How do you determine the ideal number of participants for user testing?

Based on our experience here at Miratech, ten participants are enough to detect the majority of usability problems on a website.

The ideal number of participants

People often ask us: “How many users does it take to test the usability of an interface successfully?” Based on the Nielsen & Landauer Curve and our own experience in usability testing, the answer is at least 10 participants.

The more participants you add, the less you learn

Nielsen and Landauer demonstrated that a test conducted with 10 users revealed 98% of the problems with a website! This small sample is surprising when compared with classic testing methods.

The Nielsen & Landauer Curve

The Nielsen & Landauer Curve: a test with 5 participants allows you to identify more than 80% of the problems with a site.

We know that there is no such thing as an “average” user and every user behaves differently. The goal of user testing is precisely to take a variety of behaviors into account.

Once in front of the interface being tested, participants navigate according to pre-defined scenarios, and perform tasks that correspond to the critical elements of the study. We check that they are able to get around the site easily and find information, and that the links are in the right places. Any usability problems are clearly identified.

During testing, each participant adds new information to the list of problems, but as testing progresses this new data becomes less and less significant: after 5 participants, the information gathered is nearly identical.

Increase the number of tests, not the number of participants!

Increasing the number of participants will not make the results any more pertinent, but it will increase the cost of the test. When you add extra participants, the cost far outweighs the benefit (identifying interface problems).

We recommend you focus on organizing several tests with about 10 users, and improve the interface after each session.

Find the right compromise for your requirements

In concrete terms, the ideal number of participants for user testing is a compromise between:

- the budget, which includes recruitment costs, financial compensation for participants, interface corrections, etc.

- the time frames : reducing the number of participants can save time. Your time budget must also take into account the time it takes to analyze the results of several tests.
When several tests are performed, you also need to allocate time for correcting the problems uncovered in the previous session.

- the interface : it is important to have sound knowledge of the site, its target audience and the goals that need to be achieved through testing. If the site is aimed at different types of visitors, for example, perform several tests with users from the different target groups.

In practice, you will have to take all three of these factors into account to reach a good compromise between the cost of the operation and the value of the results.

– T. Baccino, C. Bellino, T. Colombi, Mesure de l’utilisabilité des interfaces, Lavoisier 2005.
– Jakob Nielsen’s Alertbox, 2000.