Download ‘printer-friendly’ PDF version (File size: 15 KB).
Usability testing is a technique for ensuring that the intended users of a system can carry
out the intended tasks efficiently, effectively and satisfactorily.
When is Usability Testing appropriate?
Usability testing it is carried out pre-release so that any significant issues
identified can be addressed.
Usability testing can be carried out at various stages of the design process. In the
early stages, however, techniques such as walkthroughs are often more appropriate.
Usability testing is not a substitute for a human-centred design process.
Who should attend?
A usability testing session should be attended by:
- A representative user
- A test host.
It is highly recommended to also have in attendance as observers:
- At least 1 developer.
- At least 1 business representative.
Observers should ideally be in a separate observation room.
How many test sessions should be conducted?
This depends on the complexity of the system, the number of intended users, and
budgetary constraints. Even one test session is better than none, but 6 to 8 will tend to
uncover the majority of significant issues.
Do you need to video-tape?
Video-taping is useful for presenting to project members who cannot attend a test
However, video-taping adds to the complexity and cost of testing, and it is certainly
possible to carry out effective testing without such equipment, provided you log all
activities in detail.
Preparing for a usability test
Identify representative tasks. Ideally, these should derive from scenarios or use cases
developed earlier in the design process. Have a clear definition of when a task is
- Prepare a test schedule.
- Book the required room/s.
- Identify representative users and invite them to attend.
You will need the following materials:
- Video-taping equipment (if used).
- A formal script so that all participants are treated in the same way.
- A consent form for video-taping (if conducted).
A pre-evaluation questionnaire to check that your participants match the required
profile, and to check whether any effects observed are dependent on demographic
A list of tasks, together with criteria for measuring whether they have been
- Logging sheets to record timing, events, participant actions, concerns and comments.
A post-evaluation questionnaire to measure user satisfaction and understanding and to
glean any additional information participants may want to provide.
- Cash or an appropriate ‘thank-you’ gift.
- For all but the simplest and most informal tests, run a pilot test first.
Ensure participants are put at ease, and are fully informed of any taping or
observation. Attend at least one test as a participant, to appreciate the stress that
Ensure that participants have the option to abandon any tasks which they are unable to
- Do not prompt participants unless it is clearly necessary to do so.
Record events in as much detail as possible—to the level of keystrokes and mouse
clicks if necessary.
If there are observers, ensure that they do not interrupt in any way. Brief observers
formally prior to the test.
Be sensitive to the fact that developers may be upset by what they observe or what you