integrated query LINQ which proved to be particularly useful for this implementation The chapter also described the database set up class structures and provide detailed descriptions of how the main features were implemented Chapter 6 Evaluation focused on three key areas Empirical involving a small scale study subjective to gather user opinion and analytical where a HCI expert was asked to perform discount usability evaluation The chapter concluded by relating back to the hypothesis to discuss whether the new application developed for OneLilly targeting is superior to the existing system A summary of the project outputs will now be given 7 2 Project Outputs Throughout the project document the question of whether navigational redesign impacts end user experiences has been considered this investigation was the primary research aim of the project Through stakeholder s choice it has been possible to implement and therefore investigate breadcrumbs as a solution Participatory design was ut
93. SE PEON MONO Zp 1uensai uodx3 z s uejd peuinbeu jeox3 paS g 10j es olg Ly ode 3290 adh eyep i je x 6 vodwy ip eiedaug E 5 z sipe qo sseuisnq waua Apex saye 0 eu juejnse MBIA L Z sseursng 9 Figure 30 Envisions HTA 65 Comparing this HTA to those completed in chapter 3 the first thing to notice is the significant reduction in the number of steps required it is now possible to represent all sub task within the diagram and secondly it is possible to see the increased amount of flexibility A copy of this HTA was emailed to a selection of users to comment on but only positive feedback was returned since all users agreed with the task breakdown So far stakeholders have been heavily involved to complete two and in some cases three iterations though low fidelity prototyping Every design question has had at least two options provided and a decision made based on learning s from previous chapters and user expressed preferences The next part of the design process is to begin implementing higher fidelity prototypes Such prototypes are more like the final system and will allow the design decisions made so far throughout this chapter to be tested 4 5 High fidelity prototyping The purpose of this project is to develop a fully functional prototype so that it can be successfu
94. Scriees ve Aunt onu Seren 199 on Cound Xa cepo o eee 5 terore ere ca Oro 2o see go S o Gasecinc Cre Bef QoS vons AAW SEAS ASSAY goes Ex Check CM ON cA Ro dx needs No XA Ue wan QLANS bo dc We om UWS d Uke be uster ou d Cate badegqound Fnows Age not VY Aon tn vent a ade angon ve use tanq Ey Aane povided Cyr oarnra 5 conwy gt XX a Qo c5 uan coe eX AS Age Qa vied Q coo of SC ex Vw cS Moetu len Gbk esua hod e o gonm AS Found Rowe CANSA to Go bauk radiu te Stork sgan O darge ead Wo Nw ro car SANA EACH Seg Lo ec ceo Sye 4 doka CS benng Seni auio Seco Le cox en CrorackorS eke lt lt gla wsoy A WOK Loaded and whan Aus teans Coe eges OS Loqged or AVZE and COO Cu Xe Va Oc 26 exe Xx Ore groaned Maght not ed xf Stand Abate Te eo Xe 6 o COD ot Mes Sone nage gy vga ouar Fe 124 Interview Transcript D Date Zo Jo Firs hard Interviewee name and team c ess Jk ey KOM _ Suppor For yoo 1 What is good about the current system there any features you
95. Set Status to Planned for all records 12 For each position list the Position Division sales force 13 For each record generate the position name This corresponds to each sales representative data is taken from the initial spreadsheet 14 Complete the Local Start Date and Local End Date 15 Complete a manual check to ensure values are correct 16 Verify data using the provided macros 17 Export using the provided macro 18 Verify final load file Following this initial ordering the other users re arranged the cards to represent their routes through the task The steps taken by the other two users were as follows User Two 1 2 4 5 7 11 6 13 12 10 8 9 3 14 15 16 17 User Three 3 10 8 9 7 13 1 2 4 5 6 11 12 14 15 16 17 The discussion of the reasoning behind these choices highlighted the following points e The independent steps are 3 7 8 9 10 and 13 e Steps 14 17 are completed at the end of the process by all users e Steps 1 and 2 must be completed before any other step that is not independent since they provide the basis for completing other fields During the discussion the users stated that although they had given their preferred ordering of tasks they do vary them Initial solutions reflected much flexibility but this was due to the amount of manual data entry required When this was replaced with automation within the Access application those steps were removed meaning the flexibility had bee
96. Stakus Ure HOSEL fo 122 A3 CIO nO DU jaar quedo We Se 9 Ces AS Faciticanre erocass We x shoud be VE resanse Synem Feo S tle erecers a sate pte SET Se Pe incon eo Beery dS ead h AS a Suc Sree AQIS ae 4 What is bad about the current tool What frustrates you about it What features would you like to see removed _ Very AKER Atk BRO vv ve ovv C eso Xo Fo Ab oe of eal kS o Ineke No e UON CN c Coon CORPSE TONY maak Ho xec e A uraa yorr 9 2 Sma AGRON ee EAE 5 What new features would you like to see _ Nok kaving Ce ree od S e Aaa _ Wave C Soy Sys Cok x des Mw easy use eec Ceo TO buSiness Ome _ Quaxka jeadieack OF oC DANS adad SseGmgncs 45 OF eme 9 r amp s rego 5 Some A VEE re Wed aware ovn ui pe epe a Sram X gos ons vos a T Z WOR WED u S CONO fee SS Xe v Beaker US eJ NS un 5 O1 cao SO OCC GS N 0 ocesc ete 6 Do you have any comments on user experience of the current tool in general And can you suggest anything that would assist the user experience See Fresmon S 7 Do you have a list of priorities for enhancements New features 8 Any other comments 123 he
97. a B 4R B Cym Dep UK CYMBALTA Depression Cymbalta 5R C Cym Dep UK CYMBALTA Depression Cymbalta TIR C Zyp Bip UK ZYPREXA Bipolar Zyprexa B 6R C Zyp Sch UK ZYPREXA Schizophrenia Zyprexa S TR IS NOT NULL INVALID INVALID INVALID INVALID 13 R Non Target Cym Dep UK CYMBALTA Depression Cymbalta 12 R Non Target Zyp Bip UK ZYPREXA Bipolar Zyprexa B 8R Non Target Zyp Sch UK ZYPREXA Schizophrenia Zyprexa S 3 AutoNumber Figure 14 Cross reference table Running the buildRating2Objective query in the previous step the following fields in the cross reference are filled in automatically e ID This field is used only for internal purposes e FF POSTFIX This is used to distinguish between different types of sales representatives within a sales force In this particular example R is used for UK LPG RSP Appendix A Section 5 shows example positions and wish sales team they belong to e Rating This field contains the unique ratings from the imported data In this example the value IS NOT NULL appeared The cause of this was a sales representative entering data mistakenly whilst attempting to query Setting all other fields to INVALID prevents this error being passed through to the final files The remaining fields must now be completed by the user e Objective This field should contain the objective represented by the rating as previously seen in figure 11 In this example the rating A appears twice as this letter ind
98. able comments No technical knowledge required I don t need to understand the fact that a query is being ran all I need to do is click to view my data 1 feel a lot more comfortable using this system I am always worried that I am going to do something wrong within the existing system With participant 2 the main discussion focused on the cross reference table This table confused me at first The user went onto discuss how she thought everything in the drop down box was being assigned to that objective her comment was am just used to manually entering everything so at first a drop down box just confused me She noted how 155 now that she is aware of the correct process it is an improvement over the existing since no free text entry Participant 3 discussed the interface Her comments were that I think a blue colouring would be better it would match hyperlinks as well Appendix E Code Main Classes Only Importing Data Two main classes are SelectedRatingsImporter cs using System using System Collections Generic using System Data using System Data Common using System Data SqlClient namespace Excellmporters public class SelectedRatingsImporter public static SelectedRatingsImportResults ImportSelectedRatingsDataFromFileToDatabaseConnection string fileName string databaseConnectionString SelectedRatingsImportResults results new SelectedRatingsImportResults String excelConnectionString
99. ag Currency c Ojo C Sales Quota Unit Quota SoM Quota Percentage Data Objective Call Frequency Record Type OBCF Input Format Ref to DD Field Name Max Length Required Format Integration ID Default Value n a Affiliate Code 2 Y 0 0 Record Type 4 Y T 0 Source System Id 30 Y 0 0 128 Input Data Fields n a Operation 1 n a Operation Date 15 Contact Id 30 Y Y 0 Parent Objective Name 100 Y T 0 My Suggested Rating 30 0 0 0 My Selected Rating 30 Customer Call Frequency 10 0 My Suggested Call Freq 30 0 Call Frequency 10 0 My Reason for Difference 100 0 Product Id Indication Data Business Plan Record Type BSPL Ref to DD Field Name Affiliate Code Max Length Required Format ntegration ID Default Value Record Type Source System Id Operation Operation Date Name Type Description Start Date YYYYMMDD HHMISS End Date YYYYMMDD HHMISS Locked flag lt lt lt 5 Parent Business Plan Name Primary Position Flag Position Name n 3 KS a 5 2 Position Division Status LLY_Local Start Date YYYYMMDD HHMISS LLY Local End Date 129 NISIB
100. aire results also corroborated this The number of errors was reduced and again from questionnaire results it can be seen that this was due to the improved navigation design user comments stated how the breadcrumbs guided them through the task so they were aware of what their next actions should be In conclusion this shows that the results were obtained as a result of the new navigational structure 7 2 3 Methodological understanding Participatory design techniques were encouraged throughout the project lifecycle in a number of ways During requirements analysis interviews encouraged stakeholder interest in the project and enabled users to easily provide all of their comments on the existing system A card sorting session also took place thus allowing users to discuss and conclude specific reasons for their task sequencing choices The results from this fed into requirements and began the discussion of navigation design During the design phase users were recruited to the design team and participated throughout To begin users completed low fidelity prototyping to design the UI this was followed by group discussions to elaborate and share ideas A number of options were selected discussed and conclusions began to be made Finally users took part in a CARD see section 2 5 3 session this was to aid the navigational design decisions Participatory design techniques continued throughout implementation where users were available to give feedback
101. ak Rationale This will prevent a user either having to stop a task meaning a restart is required next time or rush to complete the task potentially causing an error to be made Saving state will increase the efficiency and overall productivity of the task Source This was raised during the suggestions for new features during interviews section 3 3 5 The system shall be capable of producing the combined OBJECTIVES resultant file following the OneLilly Entity Specification Section 5 Data File Preparation See Appendix A Section 6 Rationale This functionality is essential to completing the task Source Step 9 of the observation 3 5 1 discussed user comments to remove this step such that the template is not required but the functionality bought into the system 48 16 The system shall be capable of allowing the user to separate view each data type in the correct OneLilly format Rationale It is a high priority to ensure that the targeting data is sent to the OneLilly system correctly the first time therefore a user needs the ability to check the end file Source Following Nielson s guideline of error prevention literature review section 2 3 allowing such checks will minimise the likelihood of incorrect data being sent 17 The system should be capable of producing statistics on the number of records imported and produced Rationale This task is currently completed manually so incorporating this into the system will again in
102. am Interviewee 1 is the main user of the current tool It is her responsibility to generate the objectives files using this tool on a quarterly basis She is involved in the business rather than IT and does not come from a technical background However over the past year she has started to develop some skills using the current Microsoft Access tool aided by training Interviewee 2 Business Integrator Pharma IT team Interviewee 2 is part of the IT functional area and has a more technical background She has a good understanding of the current tool and her role is to suggest and provide IT solutions to the OneLilly team After the current tool was created Interviewee 2 took responsibility for it and thus has a good understanding of its structure use problems and benefits 26 Interviewee 3 Technical support RADS team Interviewee 3 has the most technical background of the three His role is to give technical support for existing systems The existing tool developed using Access requires additional support for which he is responsible He often assists with tasks such as updating SQL where required and to provide backup to the first line support 3 3 4 Analysis All three interviews were conducted on October 30 2007 and each lasted approximately 45 minutes From the responses it was clear that users viewed the tool from different perspectives the primary user discussed functionality she liked in the existing tool wh
103. an interface to enable users to monitor their actions 2 Norman uses this model to demonstrate why some interfaces cause problems to their users he describes the terms e Gulfs of execution this is the difference between what the user wants to do and whether the interface allows them to If there is no difference the interaction will be effective the interface should therefore aim to reduce this gulf e Gulf of evaluation is the distance between the physical presentation of the system state and the expectation of the user if the user can evaluate the presentation to form their goal the gulf is small The more effort required the less effective the interaction is Gulf of execution User s goal Gulf of evaluation Figure 1 Norman s Execution Evaluation Model To pursue a task with computer support a user needs to translate their real world goal into a software orientated goal also known as a system goal The simplest case is one where the system action is identical to the real world concepts the aim is to enable the user s task goal to be mapped very easily onto appropriate system features This can be very difficult to do and a number of researchers have come up with heuristics to attempt to support effective user behaviour 2 3 Heuristics A number of advocates of user centred design have presented a set of golden rules heuristics although they might not be suitable to every situation they do provide a useful checkl
104. analytic techniques for example task analysis that are used for design can also guide the design of support material It is necessary to make decisions about how the help will be presented to the user how will the help be requested How will it be displayed to effectively assist the user Help may be a command a button a function or a separate application A command will usually require the user to specify a topic and therefore assumes some knowledge a help button is readily accessible and does not interfere with the application a help button could support context sensitive help 1 Context sensitive help allows the user to get information about a specific area of a system it provides information that is relevant to the task the user is currently trying to accomplish It ensures that the information is timely but if the type and scope of the help is not tuned to the user s needs then it will not be helpful Frimm 26 completed some analysis of user needs and observation of user s interaction with a system and has indicated a requirement to offer the following types of context sensitive help e Where you are example solution breadcrumbs e What you can do both goal and command orientated e Field options values that can be applied to the current field e Help onthe current condition if one exists It is important to consider how this help will be displayed to the user it could be in a new window or it may use part of the s
105. ared server requested and loading the file to the OneLilly server In terms of future work for the application many suggestions were made during empirical evaluation The issue that seems to cause the most difficulty is the business object report that to be loaded into the application the existing system required columns to appear in an exact order with precise column names Within the new application this was improved such that columns could now appear in any order but still exact column names are required This causes user frustration since loading errors occur frequently due to the business objects report continually being updated Although this could be considered a business process issue such that the report could be prevented from being updated a user suggested adding functionality which automatically queries the correct server to obtain the data required therefore cutting out the report generation and loading phases Further enhancements were those that were discussed during requirements analysis or stated as secondary priorities but weren t implemented this includes e Providing specific information on load errors e Providing extensive statistics e Auditing functionality to track changes and monitor data sent Many user interface design enhancements were suggested and these were discussed in greater detail during iterative high fidelity prototyping and again during evaluation see sections 4 5 and 6 3 7 3 2 Research Moving Forward
106. ask Navigational choices and search queries are driven by how well the written text semantically matches the search goal One of the main challenges of written text is the distinction between what was actually written and what was intended There are resulting problems with this as there are different words that have the same meaning and some words having multiple meanings or different interpretations by different people The problem is that people often want to access information based on its meaning but individual words used within queries or used as part of a hyperlink do not uniquely or sufficiently express meaning There have been previous attempts to overcome the diversity in human word usage in information retrieval for example augmenting the user s original query terms with related terms e g from a special thesaurus It is also possible to measure information scent using semantic similarity but this is very difficult simply because it is not feasible to ask people for similarity ratings of all words that may be used in a user interface These methods are labour intensive and are not always successful Latent Semantic Analysis LSA is a technique for identifying both semantically similar words and semantically similar documents It was developed to address the problem with early information retrieval systems that performed exact word matching For example if you submit a query for car you would only be returned documents or links that
107. ations e Complete an empirical evaluation of the existing tool e Engage sample users from the UK CRM team at Eli Lilly to identify their individual requirements and recruit them to the design team e Undertake iterative user centred design e teratively develop and evaluate a new usable OneLilly targeting system e Comparatively evaluate new system against the existing system using appropriate usability matrices and methodologies e Complete a project overview to summarise and draw conclusions The next chapter will investigate the existing literature within the scope of the project Chapter 2 Literature Survey The user interface is the interactive part of a system that is used by the human It is the means by which users give commands to control its operation input data and respond to the output In principle it is relatively simple to create a user interface but it is rather more difficult to develop one that combines usability with usefulness and pleasure to use Preece et al 2 states that designing usable interactive products requires considering who is going to be using them how they are going to be used and where they are going to be used She states that there are four basic activities of interaction design 1 Identifying needs and establishing requirements for the user experience 2 Developing alternative designs that meet those requirements 3 Building interactive versions of the designs so that they can be communicated
108. b SelectedRatings orderby sr PositionLetter select new sr PositionLetter sr SelectedRatingCode Distinct var dropDownItems from obm in db ObjectivesMappings orderby obm Objective select obm As with the previous description again an instance of the database model is created note that this is a different model to the one used by Excel Importers since it contains different tables The code shows the two queries required the first queries from the data imported to return the distinct values of PositionLetter and SelectedRatingCode The second queries from the ObjectivesMapping table this table remains the same between transactions but can be updated by an administrator to return all objectives The next step is to bind the results from the first query to controls this is done using DatBind as shown here rptUpdate DataSource ratingList rptUpdate DataBind foreach RepeaterItem i in rptUpdate Items 81 Each row of the cross reference table is then dynamically updated with the unique position letter and selected rating The results from the second query are used to update the objectives drop down box and this is done by simply adding each item to the DropDownList DropDownList list DropDownList i FindControl cmbObjectiveMapping add items to the objectives drop down list foreach ObjectivesMapping m in dropDownItems ListItem 1 3 m Objective Trim m ObjectiveParent Trim new
109. ce and improved robustness Participants also commented on how useful it is that the new system generates business plan data as well previously this had to be manually completed externally Every participant also commented on how all functionality is now within one application without the need to copy and paste data into the template Comments such as It was easy to make mistakes when moving data the new system provides you with more confidence and It s so much easier when its all together show that there is a preference for the new system 6 4 Expert evaluation As discussed in section 6 1 this section is an analytical evaluation method it requires analysts or usability experts As discussed by Faulkner 10 it is a good idea to use a combination of both methods so the final part of this evaluation will be an expert heuristic evaluation As stated within the expert evaluation within the requirements chapter this method would usually require multiple evaluators but due to time constraints and expert availability this will not be possible and so a single evaluator will be used Interview with Jon Bolt HCC Masters Student Jon is completing an MSc in Human Communication and Computing at the university of Bath Jon has also gained experience within industry whilst completing a placement abroad Jon was presented with the new OneLilly targeting application and in relation to Nielson s guidelines section 2 3 he discussed his thoughts o
110. ce then the data has been consistently correct It is possible that this check is therefore being completed out of habit rather than necessity 2 Ensuring that the OBCF data is as expected for each rating This is achieved by checking that each type of selected rating from the original file has translated into the correct objectives in the generated file The inexperienced user had difficulties locating the table that the previous query had generated and thus was unable to check it After receiving assistance to locate this there were no further problems It was requested that anew system still allow users to complete these checks should they wish This can be achieved by allowing users to sort on columns and search for specific values within the data STEP 8 This step involves generating the objective parent records OBJT and verifying them The Make OBJT ONELILLY query creates a table named OBJT which must be verified Verification is performed by sorting the table on position division and then completing the following checks 38 e The position names correspond to the position division e The start and end dates correspond to the name of the objective e The business plan corresponds to the correct position name The user also takes a random selection of the business plans and checks that a business plan record exists either within the global OneLilly application or is contained in the business plan data currently being generated
111. cessfully load Click Save amp Proceed A message box will be displayed to confirm that the user is happy to progress with upload Upload results then displayed including number of records both successfully and unsuccessfully imported 3 Complete Go to the administration When update is clicked it cross screen and update the Zyp successfully saves and then displays reference Bip to be Zyp Bip2 new value within the drop down box Selected values successfully Select an Objective of Zyp Bip and call frequency of 2 for each item 4 View edit View each data type Successfully navigated to the correct data Edit one row of data dataset Successfully edit and 149 within each Delete one row of data within each updated a row Successfully deleted a row Export Export the final data file Successfully open a dialog box prompting user to save or open file 150 Appendix D Evaluation Supporting Documentation Section 1 Participant experiment instructions OneLilly Targeting applications comparison experiment Thank you for taking the time to complete this experiment if you have any questions throughout then please don t hesitate to stop and ask Instructions You will be asked to carry out a series of tasks within both the existing and the new OneLilly targeting application Please note that the experiment will be recorded on a Dictaphone these recordings will only b
112. ckedFlag ParentBusinessPlanName PrimaryPositionFlag PositionName PositionDivision Status LLY LocalStartDate LLYLocalEndDate bi E E B OBJT _ entry pk m Affiliate Code Bl Record Type LJ Source System ID F1 Operation mi Operation Date Ci Start Date mi End Date Status Target Business Plan Name Summary Locked Flag Geography Name Primary Position Flag Position Name Position Division Report Flag LLY Local Start Date Local End Date Ms ps o n ps e n OBCF entry PK El Affiliate Code El Record Type El Source System Id Operation El Operation Date Contact ID El Parent Objective N Suggested Rating My Selected Rating L Customer Call Freq My Suggested Call 8 Call Frequency My Reason For Diff Product 1d m Indication Figure 32 Database Diagram OBSQ entryPK mi AffiliateCode E RecordType SourceSystemID El Operation Ra OperationDate E ParentObjectiveName ProductID PrimaryPositionFlag Currency SalesQuota UnitQuota SoMQuota Percentage 8 88 For the purpose of this project there was no need to set up the relationships between the four data types The OBCF table is updated based on data contained within the SelectedRatings table and then following this the OBJT and BSPL tables are updated base
113. consequence is when yes or no is clicked Jason checked the additional help was but was still unsure so clicked on yes but then was unable to proceed with the task and had to reopen the system The user instructions fail in numerous places for an unfamiliar user they would not be able to complete the task from these instructions only Help and documentation 132 Section 5 OneLilly Brand Ratings Also shows which positions divisions belong to which sales team Eu s aaa Primary Care Primary Care HFR LPG Schiz LPG Mood Oncology Diabetes C1 C1M1 M1 1 1 MIC COMPASS Target Product Cialis Cialis _ f Cymbalta Yentreve OneLilly Brand Ratings OneLilly Brand OneLilly Brand Brand Cialis _ Yentreve Q3 Cia ED CymDep _Cia ED_ Stra Tt Rating Q4 Cia ED Cym Dep Cia ED Cla ED Yen SUI __ Stra T1 EXAMPLE POSITION EXAMPLE DIVISION _UK PCARE 011D UK PCARE 0111 UK PCARE 01 1L UK PCARE 011X UK PCARE 011Z UK PCARE O01E UK ADHD 011C _ UK PCARE1 __ UK PCARE CSO1 UK PCARE2 UK PCARES UK PCARE CSO2 UI PCARE HFR ___ Strattera _ Stra T2 a mt Stra T2 Sira T3 Stra T3 Strattera _Forsteo m 12 Stra T4 For Ft Forsteo For F2 Stra T4 Zyprexa Zyprexa HGH H1 UK LPG SCHIZ Zyp
114. contained that exact word and see nothing containing automobile unless it contained car also LSA considers the words that co occur in documents with car and the words that co occur in documents with automobile and given that they overlap significantly it is possible to determine that car and automobile are semantically related LSA will produce a rank ordered list of words that are semantically similar to it with exact names high on the list 48 8 If users cannot find what they are looking for they will not be able to perform their desired tasks and will lead onto the user giving up Preece et al 2 states that one of the usability goals is that is that the system should be efficient to use supporting the users carrying out their tasks Dix et al 1 lists four things to be aware of when designing each state or screen of the system Allowing the user to e Know where they are e Know what they can do e Know where they are going or what will happen e Knowing where you have been or what you ve done To assist the user to understand where they are some websites show bread crumbs at the top of the screen this is a path of titles showing where the page is in the site Figure 2 shows an example of this This will help the user get a full sense of their current location relative to the site structure and they can jump to their desired page by a single click MUSC188 Guitar Techniques and Methods MUSC188 Control Pane
115. crease the efficiency and overall productivity of the task Source This was raised during the suggestions for new features during interviews section 3 3 5 Non functional requirements Data requirements 18 The system shall remove old data once a transaction is complete Rationale This was discussed through the suggestion of providing auditing functionality stakeholders agreed that storing previous data was not required Source Discussed during project scope section 3 9 Environmental requirements 19 The system shall be compatible with the hardware and software available at Eli Lilly offices Rationale This system is intended for employees of Eli Lilly and therefore it should work on their existing hardware and software Source Interviews section 3 3 4 User requirements 20 The system shall use language and terminology that is appropriate for users with 21 22 no previous experience of targeting data Rationale Some of the OneLilly terminology and abbreviations are not obvious to a user who has had limited experience with targeting tasks Source Interviews section 3 3 4 At a user s request the system should provide context sensitive help Rationale The observation of the previous solution saw how a user may require additional help if this is not provided it may cause them to give up on the task Source This was discussed in the literature review section 2 3 and the interviews section 3 3 4 examples of wher
116. creen Alternatively help hints prompts or pop up s can be given as the user requests it The presentation style that is appropriate depends largely on the level of help being offered and the space required this will develop from the design of the help functionality 11 The physical layout of documentation can make a different to its usability for example large blocks of text are difficult to read on screen and may require the user to search for the help they require A useful style is to provide a summary of the key information prominently with further information if required An index can be a useful summary but should be organized to reflect the functional relationships between the subjects rather than alphabetic ordering 1 As already stated the help system will need to be designed just as the rest of the system is therefore the heuristics will also need to be consisted for example consistency will have a significant play within the help system Consistency is important to consider when designing the presentation of information on the user interface 2 4 4 Information presentation It is also important to consider how information will be presented on screen this issue has been around for many years long before computers and therefore interactive systems Obviously the presentation of information on the screen depends on the kind of information different purposes will require different representations 1 But obviously it is important
117. crumbs but with the comments and conclusions from each of these three designs being brought together to produce another design of the navigation structure and will form part of the next prototype iteration 4 2 2 Task Sequencing This obviously ties in with the navigational structuring since this will be the main guidance through the task but the aim of this section is to answer the question of how a user should use this structure to successfully complete the task Again Throughout this discussion it was noted that the criteria needing to be satisfied these are the following requirements 18 Treating each transaction separately 21 Help at the current location 22 Flexibility 26 Error prevention 27 Easy to navigate A good starting discussion is figure 23 where it can be seen that one of the users designed a proceed button Example of OBST CE Nakhon co prs pios V Sec 5 V cut emery SOME 3 ee penas cow Xo euu Were NS Monuatly erer a mew Co eon Dl sevece RA Figure 23 User Sketch showing proceed button This button would obviously move the user onto the next sub task making it very clear what a user s next actions or task should be Other users stated that they assumed they would be required to use the navigation structure to manually click onto the next sub task as the hyperlinks show in figure 22 Reviewing the criteria it was dec
118. d after A Query has ran Discussions with the users determined this step should not exist as it was put in place as a workaround for an unstable Business Objects report This has now been corrected and produces data in the correct format each time it is run This observation highlights the need to ensure that informative messages are meaningful and useful Nielson s heuristic help users recognize diagnose and recover from errors should be followed such that users are able to recover using only the guidance given in error message STEP 4 This step involves running the buildRating2Objective query which builds the cross reference table It works by selecting unique values of selected rating from the imported data and is required because the global OneLilly application from which the data originates contains a free text field enabling sales representatives to enter values which are not already in the cross reference The query appends all values to the cross reference table even if they already exist as noted during user interviews The inexperienced user had difficulty locating the required query After receiving help finding the queries window she noticed it was easy to run any query and thus asked for confirmation that she was indeed about to run the correct one Upon running the query the user was presented with the prompt shown in figure 3 9 and required assistance on the corrective action to take Had the user not received t
119. d boxes Call Frequency and Sales quota 02 Only display the upload box if the data type was selected on the overview page 03 Display one upload box and allow the user to add another if required The final option can be dismissed for two reasons firstly it is unclear which data type needs to be uploaded first and secondly this would require unnecessary additional clicks to be made With reference to Nielson s 5 heuristic Error prevention it has been decided that the second 57 option is the most appropriate since it minimises user confusion by displaying an upload box just for the data required SPECIFYING A MAPPING Two of the three users sketched a design for the cross reference table these can be seen in figure 25 and 26 Please Select kha obiecrive Product ndicadion forent Sheoc of A eek Geo Overt NOT NULL E SF CORE cannot Edited Figure 26 Cross Reference Design 2 Obviously figure 25 contains more detail but some similarities are shown between the designs this includes the general layout of the table would have been taken from the existing solution the inclusion of drop down boxes and notes about some fields being completed automatically Looking back at the observation step 3 section 3 5 1 it was noted that itis possible for the final three columns product indication and objective parent to be derived from the objective field and this is reflected in user s sketch
120. d on the data contained within the OBCF table 5 3 System Architecture Using the technologies just discussed figure 33 shows a high level overview of how the is structured Relational Database within SQL Server Application code NET Language C Sharp cs Presentation code Cascading Style Sheets Figure 33 Overview of System Implementation As figure 33 shows ASP NET uses code behind class files cs files this allows the presentation code to be separated from the core application code The code behind class file is compiled so that it can be created and used as an object This allows access to its properties its methods and event handlers This separating of layers is following the Model View Controller MVC architecture The model represents the information the data within the relational database the view corresponds to the elements of the user interface which is contained within the presentation code aspx and CSS Finally the controller manages the detail and communication which is contained within C Sharp classes C Sharp is an object orientated programming language again developed by Microsoft It was chosen because it offers the power of C and the simplicity of Visual Basic making it easy to write and maintain code 74 The benefits of following the MVC architecture in this way include e tis possible to provide alternative user interfaces for the same model particularly useful
121. d to establish a foundation of existing practices on which to build new requirements It is therefore essential to understand how the task is currently completed Data can be collected by an observation of the current task and this is completed in section 3 5 1 From the interview comments and initial discussions it is clear that the users still revert to their understanding of the previous manual solution the comments raised by the users suggest that there is a potential issue with how the task flow is structured differently between the two solutions Thus it is important to not only focus on the existing solution but to investigate this task sequencing issue also from the past experience of users This analysis will take place by using a card sorting session to allow users to pictorially show their task sequencing preferences this is completed in section 3 6 Once this data has been collected it will be possible to complete an analysis to look for the similarities and differences between the two solutions and make informed decisions on which processes new requirements can be built from Completing a hierarchal task analysis of each of the solutions will model the sequencing enabling easier comparison and this is completed in section 3 7 3 5 1 Observation Completing an observation will allow an initial assessment of the usability of the tool to do this a user who has never used the tool before was requested to complete the task and a record was
122. data As requested in the design as shown in figure 38 it is possible to export the results to an Excel file the next section gives a brief discussion of the implementation details 5 4 4 Export The export control works by using the StringWriter class to write each row of data into a table to form an output stream This is then put into a HttpContext object this encapsulates all information related to the request allowing the programmer to access it outside of the actual page and then writes this out HttpContext Current Response Write sw ToString HttpContext Current Response End Where sw is the StringWriter which contains the output stream of data contained within the GridView The actual GridView that is exported is not the one that is displayed instead there is a separate hidden one This is needed to prevent the Edit and Delete labels being added When the document is being served from the server it needs to prompt the user to save the file directly to the user s disk without opening it in the browser HttpContext Current Response AddHeader content disposition string Format attachment filename 0 fileName However for known MIME Multipurpose Internet Mail Extensions types such as Microsoft Excel application ms excel but is possible to use the content disposition header to override this default behaviour HttpContext Current Response Clear HttpContext Current Response Conten
123. e HCI expert whilst working through the task identifying well and poorly designed features His comments were then categorised using Nielson s set of heuristics see section 2 3 Notes were taken throughout the interview These can be found in Appendix A Section 4 The most discussed heuristic was visibility of system status On numerous occasions whilst the HCI expert was using the system it failed to provide him with any feedback Important episodes or events were identified as including completion of a query which if re run by mistake could result in incorrect data Jason advised that a confirmation prompt should be displayed prior to 30 commencing a long or critical action task and the user should be aware when the task has completed for example the results could be displayed before a user is returned to a main page The lack of clearly defined start and end points have a negative effect on the usability of the existing tool and the lack of guidance means that it is not possible for a user to follow a path to their goal These combine to reduce user awareness of the system state resulting in users needing to constantly refer to paper based documentation in order to achieve their goal The lack of a match between the tool and the real world was discussed in detail One particular example which was highlighted is that users are unable to see what data is being removed prior to removal since they are required to write an SQL query to delete
124. e PICTIVE and CARD PICTIVE Plastic Interface for Collaborative Technology Initiatives through Video Exploration is a technique that uses a combination of low tech design components with video recording facilities It is useful for increasing the direct and effective involvement of end users and stakeholders in the design Office materials such as sticky notes highlighters labels paper and scissors are used on a shared design surface to produce low level prototypes Some of the benefits of using materials like this include e Such low technical objects insures that all participants have equal opportunities to share their ideas e They are inexpensive and encourages the participants to be bold with them hopefully stimulates inventive design and creative problem resolution e They are easy to change and move around meaning that many variations can be tried quickly and easily This technique was developed by Muller 27 in 1991 he describes that PICTIVE is intended less as a means for the evaluation of an already designed interface but rather for the creation of the design of an interface its purpose is to involve all possible stakeholders meaning that participants did not have to be technology experts to assist with design A PICTIVE session may be a one to one collaboration or it might involve a small group participants will be asked to prepare before the session Typically this will be the user considering a task scenario for example ask
125. e available to me and will be used for analysis purposes Although this will have no impact on how you complete the task the measurements that will be taken include number of clicks time taken and number of errors made Tasks to be completed within BOTH applications Section 1 Generate targeting data for all sales reps within the LPG sales force The data required is call frequency and their parent objectives It will be for quarter 1 of 2009 This is split into 6 phases Overview prepare tool Import Updating cross reference Verify generated data Exporting to the OneLilly format Sending to be loaded Q9 gr MH Section 2 This task is split into a number of short tasks to test the usability of the website For this task data will be already be added to the application 1 Navigate to the cross reference table 2 Navigate to the call frequency data 3 Update the Cross Reference table and subsequently a new data set to change positions ending in with a selected rating of A to have an objective of Zyp Sch 4 Within the objectives data update the call frequency of the first record from 1 to ia 151 Section 2 Participant Questionnaire responses P OneLilly Targeting Questionnaire Thank you for taking the time to participate in the OneLilly targeting application comparison experiment if you could please reflect upon the tasks you were asked to do by completing this questionnaire On a scale o
126. e basis for high quality user contributions Therefore supplementary tools and techniques have been developed to enable users to contribute these include paper and pen exercises brainstorming storyboarding and workshops 1 Finally actively involving users allows them to reflect their own work and bring innovative ideas to the design process Henderson 21 discusses three main guidelines as an approach to participatory design e The goal is to improve the quality of work life Both the users and the developers share this objective the focus is not solely on the design of technology It may include things like reallocating tasks or moving furniture and the technology might allow or require this The focus is the work as a whole and the technology is only a component of this e The orientation is collaborative Both designers and developers are actively involved in decision making it is recognised that knowledge from both will be required for successful technology Through reflection and experience the users will become familiar with technology as designers will with work practices e The process is iterative Emerging design ideas will need to be tried out in a work environment It requires developing ways of presenting ideas to users such that they begin to get an idea of what it would be like to use this technology in their work Mock ups scenarios and prototypes are some of the ways to allow users this early exposure It is often said tha
127. e counterbalancing where the two groups will complete the experiments in a different order Although the experimenter should still be aware that it is possible to have asymmetrical transfer where transfer of learning from the existing system to the new system might not be the same visa versa It is also possible to consider the matched pair design as with between groups this involves the creation of two different groups but participants are matched on the factors that are considered to be important This particular experiment an example of factors to match could include participant experience or ability Taking each of these factors into consideration it has been decided that the within groups experimental design method will be used The main reason being due to limited participants but also as stated this will also be beneficial since participants can act as their own control The suggestion of counterbalancing will be used to prevent the transfer of learning effects The four participants will be divided into two groups each group will have both a novice and an experienced participant one of which will experience the existing system first and one which will experience the new system first This allocation is shown in table 6 3 Experienced Novice Existing system first U1 U3 New system first U2 U4 Table 5 Participant Allocation This allocation also reduces the effect of a greater amount of experience as a confounding var
128. e cross reference table and subsequently the data set 2 Access solution New application 7 I found it easy to edit the data produced Access solution New application 5 Please provide all your general comments on the task or new interface 154 OneLilly Targeting Questionnaire Thank you for taking the time to participate in the OneLilly targeting application comparison experiment if you could please reflect upon the tasks you were asked to do by completing this questionnaire On a scale of 1 5 with 1 being strongly disagree and 5 being strongly agree please rank the following for both the existing and the new system 1 Ifound iteasy to complete the first task Access solution New application 5 2 I found it easy to complete the second task Access Solution New application 3 I found it easy to locate the call frequency data Access solution 1 New application 5 4 I didn t need to rely on a user manual or help functions Access solution New application 4 5 I find the system easy to navigate through Access solution New application 5 6 I found it easy to update the cross reference table and subsequently the data set Access solution New application G 7 I found it easy to edit the data produced Access solution New application 9 Please provide all your general comments on the task or new interface Section 3 Telephone conversation notes the day after experiment Participant 1 not
129. e eos Sees eX pe asad ang more o e Marge dbs dake coe uk or exc Xe Coun jee Sed o E Une 2 Arethere any features you would like to see retained from the tool Or possibly there is a process or understanding you had from completing the task manually that you would like to be able to see do in a new system remo 6 foc cyecking Vesodead Gos fer RVG Oe j a QM AL rosy Cor qara co 9 She x re aod Ro user No ugo 3 Do you find the current tool easy to use E Rom an Wex Ground and hore rod gt oo no ay Co Aa e face lt s MO qp eA Scc dev gx Oe MENS users m be dar te If so what makes it easy to use If not can you describe why it s difficult to use s AUS ey Liming USES OF trormirg Cou a as reg wired t VS Joan oa Geechee Khe We Grown of C assise to lt lt 3 easag Aescri be we Noo a gurety Qr nao x gt F v wor be good move ewa from hrs 0 hawa a Muck more Of HA Sore Coca Fer non uca Ce eX ars ee bere Wa Sys Tea Coe Cron USer TOE Ou Vas 2 Q3 4 gt S vooxA ag ser goca e Tro hove aoge cart ct tog vC vena SOE OF Seer ANTENA
130. e last two chapters whereas summative evaluation is completed within this chapter to assess whether the new OneLilly targeting application is better than the previous one and if so estimate the degree of improvement and the implications for the workforce As the literature review stated this evaluation can be done in a number of ways 86 Empirical e Usability testing involves measuring typical users performance on typical tasks It is generally done by noting the number and kinds of errors that users make and recording the time it takes for them to complete a task Questionnaires and interviews are also used to gather user opinions e Field studies are completed in natural settings with the aim of understanding how the prototype is adopted and used by users in their working and everyday lives Analytical e Cognitive walkthrough e Heuristic evaluation e Formal Evaluation To ensure an effective evaluation this chapter will encompass all of these approaches and will be structured as follows e Empirical Evaluation Empirical evaluation will include users completing specific measurable tasks in a representative working environment at Eli Lilly offices e Subjective Evaluation This will enables users to give general feedback on the system this will include informal interviews and questionnaires the likert scale will be used e Analytical evaluation A HCI expert will be asked to evaluate the new OneLilly targeting application
131. e parent objt is for sales quota as different values req Boolean ifSQ false string isSQ parent_objt ToString Substring parent_objt Length 2 2 if isSQ SQ ifSQ true 1 ifSQ false work out the position from the parent_objt string position parent objt ToString Substring 8 11 OBJT objt new OBJT objt Affiliate Code UK objt Record Type OBJT objt Source System ID MANUAL objt Operation I objt Operation Date DateTime Now ToString yyyymmdd hhmmss objt Name parent objt ToString the objective is for sales quota if ifSQ objt Type Sales Quota objt Description quarter Sales Quota objective is for call frequency else objt Type Brand Call Frequency objt Description quarter Brand Call Frequency work out start amp end date Switch quarter case 01 objt Start Date year 0101 000000 objt End Date year 0331 235959 objt LLY Local Start Date year 0101 000000 objt LLY Local End Date year 0331 235959 break case Q2 objt Start Date year 0401 000000 objt End Date year 0630 235959 objt LLY Local Start Date year 0401 000000 objt LLY Local End Date year 0630 235959 break case Q3 164 objt Start Date year 0701 000000 objt End Date year 0930 235959 objt LLY Local Start Date year 0701 000000 objt LLY Local End Date year
132. e user assistance was required can be seen throughout the observation The system shall allow the user to complete the task in their preferred sequence order Rationale The switch between the first solution and the current solution violated the users understanding of the task structure User s who had varying experience of both solutions found that their understanding was lost when all flexibility was removed in 49 the current system The aim of this requirement is to provide flexibility such that any user will be able to easily navigate between different areas of the system Source Observation step 1 section 3 5 1 discusses user task sequencing preferences The card sorting session section 3 6 highlighted the need to maintain flexibility 23 The system shall allow the user to view data that they are required to edit or delete Rationale This will give a user a match between the system and the real world if they see what data is being removed it gives them a better understanding of current system state also Source Heuristic evaluation of existing tool section 3 4 Usability requirements 24 The system shall give appropriate meaningful and timely feedback after a user has completed a task Rationale If a user has not received feedback that a specific task has completed they may be waiting for no reason or attempt to move on potentially causing an error if the previous task is not complete Source Heuristic evaluation of existing
133. eadcrumbs Hudson 54 attempted to explain why breadcrumbs go mostly unused claiming that many sites do not use breadcrumbs at all some sites are not consistent with their implementation and breadcrumbs are often small and hidden Rogers and Chapparo 52 also found that positioning on the page affected its usage It is still inconclusive as to the importance of breadcrumbs as a contextual aid but if training and familiarity is the key to its success a familiarity that can only increase across the population of users over time and due to its simple implementation and lack of harm 55 it is likely that its usage will pick up given sufficient time To assist the users to understand what they can do you need to make it clear which text is cickable and which is not Sometimes this can be difficult when trying to improve the appearance of a site but it is important to note the trade off between the appearance and the ease of use 1 Visuals audio and multimedia serve as a powerful illustration of concepts and ideas It is therefore important for designers to carefully choose which visuals they use making sure that they are n necessary to assist the user graphics unrelated to the content can cause confusion and the designer should resist the urge to make the produce more visually exciting 36 The last two points assisting the user to understand where they are going was discussed at the start of this section but also has a link with feedback
134. ed a background for the project and discussed how it had both an academic and industrial foundation It stated the aims and objectives of the project and discussed which research areas would be considered Chapter 2 Literature review provided a background understanding of the key factors principles and methodologies in user interface design and development using a user centred approach The chapter began with a summary and discussion of human computer interaction HCI principles and the design considerations needed to support these principles The chapter then went onto consider user centred design tasks and usability evaluation This chapter also highlighted some of the current research impacting navigation design decisions including concepts such as information scent and latent semantic analysis 108 Chapter 3 Requirements began by providing an overview of the targeting process and focused on gathering requirements from the stakeholders using a variety of techniques including interviews observations and task analysis This chapter concludes by providing a prioritisation list and the resultant requirements specification Chapter 4 Design used a variety of participatory design techniques to allow informed user interface design decisions to be made Design activities included iterative low fidelity prototyping CARD using storyboarding creating an envisioned HTA and iterative higher fidelity prototyping This chapter used the QOC Questions
135. ed in the literature review the use of breadcrumbs increases the usability by providing users with a mental model of their location within the site Predictions will be tested using a variety of methods as described in section 6 1 Notes and measurements will be taken throughout the experiment as well as it being recorded on a Dictaphone to capture all user comments Measurements taken will include number and types of errors made number of clicks to locate a specific piece of information and time taken Notes will also be made on the path a user takes through the system Finally participants will be asked to complete a questionnaire and be informally interviewed to discuss their feedback and comments 6 2 4 Identifying variables A variable is a measurable factor characteristic or attribute of an individual or a system it may be present or absent or take on a range of values at different times To qualify as a variable two conditions must be met 1 There is a general property or characteristic involved 2 The characteristic is measurable it is important to specify the manner in which the variable is measured The variables that will be used within this experiment are shown in table 6 1 89 Variable type Description Variables Independent Variable systematically varied The IV is the system and has two levels by the experimenter 1 The existing OneLilly targeting application 2 The new OneLilly targeting application De
136. ed to do is click to view my data 57 98 Quotes such as this show how the new system no longer requires a technical understanding being required meaning users do not need to rely on help and documentation as much Reviewing the errors that were made within the new system is important to check that there are not any remaining usability problems An interesting error made by the experienced user was her assuming that everything in the drop down list was assigned to the position she did not realise that she had to select items During the proceeding interview this error was discussed and her comments showed that she was still relating her understanding to the existing system She stated I am just used to manually entering everything so at first a drop down box just confused me 58 Since this error did not occur in other participant s behaviour it is possible to assume that it was her previous experience that caused this rather than a usability problem Participant 2 U2 commented that her first error was just a simple mistake and she realised that she was not paying enough attention to the information on screen she was able to easily navigate back to re select the tick boxes Her second error was due to the objective she wanted to select was not in the drop down box once she was told a new one can be added by an administrator she stated that this should be made clearer on the screen or she should have be made aware of this beforehand Part
137. eed to be taken forward into the new system The first manual solution has an extensive number of steps therefore the HTA will exemplify detailed steps for just one record type the remaining will display the high level tasks only Figures 16 and 17 show the HTA for each of the solutions 42 pues uodxe eyejduiey UUM Ours 7 jeuy yea 2 eiqissod Buuepuo mm sony umop u1 eq ueo a Key mep enb gyep Aouanbayy 5 DSEO podu 9 HED 4980 edasd 5 SLET pue x ueis 200 eiejduio LLE amp ueu uogisod aup 0L uorsiAp uonisod ejeouec 6 SMEIS 195 BE sa e 0 pexo o 0 aep pua eys sau G E 1eaus jeusaxa eJede d Z seAngjueseJdaj Seres yoduy Lz SSBUISNG ejeJeuac Figure 16 HTA for the original solution 43 0590 podu 17 seiaueb Ayenuew z V4 3990 8 1780 ejqissod uodui 514L aqe aqe z 7 Bep pues uodxe eyejduie uiui ejejduie aweue 3o9uo E S janb anb uny TS ping uny 2 7 apan 12 18190 FG pio aeled L v 4990 pue 1780 uodx3 9 euy ejeex 1 8 G
138. em A draft test plan which can be seen in Appendix C Section 3 was created and participants were asked to perform these tasks at various stages throughout the implementation phase This meant that testing was informally being completed throughout where the results and feedback being obtained from this testing were guiding the next stage of development as described in section 5 5 Obviously if this system were to be fully integrated into the Lilly environment then a stronger test plan would need to be developed Examining the internal structure of the code to select test cases is known as white box testing whilst examining the expected behavior of the code is known as black box testing Exhaustive testing would include a test case for every possible path through the system however this is impractical and instead the aim was to use black box and select a set of test cases as close to the ideal as possible 5 7 Conclusion This chapter has described the technical details of the development of the OneLilly targeting application including programming language choice database design system architecture high level class structure and further descriptions of the most significant functionality The next chapter now focuses on evaluation and uses both empirical and analytical techniques to evaluate the new system 85 Chapter 6 Evaluation The two previous chapters have discussed the design process and methodologies used to implement the app
139. em but I m guessing it would now go with the cross reference What do you think Also we thought there should be an additional card between 3 and 4 if I remember rightly in our last discussions we said there would be validation during import so would we need to view feedback on the import We added a card for View import results and decide whether to terminate and restart or continue hope that s ok Apart from that we are happy with the general design of each screen Give me a call if you need anything else and see you shortly Nicola Ordering amp decisions Me 1 Decide I need OBJT amp OBCF 2 Presuming I am completing this for the current time period I would select Q2 and 2008 and tick the boxes for OBJT and OBCE 3 Guessing following on from our previous discussions that the upload box for OBSQ would not be present since I did not select this 3 5 our new card I will decide to continue 4 2 selected ratings from sales force R I will have A which will map to the objective Zyp Sch and B which will map to the objective Zyp Bip 5 I m guessing that N is only there as an example Since there are no N reps in the data you gave us I will ignore that value I would select the Zyp Sch from the drop down box again as we have previously discussed I am guessing that the remaining fields will be automatically filled in I would like click on Add objective and add Zyp Bip to that R as well 6 I w
140. ereas the business integrators discussed aspects such as maintainability and ease of support There were a number of good aspects to be noted about the current tool The primary user is impressed with the functionality of the tool since it allows her to complete her task within a few hours whereas previously it took several weeks However numerous suggestions about the look and feel of the system describing many of usability issues were made and the more technically minded interviewees criticised the tool s lack of maintainability Once all interviews were complete a brainstorming session was undertaken where all three interviewees shared their responses This proved successful since some of the responses turned into debates and many of the ideas were expanded upon Having analysed the responses it was found that answers fitted into three distinct categories e Presentation how the users view and perceive the system and their comments on its appearance e Interaction how the users use the system and how comfortable they felt navigating it in order to achieve their goals e Context what the users described as the functionality of the system Below a summary of the results is given Full interview transcripts can be found in Appendix A Section 1 Presentation Users commented that since the current tool is based on Microsoft Access the look and feel is consistent with other applications with which they are familiar For the purpo
141. erfaces CHI 90 Proceedings 8 G Grinstien A Kobsa C Plaisant J Stasko 2003 Which comes first usability or utility Proceedings of the 14 IEEE Visualization Conference 12 W Hudson Oct 2004 HCI and the web Breadcrumb Navigation 13 H Chi P Pirolli K Chen JPitkow Mar 2001 Using Information Scent to model user information needs and actions on the web CHI Proceedings 15 Allan D Saja Jan 1985 The cognitive model An approach to designing the human computer interface SIGCHI 16 N Bevan amp I Curson May 1999 Planning and Implementing User centred design CHI 99 proceedings 114 19 Muller Mar 1991 Participatory design in Britain and North America Reaching through technology CHI 91 20 C Spinuzzi Oct 2002 A Scandinavian Challenge a US Response Methodological Assumptions in Scandinavian and US Prototyping Approach SIGDOC 02 21 H Henderson Apr 1990 Reflections on participatory design CHI 90 proceedings 22 M Kyng Apr 1994 Scandinavian Design Users in product development Human factors in computing systems 23 J Aberg amp N Shahmehri An Empirical study of human web assistants Implications for user support in web information systems CHI 2001 proceedings 24 B Shneiderman May 2000 Universal usability Communications of the ACH 26 S Frimm Jan 1988 A User needs approach to context sensitive help ACM SIGCHI 27 M J Muller June 1992 PICTIVE An exploration in participatory
142. ers stated that a prompt would be useful to remind them that data will be removed and to ensure that they wish to terminate without sending or saving any data As discussed during the interviews section 3 3 4 many of the common errors were caused due to insufficient guidance through the task meaning users had the freedom to navigate to any sub task and potentially incorrectly run a query Although not every error is related to the navigational structure it is a good idea to review the errors made during the observation section 3 5 1 to understand how they could be prevented These errors along with potential solutions are summarised in table 2 Common Error Possible Design solutions Incorrect pre editing of business 1 Prevent pre editing being required by importing based on objects report column name rather than location Incorrect options chosen during 1 Minimise the number of options required import these were access specific 2 Pre set all options enabling the user to simply add a file only Duplicate rows present in the cross 1 Prevent duplicate rows being added reference table 2 Prevent saving or proceeding whilst duplicate rows are present in the table Attempting to view data without first 1 Disabling the proceed button from the cross reference page specifying a mapping between each 2 Allow the user to move to the correct page but with no data selected rating present Running queries inan inco
143. essfully completed and saved the cross reference table Another comment raised here was about the layout of the table She discussed that she thought having two records for multiple objectives for one rating was confusing She suggested that she would like the system to follow her thinking being able to add another objective to a rating 37 rather than having the same rating with a different objective Finally there was a discussion around user guidance and assistance to aid the user in completing this cross reference It was decided that the new system should allow the cross reference table to be sorted by any column Since there is no need for the user to view the ID column this should not be visible to them and fields requiring specific values should use drop down boxes The user has requested that automatically generated fields still be visible to them STEP 6 This step is to delete the previous OBCF ONELILLY table However the user prefers to rename it and then run the Make OBCF ONELILLY query to generate all records for the call frequency objectives The user commented that if she is generating data that is not for the fourth quarter she has to edit the SQL code since it is hard coded and so preset to 04 For the example used in this observation the data was for the fourth quarter However the inexperienced user commented that had it been for another quarter she would have been unable to complete the step since she ha
144. etter Text Trim current db CrossReferences SingleOrDefault p gt p PositionLetter r UniqueSelectedRating Label i FindControl lblSelectedRating Text Trim r CallFrequency Int32 Parse DropDownList i FindControl callFreq Text Trim 162 DropDownList myDropDownList DropDownList i FindControl cmbObjectiveMapping String s i Controls OfType lt DropDownList gt First SelectedItem Value r ObjectiveMappingID Int32 Parse DropDownList i FindControl cmbObjectiveMapping SelectedItem Value Trim db CrossReferences InsertOnSubmit r db SubmitChanges now can update OBCF OBJT and BSPL tables var ql from sr in db SelectedRatings join T in db CrossReferences on sr SelectedRatingCode equals r UniqueSelectedRating where sr PositionLetter r PositionLetter select sr var q2 from cr in db CrossReferences select cr to update the CRMapping in SelectedRatings foreach SelectedRating sr in ql int id id 0 foreach CrossReference cr in q2 if sr PositionLetter cr PositionLetter amp sr SelectedRatingCode cr UniqueSelectedRating id cr entry pk sr CrMapping id get variables from overview page string quarter String Session Quarter string year String Session Year Boolean BSPL required Boolean required Boolean OBJT required Boolean Session OBJT required Boolean
145. eve unh Oe e Ux ax end Of po ete 6 Do you have any comments on user experience of the current tool in general And can you suggest anything that would assist the user experience SASSO Sew tno wen E Oe SM Wax A Go CN Coreg red 0 0 d Cose Siege onw yor Gok your Week omurn emery ce on an im 3 mo bux neem C voC of eo rk Fer Wren Sex peers y ws ery EAS e RAf4S gt VJOLA gt Sane g OS ot ex Joo ver 3 Nd we See goen ECON 7 Do you have a list of priorities for enhancements New features E CC Oa eos w koexS of Corn oral XE Ys D XP ACA Sens 3 CX OEY Ry UO Ne hoses bere guido CA 2 Sec EF bod ASRS lc OUS COMAS 3 s vex Ex cox o0 z aniwa ko SLeruer e Sendong 8 Any other comments No Qe Contour ak Mowe 22 GY SYAkS wao mom pecorcds voe eo cu d i auu vrerord e 121 Interview Transcript Date 3o q Jot Interviewee name and team BoaaSmess Ur e croscoc AL 1 What is good about the current system Are there any features you like about the current tool look x Peer LAS Fontar Pecess LS A cammon Loo FA icy Oen maros Qro ex E users oe sed Xo navt Oe e no Car Ss Sten across eo groan oe e C onoga EX 9 00 0
146. f 1 5 with 1 being strongly disagree and 5 being strongly agree please rank the following for both the existing and the new system 1 I found it easy to complete the first task Access solution New application 2 I found it easy to complete the second task Access Solution 3 New application 3 I found it easy to locate the call frequency data Access solution 9 New application 4 didn t need to rely on a user manual or help functions Access solution 3 New application 5 I find the system easy to navigate through Access solution Y New application 4 C 6 I found it easy to update the cross reference table and subsequently the data set Access solution 7 New application Pry 7 1 found it easy to edit the data produced Access solution Newapplication 5 Please provide all your general comments on the task or new interface NN w MKS NN Poved va Whe prrone Uma Se boo n 152 PL OneLilly Targeting Questionnaire Thank you for taking the time to participate in the OneLilly targeting application comparison experiment if you could please reflect upon the tasks you were asked to do by completing this questionnaire On a scale of 1 5 with 1 being strongly disagree and 5 being strongly agree please rank the following for both the existing and the new system 1 Ifound it easy to complete the first task Accesssolution New application 4 2 I found it easy to complete the second task Access Solution
147. f involving stakeholders throughout the project It is now possible to begin bringing all stakeholders ideas together to form directions on the new implementation of the OneLilly targeting system The next chapter focuses on building up an understanding of the stakeholder s requirements It is now important to continue to refer back to this chapter throughout the project ensuring the research is put effectively to use 22 Chapter 3 Requirements In Chapter 2 an overview of the concepts and processes of which need to be considered throughout the project were given included a description of the process of user centred design The first step in this process is to identify user needs Preece et al 2 state that this is to understand as much as possible about the users their work and the context of that work so the system in development can support them in achieving their goals The aim of this chapter is to focus on gathering requirements from the stakeholders of the OneLilly targeting system and documents the approaches taken This will include gaining a greater insight into the targeting process completing interviews and an observation with a selection of the stakeholders and completing task analysis The chapter will conclude by discussing the stakeholders chosen priorities and detailing the requirements specification 3 1 Data Gathering A requirement is defined as a statement about an intended product that specifies what the product
148. filiateCode UK bspl RecordType BSPL bspl SourceSystemId MANUAL bspl Name bPlan ToString bspl Type Territory bspl Description bPlan ToString Business Plan 165 Switch quarter case 01 bspl StartDate year 0101 000000 bspl EndDate year 0331 235959 bspl LLY LocalStartDate year 0101 000000 bspl LLYLocalEndDate year 0331 235959 break case Q2 bspl StartDate year 0401 000000 bspl EndDate year 0630 235959 bspl LLY LocalStartDate year 0401 000000 bspl LLYLocalEndDate year 0630 235959 break case Q3 bspl StartDate year 0701 000000 bspl EndDate year 0930 235959 bspl LLY LocalStartDate year 0701 000000 bspl LLYLocalEndDate year 0930 235959 break case Q4 bspl StartDate year 1001 000000 bspl EndDate year 1231 235959 bspl LLY LocalStartDate year 1001 000000 bspl LLYLocalEndDate year 1231 235959 break no default required as it can only be a value from drop down box bspl LockedFlag Y bspl PrimaryPositionFlag Y bspl PositionName position string positionLetter switch positionLetter UBI bspl PositionDivision break UN bspl PositionDivision break bspl PositionDivision break PENIS bspl PositionDivision break default as bspl brea case U case U case U case y division is required Posi
149. focused on gathering requirements from stakeholders including gaining a greater insight into the targeting process completing interviews an observation and a task analysis Discussions of priorities and project scoping have taken place now the next chapter builds upon these requirements as a foundation to begin design It is important to note that the requirements may still evolve during the iterative design process 50 Chapter 4 Design The aim of this chapter is to work alongside stakeholders to provide a means to satisfy the requirements stated in the previous chapter and to explore potential navigational design structures Preece et al 2 states that there are two types of design conceptual and physical The former is concerned with developing a conceptual model that captures what the product will do and how it will be behave whereas the latter is concerned with details of the design such as screen and navigational menu structures both of these will be considered throughout this chapter The design emerges iteratively through design evaluation redesign cycles continually involving stakeholders as discussed in the literature review participatory design will be used throughout this chapter this was discussed in greater detail in section 2 5 To allow informed user interface design decisions to be made this chapter will include the following participatory design activities e Iterative low fidelity prototyping section 4 3 e CARD
150. framework simplifies data access further This project will use SQL Server to create a relational database and LINQ to SQL for accessing this data 5 1 2 LINQ to SQL Language Integrating query LINQ gives programmers the ability to query and transform data using any NET language LINQ can be used with any data source for example from XML documents relational database tables or inside a collection of objects LINQ to SQL is the NET Language Integrated Query for relational data It allows the programmer to model a relational database using NET classes It is then possible to query the database using LINQ as well as update insert delete data from it Section 5 3 provides more information on how this is set up 5 2 Database Structure The relational database structure was designed based on the learning s from previous chapters and it was found that the system needs to include tables for imported data cross reference information final output and storing user information Therefore following tables are required e A table for each of the data types that can be included in the generated objectives file e BSPL OBJT e OBSQ e OBCF e A table for the selected ratings data to be imported into SelectedRatings e A table to store selected rating mapping to objective CrossReference e A table to store the mapping between an objective and its related fields product indication and parent objective ObjectiveMappings e Tables to store
151. g phase with interviews will give a good introduction to the current process and allow us to see where potential issues lie Two operational data stewards are the main users of the current tool one as the primary user and the other as a secondary user In addition three business data stewards in the team could potentially be asked to use the tool to support the main users Users have a range of experience and technical abilities Besides users other potential stakeholders are e Business integrators who directly support and assist users with the tool e RADS Regional Applications Development and Support a technical group with which business integrators liaise This group that will support the solution and carry out any technical fixes required For the purpose of these initial interviews a selection of stakeholders will be used the primary user a business integrator and a member of the technical support group 25 3 3 1 Interview Process Interviews will be semi structured and informal This will allow preparation of a basic script to ensure necessary points are covered and ensure some degree of consistency in the questions put to each interviewee This interview style will also provide opportunity for interviewees to explore avenues that had not previously been considered Our aims in conducting the interviews are e To gain an understanding of what users consider to be the good and bad points of the existing system e Toundersta
152. gn data collection method and controlling confounding variables 9 Deciding on the appropriate statistical or other analysis 10 Carrying out the experiment Although this is only a small scale study it is still important to consider each of these steps for the experimental evaluation 6 2 1 Hypothesis A hypothesis is a testable statement that can be verified or falsified with an experiment It is tested by manipulating one or more variables involved The variable that is being manipulated is known as the independent variable because the conditions to test the variable are set up independently before the experiment starts these will deliberately invoke a change in variables known as the dependant ones 2 Murphy 41 states that the most common statistical procedure is to put a null hypothesis Ho against an alternative hypothesis Hi Null hypothesis usually refer to no difference or no effect The aim of this experiment is to test the effect of the independent variable on the dependant variable this is done by supporting a null hypothesis Null hypothesis Ho There is no difference between the new and existing interface Hypothesis H The new application developed for OneLilly targeting is superior to the existing system Thus the alternative hypothesis is directional if the null hypothesis has been rejected and the alternative is supported it will be possible to attempt to understand why this is the case The resul
153. gure 21 and 22 show the remaining two user sketches both of which adopted a breadcrumbs style this compares to comments raised during the interviews section 3 4 5 ees 21 Navigational structure designed by a user with minimal experience xa en gt wc en ien Kher pa is Lee us Figure 22 Navigational structure designed by an experienced user The design concept of breadcrumb items representing each sub task remains the same in both of these sketches As indicated in the literature review this will guide and indicate the users current location within the system One of the noticeable differences between each of these designs is the content of the breadcrumb items figure 4 4 has listed all the data types which are obviously that particular user s preference whereas figure 4 5 has moved all verification tasks within one breadcrumb item Data and then allowed the user to navigate to each data type from within this page Again it is important to remember that different users have a difference preference of the order in which they verify each data type meaning the breadcrumb labels in figure 4 4 would not be appropriate for all Other comments were raised during this discussion firstly one designer added a validate section and the other did not The designer of figure 21 discussed how she wanted each sub task to be assigned to a label It is possible to see that this is an experienced user since she is transferring he
154. h the most substantial difference being increased productivity due to the amount of time the system saves Some of the additional benefits users described the system as bringing included e Reducing the overhead on the business data stewards since no technical knowledge is now required to complete the task e This also reduces the overhead on IT staff since tedious help and training on Access is no longer required 111 e The data quality and integrity will also be significantly improved leading to a reduction in load errors again leading to increased productivity e The process has been simplified now that all aspects of the task are completed within one system 7 3 Future Work This section will consider further work to the OneLilly targeting application including potential enhancements or additional features mentioned by users during the empirical evaluation Finally concluding with a discussion on how the current research may be moving forward 7 3 1 System Further Enhancements As stated in section 3 9 the project was scoped to complete implementation for just a single sales force this was due to limitations of implementation time Therefore the most significant further extension would be to expand the application to all eight sales forces Due to the fact that the system was developed externally to the company there were also some limitations for the export functionality this included the ability to place completed files onto the sh
155. he annotated screen shots from the implementation section see appendix C Section 2 during the telephone conversation 58 there were some additional comments and suggestions raised these included 102 e tis not immediately obvious that the breadcrumbs are clickable one participant thought it was just a display of your status It was suggested that the text should be blue meaning the breadcrumb design should no longer be green e When verifying data the breadcrumbs do not change one participant suggested that there be another status bar that pops up below that updates as you work through each data type This would also require a save button at the end of each data type e It was suggested that there is a Finish breadcrumb that they should be redirected to once they have sent the final file The participant also suggested that if they click the cross or log out they should be prompted reminded to send the file Another good point to note is that 3 of 4 participants discussed how the new system is structured in the same way as they think comments such as I like how steps are more clearly defined and the system guides me through 57 Reasons for such comments are due to not only the use of breadcrumbs but also because when the user clicks save at each stage the system automatically redirects them to the next step Comparing these comments to those that were raised during the requirements gathering see section 3 3
156. he author Signed Abstract This project presents a user centred design approach to the redevelopment of an application used to prepare data for a Customer Relationship Management system new system could potentially take minutes to support tasks that currently take hours With an aim to improve the navigational efficiency through the user interface this project conducts background research followed by requirements analysis and design and development Finally an empirical evaluation is used to measure the success of the new interface by completing a comparative evaluation IV Contents CONTENTS sac dm EAR CER a BA DA a i C cO Seco dE naian kada andana bonas ai aar dadina itrabi ania vainikai II LIST UUE V LIST OF TABLES VI ACKNOWLEDGEMENTS saa aa a e a eba VII INTRODUCTION aa a 1 1 1 AIMEE LR RE E A EE LA A AS ES 1 2 OBJECTIVES LITERATURE SURVEY Sd aa es 3 2 1 INTRODUCTION TO GMs s SAREE AROS 2 2 WHAT IS COGNITION 2 3 HEURISTICS sees m 2 4 DESIGN CONSIDERATIONS atsan 2 41 Navigation design oe e P ED e RERO RE lo Ce a 2 4 2 Designing for both novice and expert users 2 43 Providing user support 2 4 4 Information presentation 2 5 USER CENTRED DESIGN testet tette siet asset e ebd sri
157. he next and final chapter of this project conclusions will be drawn about the development of the new OneLilly targeting application what has been achieved and whether the objectives stated at the beginning have been met This chapter has already suggested some improvements and potential further developments but the final chapter will summarise and extend upon these 107 Chapter 7 Conclusions The aim of this project was to evaluate and redevelop the OneLilly targeting application with particular attention to improving the navigational efficiency through the user interface The research aim of the project was to investigate how navigational design can influence user experiences The question to now consider is whether the aim and objectives stated in chapter 1 have been achieved To answer this question this chapter will explore what has been achieved by discussing the main learning s from the project and an analysis of whether the research aim has been met The chapter will then move onto a discussion of potential enhancements or future work which could be undertaken to extend the development and finally concluding by discussing ideas for improving or moving the current research forwards 7 1 Project Overview This project comprises of six chapters which encompass the different stages undertaken to achieve the redevelopment of the OneLilly targeting application Now follows a summary of each of the chapters Chapter 1 Introduction provid
158. his assistance she would have been unable to continue She was guided back to completing the previous step enabling her to complete step 4 successfully This observation again highlighted problems with navigating the current tool Although tasks need to be completed in a precise order for the tool to succeed it gives the users freedom to complete them in any order It also highlighted the lack of information on how to proceed especially if problems were encountered Section 2 3 1 of the literature review discussed navigational design and suggested that each screen of a system should give the user enough information to allow them to move closer to their goal Another problem that was highlighted during this step is that of the current tool adding duplicate values to the cross reference table All of these observations must be taken into account when forming the requirements specification and for a new system STEP 5 This step is to update the cross reference table The user explained that this step is important to ensuring that generated data are correct Figure 14 shows a completed cross reference table 36 Rating20bjective Table ID FF POSTFIX Rating Objective Product Indication Objective parent b R A Zyp Sch UK ZYPREXA Schizophrenia Zyprexa S 9R A Cym Dep UK CYMBALTA Depression Cymbalta 2R Zyp Sch UK ZYPREXA Schizophrenia Zyprexa S 3R B Zyp Bip UK ZYPREXA Bipolar Zyprexa B 10R B Zyp Bip UK ZYPREXA Bipolar Zyprex
159. iable since each group contains both an experience and novice user 92 The data collection method will be through the use of questionnaires and written notes throughout the evaluation sessions each of these can be found in Appendix D For further analysis purposes recordings of both the experiments and interviews will be taken using a Dictaphone A stopwatch will be used to record timings and these will noted on paper but also spoken into the Dictaphone as a backup recording 6 2 8 Experimental results The experiments were conducted across two days 26 27 March 2008 and were held at the Eli Lilly offices participants were given the same equipment set up and brief before the experiment took place Measurements taken are those described in table 6 2 2 furthermore non statistical results were collected at the end of each experiment via informal interview Each measurement will complete additional calculations on each task these include Mean average n Standard Deviation spread of values O n 1 Where X is the individual value e g time taken and n is the number of participants This section will now discuss each measurement type and conclude by drawing all analysis together Time taken The aim of this measurement is to measure how long it took each participant to complete individual tasks within both the old and the new system The results are shown in tables 6 and 7 Time Taken to co
160. icates that a customer will be targeted for both Zyprexa and Cymbalta e Product This field should contain the product identification for the given rating which must match the identifier used in global OneLilly application It is possible that the user can enter any value so the user must complete an extra check to ensure no mistakes have been made e Indication This is the illness at which the given objective is targeted For example the objective Zyp Sch has the indication Schizophrenia Again it is possible for the user to enter any value meaning an addition check for errors must be completed e Objective parent This field is used to form part of the objective name as seen in the second layer in figure 6 More information on how the objective parent forms part of the targeting data was discussed in section 3 3 Further discussions with the user made it apparent that it is possible for the final three columns to be derived from the objective field therefore users should not need to manually enter these values When the inexperienced user reached this step she was immediately confused by the additional rows added She received help to understand which ones existed already and thus could be deleted The experienced user discussed how she normally sorts on the position letter however this was not necessary since only with one sales representative is being used here Once the user was familiar with which fields required updating she succ
161. icipant 3 made no errors within the new system and participant 4 s error was due to data being contained within the wrong worksheet second instead of first therefore an error would have occurred whichever system was being used It was seen that although this error could have been prevented by allowing the user to chose which worksheet their data is in this could also have opposite effect of confusion to what this means as has happened in the existing system The final error made whilst using the new system was that only one data type was exported rather than all the participant commented that since he has now experience of the task he did not realise that all was required and he simply clicked the first link on the page During discussions it was mentioned that an improvement might be to add helpful comments for example This will open a file containing business plan data in the OneLilly format to this screen although it is useful to note that this is discussed in the help video Number of clicks This experiment was conducted to investigate the general usability and to attempt to answer the question of whether navigational redesign has an impact on end user experiences by lowering the number of commands and eliminating tedious message prompts and unnecessary clicks To assess the efficiency of the interaction design the number of clicks were recorded These results are taken from part 2 of the experiment Participants were asked to locate a
162. ided that a combination of both of these approaches would be suitable since it will be easy to navigate and provide flexibility such that a user is easily able to navigate to any sub task from any single page The remaining criteria were yet to be considered in these low fidelity prototypes instead these were discussed and comments will be built into the next prototype iteration The task sequencing will affect how each transaction is treated separately since an end point needs to be reached and the user needs to be aware that once this is reached all data will be 55 removed Users discussed how this could be achieved and decided it would be useful to make it clear within the navigational structure that there is an end point This could be achieved with the following two options 01 A finish breadcrumb item 02 A finish button on the final screen Once the transaction is complete all data needs to be removed this can be added to a finish button Having an additional breadcrumb item would bring no extra benefit since there is nothing to be placed in this section therefore the second option is the most appropriate Additionally users discussed need to be able to terminate the task without necessarily completing it e g last minute business decision to postpone objective creation or if there was an error with imported data This can be completed using a button or option that when clicked will terminate the transaction and remove all data Us
163. idered how these affect designers decisions I will now move onto the process of designing User centred design puts the users at the centre of its design and development they should take part in the decisions that affect and the way that itis designed and used As stated in the introduction there are four stages to interaction design these stages are carried out in an iterative fashion with cycle s being repeated until usability objectives have been attained this process is shown in figure 3 12 X 1 Plan the user centred approach GNE Meets requirements 2 Specify the context of use 5 Evaluate designs 3 Specify user and against user organisation requirements requirements 4 Produce design solutions Figure 3 The User Centred Design Process Early in design phase the design solutions are likely to be mock ups and prototypes as the design progresses higher fidelity designs will be evaluated against more specific requirements 16 The following table shows a high level characterization of the most popular user centred design methods 17 Method Focus groups Contextual inquiry Usability testing Card sorting Participatory design Questionnaires Interviews Task analysis Brief description Facilitated discussions to share ideas amp opinions about the system Field study where designers visit the users workplace to analyse habits activities amp environmental factors Users work wi
164. ies within each data type Rationale The HTA s figure 16 and 17 shows that a user is currently able to update any information within the generated data Users requested that this functionality be maintained since there may be a need to edit data for an individual sales representative Source The final step of each HTA model figures 16 and 17 and the context section of the interview section 3 3 4 The system shall allow the user to save generated data to an Excel spreadsheet Rationale As discussed with stakeholders providing an audit trail is not part of the scope for this project for this reason it was requested that data can be saved externally to the system Excel is an appropriate file type since this was already used within the existing system Source Auditing raised during interview comments section 3 3 4 and compromise discussed within project scope section 3 9 The system shall complete verification on any data directly edited Details of verification required for each data type is specified within the OneLilly formal input specification see Appendix Section 2 Rationale This will prevent errors occurring and will result in increased data quality Source Following Nielson s guideline of error prevention literature review section 2 3 Data quality and accuracy was raised throughout the requirements analysis the observation highlighted this section 3 5 1 The system should allow a user to return their task after a bre
165. iews section 3 3 5 The system shall allow a user to add call frequency and sales quota data by providing an Excel spreadsheet Rationale This is basic functionality of the system without this data the system would be unable to fulfil its purpose of generating targeting data for the sales force Source Step 2 of the observation section 3 5 1 The system shall validate the data added via Excel spreadsheets Rationale The global OneLilly application contains a free text field allowing sales representatives to enter any value for their selected rating but as stated during the observation only certain values are valid and correspond to an objective Source Following Nielson s guideline of error prevention literature review section 2 3 This was also raised as an important feature during the interview analysis section 3 3 4 The system shall should remove invalid values and inform the user Rationale The current tool automatically removed invalid entries as raised in the interview it is essential that this functionality remains in place Source User s regard the removal of invalid values as a required feature of the system section 3 3 4 HCI Expert evaluation section 3 4 highlighted the importance of providing user feedback The system shall be capable of allowing a user to specify a mapping between a selected rating and an objective Rationale Again this is basic functionality of the system without this mapping the system would
166. ificant amounts of data therefore as figure 47 shows usability can be improved with the addition of number of pages along with Previous and Next Context Every participant made a comment about the fact that the new system does not require them to have any technical knowledge and how much time and effort it saves not having to work these details out Comments such as Just being able to select the quarter from a drop down box is so much easier than having to understand how and where to update the SQL were made In terms of functionality there were a number of key areas that participants were particularly impressed by e The verification on import comments like Its fantastic that I don t need to worry about errors moving through each stage 57 e The information provided when there is an import error is extremely useful to users e Being able to verify edit and export data from within the tool saves users a lot of time Overall participants were very happy with the functionality that this prototype provided them with The main comment was that currently the tool is only for the LPG sales force and if it were to be put to its full use the main enhancement would be to extend it to all sales forces There were also some suggestions for usability improvement e The results from the import verification could provide more information e g the nature of the error rather than just its location e It would be useful if the
167. ign decisions e Useful to access at later stages to understand critical decisions and avoid incorrect assumptions being made e Accumulated knowledge can be reused to transfer what was worked in one situation to another situation with similar needs MacLean et al 39 proposed the Questions Options and Criteria QOC notation as an approach to design rationale QOC uses a semi formal notation which uses questions to identify key design issues each of which is linked to at least two options providing possible solutions to the question All options are comparatively assessed either positively or negatively against a set of design criteria to determine the most favourable option 38 Figure 18 pictorially shows the QOC notation an option is favourably assessed in terms of a criterion linked with a solid line whereas negative links are dashed Option Criteria Question Option Criteria Option Criteria Figure 18 The QOC Notation The key to an effective design space analysis is deciding the right questions to use to structure the space and the correct criteria to judge the options The initial questions raised must be sufficiently general that they cover a large enough portion of the possible design space but specific enough that a range of options can be clearly identified Questions which isolate single issues are easier to understand and resolve A bad example might be How should menu s work
168. ileName HttpContext Current Response ContentType application ms excel using StringWriter sw new StringWriter using HtmlTextWriter htw new HtmlTextWriter sw Table table new Table table GridLines gridview GridLines if gridview HeaderRow null GridViewExportUtil PrepareControlForExport gridview HeaderRow table Rows Add gridview HeaderRow foreach GridViewRow row in gridview Rows GridViewExportUtil PrepareControlForExport row table Rows Add row if gridview FooterRow null GridViewExportUtil PrepareControlForExport gridview FooterRow table Rows Add gridview FooterRow table RenderControl htw HttpContext Current Response Write sw ToString HttpContext Current Response End private static void PrepareControlForExport Control control for int i 0 i lt control Controls Count i Control current control Controls i if current is LinkButton control Controls Remove current control Controls AddAt i new LiteralControl current as LinkButton Text else if current is ImageButton control Controls Remove current control Controls AddAt i new LiteralControl current as ImageButton AlternateText else if current is HyperLink control Controls Remove current control Controls AddAt i new LiteralControl current as HyperLink Text else if current is DropDownList control Controls Remove current control Co
169. ilised as a methodological tool successfully within the development process This section will now draw conclusions by summarising the results of the theoretical empirical and practical work completed during the project 7 2 1 Theoretical The project has given a good insight into the current research areas impacting navigation design The initial research gave an introduction into how the navigational structure has a direct impact on how users typically forage for information and therefore how quickly or easily they reach their goal Design lessons for information scent were stated by Nielson 14 and were considered throughout the design An investigation into current research surrounding the usage and effectiveness of breadcrumbs as a navigational aid also took place It was clear that many 109 researchers have different opinions of their utility meaning that currently it is still inconclusive as to the importance of breadcrumbs as a contextual aid During the design phase of the project users made the decision to use breadcrumbs as the navigational aid in the new system this therefore provided the opportunity to further demonstrate or extend the current research in this area Throughout design and implementation phases user discussions and conclusions shaped the design of the breadcrumb navigation Design decisions such as colouring for pop out and status reasons and item ordering for task sequencing reasons was particularly important in
170. iment 2 Counterbalance the variable by including all of its variables equally often in each condition 3 The systematic relationship between the confounding variable and the independent variable may be destroyed by randomisation For this particular experiment time of day and environment will be kept constant The participant mood is the most difficult to control One way to consider this is to link it in with the time of day if experiments are completed at the start of the day participants would not have been pre exposed to work related stress The confounding variable of experience can be dealt with using counter balancing and matching This can be completed by having an experienced user in each group and each group beginning with a different system a greater discussion on this can be found in section 6 2 7 6 2 5 Experimental task and method This is a comparative evaluation therefore the experiment will consist of the same representative tasks being completed both within the existing as well as the new OneLilly targeting application Participants will be asked to complete two main tasks the first being the standard task of generating targeting data for a sales force and the second task is designed to test the usability of each of the applications This is done by asking the participant to complete a 90 variety of low level tasks or actions for example navigating to a specific area or updating a call frequency value Full details
171. in a uniform manner The first step is to establish a connection string string excelConnectionString String Format Provider Microsoft Jet OLEDB 4 0 Data Source 0 Extended Properties Excel 8 0 HDR YES fileName This is requesting to open an Excel 8 0 file location to file is stored in variable filename which is passed into the method using OLEDB The HDR YES in extended properties states that the first row is to be used as column headings these can then be used in select statements The next step is to open the connection and read data using a SELECT statement to choose the required columns from the Excel sheet using DbCommand command connection CreateCommand Main comes from the name of the worksheet command CommandText SELECT Division Position Con Integration ID My Selected Rating FROM Main try connection Open catch Exception This is if the file cannot be opened throw new CannotOpenExcelFileException Using such a SELECT statement means it does not matter which order the columns appear within the spreadsheet although there will still be an issue if the column header names are changed 78 The code then continues onto read each row of data and as concluded within section 4 3 1 this data needs to be validated therefore each row is checked against a set of constraints these are discussed in Appendix B Section 3 For
172. in how the performance will be after regular usage Preece at al 2 also describes some practical issues these include things like access to appropriate users this might be dependent on schedules or budgets There might be situations where the evaluators have experienced surprise events that require a decision to be made there and then You also need to consider users themselves for example they might not be a valid representation of the user population the end product is aimed at The evaluator should also consider and be aware that some users might be anxious and uncomfortable when they make mistakes this could potentially lead them to not acting in their usual way Therefore it is important that evaluators put the users at ease 2 7 Conclusion This chapter has focused on some key aspects of HCI usability guidelines affecting design considerations and design and evaluation techniques each of these will need to be considered when moving through the project This literature review has also looked at the importance of following the HCI principles and how these guidelines can influence the design considerations for incorporating user support and effective navigation Breadcrumbs has been highlighted as a potential solution to aid the user navigating the system this chapter has also covered some guidelines to support designing interfaces for both novice and expert users 21 The entire literature review had a significant focus on the importance o
173. ing them to prepare what they would like the system to do for them and the developer considering an initial or preliminary set of components based on prior discussions Each participant will then share their expertise at the design surface to produce their potential designs together Since sessions are recorded on video 16 tape it prevents there being a need for a note taker and so allows the participants to work more effectively together 7 This design technique has been effective but can be time consuming for larger systems 29 CARD Collaborative Analysis of Requirements and Design is another participatory design technique It is primarily used for analyzing and redesigning task flows within a system Like PICTIVE it can be used one to one or in groups It is an informal or semi structured card game that supports analysis and critique of a system As described by Tudor 35 although the PICTIVE technique has been effective sessions are time consuming and tend to dive immediately into detailed design the task flow is often assumed to be correct rather than critically examined Whereas with CARD it is intended to focus on the flow of the users task detailed design work can then be done using PICTIVE these two techniques therefore complement one another With the growing recognition that an understanding of the users work practice would be useful in the design of a new system there has been a surge of interest in another technique for
174. intention is to identify any areas that are likely to cause difficulties because they violate known cognitive principals or ignore accepted empirical results These methods can be used at any stage in the development process making them flexible A number of methods have been proposed 1 The two most used are Cognitive walkthrough was originally proposed as an attempt to introduce psychological theory into informal and subjective walkthrough technique It involves one or a group of evaluators these are experts rather than users for example HCI experts software developers or designers inspecting a user interface by going through a set of tasks and evaluating understandability and ease of learning Walkthroughs require a detailed review of a sequence of actions these refer to the steps that an interface will require a user to perform in order to accomplish a particular task The evaluators simulate users by stepping through that action sequence to check for potential usability problems They will provide a story about why that step is good or bad for a new user To do a walkthrough you need four things 1 Aspecification or prototype of the system 2 Adescription of the task the user is to perform using the system 3 Acomplete written list of the actions needed to complete this particular task 4 An indication of who the users are and what kind of experience and knowledge can assumed about them Given this knowledge the evaluators will no
175. ion phase there were many more user comments which continued to feed into the design These included comments such as colour schemes number of rows visible placement of buttons and so on 4 6 Conclusion All design questions have now been answered with sufficient evidence that the appropriate decision was made Now follows a summary of the QOC notation used throughout this chapter 66 Questions Navigation Q1 How should the user navigate through the system Task Sequencing Q2 How should the user progress through each sub task Q3 How should the user help be provided Q4 How prevented should errors be Task Support Q5 How should data be added to the system Q6 How should mappings be made between ratings and objectives Q7 How should data be generated Q8 How should this data be validated Q9 How should the resultant file be exported Options 01 Mind Map 02 Breadcrumbs 03 Side Menu 01 Manually click through menu items 02 Proceed button redirecting to the next most logical sub task 03 A combination of both ofthe above 01 Written instructions 02 Video O All options in table 4 1 01 Written instructions 02 Video 01 Using a drop down box only editable by an administrator 02 Using drop down box editable by any user Free text entry 01 When the Proceed button is clicked 02 When the user wishes to view that data set 03 Buttons t
176. ious uses of the tool if itis being used as specified The current work around is to rename tables to allow for data retention however this causes problems with relational integrity and thus is not deemed a reliable mechanism e tis not possible to compare data generated with a given run of the tool to data generated previously This was being noted as useful in investigating errors in generated data Users commented on the robustness of the tool noting that it is easy to break Some of the comments included Incorrect results are produced for some values because of inflexible or incorrect SQL If any tables or queries are re named to preserve old data relational integrity of the data is compromised and thus incorrect output can be generated It is possible for any user to delete any table or query The tool has no security and can be edited by anyone since it is located on a shared server In the next section a summary of the findings is given 3 3 5 Summary of findings The crux of the suggestions made about usability can be summarised as follows Removing the visible SQL code and enabling the entire process to be carried out through a graphical user interface Navigational assistance is required The current tool is only suitable for experienced users this limited its uses and can hold back deadlines if specific users are away Therefore it is important that the new systems includes user guidance at all appropriate point
177. irst option will reduce the number of clicks that are required without loosing any functionality but the second option is following the same process as the existing system After a small debate users chose the second option simply because this is what they are used to A first iteration of design decisions is now complete and as stated throughout this section the next step is to work through another iteration which will combine all comments and conclusions that were made here This iteration will use storyboarding the reason for this is to allow users to attempt to interact with the prototype to allow them to see how they might progress through an example task It will encourage users to explore new ideas suggest modifications or notice something that may be missing or incorrect within the current designs 60 4 3 CARD using storyboarding As described in the literature review section 2 5 3 CARD Collaborative Analysis of Requirements and Design is a participatory design technique used to explore workflow options This will offer users a chance to role play with the prototype and attempt to interact with it by stepping through a particular task Although earlier the decision was made to use breadcrumbs as the navigational structure the ordering of breadcrumb items is yet to be set Understanding the workflow particular users follow is important to complete the design decision of how breadcrumb items should be ordered It is essential to make sure
178. ist The most well known and used are Nielson s ten heuristics and Shneiderman s eight golden rules 1 Nielson s guidelines are 1 Visibility of system status The system should always keep users informed about what is going on through appropriate feedback within reasonable time 2 Match between system and the real world The system should speak the users language with words phrases and concepts familiar to the user rather than system oriented terms Follow real world conventions making information appear in a natural and logical order 3 User control and freedom Users often choose system functions by mistake and will need a clearly marked way to leave the unwanted state Support undo and redo 4 Consistency and standards Follow platform conventions Error prevention Either eliminate error prone conditions or check for them and present users with a confirmation option before they commit to the action 6 Recognition rather than recall Minimize the user s memory load by making objects actions and options visible The user should not have to remember information from one part of the dialogue to another Instructions for use of the system should be visible or easily retrievable whenever appropriate 7 Flexibility and efficiency of use Accelerators unseen by the novice user may often speed up the interaction for the expert user such that the system can cater to both inexperienced and experienced users Allow users to
179. itted to the database and the number of records successfully added returned to be displayed on the user interface The list containing records that did not pass the validation is looped through and added to a table displayed to the user Figure 36 shows an example result 79 Lilt Overview Import X Ref Export Import You are logged in as sam cooper Call Frequency OBCF Please upload the Business Objects report for sales representatives selected ratings Admin Upload an Excel file Browse Help Validation Errors Number of records imported successfully 27 Restart Number of records with validation errors 1 Log Out Row Number Division Position Contact Integration ID Selected Rating 26 UK LPG BP UK LPG 012B UK 157639 NB Sales Quota OBSQ Please upload the sales quota data Upload an Excel file Browse Import Save amp Proceed Figure 36 Print screen of the import screen after file upload As can be seen from figure 36 there is an option to restart within the left menu if the user wishes to make appropriate changes to the spreadsheet and import again As discussed in section 5 4 1 this menu is contained within the master page meaning the code for restart is not part of the import page If restart is clicked it simply removes all data from every table within EISelectedRating ready to start a fresh transaction Finally clicking the Save amp Proceed button here
180. l D Course Menu Basic Control Panel gt Manage Course Manage Course Manage Students Backup Course j can manage studer nd restore 1 Add or Import students View some students Search records breadcrumb Change settings P Manage columns links Share Designer Access Show update log U car sve ar Create backup Upload backup current page j can give other designe Figure 2 An Example of Breadcrumbs They provide a trail for the user to follow back to the starting entry point of the system website Although one of the benefits of breadcrumbs is the fact that they are small and therefore take up minimal space 11 Hudson 12 discusses the fact that users do often ignore them he lists some reasons why this might be the case e Designers seem to be ashamed of their breadcrumb navigation they make the text too small or hide them in the heading area or fail to use underlines to make the links obvious e Many sites do not include them Breadcrumbs are often used in addition to global and secondary navigation systems There are three types trail path attribute and location 49 Path trails convey the path the user has taken through the site and are typically represented and used with the back and forward buttons on any standard web browser Location trails simply show the path down the website hierarchy leading to the page Attribute trails present a particular website as a set of attributes that it conforms to
181. l be analytic and informal techniques that will be used this section will now highlight some of these 1 Rosson and Carroll 3 state there are two types of evaluation and both of which will be used e Formative takes place during the design process each developed prototype will be evaluated The goal of formative evaluation is to identify aspects of the design that can be improved and set priorities e Summative is completed to access the design result and therefore is most likely to happen atthe end of development although it can be completed at critical points during development Its goalis to answer questions like does the system meet its goals and is the new system better than the previous There are two general classes of evaluation methods analytical evaluation and empirical evaluation Analytical evaluation does not involve users instead analysts or experts will do the evaluation where as empirical evaluation will involve users These are complementary to formative and summative goals 2 6 1 Analytical Evaluation Analytic methods do not involve users An important motivation for analytical methods is that they can be used early on in the development phase The first evaluation of a system should ideally be performed before any implementation work has started If the design itself can be evaluated expensive mistakes can be avoided since this allows the design to be altered prior to any significant resource commitments The
182. l it be a one off problem or persistent 4 How seriously will the problem be perceived by the user 19 Even though severity has several components it is common to combine all aspects into a single rating as an overall assessment to facilitate prioritising This is usually within a scale of 0 4 I don t agree that this is a usability problem at all Cosmetic problem only need not be fixed unless there is extra time available Minor usability problem Fixing this should be given a low priority Major usability problem Important to fix 4 Usability catastrophe Imperative to fix this before the product is release Nielson 6 states that it is difficult to get a good severity estimate from evaluators since they are more focused on finding new usability problems Each evaluator will only find a small number of the problems and therefore their decisions on severity are based on just the set of problems they identified Instead it is a good idea to set the severity ratings once the evaluation session is complete this can be done by sending out a questionnaire listing the complete set of identified problems and asking then asking each evaluator to rate each usability problem It is important to note that each evaluator should produce individual ratings independent of other evaluators Nielson also states that severity ratings from a single evaluator are too unreliable to be trusted As just described above the more evaluators asked to judge the
183. lication this chapter now brings all this together by completing a comparative evaluation between the existing and new system As Dix et al 1 states evaluation has three main goals 1 Toassess the extent and accessibility of the systems functionality it must meet the user s requirements therefore meaning that the design should enable the user s to perform their intended tasks more easily 2 Toassess the user s experience of the interaction this may include aspects such as how easy itis to learn and use and the user s satisfaction 3 To identify any specific problems with the application this maybe aspects of the design which when used in their intended context cause unexpected results or confusion amongst users However a further goal is the compare systems to establish if there have been improvements made in the redesign As noted in the literature review and discussed throughout chapter 4 evaluation occurs throughout the design process therefore the purpose of this chapter is to complete a final evaluation to assess the usability benefits of the final prototype compared to the existing system and ensure that all user requirements have been met by collecting information about users experience when interacting with each of them 6 1 Evaluation activities As discussed in the literature review section 2 6 there are two general classes of evaluation methods formative and summative Formative has been used and discussed throughout th
184. lly handed over to the company to complete longer term testing following their business processes Therefore it has been decided to iteratively develop the prototype using the NET framework rather than common high fidelity prototyping tools such as Visual Basic or Macromedia At this stage both design and implementation are being carried out simultaneously The process will again entail iterative development continuously gathering user feedback This section of the design document will only note any further design decisions that arise from this process the remaining details of iterative development can be found in the next chapter where implementation details are discussed in greater detail There was only one significant finding during user experimentation of a high fidelity prototype this was a potential issue was found with the cross reference table Currently if a user saves the cross reference and moves onto view data but finds a mistake they are unable to re update the cross reference table successfully This is because the design had only considered this to be updated once only Further discussions with the user it saw that an additional requirement was needed to specify that the system should allow multiple updates of the cross reference table This will prevent the need to for the task to be restarted and so increases productivity Although there were no other findings that impacted the requirements specification throughout the implementat
185. lution and consequently led to a negative transfer of existing user knowledge This in turn caused a great deal of user frustration with some users abandoning the system and resorting to completing the task manually The CRM system named OneLilly was deployed to three countries in May 2007 The system to be redesigned is used for the preparation of data to be loaded into OneLilly and will be scoped such that the tool will be used by the UK team only The system generates targeting data this is information that will define which healthcare professionals a particular sales representative will visit the number of visits that should be made and finally the sales quota they are aiming to achieve The role of the system is to generate and output a combined objectives file to be sent to OneLilly using data from a variety of sources including input directly from a user 1 1 Aim With users as design partners the aim of this project is to evaluate and redevelop the OneLilly targeting application with particular attention to improving the navigational efficiency through the user interface The research aim of the project will cover an investigation into how navigational design can influence user experiences 1 2 Objectives The key objectives of the project include e Complete literature review to gain a foundation of knowledge in the relevant domain including usability standards participatory design methodologies and navigation design consider
186. mad fo uere Svo e mE S Cay C 2 Youn Coe Tusry on e Qu v3 QoS mont Wen SOMO CR Ce OYE PK to onswe mak OR S Qe QoS me Ore Gin gt Xe 5 a 8 amity choke Ne sani your Stake corren Ws 2 ARR Cue ba k khane Nna Qc olar aie Do you have any comments on user experience of the current tool in eneral And can you SS ok 6 y y 8 Y Gre suggest anything that would assist the user experience Aer nck Seung 56 S Vox dead ya is cover UR aS odere e bo caus Suain SPO Fives c s neak cM abu to Urargr 2 Eases add Bo e v x 2209 e Parts For ee k wee task MEET Sekt QUE L gt videos oF 6 bar Jou Dak Sege Bour 7 Do you have a list of priorities for enhancements New features gt gui War ace Uf L Coron Lj ve edes SK en SUS 8 Any other comments ne e pe ce YO Vos COS Y cored ees ob Sares gu oe Ton 2 Dn De Ea S A Busiress plans for aii Adanman AA Moo now Q ANO ow So erc 126 Section 2 OneLilly Objective Input Specification Data Objective Record Type OBJT Ref to DD Field Name Affiliate Code Max Length Required Y Format
187. meeting requirements related to providing guidance through the task but at the same time allowing flexibility Small status details such as shading to demonstrate progress through the task to aid location understanding had to be reconsidered to ensure previous breadcrumbs were still seen to be clickable During research and experimentation of such designs some key findings were highlighted e Keep the navigation as simple as possible with minimal links during the design phase particular user attempted to add every action to the breadcrumb items this design was rejected by remaining users due to its complexity and potentially preventing flexibility e Consistency is important such that the same navigation structure appears on every page this was found to be an issue when the administration page initially had a slight variant in the structure this was changed as a result of user feedback e It needs be clear that breadcrumb items are clickable e Each breadcrumb item should use short clear and precise words that are familiar to the user e Keep similar links together It was seen that only links relating to the task formed part of the breadcrumbs remaining links such as accessing help or administration were contained in the left menu e Allow for flexibility For this project this was particularly relevant and as highlighted within the literature the interface must allow the users to recover from errors As seen earlier this i
188. mplete task Experienced Novice Task 1 Task 2 Task 3 Task 4 Task 5 Task 6 TOTAL 12 25 1 50 8 00 26 75 8 80 0 20 57 50 6 50 1 25 10 50 28 40 12 25 0 16 59 06 Table 6 Time taken using existing system 93 Time Taken to Complete Task Minutes Task 1 Task 2 Task 3 Task4 Task5 Task6 TOTAL 1 25 2 00 2 90 4 25 0 23 0 22 10 85 0 26 2 50 2 30 2 90 0 16 0 18 8 30 Experienced Novice Table 7 Time taken using new system Further to this collection of data figure 39 was produced to show the mean average time taken to complete each task Mean Time Taken B Existing Time Taken Minutes Figure 40 Bar Chart Displaying Mean Average Time The most considerable aspect of these results is the total time taken between the two systems Immediately it is possible to see that there is a vast improvement with a mean total time taken of 71 22 minutes compared to 9 21 whilst using the new system It can also been seen that the standard deviation is only 1 38 on the new solution compared to 20 04 on for the existing It is possible to highlight some of the interesting points about this set of results Firstly it can be seen that as expected the most experienced user U1 completed the task within the quickest time for the existing application However note that she actually took the longest within the new
189. n and task this could potentially lead to confusion when returning His suggestion was to move each help video to the relevant page making it clear that it is a video and not written instructions since this might encourage them to view it more This comment was also raised by a user See annotated screens in Appendix C Section 2 therefore this feature would be changed if there were to be another development iteration 106 The discussion also highlighted the fact that users may ignore and therefore not even notice the functionality contained within the left hand side menu the suggestion was to move the restart button to where it would be required import screen This again was raised by users also Appendix C Section 1 so again would be an additional change to make This therefore leaves only basic functions like the logout contained within the left hand menu A final suggestion that Jon made was to only display the admin links both the button on the right hand side and the link on the cross reference page to admin users only Although currently if an admin user attempts to access it they receive an informative error message Jon suggested it would be better to remove the functionality entirely This is a valid point and if the system were to be taken forward this feature would be added in 6 5 Conclusion This chapter has completed an informed evaluation by focusing on three key areas Empirical involving a small scale study subjective t
190. n convention for distinguishing between mandatory and desirable requirements the word shall for mandatory and should for desirable 46 Functional requirements 1 The system shall only allow the UK OneLilly team members to access the system Rationale This is required to prevent unauthorised employees to accessing the system and potentially viewing confidential information Security is not a main concern for this project and if this prototype were to be taken further Lilly would integrate this with their sign on system Therefore the login functionality should be kept simple with the key requirement being usability Source The need for security was raised during the interviews section 3 3 The system shall have an administration area that enables objectives and their corresponding fields to be updated Rationale During the observation we learnt that allowing the user to enter free text will cause errors to occur to prevent this from happening it will be replaced with a drop down box which requires administering Source Step 5 of the observation section 3 5 1 The system should only allow specific users access to the administrator area Rationale As discussed in requirement 2 allowing free text entry causes errors this can be prevented by allowing only specific users to update objectives Source Following Nielson s guideline of error prevention literature review section 2 3 and suggestions for new improvements during interv
191. n lost The new system will also contain automation meaning many of the steps listed in this card sorting session will no longer be required The reason for completing this is to understand how much flexibility the users require and the level of desired flexibility in the new system should have As this session has shown the user is aware that some steps need to be completed prior to others can in the Access solution users found it more difficult to know the exact ordering without referring to the paper based guide After discussions with the users it was clear that the 41 new system should maintain some flexibility in terms of preparing the data but the new system should prevent errors such that it does not allow a user to proceed to another step if there is a pre requisite Using the results from the observation and card sorting session it is now possible to model task within a hierarchal task analysis 3 7 Hierarchal Task Analysis The CARD session has provided a useful insight into the flexibility of the task flow for the previous solution it is now important to bring together both the observational findings and the card sorting session to enable us to analyze the underlying rationale and purpose of what the users are doing This will provide an informed basis for the new system As discussed in section 3 6 creating a HTA for each solution will be of benefit such that the two models can be compared to find similarities or differences that n
192. n the new interface The majority of comments that Jon raised fitted into Nielson s fourth guideline consistency and standards Due to the fact that the screens were designed solely from user sketches some standards were not followed these included the design of hyperlinks breadcrumbs button design and colour matching The standard for hyperlinks is blue and underlined but the new 105 system simply underlines the text only Jon discussed how this could prevent users from immediately recognising that the text is clickable Jon commented that although he cannot see any issues with the current breadcrumbs he suggested it might have been more sensible to follow the standard for breadcrumbs as shown in figure 48 Glossary Categories gt Website Design gt Navigation Figure 48 Standard breadcrumb design He also commented on the good aspects of the systems breadcrumbs these included how it is useful that the current page is displayed in bold and previous pages are shaded green to indicate they are completed as shown in figure 49 Overview Import X Ref Data gt Export gt Figure 49 Breadcrumb design Again his comments touched on how hyperlinks need to be standard this is because it is not immediately obvious that the breadcrumbs contain links Even if the user was aware that they are links the green shading could fool users into thinking that the pages are no longer accessible since they are completed this was also
193. nally the chapter will conclude by detailing the testing that took place throughout 5 1 An Introduction to the NET Framework The NET framework created by Microsoft is a general purpose software development platform that allows different programming languages and libraries to work together seamlessly to create windows based applications The framework consists of two main components the NET framework class library and the common language run time CLR The class library consists of an object orientated collection of reusable classes that can be used to develop a variety of applications The CLR is the foundation of the framework which manages code at execution time figure 31 shows the visual overview of this 70 C VB NET J code code code Compiler Compiler Compiler NET compatible languages compile to a Common second platform neutral language called gree Common Intermediate Language CIL anguage The platform specific Common Language pee Runtime CLR compiles CIL to machine readable code that can be executed on the current platform 01001100101011 11010101100110 Figure 31 Overview of the NET Common Language Runtime It can be seen that high level language compilers for C VB NET and C are provided to turn source code into common intermediate language CIL the CLR then compiles this into machine readable code that can be executed 5 1 1 ASP NET ASP NET is used to build web based applications I
194. nce which you will need to since there is only one card for viewing data but you will have more than one You can also create your own cards if you feel there is something missing Please either write on the cards or paper if you prefer any decisions you made at a particular step e g you selected OBCF and OBJT or you decided that editing was not required I understand that you have limited time but if you could also consider anything that you feel is missing or not right and take any notes on this Maybe you have a new idea Or thought of something that is missing What Information you need to send back to me The CARD ordering you choose either take a photograph or record the card numbers All notes on the decisions you made Any additional thoughts you had whilst completing the task If there is anything you are unclear about or if you have any questions at all then please feel fr to call me Thanks very much for your help Kind Regards Sam Reply email On 27 March 2008 at 17 18 lindsay nicola e lilly com wrote Hey Sam 141 I have added the both mine and Carol s ordering and decisions to the bottom of this email I noticed something that we could be missing or maybe it just didn t fit on the card Card 5 does not have a column to select the call frequency Not sure how we missed this in original designs but obviously as you know this is just added to the SQL code in our current Syst
195. ncluded allowing the user to return to a previous section to make corrections e Where possible provide the user with information on their current location this decision came out during design where the decision was made to make the current link bold Chapter six provided evidence that the implementation of breadcrumbs as a new navigational design was effective in supporting the user to complete their task guiding their journey through the task and as a result the null hypothesis was rejected The evidence found within this project has advanced the state of the current research on navigation to show that with user involvement during design will have a positive effect on the usage and effectiveness of breadcrumbs User choice to include breadcrumbs and the users role in design were possibly pivotal in realising a positive impact of breadcrumb design features 7 2 2 Empirical The empirical project output is the observable evidence from the evaluation As stated within the chapter six conclusions there was sufficient evidence from a variety of sources to conclude that the new targeting application is superior to the existing system The supporting data included time taken number and types of errors made and number of clicks required to reach a specific location or goal A reduction in the number of clicks required in the new system shows that the navigational structure was used to quickly and effectively support the user 110 questionn
196. nctional what the system should do and non functional the system properties and constraints Non functional requirements cover a broad range of different requirements types consequently Preece et al 2 proposed dividing them further into categories as follows e Data requirements capturing the type volatility size amount persistence accuracy and value of the amounts of the required data e Environmental referring to the circumstances in which the interactive product will be expected to operate these include the physical social organisational and technical environments e User requirements capturing the characteristics of the intended user group e Usability requirements capturing the usability goals and associated measures for the new system These categories will be used to define the requirements for the new OneLilly targeting system Preece et al 2 states that One of the aims of the requirements activity is to make the requirements as specific unambiguous and clear as possible to ensure the requirements specification follow this they will follow the Volere process Robertson and Robertson 1999 the template suggested includes the following e Requirement number to uniquely identify the requirement e Description a statement of the intention of the requirement e Rationale an explanation as to why the requirement is necessary e Source where the requirement was raised The listed requirements will follow the commo
197. nd quick to recover from errors Nielson 9 defines usability by five quality components 1 Learnability how easy is it for users to complete basic tasks the first time they use the system 2 Efficiency once users know the system how quickly can they perform tasks 3 Memorability when users return to using a system after a period of time how easily can they establish proficiency 4 Errors How many errors do the users make Are the severe and can they easily recover from them 5 Satisfaction How pleasant is the system to use Nielson also discusses utility which refers to the design functionality does it do what the users need Usability and utility are equally important it does not matter if something is easy to do if it is not what the user wants just as it is no good having a system that could perform the right tasks but the user cannot make it happen because the user interface is too difficult to use Ben Shneiderman 8 makes the same point by stating that usability is about understanding stating and serving users needs Since the needs are the requirements that shape tool they will determine the utility This is stressing the fact that for a system to be usable it must perform exactly what the user desires This project is focused on improving the users experience and therefore ensuring that the system enables the user to perform exactly what they need to As detailed by Preece et al 2 the best way to accomplish
198. nd the reasons behind these thoughts e Tounderstand the changes users would like to see in a redeveloped system All interviews will be recorded using a Dictaphone for later review All interview participants were sent a brief email prior to the interviews outlining the questions and informing them that the interviews will be recorded a signed copy of this along the supplementary notes that were taken can be found in Appendix A Section 1 3 3 2 Questions Below is a list of the questions that were asked of each interviewee 1 What is good about the current tool Are there any features you like about the current tool 2 Are there any features you would like to see retained from the tool Alternatively is there a process or understanding you have from completing the task manually that you would like to be able to see in a new system 3 Do you find the current tool easy to use If so what makes it easy to use If not can you describe why it is difficult to use 4 Whatis bad about the current tool What frustrates you about it What features would you like to see removed 5 What new features would you like to see 6 Do you have any comments on user experience of the current tool in general Can you suggest anything that would enhance the user experience 7 Can you prioritise enhancements and new features 8 Do you have any further comments 3 3 3 Background of interviewees Interviewee 1 Operational data steward OneLilly te
199. nothing was imported T5 Exported one data type instead of all Table 8 Types of errors made Number Of Errors Made Task 1 Task 2 Task 3 Task 4 Task 5 Task 6 TOTAL Table 9 Number of errors made using existing system Number Of Errors Made Task 1 Task 2 Task 3 Task 4 Task 5 Task 6 TOTAL Table 10 Number of errors made using new system 97 Further to this collection of data figure 43 was produced to show the mean number of errors made for each task Mean Number of Errors New N E e E E El 2 E 3 z Existing Task Number Figure 43 Bar Chart displaying the mean number of errors The most considerable aspect of these results is the clear reduction in the number of errors for the new system with no errors being made in both tasks 4 and 6 Also note how the standard deviation is again reduced for the new system this continues to support the assumption that the users performance is better whilst using the new system It is possible to categorise these errors into two main types Validation where the system should prevent the user from completing this particular action and insufficient guidance which has led to users lacking understanding in terms of not being aware of what actions are possible or where they should or can move onto next Such errors have been reduced through the use of breadcrumb navigation within the ne
200. nt of failure since an incorrectly named table will produce incorrect data or not work at all The inexperienced user commented that it was difficult to find which table to rename when following the step by step guide Help was provided to enable her to do this correctly She noted that some guidance within the tool would have made this easier Whilst following the import wizard the user had to ask for assistance as she was unclear on which options to choose She commented that although the wizard as shown in figure 12 provided some guidance it was general rather than specific to her task The user commented that The user described how she the field Haie rep bod 15 remembered being told 3a Import Spreadsheet Wizard z not required in the final about column headers and You can specify information about each of the elds you are importing Select fds in the area below You can then modify fid objectives data so why is it information in the Field Options area bei i eing imported The could not remember EN Field Options Pe e 8 P whether it should have been Field Name BEEN Data Type Text i experienced user edited have Indexed No E Do not import field Skip commented that it is imported for ease of use since it is already within the business objects report but is ignored in the rest of the tool Sales Rep Name Jiv Barrow Neither of the users understood what Indexed meant so just ign
201. ntegrity Due to time constraints only one user experienced completed screen designs for every section the other two users completed 2 4 screen designs Due to this some of the design options were discussed from just one or two sketches ADDING DATA As shown in figure 24 all user sketches followed a consistent design using an upload box to add data into the system Me oC oes p a Bu load a Ce Rte V oe ct 5 G C ENDE emm Wesce P OAT qase A Af witked on evs ec ee lt s Find pex More amp cout EXE ONE eqs e CH CN m ececo Figure 24 User sketches for upload functionality There are a couple of things to note here firstly two of the designs do not contain an upload button and secondly the varying designs of how the upload boxes are displayed One user commented that she simply did not think about the details of data being added to the database and the other user stated that her design would add the data contained within the Excel sheet to the database once the proceed button was clicked In order to satisfy requirements 5 and 6 validation of data contained within the spreadsheet it would be more appropriate to add the data whilst still within the import section so that the feedback is returned to the relevant section Currently there are three designs options for the display of upload boxes 01 Always display both uploa
202. ntrols AddAt i new LiteralControl current as DropDownList SelectedItem Text else if current is CheckBox control Controls Remove current control Controls AddAt i new LiteralControl current as CheckBox Checked True False if current HasControls 159 GridViewExportUtil PrepareControlForExport current Import code behind file Import aspx cs using System using System Collections using System Configuration using System Data using System Ling using System Web using System Web Security using System Web UI using System Web UI HtmlControls using System Web UI WebControls using System Web UI WebControls WebParts using System Xml Linqg using ExcelImporters using System Collections Generic public partial class LillyTargeting Import System Web UI Page protected void Page Load object sender EventArgs OBSQPanel Visible SharedState shouldDisplayOBSQ lblFileError Visible false lblValidationErrors Visible false protected void saveButton Click object sender EventArgs Response Redirect Cross Reference aspx protected void cmdUploadCallFrequencyFile Click object sender EventArgs string filename C ToUpload CallFrequencyData xls TODO implement checking of the file type here if fulCallFrequency HasFile fulCallFrequency SaveAs filename SelectedRatingsImportResults results SelectedRatingsImporter ImportSelectedRatingsDataFrom
203. o gather user opinion and analytical where a HCI expert was asked to perform discount usability evaluation Each technique has provided different but complementary data to comparatively evaluate the new system against the old Results and discussions have also enabled the question of whether breadcrumbs as a navigational aid have improved end user experiences To conclude the section will begin by relating back to the hypothesis Null hypothesis Ho There is no difference between the new and existing interface Hypothesis H The new application developed for OneLilly targeting is superior to the existing system Through each of the evaluation methods and results described throughout this chapter it is possible to reject the null hypothesis and find sufficient evidence from a variety of sources both qualitative and quantitative data to accept the alternative hypothesis It is possible to make the conclusion that the new OneLilly targeting application is superior to the existing system and through analysis and discussions throughout this chapter it is also possible to say that the approach adopted to improve navigational efficiency was successful Although as stated throughout the evaluation this was a small scale study only since there were limited participants available Further studies would need to take place to be able to draw more extensive conclusions but clearly not possible with target population as they are limited in numbers In t
204. o invoke the query to be ran 01 Validate on Import O2 User specify at cross reference 01 When user clicks Export 02 When user clicks Send to OneLilly 67 Criteria requirement number 22 Flexibility 27 Easy to navigate 18 Treating each transaction separately 21 Help atthe current location 22 Flexibility 26 Error prevention 27 Easy to navigate 4 7 9 10 11 15 Specifying core functionality 1 2 3 5 6 8 13 18 Supporting data integrity As stated in the introduction and found throughout this chapter the previously stated requirements specification is open to changes Here follows a list of those additional or modified requirements Modification to requirements 11 The system shall allow a user to edit or delete any entries within each data type Rationale The HTA s figure 3 12 and 3 13 shows that a user is currently able to update any information within the generated data users requested that this functionality be maintained since there may be a need to edit data for an individual sales representative Source The final step of each HTA model figures 16 and 17 and the context section of the interview section 3 3 4 Additional functional requirements 29 The system shall allow a user to specify which data types they wish to be generated Rationale This provides the user with flexibility making the task easier to complete Source User designs section 4 3 sketched functionality
205. o many spontaneous suggestions for improvement Sometimes if two participants are working together produces more talking as one participant explains procedures and decisions to another 4 This project will use field studies these are conducted within the user s everyday working lives they are typically conducted to find out how a product or prototype is adopted and used by users in their normal circumstances Evaluating how people think about interact and integrate the product within the setting it will ultimately be used in will give a better sense of how successful the product will be in the real world 2 The nature of this situation means that you can observe interactions between systems and individuals that would have been missed in a laboratory study Field studies can be used to 5 Help identify opportunities for new technology 6 Establish the requirements for design 7 Facilitate the introduction of technology or deploy existing technology in new contexts 8 Evaluate technology The basic techniques for data gathering in field studies are interviews questionnaires and observations The studies to be reported here relate to the improvements in the user interface design Both of the methods described above have limitations Shneiderman 4 highlights some of these Testing in this way does emphasise first time use and can have limited coverage of the interface features Since usability tests are often short it is difficult to ascerta
206. of final data set e Exporting final file directly to the required server e Userroles with the functionality to add new users and update their role 45 e Saving the current state and when logging back in being returned to that particular users last state e Providing detailed statistics e Provide user guidance throughout the task 3 9 Project scope The benefits of the new OneLilly targeting system will include increased navigational efficiency improved business processes such that all tasks are completed within a single application and an easier to use interface requiring no technical knowledge These benefits will then lead onto increasing the number of potential users and overall productivity of the targeting process The aim of the project is to produce a working prototype of the OneLilly targeting application this can be satisfied by developing the application for a single sales force within the UK this will be LPG Section 3 4 4 of the interviews highlighted some concerns for viewing previous data and keeping an audit trail of previous data that has been produced Importantly the aim is to develop a new usable interface rather than auditing functionality so this is not a major concern since stakeholders stated that this can be fulfilled by allowing a user to save every file produce and storing the file elsewhere 3 10 Requirements specification In software engineering two types of requirements have typically been identified fu
207. of the data however The corresponding OBJT records for the sales quota data are appended to the records exported from the targeting tool When all data has been imported and verified Export button is used to format the data for loading into OneLilly The user is informed of the location in which the generated file is placed Two final verification checks are performed prior to data loading e That the expected number of records appears in the file e That the end character is present at the end of every record The inexperienced user used a different method to export the tables as she has previous experience of Microsoft Access The remaining steps were completed as expected It was decided that a new system should allow for this step to be removed Users should import the sales quota data and all verification checks should be completed within the system It was however requested that it be possible to export the data in the Microsoft Excel format or view the final file in the format it is as sent as This will enable users to complete any remaining checks they wish as well as enabling them to retain all data files produced 3 5 2 Observation summary Observing users with the existing tool has given an insight into the targeting process and has highlighted the following things 39 e The users relied heavily on the step by step guide and were confused when it did not correspond to the tool e Navigation of the system was a problem fo
208. or delete a particular entry Secondly she stated that she does not believe that there is a need for an add button since all entries are based on data added and therefore if data is missing it should be generated from the beginning of the process Users agreed with the comment raised and decided that there is only a need for Delete and Edit this new decision will impact requirement number 11 since the add functionality needs to be removed In terms of how the data itself should be validated there are two options 01 Only allow valid data to be added to the system therefore removing invalid during import 02 Allowing the user to specify which entries are invalid at the cross reference phase as the existing system does The second options provides additional flexibility such that if there was a special case selected ratings or the standard ones were to be edited then it would still be possible to use the system 59 When this was discussed with the users it was clear that such a circumstance would be very rare and if the standard selected ratings were to be edited then this would be a one off case only meaning the system itself should be updated The first option makes the task simpler for a user and follows Nielson s heuristic to prevent errors were possible EXPORTING AND SENDING RESULTANT DATA Again two of the users sketched designs for the export functionality as figure 28 shows there is a slight variation in the use
209. or partially complete the final task cannot be completed until all are green Error prevention Erroneous tasks should not be possible for example running a query before its prerequisite should not be possible or exporting cannot be done until all other steps have been completed Recognition rather than recall Flexibility and efficiency of use During the interview Jason discussed how the help was good in some parts of the system i e before running a query it displays the following prompt Microsoft Office Access You are about to run append query that will modify data in your table Are you sure you want to run this type of action query For information on turning off confirmation messages for document deletions click Help Show Help gt gt ves ve nee This is useful as an accelerator since there is additional help available for users that need it Aesthetic and minimalist design At numerous locations during the system message prompts contain information which is irrelevant or rarely needed Messages should be relevant and only shown when required Help users recognise diagnose and recover from errors At start up ofthe system the following prompt is displayed 131 y y I A Security Warning Unsafe expressions are not blocked Do you want to block unsafe expressions w o e Itis unclear to the user what this is and what the
210. ored this option Figure 12 Prompt for Importing Data The need for user guidance was raised during the interviews Again this features in Nielson s heuristics and will be taken into account when forming new requirements STEP 3 Because the Business Objects report is not always in the same format it is important to check that the imported data has the same columns as the existing table This is made easier by retaining previous data All columns must be the same since queries refer to specific fields The experienced user demonstrated how she would correct it if the fields did not match which required her to use the design view for the table The inexperienced user forgot to complete this step as it was not on her guide sheet Consequently the next step failed and the tool displayed the prompt shown in figure 13 which did not help the user diagnose the problem 35 Bopen We X 22 Objets Name Description Modified Cj Tables create query in Design view a Create query by using wizard AGA PR CS Forms 1 Buld Rating20bjective SelectedRatings Con Integration Id i Objectives Reports 81 Create OBJT Parent objective h Pages MAKE LPG OBJT Es amp Macros MAKE ONELILLY OBCF E Sb oscr 16 10 2007 15 45 1 g Modul Gb Rating Q 31 07 2007 15 01 2 Groups Figure 13 Prompt Displaye
211. ould chose OBJT first 7 Looking at my results 8 No editing required Go back to card 6 Now choose OBCF 7 Looking at my results 8 No editing required 10 Open my file I would save it and open it from there 11 Export file Carol 1 OBJT amp OBCF 2 I would select Q3 and 2008 and tick the boxes for OBJT and OBCF 3 Add my file New card to view upload feedback Decide to continue 4 A gt Zyp Sch and B gt Zyp Bip 5 Select the Zyp Sch from the drop down box will the rest the fields b automatically added I would then click on Add objective and add Zyp Bip to that R as well 6 I would chose OBCF first 7 Look at my results 8 No editing required o back to card 6 Now choose Look at my results No editing required G 7 8 142 10 Open my file 11 Export file Nicola LINDSAY BSc Hons OneLilly Data Operational Steward Lilly UK Embedded image moved to file Lilly House Priestley Road pic22532 gif Basingstoke Hampshire RG24 9NL UNITED KINGDOM Tel 44 1256 775977 Fax 44 0 1256 775858 www lilly co uk Section 3 Email communications for selected ratings validation requirements On 07 March 2008 at 13 18 lindsay nicola e lilly com wrote The checks required on imported selected ratings are The position division and contact integration ID fields all have to be less than 15 characters and cannot be empt
212. ource System 20083821 083800 20083821 083800 20085522 125526 20085522 125526 20082924 022907 20082924 022907 MANUAL Q1 2008 Terr MANUAL Q1 2008 Terr MANUAL Q1 2008 Terr MANUAL Q1 2008 Terr Edit Delete MANUAL Q1 2008 Terr Q1 2008 Terr Edit Delete MANUAL Nos i bes nik MEY d rf Yo Yo ES pos tor Flag isl foo T Shu ud be a nrt po pori ume e 7 ow O AM Ex cA y prev ir dus Lier along et en tle Screen Moye eusse some of Ue ima columns they emin Uten po ol gt Excl Jive wwe room on we sueann 148 Export BSPL You are logged in as sam cooper Export OBJT Export OBCF Admin Help Restart Log Out x Export OBSQ Export All Section 3 Draft Test Plan 0 ex Shrouds Say vi some users MJ Sore Send to OneLilly Qer Sent uun lt 1S baog ovt AN serye as Je export Vei sH Mgune logout Co This is the test plan that users were provided with during and at the end of implementation phase Step Task Test data Expected Result 1 Complete Select 2009 Q1 Successfully selected EEE Choose to generate Successfully selected OBES OEE When click move to Import screen Click Save amp Proceed 2 Import Click Browse amp locate test data On clicking browse a file open box Click upload will suc
213. ow participatory design techniques have successfully been used to appropriately and accurately answer design questions All user feedback has directed further 68 iterations and this will now continue throughout the implementation stage The next chapter will now describe the development environment in which the new OneLilly targeting application is being developed in and give details of the technical background 69 Chapter 5 Implementation and Testing The aim of this chapter is to describe the implementation details of how the findings from previous chapters were used to develop the new OneLilly targeting application It will begin by giving an introduction to the NET framework and discussing why ASP NET was chosen as the development environment It will then move onto an exploration of a recently added extension to the framework language integrated query LINQ which proves to be particularly useful for this implementation The chapter will also describe the database set up class structures and provide detailed descriptions of how the main features were implemented As discussed numerous times development is an iterative process so participatory design techniques will continue to be used allowing the interface to be constantly evaluated The feedback obtained will continually be fed back into the design and new iterations A brief discussion of the feedback obtained and subsequent changes made in the interface will be discussed and fi
214. owing text shows the instructions that were sent to each user Imagine the cards are now the system and you need to generate OBCF and the corresponding data only for the 192R sales representative Simply arrange the cards in the order you would expect like to s screen whilst you complete the targeting data you may use a card more than once which you will need to since there is only one card for viewing data but you will have more than one You can also create your own cards if you feel there is something missing Please either write on the cards or paper if you prefer any decisions you made at a particular step e g you selected OBCF and OBJT or you 62 decided that editing was not required I understand that you have limited time but if you could also consider anything that you feel is missing or not right and take any notes on this Maybe you have a new idea Or thought of something that is missing A full copy of the email sent detailing all instructions the users were given can be found in Appendix B Section 2 4 3 1 Results A copy of the results returned via email can be found in Appendix B Section 2 As seen within this reply the completion of this activity has resulted in the findings of a new requirement one of the users found that the design of the cross reference currently does not include a column to specify the call frequency Comparing these results with the discussions around breadcrumb
215. pendent Variable in which the Preference ratings participants performance is Number and types of errors made per task measured Number of clicks to locate a specific piece of information Number of users completing a task successfully Time taken to complete a specific task Table 3 Experiment variables Data collected will be both qualitative collecting detailed information and quantitative which involves the analysis of numerical data Together both data types will provide collective evidence to test the hypothesis A range of additional variables may exist which could affect the experiment these include participant experience and mood the time of day or the environment in which the experiment is completed within Even the ordering in which the applications are tested could have an affect Although these will not have an adverse affect on the results it is important to be aware and control them where possible Johnson 42 states that the strength of an experiment depends on the experimenter s ability to guarantee that only the manipulated variables are permitted to vary systematically if there is an uncontrolled variable varying along with the intended one it is known as a confounded variable Many researches have suggested ways to deal with the effects of these variables Keppel 43 states that the three most common ways and many experiments include combinations of them all are 1 Hold the variable constant throughout the exper
216. r knowledge of the existing system where data is viewed and validated separately The designer of figure 21 commented how she presumed the system would complete basic verification on behalf of the user and any edits being required from the user can be done whilst viewing each data type This discussion concluded with an agreement that user validation would be completed by the system and user verification can be completed whilst viewing each data type this meant that this section could be removed There was also a comment raised about the need for an overview section a sketch of this section can be seen on the left hand side of figure 19 after all users reviewed the design for this page it was agreed that the system should allow the user to specify which data types the system will generate This is additional requirement which will need to be added to the requirements specification 54 Finally it is useful to note that all users designed the links to the sub tasks on the top of the page and auxiliary functions were contained within a left hand side menu User s were asked whether they considered having sub tasks contained within the side menu but all agreed that having it on the top of the screen made more sense since it is a guidance through the task whereas a side menu may be thought as simply options In summary the options discovered here were 01 Mind map 02 Breadcrumbs 03 Side Menu The current decision is to progress with bread
217. r the inexperienced user e Messages boxes or prompts are often ignored as they are not understood or do not contain relevant information e The task depends on the cross reference table being completed properly but this is a complex task with which users struggle It is clear that users of the existing tool rely on their understanding of the manual system that was in place prior to it and the fact that task order is different means that their knowledge does not directly map onto the tool Consequently an analysis of the difference in task flow between the two solutions will now be completed 3 6 Card sorting session Because the targeting process is so extensive and flexible it was decided to scope this session to deal purely with generation of business plan data Each participant was asked to compile a list of the steps required to complete the BSPL data prior to the session An informal discussion session was then held to enable participants to compare describe and discuss the reasoning behind their choices Three users with various levels of experience with the task participated in the session One of these was the primary user who also participated in the observations documented previously 3 6 1 Results and analysis Each participant produced a different ordering of tasks Immediately each user stated how flexible the manual solution was and how each time they complete the task they alter the task flow Although it is important to
218. ractions to be reduced where as novice users would like to be able to see more of the system allowing them to gain skills and knowledge about the system to help them become experts Novice users require 10 the interface to be easy to learn and guessable this means that the users who do not have previous experience can use the interface Some of the heuristics above support this for example recognition rather than recall and consistency J Wu 25 also provides some guidelines and recommendations for accommodating for both novice and expert users e Design the top level of your interface to be simplified as much as possible e Provide accelerators for experienced users e Show the corresponding short cut of a normal menu item within the menu item to enable users to get to know them e Provide online help to assist novice users without getting in the way of experts e Allow experts to change default settings where appropriate Novice users may require additional support so the next design consideration is to gain a background understanding of how this can be done 2 4 3 Providing user support There are many ways of providing user support and it is up to the designer to decide on the most appropriate way for any given system There are a number of things to be taken into account e Design of the user support should not be seen as an add on to system design it should be designed with the rest of the system e The same modelling and
219. rexa Cymbalta Cymbalta f Cymbalta Zyprexa UK LPG 011H__ __ UKLPG 0113 _Gemzar Semzar Gemzar Humalog HLOGT HLOG T2 jUKDIAB O11M _ HMIX25 T1 HMIX25 T2 HMIX25 T2 UK DIAB 011W UK ONCO 011G UKDIABETES UK DIABETES2 Humalog 25 Humalog 50 HMIX50 T1 HMIX50 T2 HMIX50 T2 UK DIAB 0118 UK DIABETES3 Byetta 1 2 T BTA T2 Humalog Mix50 Humalog _ HLOG T2 HLOG T3 HLOG T3 Humalog Mix25 HMIX25 T2 HMDQS T3 HMIX25 T3 HMIXS0 T2 HMIX50 T3 HMIX50 T3 Byetta BTA T2 BTA T3 BTA T3 HLOG T3 HLOG T4 HLOG TA Humalog 25 HMIX25 T3 HMbQS TA Humalog _ Humalog Mix50 Hu HMIX50 T3 HMIX50 T4 HMIXSO T4 Byetta BTA T3 133 BTA T4 BTA T4 Section 6 OneLilly Entity Specification Objectives 5 Data file preparation 5 1 General file specification The file is an ASCII delimited one with Field separator End of record Separator plus standard CR or and LF character depending on source Operating System MQ and Informatica will take of conversion if required The Objective file will contain 4 record types OBJT OBTM OPRD and OBCF see section 5 3 for more details
220. ro massages wer Sark Sie oX hore Ute ed yo We boan oct e hE Kok POSS vu ev o this 4 yee C S Koes vx Mace sense bec ose N wow SRO 7 Slim sd A Leas OQ SX rk Wan ObcE rey COM oaa tbe ogpoluce oven Nso bk Aea inStruaccion varg Sig Wey Qo CHEM Cr wwe 5 120 4 Whatis bad about the current tool What frustrates you about it What features would you like to see removed More wro Or Mong DEE eran AOAMOASES C oux SINS Wank To be able NO Norwmguakon very har As ow c S A Sme SQL access 9a 0o X o0 m verg UO oS Oe o vori 2 5 name AS Ernved pecawrse tte SOL cede Ardea doae wo SEC Oe VOT Au Length dwikon nomas X 2 ajase ode o borcom Sac SORES Aree Ska Sx AnaS te Ya cou TO See tas x cae bua aso qo Xo 5 mass changes 5 What new features would you like to see RAS BINS tas So wv ACLS orrek fer uos of ere Cose Owke gesso lec perry EEK KNOY aer Woo ny read ve Jot d Cr gt on S c fx RON 7 CoA ono co Ayax Amonun vent T7 Oeecef oF Gee s cho bormec gesino cl NO rox QOSS wvle cO Ves a CC cee c WFQ AGS min BOF SO Xu POW aoa FO o OR MASA nS _ E gorvking nema Sees L r
221. rowse for a file lblFileError Visible true protected void cmdUploadSalesQuotaData Click object sender EventArgs e Cross Reference code behind file Cross_Reference aspx cs using System using System Data Linq using System Collections Generic using System Ling using System Web using System Web UI using System Web UI WebControls public partial class LillyTargeting Cross Reference System Web UI Page protected void Page Load object sender EventArgs e to prevent the page from being reset if Page IsPostBack return TargetingDataContext db new TargetingDataContext unique position amp letter items needed in x ref var ratingList from sr in db SelectedRatings orderby sr PositionLetter select new sr PositionLetter sr SelectedRatingCode lt Distinct var dropDownItems from obm in db ObjectivesMappings orderby obm Objective select obm rptUpdate DataSource ratingList 161 rptUpdate DataBind for each record returned in ratingList query foreach RepeaterItem i in rptUpdate Items String posLetter Label i FindControl lblPositionLetter Text Trim String selRating Label i FindControl lblSelectedRating Text Trim String callFreq DropDownList i FindControl callFreq Text Trim CrossReference current null try posLetter amp amp p UniqueSelectedRating selRating catch Exception current null DropDownLis
222. rrect order 1 Do not give users direct access to queries Specifying an incorrect quarter 1 Allow user to select quarter 2 Automatically set quarter based on current date Table 2 Common errors and possible design solutions Finally although the design of the help is not contained within these initial prototypes it was considered that this could either be in written form or using videos As previously discussed during the interviews users preference is video format In summary the options discovered here were Q2 01 Manually click through menu items 02 Proceed button redirecting to the next most logical sub task 03 A combination of both of the above Q3 O1 Written instructions 02 Video 56 The options for question four were specified in table 4 1 As with the previous section the conclusions drawn from this section will now be taken forward to the next prototype iteration 4 2 3 TaskSupport As described throughout this project it is absolutely vital that the system design effectively supports all tasks For this reason the questions to be asked are split into each of the sub tasks required to enable a user to reach their end goal of delivering the generated data in the correct OneLilly format Again Throughout this discussion it was noted that the criteria needing to be satisfied these are the following requirements 4 7 9 10 11 15 Specifying core functionality 1 2 3 5 6 8 13 18 Supporting data i
223. rs but the HCI expert discussed how they could be used more appropriately See figure 9 for an example the image shows different query types and their corresponding icons Name n Create query in Design view 2 Create query using wizard Build Rating2Objective EF Build Objectives 1 Create OBIT Parent objective Gr MAKE LPG OBJT Cr ONELILLY OBCF E el Rating Q Figure 9 An Example of Query Icons 31 The first thing to note is whether or not the icons give an accurate reflection of the associated activity associated with it Where possible icons should be based on existing traditions or standards However care should be taken not to confuse users by using a standard icon where a non standard activity is likely to be performed An example of this is the append query icon third from the top Upon first examination and without knowledge of the icons Microsoft Access uses it would be natural to assume that a plus sign means creating a completely new data set rather than appending to an existing one It was pointed out that this particular screen could have been improved through grouping queries by type rather than ordering them by name This is especially important since some query names can be very similar Now that the initial data gathering phase is complete it is now possible to begin modelling this is done using task analysis 3 5 Task Analysis As stated in chapter 2 a task analysis is use
224. rs before making the decision on which to use Within groups will use the same participants for all conditions This experiment will have two conditions one is the control factor which will be the existing system and the second will be the experimental treatment being the new system Between groups design will use different groups of participants in each condition meaning that participants will receive only one of the different conditions The advantages of between groups design include the fact that they are simple to understand easy to design and analyse and due to the fact that there are different participants transfer of learning effects is also prevented Although it must be possible to match participants to reduce the effects of individual differences related to their experience A limiting factor of between groups is that it requires more participants meaning that this option has to be chosen when there are insufficient participants Sometimes it is necessary to complete a within groups experiment since participants act as their own control this enables a very good baseline from which to judge differences between the systems that does not relate to individual differences between the participants It should be noted that there are ways to overcome these transfer effects this could include giving participants a break or completing some other distraction task between the experiments this will reduce the memory for transfer Another option is to us
225. rs preferred options EARRORTY en Suet Ee cn ougan Erg Cuck te eeu Riess Lomoered Aaka torfe Cocmex lt Kk hefe Exe do SN Se e One ESS Figure 28 Export designs Comparing the two designs there are two differences firstly the second design allows the user to separately view each of the data types in the correct format and secondly the first design contains an option to view the final file within Excel From further discussions it was seen that the ability to view each data type separately will not being any additional benefit than viewing the complete file The ability to export the completed file to export turned in a debate since one user the main user of the existing system stated that this functionality is required to allow further checking to take place i e it is possible to sort filter within Excel whereas the other users stated that they would not use this functionality since they have already had the ability to export and check data in previous sections In terms of the question of how data should be exported there are two options 01 When Export is clicked the completed is placed on shared server AND automatically sent to OneLilly since the user has the ability to view what will be sent before export is clicked 02 Keep functionality separately such that it is saved to server when Export is clicked and sent to OneLilly when Send is clicked The f
226. s using System Collections Generic namespace Excellmporters public class SelectedRatingsImportResults private List lt KeyValuePair lt int EISelectedRating gt gt ratingsWithValidationErrors private int numberOfRatingsSuccessfullyValidated private int numberOfRatingsSuccessfullyImported public List lt KeyValuePair lt int EISelectedRating gt gt RatingsWithValidationErrors get return ratingsWithValidationErrors set _ratingsWithValidationErrors value public int NumberOfRatingsSuccessfullyValidated get return numberOfRatingsSuccessfullyValidated set _numberOfRatingsSuccessfullyValidated value public int NumberOfRatingsSuccessfullyImported get return numberOfRatingsSuccessfullyImported set _numberOfRatingsSuccessfullyImported value Exporting Data GridViewExportUtil using System using System Collections Generic using System Data using System Configuration using System IO using System Web using System Web Security using System Web UI using System Web UI WebControls using System Web UI WebControls WebParts using System Web UI HtmlControls using System Runtime Serialization Formatters Binary public class GridViewExportUtil 158 public static void Export string fileName GridView gridview HttpContext Current Response Clear HttpContext Current Response AddHeader content disposition string Format attachment filename 0 f
227. s never worked with SQL previously It is also useful to note that both users said that they do not read the message boxes that are displayed before the query runs commenting that either they did not understand them or the messages were uninformative At this stage the users discussed how they rename the table to enable them to save the previous data This was also discussed during the interviews They stated that the reason it is desirable to save previous data is because it can be used to check the generated data helping them verify that the query ran correctly and that this functionality should be maintained in a new system A new system should be entirely GUI based and not allow users direct access to queries This means that there will be a requirement to select for which quarter data is being generated STEP 7 This step is to perform high level checks on the generated OBCF table in order to look for significant problems with the data The user describes it as giving her a sense of security that the tool has completed what it is meant to There are two checks carried out by every user at this stage 1 Ensuring that the divisions correctly correspond to the positions For example positions ending in the letter R must have the division UK LPG RSP This is done by sorting on the position field and completing a visual check e Itis useful to note here that this check was completed before the current tool existed and in the time sin
228. s such that a non experienced user is able to use the tool e Include guidance through the task possibly having a bar across the top informing users of the stage they are currently at like shopping online checking payment e Include online help facilities throughout the tool Enable a user to get context sensitive help if they desire e Include training videos illustrating how to complete various tasks for users who desire this Suggestions for new features can be summarised as follows Include the ability to complete the necessary tasks for all sales forces within one tool rather than having multiple instances Improve output options for example sending generated data directly to the server where it is used Provide the ability to save work in progress and return later to a task 29 e Provide the ability to keep an audit trail of what data was sent to OneLilly and by which user e Where possible prevent free text entry and potentially only allow specific users to edit static fields e Provide key statistics about generated data upon completion of the process this could include a record count for e Each data type e Invalid records e Each division e Each sales force Suggestions and comments about maintainability and support can be summarised as follows e Allow the output file specification to be easily modified to accommodate new software releases e Either integrate the tool with Lilly single sign on
229. s took a greater amount of time at each task Although the exception being at step 4 verifying where novice took longer this could be due to experienced users being more familiar with the data and checks required 95 Another highlighting aspect of U1 results is the time taken to complete the first task the overview or preparation of data when this was discussed during the interview she commented that she takes her time atthe start to check all variables thoroughly She described that from her experience spending greater time checking at the start prevents errors running through the task and saves her time in the long run This would make sense since a novice user would not be aware of what potential errors could occur and therefore what they should be checking so they will simply complete the task and move on It can be seen that task 4 verifying generated data in the existing system takes significantly longer than the rest of the tasks As shown as part of the observation in chapter 3 and comments from interviews data integrity and validation was an issue due to the fact that the system did not have any built in validation and it allowed a user to freely access and edit queries This is a key area that has been improved within the new system since users have restricted access and all data is validation on import and during user editing This has provided users with more confidence in the data integrity which has led to the reduced verification
230. se of this tool this does not support the usability however A comment was made about Access allowing a diagrammatic view of SQL queries This is useful because users and business integrators are required to manually update SQL this is difficult and the diagrams assist them The currently the tool is made up of seven different databases for each sales force each of which is set up slightly differently causing confusion for users Interaction Every interviewee commented on the navigation of the tool There is no logical sequence of actions and this makes it extremely difficult for new users to understand and use Some of the comments made regarding this included 27 e Thesequence does not follow that of the previous solution e Undesirably the system allows you to navigate anywhere and edit anything in any order e tis very difficult to navigate around or understand what the next action should be and users always need instructions in front of them e tis not intuitive e A good technical understanding is needed to follow it through All of these comments suggest the usability of the existing tool is very poor Users communicated that is extremely difficult and near impossible to work out the process which must be followed to produce the correct output without being given a considerable amount of guidance and training thus restricting the number of people who can support the primary user One of the interviewees made the commented that
231. sesnsesesesnsssssesassasseeassacansassaeaneaceaeanes FIGURE 9 AN EXAMPLE OF QUERY ICONS FIGURE 10 OVERVIEW OF THE TARGETING PROCESS ssssssssssssssssessesscsecsssscsscssecssesusacsscsnsesescsnsesesesnsssesesasssesesassacsesassacansassacseaeeseanes FIGURE 11 NOTATION USED FOR LPG SALES FORCE ssssssssssssssssessesscsecsssscscscsscsucsusecsscsnsesescsnsesescsnsssesesusssesesussassnsassscansessataeaseaeanes FIGURE 12 PROMPT FOR IMPORTING DATA FIGURE 13 PROMPT DISPLAYED AFTER A QUERY HAS RAN ussesssssessessssseesseseesnseseeneeseeaeessessesnseseenseseeaeeaesteaneesesneeaeeneeatenteaeeseaneeaeenss FIGURE 14 CROSS REFERENCE TABLE AA aaa E a T a EE AEE FIGURE 15 USER S RESULTS FROM CARD SORTING SESSION eee eee eere ee rten teretes tates FIGURE 16 HTA FOR THE ORIGINAL SOLUTION eene rne nnt tnntn retentu tentata stata ta tta taa sa Do Posso Do FIGURE 17 HTA FOR THE ACCESS SOLUTION ssssssssssssssssssssssessesscsessesscsucsnsssssesusacsscsnsasescsnsesescsnsesesesnsesesesnsssesesassaesesassatansassacaeaseaeanes FIGURE 18 THE NOTATION ec FIGURE 19 EXAMPLE USER SKETCH acs tok ree noe neo aS aed dean o sac FIGURE 20 NAVIGATIONAL STRUCTURE DESI
232. specific piece of information or complete a simple task within the application Results are shown in table 11 Note The instruction sheet see Appendix D Section 1 listed 4 tasks but the results do not contain data for the final task the reason for this is because it is not possible to edit an individual record within the existing system meaning it is not possible to complete the comparison for this particular task 99 Number of Clicks Existing Task 1 Task 2 Task 3 Task 1 Task 2 Task 3 4 11 17 21 28 Table 11 Comparing the number of clicks required Again further to this data collection figure 44 was produced to show the mean number of clicks required to reach their goal Mean Number of Clicks Existing Number of clicks Task Number Figure 44 Bar Chart displaying the mean number of clicks Although it should be noted that this is a small selection of tasks itis possible to see that generally a participant was able to reach their goal quicker easier using the new system As figure 6 3 shows the most significant difference is task 3 updating the cross reference the main reasons for this difference is due to the fact that within the existing system the user is required to manually delete existing data and re run queries In the new system this is completed automatically when the save button is clicked Again it is possible to see that the standard deviations are larger for the existing s
233. t is part of the NET framework and is built on the common language runtime allowing programmers to write ASP NET using any language which can be compiled to Common Intermediate Language Using the NET framework drastically reduces the amount of code required to build applications because of the amount of functionality available for free in the framework such as user and role management In addition to this the availability of rapid application development tools Visual Studio with visual designers for initial layout supports iterative prototyping It has therefore been decided that this is appropriate for the development required Some of the additional advantages of using the NET framework and in particular ASP NET include e ASP NET makes development simpler and easier to maintain with an event driven server side programming model e The source code is compiled the first time the page is requested Execution is fast as the Web Server compiles the page the first time it is requested The server saves the compiled version of the page for use next time the page is requested e Easy access to ODBC data sources for working with Excel files e Easy integration with existing technology in use at Eli Lilly 71 ASP NET also contains powerful database functionality in order to facilitate interfacing with a database and has particularly strong links with SQL Server The recently added LINQ to SQL Language Integrated Query extensions to the
234. t list DropDownList i FindControl cmbObjectiveMapping add items to the objectives drop down list foreach ObjectivesMapping m in dropDownItems ListItem i new ListItem String Format 0 1 1237 m Objective Trim m Product Trim m Indication Trim m ObjectiveParent Trim m entry pk ToString Trim if current null if current ObjectiveMappingID m entry pk l Selected true list Items Add 1 list EnableViewState true call frequency drop down list DropDownList callFreqList DropDownList i FindControl callFreq ListItem item new ListItem 1 ListItem item2 new ListItem 2 ListItem item3 new ListItem 3 ListItem item4 new ListItem 4 callFreqList Items Add callFreqList Items Add callFreqList Items Add callFreqList Items Add item item2 item3 item4 callFreqList EnableViewState true protected void cmdSaveCrossReference Click object sender EventArgs e TargetingDataContext db new TargetingDataContext bool crossReferenceAlreadyHasData db CrossReferences Count gt 0 to prevent duplicates as the user is re saving if crossReferenceAlreadyHasData db CrossReferences DeleteAllOnSubmit CrossReference db CrossReferences foreach RepeaterItem i in rptUpdate Items int test i ItemIndex CrossReference r new CrossReference r PositionLetter Label i FindControl lblPositionL
235. t users cannot tell you what they want but when they see something and get to use it they are soon able to explain what they want 2 Therefore as the final bullet points suggests once initial information has been collected about the task and the user s opinions it is worth building iterative prototypes A prototype can be a paper based outline of a screen or a set of screens an electronic picture a video simulation a cardboard mock up or a piece of working software It allows stakeholders to interact with the envisioned product to gain some experience explore imagined uses and suggest improvements Lichter 32 defines prototyping as an approach based on an evolutionary view of software development affecting the development process as a whole It involves producing early working versions prototypes of the future application and experimenting with them It provides a communication basis for discussions among all groups involved in the development process 15 especially between users and developers Floyd 33 discusses the different goals of prototyping e Exploratory prototyping is used when the problem is unclear initial ideas are used as a basis for clarifying what users and management want It helps the developers gain an insight into the application area and the user tasks e Experiment prototyping focuses on technical implementation of the development goal it helps ensure feasibility of a particular application e Evolutionar
236. tType application ms excel The same concept is used when exporting the end file but the ContentType is set to text plain rather than Excel 83 Appendix E contains the full listings of each of the main classes to see the full set of code please refer to the attached CD A description of how participatory design was used throughout this iterative implementation will now be given 5 5 Participatory Design at the Implementation Phase Due to time and location constraints it was not possible to complete participatory design directly with users instead email and telephone communications had to be made throughout the implementation phase During the development the application was locally hosted to enable users to freely access and test it throughout next section describes this in greater detail This provided continual user feedback which successfully guided iterative development ensuring that the system worked as expected Feedback obtained in a very of ways including telephone calls whilst the user was using the system email feedback an example can be seen in Appendix C Section 1 and annotated print screens Figure 39 shows an example of one of the ways in which users provided feedback further sketches can be seen in Appendix C Section 2 You are logged in as sam cooper Please upload the Business Objects report for sales representatives selected ratings Admin Upload an Excel file Thus s useful Help valida
237. tanding the user s expectations means the designer needs to understand the role of the computer system This can be done using task analysis it allows the major task to be broken down into sub components this will continue until the task can no longer be broken down into individual sub components The process of user centred design obviously requires interaction with a number of stakeholders and as discussions earlier in this chapter has already shown the design of a usable system can only be successful by learning interpreting and understanding the user s needs There are a number of methods that can be used in the requirements phase to begin to understand users one of the most commonly used techniques is task analysis 2 5 2 Task Analysis Task analysis is mainly used to investigate existing solutions and is used to establish a foundation of existing practices on which to build new requirements or design new tasks It analyses the underlying rational and purpose of what users are doing trying to achieve and how they are going about it Task analysis is a term that covers techniques for investigating cognitive processes and physical actions at a high level abstraction and in detail The most widely used version is Hierarchical Task Analysis HTA another well known technique is GOMS Goals Operators Methods and selection rules 2 HTA was originally designed to identify training needs itis a hierarchy of tasks and sub tasks It describes
238. targeting system since it is a separate application For the purposes of this observation the report was already generated and provided to the inexperienced user As the inexperienced user was introduced to the tool she commented on the fact that data was not produced in a logical order Referring back to Figure 3 2 data was generated from the top down when performing the task manually whereas with the current tool flow begins with loading and generating data for the call frequency and then working from the bottom up to create the targeting data The sequence of tasks not following that of the manual system was also raised during the interviews the current tool has a very strict sequencing of tasks with little flexibility These comments along with Nielson s third heuristic of user control and freedom have highlighted a usability issue and will be used in forming new requirements 1 A Lilly drug used to treat a range of mental disorders including schizophrenia and bi polar disorder A Lilly drug used for treating depression 34 STEP 2 Deleting old data and importing new data from the provided business objects report The experienced user explained that she prefers to keep previous data so she renames the table containing the previous data rather than deleting it New data is then imported by Import command on the File menu and following the wizard ensuring the destination table name is SelectedRatings This provides a poi
239. th the prototype to perform given tasks see section 2 6 Information is written on individual cards users then sort depending on some given criteria Users join the design team See section 2 4 3 Users are asked a standard set of questions Semi structured questions are asked to stakeholders content experts support staff and users themselves Identify steps required through interview or observation See section 2 4 2 Table 1 User centred Design Methods Sample size Low Varies Low Medium High Low High Low Low Table 1 is an important summary and highlights the key methods that will be used throughout this project Although a variety of these methods will be used throughout the project this next section will take a greater look at the two most significant methods involved in user centred design task analysis and participatory design 13 2 5 1 Why user centred design As pointed out by Ellen Bravo at a conference in 1990 leaving users out can have disastrous effects not just on the task goals but on the users themselves 7 Alison Black 18 states that the most successful designs come from understanding the needs of the people that use them Early involvement of users with the future system may lead to adjustment of their expectations making the eventual acceptance more likely 7 This in turn could potentially lead to cost reductions due to early acceptance and reduced development costs The process of unders
240. the data from the global CRM system The purpose of this is to extract the rating that each sales representative has selected for each health care professional A sales representative will give each of their customers a rating which determines how each of their customer s are targeted for particular drugs The rating is selected by each sales representative from a drop down box within the CRM system and has a value taken from the following values A A B B C C and Non Target Each letter refers to a particular rating which may be different for each sales force Figure 11 shows the notation used for the LPG sales force For each target customer a value needs to be entered in the Selected Rating field EE A Zyp Sch Only A Zyp Sch Cym Dep B Zyp Bip Only eee pa Zyp Bip Cym Dep 5 2 Cym Dep Only Zyp Sch Zyp Bip Non Target Zyp Sch Zyp Bip Cym Dep tof pecially Diabetes 8 Data Status Vasdsted Figure 11 Notation used for LPG Sales Force Zyp Sch refers to Zyprexa for the purposes of treating schizophrenia Zyp Bip refers to the same drug for use in treating bipolar disorder Cym Dep refers to Cymbalta for the purposes of treating some forms of depression This coding system forms the cross reference which has been mentioned during interviews The generation of the business objects report is completed outside the
241. the standard log on for all systems across the Lilly environment for authorisation purposes or implement a proprietary login system and support tools for the tool e User control of data versus the likelihood of breaking the tool e Deployment installation on a shared server is preferred than deployment to multiple machines standard office machines run Internet Explorer V7 Now interviews are complete the next part of the data gathering process is to complete a heuristic evaluation of the existing tool 3 4 Heuristic evaluation of the existing tool As part of the requirements analysis for the new tool it is important to identify the underlying usability problems with the existing system As discussed in chapter 2 there are a variety of techniques for evaluating an interface Heuristic evaluation is an informal method of usability analysis whereby evaluators are asked to comment on the design within a defined framework A HCI expert was asked to complete this evaluation Interview with JC JC is completing an MSc in Human Communication and Computing at the University of Bath JC has experience in the pharmaceutical industry as well as completing a dissertation within the HCI field Method The HCI expert was presented with the system and asked to perform the task following the same instructions provided to a user A copy of these instructions can be found in Appendix A Section 3 A concurrent protocol was provided by th
242. tion will lead onto modifications of the design by amending it to suit users needs Without evaluation designers cannot be sure that the system is usable and what the users want the main requirement for an exemplary user experience is to meet the exact needs of the users without fuss or bother 2 This therefore highlights that evaluation is a key area to ensuring this is possible and therefore having a successful end system There is a close link between evaluation and prototyping techniques as previously described As Dix et al states 1 evaluation has 3 main goals e assess the extent and accessibility of the systems functionality e assess the user experience of the interaction e To identify any specific problems with the system As already stated above the systems functionality is extremely important since the user must have all that they requested secondly this functionality must be clearly reachable by the users in terms of the actions that the users need to take to perform the desired task Aspects like how easy the system is to learn the users satisfaction and enjoyment is important to assessing the 17 user experience Finally any problems with the system will have a direct impact on the usability and therefore these also need to found as early on in development as possible Since the aim is to continuously evaluate it will not be possible to carry out extensive experimental testing throughout the design therefore it wil
243. tion Errors Number of records imported successfully 27 Number of records with validation errors 1 Restart Save amp Proceed Lesart 1 bf on Mun poge enough 9 gun e T ek VERN NSCFUL SUA no be beker MC Feld aes 5 OS Woh hea dr S Ue SWK Nee issue oC Columna NAMES comes We Yo Ss ys ask gt bu ness P x we showd nos a m vos ress obje ks Figure 39 An example of user feedback during implementation 84 Although informal testing was completed throughout development once the application had met all mandatory requirements and all users were happy with the interface design a final testing phase took place the test plan can be seen in Appendix C Section 3 The next section will now provide further details on this 5 6 Testing Testing is defined as the process of exercising or evaluating a system by manual or automatic means to verify that it satisfies specified requirements or to identify differences between expected and actual results 56 This means that testing is the determination of how close to ideal the systems behavior is and any deviation indicates the presence of a bug As previously stated this project is developing the application as a proof of concept only and for this reason there was no need to follow a strict test plan Testing still took place but this was just on a smaller scale than that of a typical syst
244. tionDivision y k bspl Status Planned db BSPLs InsertOnSubmit bspl db SubmitChanges Response Redirect Complete data aspx 166 position Substring 10 1 LPG BP LPG MOOD LPG SCHIZ LPG RSP make sure there is something if no match NKNOWN Appendix F Ethics Checklist UNIVERSITY OF BATH Department of Computer Science 13 POINT ETHICS CHECK LIST This document describes the 13 issues that need to be considered carefully before students or staff involve other people participants for the collection of information as part of their project or research 1 Have you prepared a briefing script for volunteers You must explain to people what they will be required to do the kind of data you will be collecting from them and how it will be used All participants will be provided with information on what they will be required to do the data that will be collected and how this will be used in advance for each event Interviews observations participatory design and user evaluations 2 Will the participants be using any non standard hardware Participants should not be exposed to any risks associated with the use of non standard equipment anything other than pen and paper or typical interaction with PCs on desks is considered non standard No 3 Is there any intentional deception of the participants Withholding information or misleading participants is unacceptable
245. ts and analysis will then enable a discussion as to whether there is an impact on end user experiences after a navigational redesign The next section will consider what measurements need to be taken and what these mean to effectively complete the experimental design 6 2 2 Measurements To test the hypothesis itis important to consider what properties of the application would make it superior these properties will then be those that are tested User preference is a clear property that will indicate this and the problem is that this is a very subjective measurement and alone will not allow the hypothesis to be tested Therefore this needs to be divided further to collect quantitative data User preference will be influenced by the usability of the application Section 2 6 2 of the literature review states some quantitative performance measures that can be taken to test usability The most appropriate of these measurements will be used to form the experimental design It is still possible to gather quantitative data on user preference for example using a likert scale where participants can specify a rating between 1 and 5 to specify their level of 88 agreement with a particular statement Therefore additionally to performance measures user preference will be tested using the likert scale on a questionnaire It is now possible to develop predictions of what these measurements will show and consider how these predictions can be tested 6 2 3
246. ty These have been categorized into Nielson s heuristics Heuristic Comments Visibility of system Difficult to know where to go once system has loaded there is no defined start status or completion screen Many times throughout the system there is a lack of feedback for example once a SQL query has been ran it does not inform the user it is complete Due to the amount of data the system could produce without the visibility of the user the system should provide summary information to make sure the user is aware of what the system has produced Match between system and real world Users are unable to see what data is being removed since they are required to write an SQL query to delete rows This is a concept that is not familiar with non technical users it should match the real world in terms of viewing selecting and deleting data through an interface User control and freedom There is no defined structure of goals meaning that users are able to navigate to any part ofthe system at any time this means they have the freedom to delete any data or even entire queries Consistency and standards Comments arose around the icons Access uses to represent different types of queries Jason advised that these should give a better reflection and made comments about how to use colour effectively for example when a user is working through a task a potential menu on the top could be highlight green for complete and orange f
247. urrent tool What frustrates you about it What features would you like to see removed enon 7 What new features would you like to see 8 Do you have any comments on user experience of the current tool in general And can you suggest anything that would assist the user experience 9 Do you have a list of priorities for enhancements New features 10 Any other comments Interviews will be recorded for future use but will remain anonymous will not used for any other purpose and will not be included in the final project hand in Participants can leave at any stage after all interviews have been conducted a debrief session will be held where all participants will come together to review brainstorm and have an opportunity to ask any other questions Participants have read and agreed with the interview process 118 Interview Transcript lane Date 2 Jo Interviewee name and team Ogura Daxa CNOA usar of te N L 1 Whatis good about the current system Are there any features you like about the current tool Wy V Swaen Ewwwoc Cues Crom Wren Seas regs hore ae wo Ove Lo 2 j wet VAS Gu Coe Ce DN CAM SA ox 467 few dida CON Be dite VA mgoa eO Bo 2O0wiecx oe gt Co c onec 2 AAS gd A X x z Checks the oona is Vaid Suni xs v ux Or Oque Funa dor Sabres regrnce roves to ce iro ALRMOVEWN luks guarry
248. user information e User details Users e Role details Roles e Mapping between a user and their role UserInRole Note there is no table for the sales quota data to be imported into this is because there is no intermediary format Data is loaded straight into the OBSQ table The data types to be included in the generated objectives file need to follow the OneLilly input specification which can be found in Appendix A Section 2 The SelectedRatings table was designed based on the information provided in Appendix B Section 3 and the remaining tables were designed based on the findings from the observation in chapter 3 see section 3 5 1 Figure 32 shows the database diagram 72 Users x UserId mi UserName Bl PasswordHash E PasswordSalt gl Email E Comment Enabled E DateCreated E DateLastLogin _ DateLastActivity Ej DateLastPasswordChange UsersInRoles x HashId E UserName Ej RoleName E RoleName ObjectivesMapping 9 entry pk By Objective Product 1 Indication 1 ObjectiveParent CrossReference 9 entry pk El PositionLetter UniqueSelectedRating ObjectiveMappingID CallFrequency SelectedRating x entryPK E Division Ei Position E ConIntegrationID SelectedRating PositionLetter E CrMapping BSPL 9 entry pk AffiliateCode RecordType El SourceSystemId El Operation mi OperationDate 1 Description StartDate EndDate Lo
249. w critique the system by trying to answer the following questions at each step 18 Will the users try to achieve the correct effect Will the user see that the correct action is available Once the user has found the correct action how will they know it s the one they need Pe wn If the correct action is performed will the user see that progress is being made towards the solution of the task Are they able to understand the feedback they are receiving It is important to document the walkthrough to keep a record of what needs improvement and what works well it is therefore a good idea to produce standard evaluation forms It can also be useful to indicate the severity of the problem whether the problem could occur often or how serious it could be for the users 1 The second approach is Heuristic evaluation which uses the heuristics discussed earlier in the chapter It is an informal method of usability analysis where a number of evaluators are presented with an interface design and asked to make comments It was first developed by Nielson and Molich The general idea is that several evaluators will independently critique a system to come up with potential problems Nielsen and Molich 5 conducted four experiments to investigate how many evaluators are sufficient to find the maximum number of usability problems He found that individual evaluators were quite bad at doing evaluations individually since they only found between 20
250. w system Between all users there were 22 errors made within the existing system the majority of these should not be possible Nielson s guideline for Error prevention states that error prone conditions should be eliminated thus errors like incorrectly spelt products and indications can and have been within the new system eliminated by using a drop down preventing an incorrect spelling to be made The same concept is relevant for manually written queries or table set up setting primary keys within the new system all technical aspects are now hidden and completed automatically The remaining errors mainly by inexperienced users fall into the category of insufficient guidance and included errors like incorrectly moving data or being unaware of how to run a query or export data As Nielson states having help and documentation maybe necessary but it should be kept short and focus on particular tasks as the design chapter discusses the help function within the new application follows this guideline by using videos specific to each step This functionality along with the fact that everything is now contained within one system prevented such errors occurring in the new system Comments from inexperienced users showed that help functions and guidance are no longer as important since the system structure supports them better than the existing system One particular comment stated I don t need to understand the fact that a query is being ran all I ne
251. what order and under what circumstances sub tasks are performed HTA focuses on the physically and observable actions that are performed it also includes actions that are not related to the software at all The starting point is the user goal the tasks associated with this are identified and then sub divided as appropriate 2 GOMS was developed in the early 1980 s as an attempt to the knowledge and cognitive processes involved when users interact with a system Goals refers the goal that the user is trying to accomplish Operators refers to the cognitive processes and physical actions that need to be performed in order to reach that goal Methods are the learnt procedures for accomplishing the goal they consist of exact sequence of steps required Selection rules are used to determine which method to select when there is more than one available for a given stage of a task After the initial practices and processed have been understood the next most significant user centred techniques is participatory design Users will participate during the initial exploration of the problem identified to help define and focus ideas for the solution right throughout development through to evaluations of proposed solutions This is the most important part of ensuring the solution is exactly what the users want and need 14 2 5 3 Participatory design Computer systems development will always be accompanied by the problem of how to define requirements for
252. when dealing with individual or specific user interface requirements e Will make maintenance and code re use easier user interface changes will be much easier e Theclearseparation makes the testing of the application much easier One of the drawbacks of using such a model is the rigorous separation between the model and the view can means that there are more files to manage and could potentially lead to increased complexity making debugging more difficult This Model View Controller framework was originally proposed in the 1980s and is one of the most widely used for GUI design 44 More recently there have been further research projects that have come up with alternatives for example Joelle Coutaz developed the PAC model It is now possible to begin implementation following this architecture and using the conclusions made during chapter 4 5 4 System Development The main sections overview import cross reference each data type and export required have already been thoroughly discussed in chapter 4 each of these will now form pages with a code behind class file Additional classes will be required for importing and exporting Figure 33 shows the high level class structure for each of these including the LINQ objects required Note light blue shades reflect aspx pages with a code behind file cs and green shades reflect classes cs only Following this a description of the most significant sections of code are described 75
253. www useit com papers heuristic severityrating html 9 J Nielson 2005 Introduction to usability online Available at http www useit com alertbox 20030825 html 14 J Nielson 2005 Information foraging online Available at http www useit com alertbox 20030630 html 17 University of Washington usability research group User centred design methods online Available at http www indiana edu usable presentations ucd_ methods pdf 18 Alison Black Nov 2006 User centered design online Available at http www designcouncil org uk en About Design Design Techniques User centred design 25 Jing Wu Apr 2000 Accommodating both experts and novices in one interface online Available at http www otal umd edu UUGuide jingwu 46 NIELSON J 2003 Information foraging Why Google makes people leave your website faster online Available from http www useit com alertbox 20030630 html Accessed 12 04 08 49 INSTONE K 2002 Location Path and Attribute breadcrumbs online Available from http instone org files KEI Breadcrumbs IAS pdf Accessed 12 04 08 51 BERNARD M 2003 What is the best way to arrange menus Criteria for optimal web design Designing for usability online Available from http www optimalweb org Accessed 12 02 08 52 ROGERS B L CHAPARRO B 2004 online Available from http psychology wichita edu surl usabilitynews 61 breadcrumb htm Accessed 12 04 08
254. x thought everything within it the wrong one would be assigned to that position U2 T1 Written delete query incorrectly T1 Didn t select the 3 data types needed T2 Did not select tick box for first rows as T3 Chose incorrect objective didn t know column names more could be added T2 Inserted into a new table instead of existing T3 Incorrectly spelt product T3 Did not remove duplicate rows T3 Missed the INVALID fields in the remaining columns T4 Incorrectly copied data to template put into wrong sheet T4 Macro within external template failed there 96 were less rows than last time ran meaning there were null s verification could not be completed T5 Exported from Access but did not understand how to export to correct format using the template U3 T1 Written delete query incorrectly T1 Did not run the query just saved it instead T2 Let Access chose a primary key T6 Clicked Export only participant thought this would send it as well None U4 T1 Forgot to update SQL to the correct quarter T2 Imported the incorrect worksheet T2 Let Access chose primary key T3 Spelt the indication incorrectly T3 Did not remove duplicate rows T4 Accidentally re ran the query instead of opening the table of data generated T5 Exported the incorrect table forgot to run the distinct query first T2 System only looks at first worksheet but data was within second so
255. y The selected ratings that are allowed are A A B Bt C C Non Target 143 Appendix C Implementation Supporting Documentation Section 1 Email feedback during implementation On 17 March 2008 at 13 18 goddard card lilly com wrote Hi Sam I looked over what you have done so far and it s looking great Here are just a few comments At the moment even when I don t have a file in the upload box I can still press upload When I did upload the file it said 23 rows were imported but there was only 22 When your viewing data the data link in the navigation is not bold None of the export functionality appears to be working yet But I guess you not finished yet Section 2 Feedback via annotated print screens during implementation Quz Yeyt s clearer blak S 00 Overview Import X Ref gt Data gt Export Baid highlight Your locahen Is good ly Veic b You are logged in as sam cooper Please select sales force Please select quarter Please select year Help BSPL Business Plans Restart dx OBJT Parent Objectives OBCF Call Frequency Log Out OBSQ Sales Quota Save amp Proceed 144 You are logged in as sam cooper Call Fre Please upload the Business Objects report for sales representatives selected ratings Admin Upload an Excel file Browse Import Save amp Proceed Restart 1 T alae a pu on on the efk hand Side
256. y diverse tasks Research in the future is likely to design and evaluate how interfaces and navigation can be improved to understand and respond to individual user needs The project has been of great success and has benefited both the company and the author understanding as well as considering and enhancing current research 113 Bibliography Books 1 Dix et al 2004 Human Computer Interaction ed Pearson Prentice Hall 2 Preece et al 2007 Interaction design 2 ed Wiley amp son 3 M Rosson and J Carroll 2002 Usability Engineering Academic Press 4 Shneiderman and Plaisant 2005 Designing the user interface 4 ed Pearson Education 7 D Schuler and A Namioka 1993 Participatory Design LEA publishers 10 C Faulkner 1998 The Essence of Human Computer Interaction Prentice Hall 11 J Nielsen 2000 Designing Web Usability New Riders Publishing 41 K Murphy 2004 Statistical analysis Lawrence Erlbaum Associates publishers 42 PJohnson 1992 Human Computer Interaction Psychology Task Analysis and Software Engineering McGraw Hill Book Company 43 G Kepple amp T Wickens 2004 Design and Analysis A Researchers Handbook Pearson Prentice Hall 44 Sommerville 2004 Software Engineering Seven Pearson Addison Wesley 56 E Stiller amp C LeBlanc 2002 Project based software engineering Addison Wesley Conference proceedings 5 J Nielson amp R Molich 1990 Heuristic Evaluation of User Int
257. y prototyping is a continual process for adapting an application to organizational constraints and requirements it allows the developers to work in close cooperation with the users to continue to improve the system Firstly a low fidelity prototype will be produced these are generally limited function they are constructed to view concepts design alternatives and screen layouts rather than model the user interaction with a system In general they are constructed quickly and provide limited or no functionality they demonstrate the general look and perhaps feel of an interface They are generated to communicate educate and inform but not to train test or serve as a basis on which to code Rudd et al 34 discuss some of the advantages of low fidelity prototyping e Huge value in the early requirements gathering and analysis phase it can act as a communication medium by which requirements can be developed e They can be constructed early in the development without a large investment in cost and time e Since they are constructed on paper they require little or no programming skill e They are easily portable and easily presented it can be easily demonstrated to potential users to quickly obtain feedback on how well the design meets their needs Paper based prototypes or screen mock ups offer a way to make effective use of the users experience two specific examples of this technique have been developed for participatory design and these ar
258. ystem but one of the main differences for this particular measurement is that both tasks 1 and 2 within the new system have a standard deviation of zero Although as previously noted this is a small selection of tasks and participants but this data is supporting evidence and therefore a good indication that the usability has been improved through the navigational redesign 100 6 3 Subjective user evaluation After completing the experiment each participant was asked to fill in a questionnaire which included 7 questions using the likert rating scale The final question involved an informal interview to gather the participant s general opinion on the new system Due to time constraints these interviews were conducted on the telephone the day after the experiment took place to assist participants remember their experiences both systems were being used during the telephone conversation Written notes and completed questionnaires can be found in appendix D sections 2 and 3 The Questionnaire contained the following statements and asked users to provide a rating for the new and the existing system Using a scale of 1 5 with 1 being strongly disagree and 5 being strongly agree 1 I found it easy to complete the first task 2 I found it easy to complete the second task 3 Ifound it easy to locate the call frequency data 4 I didn t need to rely on a user manual or help functions 5 find the system easy to navigate through 6 Ifound it easy
259. ystem from the user As highlighted earlier there is still a concern with the ordering of data production with reference to figure 6 the number of steps where an alternative ordering is possible is significantly higher in figure 13 This concern does need to be considered when designing the new system if it provides flexibility then it will be a user s choice on whether they wish to work on data top down or bottom up 3 8 Setting priorities The next stage is to discuss the findings from this phase with the stakeholders prompting them to consider the priorities of the problems identified 3 8 1 Main priorities As requested by users during interview sessions e Ability to export all file types to Excel e Little or notraining should be required for the usage of the new targeting system e Provide navigational assistance to the user e The application should have appropriate security such that only OneLilly team members have access e Remove all visible code from the user e Preventing users from updating table structures and queries e Improve the usability of the cross reference table e Prevent duplicated data being inserted into the cross reference table e Provide users with feedback particularly when invalid data has automatically been removed 3 8 2 Secondary priorities Users made the decision to place these items as secondary priorities because they do not form part of the core functionality supporting the task e Verification

Cooper-S-Dissertatio.. - Department of Computer Science

Contents

Download Pdf Manuals

Related Search

Related Contents