Usability testing is the practice of watching real people try to complete specific tasks on your website, app or prototype, so you can see where they hesitate, misread something or give up. It shows you why visitors struggle, which analytics on its own cannot do.
How usability testing works
A basic round has five steps:
- Choose the tasks that matter most to the business, written as realistic goals. “Find out whether this clinic treats children and book a first appointment” is a task; “look around the site” is not.
- Recruit participants who resemble your customers. A handful of people per round is a common rule of thumb, because the biggest problems tend to show up in the first few sessions.
- Run the sessions, either moderated (you watch live, in person or on a video call) or unmoderated (a testing platform records participants working alone). Ask people to think aloud as they go.
- Note every point of trouble: where they paused, what they clicked by mistake, what they said they expected.
- Rank and fix the issues, then test again to check the fixes worked.
It complements other research methods. Session recordings and heatmaps show what large numbers of real visitors did, but not what they were thinking. A/B tests tell you which version performs better, given enough traffic. Usability testing fills the gap by explaining behaviour, with very few participants.
Why it matters
Most UK small business websites do not have the traffic for reliable split tests, but anyone can watch five people try to book an appointment. Those sessions regularly expose problems that the owner and the web designer cannot see because they know the site too well: a menu label that means nothing to customers, a price that is hard to find, a form that rejects a perfectly valid UK postcode. Fixing problems like these before spending more on ads or SEO means the extra visitors arrive at a site they can actually use.
It also catches accessibility barriers, especially if you include participants who use screen readers, magnification or keyboard navigation. If you record sessions, remember that recordings of identifiable people are personal data under UK GDPR: tell participants what you are recording, why, and how long you will keep it.
Common mistakes
- Testing with colleagues, friends or family, who already know the business and try to be helpful.
- Asking opinions (“Do you like the design?”) instead of watching behaviour.
- Leading the participant, for example by naming the button they should be looking for.
- Testing only on a desktop when most of your visitors arrive on a phone.
- Waiting until the site is built and launched, when changes cost the most.
- Writing up findings and then fixing none of them.
How to act on it
Pick the two or three tasks that earn you money, such as making an enquiry, finding a price or completing a purchase. Write a short script that sets the scene without giving hints. Recruit people who match your customers, offer a modest voucher for their time, and run the sessions on the devices they would really use.
Afterwards, list each problem with how many participants hit it, fix the ones that stopped people completing the task, and run another small round. Testing a prototype before development is the cheapest point to do this, which is why I build it into the early stages of a website redesign rather than leaving it until after launch.