The Five-User Test Works. You're Probably Running It Wrong.

Every few months, someone on a product team tells me usability testing with five users is not enough. Five people cannot represent our market, they say. And they are right, if the goal is to represent the market. That was never the goal.

What five users are actually for

Small-sample usability testing is a tool for finding problems, not measuring them. If a significant share of people struggle with your checkout flow, there is a good chance at least one or two of five participants will hit that struggle in front of you. You will not know the exact percentage. You will know the problem exists and roughly why.

That is useful. But only if you do the next part.

The part most teams skip

The real idea behind small tests is to run several of them in a row, fixing what you find between rounds. Five users, fix, five more users, fix again. Instead, many teams run one round of five, write a 40-page report, and move on.

Here is how I structure a typical study now:

1. Round one, five participants. Find the biggest, most obvious problems.
2. Fix the top three issues in the prototype, even roughly.
3. Round two, five new participants. Check whether fixes worked and discover problems that were hidden behind the first ones.
4. Repeat until new rounds stop revealing serious issues.

On a recent onboarding project, round one found that 4 of 5 participants missed the skip button entirely. We fixed it. Round two revealed that once people skipped ahead, they did not understand the second screen at all. We would never have seen that problem if we had stopped after round one.

When five is not enough

There are times you genuinely need more participants:

• Distinct user groups. If admins and end users use the product very differently, test five of each.
• Measuring rates. If a stakeholder needs to know what percentage of users complete a task, you need a much larger quantitative study.
• Comparing designs. Choosing between two options with small differences requires numbers, not observation.

Signs your small tests are failing

You are probably running five-user tests wrong if:

• You only ever run one round per project.
• Your findings list has 30 items with no priority.
• The same issue appears in three consecutive studies, months apart.
• Nobody from engineering or product has watched a session.

A usability test is not a verdict. It is a conversation with your design, and it needs more than one turn.

The fix is rarely a bigger sample. It is a shorter loop. Smaller studies, more often, with changes in between, will teach you more than a single large study that arrives after the design is already built.

Cover: Photo by UX Indonesia on Unsplash.