Quick answer: User testing means watching real people try to complete tasks on your website while they think aloud. You do not need a lab. Testing with around five people from your target audience will usually reveal most of the serious problems on a page or flow, because the same issues come up again and again. Write three to five realistic tasks, recruit people who resemble your customers, watch without helping, note where they hesitate or fail, then fix the biggest problems and test again. Several small rounds teach you more than one large study.
Why user testing is worth doing
Analytics tells you what visitors do: where they drop off, which pages they view, how long they stay. It rarely tells you why. Heatmaps and recordings get closer, but you still have to guess at intent.
Watching a person try to use your site fills that gap. You hear them say “I expected the price to be here” or “I am not sure if this button will charge me”. Those moments are hard to discover any other way, and they often point to simple fixes.
Business owners and designers know their sites too well to see them as a newcomer does. Even a handful of sessions usually changes how a team thinks about a page.
Why five users?
The idea comes from usability research by Jakob Nielsen and Tom Landauer, popularized in Nielsen’s well known article arguing that small tests with about five users find most usability problems. The reasoning is simple: the first user shows you many problems, the second shows some of the same plus a few new ones, and by the fifth you are mostly seeing repeats.
A few caveats matter:
- Five per distinct user group. If you serve very different audiences, such as buyers and suppliers, test each group separately.
- Five per round, not five forever. The point is to test, fix and test again, rather than spend the whole budget on one big round.
- It is about finding problems, not measuring them. Five users will not give you reliable statistics like a conversion rate. For that, you need analytics or an A/B test.
Planning a small test
Decide what you want to learn
Pick a focus. Testing “the whole website” produces vague findings. Better examples:
- Can visitors find and request a quote for our main service?
- Can shoppers find a product in the right size and complete checkout?
- Do people understand our pricing page well enough to choose a plan?
Write realistic tasks
Tasks should describe a goal, not the steps. Avoid using the exact words on your buttons and menus, or you will be testing word matching rather than understanding.
Weak task: “Click on Services and then click Request a Quote.” Better task: “You need your office carpets cleaned next month. Find out whether this company can help and ask them for a price.”
Three to five tasks fit comfortably into a 20 to 30 minute session. Order them from simple to more involved.
Recruit the right people
Testers should resemble your customers in the ways that matter: their knowledge of your field, their goals and their comfort with technology. Avoid colleagues, family members who know your business, and web professionals.
Low-cost recruiting options:
- Existing customers, contacted by email with a small thank-you
- Your newsletter or social media followers
- People in local groups who fit the profile
- Short screening surveys on your own site
- Remote testing services that supply participants for a fee
A modest gift card or discount is usually enough as a thank-you. Over-recruit by one or two, because some people will not show up.
Running the sessions
In person or remote
In person lets you read body language and is easy to set up: a laptop, a quiet room and screen recording software. Remote sessions over video call with screen sharing work just as well for most websites and widen your pool of participants. Some services also offer unmoderated remote testing, where participants record themselves completing your tasks without you present.
Moderating well
The hardest part is staying quiet. Some guidelines:
- Explain that you are testing the site, not them. Nothing they do is wrong.
- Ask them to think aloud. Remind them gently if they go quiet.
- Do not help or explain. If they get stuck, ask “What would you do if you were on your own?” Only step in if they are completely blocked, and note it.
- Ask neutral questions. “What are you thinking now?” rather than “Did you find that confusing?”
- Avoid defending the design. Your job is to learn.
Record the screen and audio if participants agree. Recordings help when you share findings with people who were not there.
Test on the right devices
If most of your traffic is mobile, test on phones. Mobile users often struggle with things desktop users never notice, such as tiny tap targets, long forms or menus hidden behind icons. Our guide on mobile conversion rate covers common issues.
Making sense of what you saw
After the sessions, list every problem you observed. Then group them:
- Critical: stopped someone from completing a task
- Serious: caused long delays, errors or frustration
- Minor: small hesitations or cosmetic issues
Count how many participants hit each problem. An issue that tripped up four out of five people deserves attention first, even if it looks small.
Look for patterns in what people expected. If several participants looked for pricing in the menu, that tells you something about where pricing belongs.
Write findings as short, specific statements with a suggested fix:
- “Three of five participants did not notice the quote button below the hero image on mobile. Move it above the image and make it full width.”
- “Four of five participants were unsure whether the price included installation. Add a line under the price that says what is included.”
Turning findings into fixes
Fix the critical and serious problems first. Many will be quick changes to wording, button placement, form fields or navigation labels. Then run another small round to confirm the fixes worked and to find the next layer of issues.
User testing fits well with other research. Use analytics to find pages with problems, heatmaps to see behaviour at scale, onsite surveys to ask visitors directly, and user testing to understand the why. Our CRO audit guide shows how these pieces fit together.
Budget-friendly tools
- Video calling software with screen sharing for remote moderated sessions
- Free or low-cost screen recording software for in-person sessions
- Remote testing services that recruit participants and record sessions, useful when you cannot find testers yourself
- Session recording tools, such as those covered in our Hotjar vs Crazy Egg comparison, to supplement live tests
FAQ
How much does a small user test cost?
If you recruit and moderate yourself, mainly your time plus thank-you gifts. Paid remote testing services cost more but save recruiting effort.
How often should we test?
Before launching a new site or major page, after big changes, and whenever analytics show a problem you cannot explain. Small, regular rounds work better than rare large ones.
Can I test a design before it is built?
Yes. Clickable prototypes or even paper sketches can reveal confusing layouts and labels before any code is written, which is the cheapest time to fix them.
What if participants disagree?
Focus on what they did, not just what they said. Behaviour is more reliable than opinion.
Small tests, big lessons
You do not need a big budget to see your site through your customers’ eyes. A few hours with five people can show you where visitors get lost and why. If you want help testing and improving the pages that matter most to your sales, see our Turn visitors into customers service.
