Upload
britney-ellis
View
214
Download
1
Embed Size (px)
Citation preview
#17 - Involve Users in the Development Model of Multinational Corporations - Is it worth it?
Experience Report
IRCSE '08: IDT Workshop
Friday 31 October 2008
Martin Olausson
Page 2
Background
Several project results end up in unsatisfied users
To be able to deliver a successful software product that suits the users' requirements and expectations You have to fully understand these requirements and
expectations
To be able to fully understand these requirements and expectations You have to involve the actual
users in the development process
Page 3
Background
Users are mostly indirectly involved in the software product design through indirect links These links are normally few
Studies shows that successful projects consist of several direct links between users and developers To achieve these direct links users need to be directly involved
with the software developers
Page 4
Problem Definition
Be able to shorten the distance between theusers and the R&D organization
Project managers usually critic user involvement to be expensive and time consuming If we would be able to prove the incorrectness of these
prejudgments we could move the organization towards a more user centric setup
Thus a project were assigned where focus was to involve the users in the design process. Targeting configuration of a control system
Page 5
Identify User to Involve in the Project
People from product management, line management and R&D upper management with business knowledge to know where they wanted to gain or keep market shares did provide a user profile
User profile’s attributes were among others Years of control system experience, age, culture, professional
roles, industry and company background
Targeting novice user in the age of 40 to 50 years Novice = less than one year control system experience
Page 6
Methods - Interviews
Totally 16 open ended questions, each interview 2 hours duration “Please describe what type of continuous maintenance you perform
on the distributed control system”
Results and lessons learned After 10 interviews the number of
unique findings was down to ~4 per interview
Ended up to be dialogs, not monologues. Clearly inform the users about our level of knowledge to avoid incorrect expectations
Compile information directly afterwards as it often are ambiguous and vague. ~2 hours per interview
No culture differences in work flow However the Chinese users did only want to participate in email
Page 8
Methods - User Brainstorming
As the users' daily work is focused of using the distributed control system we believed that the users had great ideas of how their problems best could be solved
Results and lessons learned The brainstorming session resulted in new
requirements instead of what we expected; a pale of solutions ideas
Page 9
Methods - Paper Prototyping
Used After the interviews but before starting the actual implementation of the prototype Printed out and showed on face 2 face
meetings
Time to redesign and test on meetings
A fast and effective method for allowing users to gain insight in the proposed design and functionality
Results and lessons learned Users' adaptation to this method showed out to be quick
As the users saw that their involvement did change the design immediately during the test, they were positive to get involved in further tests
Page 10
Methods - User Tests
When paper prototype tests were finished a prototype of the software product were implemented Purpose to test design and functionality
and not to be used within the final software product
No architecture documents, function descriptions or design document were written
Results and lessons learned Conducted 37 user tests (20 for
design)
The average number of iterations was 2
Page 11
Results
To verify results 20 users were identified for a test,10 novice users and 10 advanced users List of 10 tasks to complete without any time limits
Page 12
Results
We identified 5 similar projects with the same planned user profile, project schedule and project costs
Main reason for the increased costs was that the interpretation of some requirement was changed during the development
Page 13
Summary and Conclusions
Establish contacts with users is a time consuming process It is important to start early in the project to identify the user
profile and to establish contacts with the users
Invaluable input as project gained much credibility towards both steering committee and stake holders The proposed design and functionality were entirely based on
and confirmed by the users
Great use of detailed log. Statistics were very valuable when defending design and functionality decisions
To avoid redesign of the good parts of a products it is important to present users’ positive feedback as well
Page 14
Summary and Conclusions
Take care of the users. Too many times the users have invested time in projects by sharing their opinions without getting any further updates of what come out of their contributions Let them know of project outcome
Even if your project schedule seems too compressed to involve the users you can still do something Any single interviews, paper prototyping or user tests are so
much more than none