Running Usability Tests on a Startup Budget

The research industry has convinced a lot of founders that usability testing means a research lab, a moderator, a note-taker, and weeks of recruiting. In practice, the single highest-value research method for an early-stage product is a five-user hallway test run in an afternoon - and it costs almost nothing.

Five users is not a compromise

Nielsen Norman's long-standing research holds up in practice: testing with five representative users surfaces roughly 80% of the usability problems you'd find with fifteen. The value of users six through fifteen is mostly confirmation, not discovery. For a startup deciding between "test with five this week" and "test with fifteen next month," five this week wins almost every time.

Recruit from your actual funnel

Skip the recruiting agency. Pull five people from your waitlist, your existing free-tier users, or - for a genuinely early product - friendly acquaintances who match your target user's rough profile. The signal you lose from an imperfect panel is smaller than the signal you gain from testing this week instead of next month.

Give them a task, not a tour

The most common mistake in a DIY test is narrating the interface instead of handing over a task and going quiet. Say "you want to do X, go" and then watch - the moments where someone hesitates, misclicks, or asks "wait, where do I..." are the entire point of the exercise. A test where you're talking more than the participant isn't a usability test.

Fix the pattern, not the instance

When two of your five users stumble on the same step, that's not a coincidence to note and move past - it's a pattern that will hit a meaningful share of your real user base. Prioritize fixes by how many of your five hit the same issue, not by how easy the fix is to ship; the easy fixes are rarely the ones actually costing you activation.