You can run website usability testing yourself. No lab, no expensive equipment. You need two or three tasks, a few people from your target group and solid notes on what you observe. When? Before a redesign, or when the number of inquiries suddenly drops. A test like this shows where customers really get lost, and it does so before you pay someone for a new design. Below you will find the steps in order: goal, tasks, participants, test format, note taking and the order of fixes.
Table of contents
What is website usability testing and what will it show before a redesign?
It means watching real people do specific things on your website. You do not ask them for an opinion, you just watch what they do. Analytics will show you where people drop off (a typical signal is a high bounce rate). But it will not tell you why. That is the job of a test. With a company website, the same set of problems usually comes up in practice: muddled navigation, hidden contact details, a form that scares people away and an offer written in language the customer simply does not get. And this is exactly where you see how big the role of UX and UI is in whether anyone sends an inquiry at all.
How to prepare a usability test script step by step
A script is a short list of realistic tasks that are tied to the business goal of the website. A few sentences of introduction for the participant at the start, and that is it. You put it together in five steps:
- Set one goal for the test, e.g. checking the path to a request for a quote.
- Pick the most important paths: finding the offer, getting in touch, sending an inquiry.
- Describe the tasks as real-life situations, not as “click” style instructions.
- Write down what success means in each task.
- Rehearse the whole thing with someone from your team and fix whatever sounds unclear.
A bad task: “Go to the Services tab and open Audits”. A good one: “You want to check whether this company can help with a slow website. Find out how to book a meeting with them”. Feel the difference? If you give away the names from the menu, you are testing reading comprehension, not navigation. After each task, ask what was difficult and what the participant was actually looking for.
Who should you invite to a test with users?
People similar to your customers. Not employees and not friends from the industry, because they know the company jargon and will find their way around from memory. With no budget, the easiest option is to reach out to current customers, newsletter subscribers or people from your target group, whom you thank with a small gift.
The W3C recommends including people with disabilities in research as well, because they catch barriers that everyone else does not notice at all. The principles described in designing accessible user interfaces will come in handy here, and if more problems come up, WCAG-compliant websites may be the next step. Oh, and do not look for the perfect number of participants. In my opinion, it is better to run a few short rounds than one huge one.
Moderated or unmoderated test: which to choose on a small budget?
For your first study I recommend a moderated one. It will give you the most insights. An unmoderated test will do just fine for simple tasks and repeat rounds. In the moderated version, the facilitator sits next to the participant or connects with them over video. They use the think-aloud method, which means asking the participant to comment on every step as they go. They ask follow-up questions. But they do not help (and that is harder than it seems). In the unmoderated version, the participant works alone, and you analyze the screen recording or the session recordings from the website afterwards. One thing is not up for discussion: before recording, you always inform the participant and get their consent.
How to take notes on usability study results
For each task, write down three things: whether it succeeded, where someone hesitated and what they said. You note the behavior, not your interpretation of it. A plain spreadsheet with these columns is enough: task, participant, problem, quote from the session, severity of the problem. In a moderated test, let a second person take the notes while the facilitator focuses on the conversation. After all the sessions, you group the observations and mark the ones that come back with different participants. Because those, and not one-off slips, show the real weaknesses of the website.
What to fix first after website usability testing?
First, whatever blocks contact or a purchase and affects many people. Cosmetic changes can wait. All in all, the order almost always looks similar:
- blockers on the path to an inquiry, e.g. a hidden form or a button that does not work,
- unclear offer content,
- navigation and the names of menu items,
- small visual details.
A lot of these fixes do not require a new design. A redesign only makes sense when the problems are structural and sit in the architecture of the whole site. After the changes, run a short retest on the same tasks. And if you want to take a broader look at your website, website audits will help by adding a technical analysis to your observations.
Regular website usability testing lets you spend your redesign budget where customers actually get lost. Not on a hunch. Start with one goal and three tasks, and book your first session this week.
Frequently asked questions
How long does one usability test session take?
Not long. As long as it takes for a few tasks and a conversation after them. It is better to plan a concise session than to wear the participant out with a long script. A tired person starts guessing, and then the results are worthless.
Can I test my website myself, without an agency?
Yes. On one condition: you do not give hints and you do not defend the design when someone runs into trouble. You are there to observe, not to explain. It is good to have a second person taking notes so that nothing slips away.
How is a usability test different from a survey?
A survey collects declarations, meaning what people think they do. A test shows what they really do when faced with a specific task. And that is why it often brings out problems that participants would not even mention themselves.








