This document discusses the benefits and process of usability testing for the Reactome website. It notes that usability testing helps identify user issues early, informs design improvements, and reassures stakeholders. The process involves recruiting intended users, observing them complete typical tasks, and identifying difficulties. Issues are then prioritized and addressed by developers to improve the user experience. The usability testing described here identified unexpected problems and helped rank development activities.
2. Why do usability testing?
Learn from user behaviour
Test assumptions to inform the design process
Involve users in design
Identify unnecessary features and spot issues
Fix problems early (when it’s easy)
Objective evidence, not opinions
Reassure stakeholders
Save a lot of work later on....
3. Who?
Identify the right users to test
Who are the intended audience for the features you want to
test?
Do you have access to a sample set?
4. Recruiting
We need volunteers to test our website – biologists who
occasionally use Reactome are ideal.
Testing involves working through typical usage scenarios,
e.g. ‘Is Fyn in Reactome, and if so in what pathways?’.
Duration ~1 hour. All participants receive a gift!
Purpose is to record when the user has difficulties with
the task – not a test of the user!
Results will guide our next design /improvements phase.
Contact: sjupe@ebi.ac.uk
5. Test scenarios
Write scenarios and tests
Guide (not show) the volunteer to key views or functions
Open-ended tasks, no answers provided
You want to identify when the user finds it difficult to
complete the task – and how you could address this
6. Resources required
A quiet place with some privacy
Best if a familiar setting for the user
Privacy to avoid disturbance and relax test subject
PC with software or web access to site
Optional: webcam, microphone, keylogger software such as
Silverback (Mac) or BB FlashBack (free!)
Observer with notepad
To note issues as they arise
Ideally observer and tester
Tester interacts with test subject
Observer can focus entirely on observing, may ask follow-up
questions
7. The testing process
Use a Tester and an Observer
Explain the testing process – use a script if you need to
Explain that the site, not the user, is being tested
User background – biologist or programmer? Helps to
relax the volunteer
Ask volunteer to give a running commentary
Explain scenarios. Start tasks by reading them out, give
a reference paper copy.
End testing when time runs out, or tests becomes
frustrating
8. During the test...
Remind the test subject to keep up the running
commentary
Guide the test subject only when needed - don’t train
Try not to directly answer questions, guide.
Ask questions that draw out answers:
What are you thinking now?
Is that what you expected to happen/see?
If there is an issue, note it but also investigate, ask test
subject as above, try to find out why there is an issue,
what would avoid it/improve it.
10. Drawing Conclusions
Don’t leap to conclusions – 5 or 6 people can be
enough to identify common issues, 3 is not.
Look for common themes in your observations...
Video is for deeper analysis but often the top problems
become obvious...we didn’t use our videos
11. More info
EBI Usability experts (Jenny Cham, Francis Rowland,
Sangya Pundir)
http://ebiinterfaces.wordpress.com/category/usability/
http://www.ebi.ac.uk/~frowland/dev/usability_testing/us
ability_resources.php
Lots of online books & videos
12. Acting on the results
Reactome developers brought in to group issues by
technology and difficulty
Usability identifies potential issues - there may be more
than one solution
Resolution depends on difficulty/resources
Find and fix early!
Project Prioritization Poker used to prioritize usability
issues against other objectives (i.e. our wish-list)
Fun - but serves it’s purpose, ranking with scores
Reactome Management adjustments and sign-off
Final ranking used to generate task list(s)
16. Summary
Usability testing has proven to be:
Easy to set up and run
Useful:
Identified unexpected issues (e.g. Zoom level)
Used to avoid guesswork on known issues and help rank
developer activities
Usability testing planned for:
Alternate pathway illustration styles
Redesigned layout for Homepage
Diagram key alternatives