HowtoRunaWebsiteUsabilityTestWithoutaBigBudget

You don't need a research lab, a panel provider, or a five-figure budget to find out why users are struggling with your website. Usability testing on a budget of essentially zero — just your time and a free screen-recording tool — will still catch the majority of what's broken. This guide is the exact method we use before client launches: five participants, a handful of realistic tasks, one afternoon, and a short synthesis session afterward to turn what you saw into an action list.

Why five users is the number that matters

This isn't a budget compromise — it's the research-backed sweet spot. Nielsen Norman Group's long-running analysis of usability testing found that testing with five users uncovers roughly 85% of usability problems, and each additional user beyond that starts to reveal the same issues you've already seen. The efficient move is running several small rounds of five rather than one large round of twenty. The full methodology is documented on nngroup.com, and it holds up whether you're testing a SaaS dashboard or a checkout flow.

Five also happens to be a number you can realistically recruit without a research panel. Ask existing customers, post in a relevant community, or simply grab five people who match your target user but have never seen the site — colleagues in a different department count, as long as they're not designers who'll critique your Figma choices instead of using the product.

Write task scenarios, not leading questions

The single biggest way people ruin a DIY usability test is by asking questions instead of giving tasks. 'Do you like this navigation?' produces polite, useless opinions. A realistic scenario produces behaviour, which is what you actually need to see.

• ◆Weak: "What do you think of our pricing page?"
• ◆Strong: "You're deciding between our Starter and Pro plans for a 5-person team. Find out which one you'd need and how much it costs per month."
• ◆Weak: "Is our checkout easy to use?"
• ◆Strong: "Add this product to your basket and complete checkout using this test card number."

Write 3 to 5 scenarios covering the core jobs your site needs to do — usually some mix of finding information, comparing options, and completing a transaction or form. Keep each one a single sentence with a clear end state, so you know exactly when the task is 'done' and can time it.

Running the session remotely for free

You don't need participants in a room. A video call with screen sharing, recorded, does the job. Most video conferencing tools include free recording; if you want something more purpose-built, free-tier screen recorders capture both the screen and a webcam thumbnail of the participant's face, which is genuinely useful for reading reactions later.

Keep sessions to 20-30 minutes. Start with a one-line reminder that you're testing the site, not them — there are no wrong answers, and confusion is useful data, not a personal failing. Then read out each scenario one at a time and go quiet. The instinct to jump in and explain is the thing you have to fight hardest.

What to watch for versus what people say

There's a well-known gap between what users say and what they do, and it's the reason think-aloud, task-based testing beats surveys. People will often complete a task with visible difficulty and then tell you afterward that it was 'fine' — politeness and self-justification kick in fast. Trust the behaviour over the verdict.

• ◆Hesitation: a pause before clicking, especially on navigation or a CTA, usually means the label or placement didn't match their expectation.
• ◆Backtracking: clicking into a page and immediately hitting back is a strong signal that page didn't deliver what the link promised.
• ◆Mis-clicks: tapping near but not on the actual target suggests the tap area or visual affordance is wrong.
• ◆Re-reading: eyes or cursor moving back over the same text repeatedly usually means the copy is unclear, not that the user is careless.
• ◆Verbal confusion cues: "wait, where did that go" or "hmm, I thought that would..." are gold — write them down verbatim.

Turning five sessions into an action list

Immediately after each session — not at the end of the day, when details blur — jot down every point of hesitation, backtrack, or confusion against the specific task it happened on. After all five, look for repeats. If three or more of five participants stumbled on the same step, that's a real problem, not noise. If only one person struggled, note it but don't redesign around it yet.

Group the repeated issues into a short, prioritised list: what broke, how many of five hit it, and a proposed fix. This is deliberately similar in spirit to the structured list format in our article on

7 UX mistakes killing your conversion rate — most DIY usability tests end up surfacing two or three of those seven mistakes directly, which makes them fast to action once you've seen them on video.

If you're testing before a redesign rather than after one, it's worth first reading our explainer on the difference between UI and UX design — a five-user test is fundamentally a UX exercise, and knowing what you're actually measuring keeps the findings focused.

When to bring in outside help

This method scales down to zero budget, but it has limits — recruiting genuinely representative users, running moderated tests at volume, and turning findings into a shipped redesign is a bigger lift. If you've run the DIY version and found real problems but need help fixing them properly, our UI/UX design services pick up exactly at that point. And if budget for a fuller project is the open question, our guide on how much a website costs in 2026 will help you scope realistically. Otherwise, get in touch and tell us what your five users found — we can usually tell you fast whether it's a quick fix or a structural one.