Why five users can be enough
The idea in brief
Nielsen Norman Group published the article “Why You Only Need to Test with 5 Users”. Its commonly cited model suggests that about 85% of usability problems appear with five users, because each new participant increasingly repeats problems already seen and reveals fewer new ones.
This makes small tests worthwhile: a few sessions give most of the insight at a fraction of the cost of a large study.
The real advice: several small rounds
The point is not to test with five people once and stop. The article recommends running several small tests during a project, fixing the problems found between each one.
- Test a prototype with about five participants.
- Group the problems and fix the most serious ones.
- Test the revised design with a new set of participants.
- Repeat while the project continues.
Fixing between rounds matters because a problem seen by every participant does not need to be seen again by more people; it needs a fix.
Limits to remember
Different groups
If your product serves distinct user groups, each group needs its own small sample.
Quantitative studies
Measuring task times, success rates or comparing designs with numbers needs many more participants than five.
Not a guarantee
Five users show most common problems, not every problem. Rare issues can remain.
Use five as a practical size for qualitative rounds, not as a rule for every kind of study.
Putting it into practice
Small rounds fit well into regular project work because they are cheap and quick to schedule. A few habits make them effective.
- Test early with rough prototypes; changes are cheapest before development.
- Write the tasks and success criteria before the first session so all participants get the same conditions.
- Recruit people who match the real audience; five wrong participants teach little.
- Rank the problems by severity and fix the serious ones first.
- Share short clips or notes with the team so findings are seen, not only reported.
Treat the number as a budgeting aid. If a round shows serious problems, extra participants add little before those problems are fixed.
Sources
- Nielsen Norman Group. Why You Only Need to Test with 5 Users. https://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/
This guide was drafted with AI assistance from public standards and references. For the exact wording of a standard, use the sources listed above.