Most guides to user testing describe the study you would run if you had a UX research function, a recruited panel and six weeks, and you probably have none of those; you have a product manager, a designer who is already busy, and a release next Thursday.
That gap is why so many teams read about usability testing, agree it sounds sensible, and never do it. The advice is not wrong; it is written for somebody else.
So this guide answers a narrower question: what is the smallest test that still tells you something true? Five people, three tasks, one week, no budget line: it is less than the ideal study and enormously more than nothing, which is what most websites and apps get.
Half of this article is about what happens after the session, because that is the half that gets skipped. Watching somebody struggle is interesting, while deciding what to fix, in what order, and proving the fix worked is the part that changes the product.
One thing you will not find here: a number of users stated as a law. You have probably seen the figure that a handful of people find most of the problems, and it comes from a model with assumptions attached rather than from a measurement of your product, yet it gets repeated as though it were physics, and what small tests can and cannot tell you is covered properly further down.If the underlying method is new to you, our guide to covers the thinking that testing serves.
