Seven simple steps to run a usability test

To run a usability test, define what you want to learn, choose realistic tasks, recruit five to eight participants who match your users, prepare a short script, run sessions where participants think aloud while you observe without helping, analyse patterns in what you saw and turn the biggest problems into prioritised fixes.

Why usability testing matters

Usability testing means watching real people try to complete real tasks with a product or prototype. It is one of the most reliable ways to find out whether a design works, because it shows what people actually do rather than what they say they would do. Teams are often surprised by what they see: labels that seemed obvious are misunderstood, important buttons go unnoticed and flows that looked simple turn out to be confusing.

Testing does not have to be expensive or time-consuming. A small number of short sessions, run regularly, usually reveals the most important problems. The seven steps below describe a simple process that any team can follow, whether testing a rough prototype or a live product. Our post on designing for usability explains the principles behind what you are testing for.

The seven steps at a glance

1. Define your goals

Start by deciding what you want to learn. Good research questions are specific, such as "Can first-time users book an appointment without help?" or "Do users understand the difference between our two pricing plans?" Clear goals keep the test focused and make it easier to decide which tasks to include and what to measure.

2. Choose realistic tasks

Tasks should reflect what real users want to do, framed as scenarios rather than step-by-step instructions. Instead of "Click the Book button", say "You need to see a dermatologist next week. Book an appointment that suits you." Avoid using the exact words from the interface, which can give away the answer. Three to five tasks usually fit into a 30 to 45 minute session.

3. Recruit the right participants

Participants should match the people who will use the product, including their experience level, language and devices. For most qualitative tests, five to eight participants per user group reveal the main problems. Recruit from your customer base, website visitors, research panels or specialist agencies, and offer a fair incentive. For Gulf audiences, include both Arabic and English speakers where relevant. Our guide to UX research agencies covers recruitment options.

4. Prepare a script

• A friendly introduction explaining that you are testing the product, not the person
• A request to think aloud while working
• Short warm-up questions about the participant’s background
• The task scenarios, read in the same way each time
• Follow-up questions after each task and at the end
• Consent for recording and how data will be used

5. Run the sessions

During each session, the facilitator reads tasks, encourages the participant to think aloud and observes without helping or leading. If participants ask for help, respond with neutral prompts such as "What would you do if you were on your own?" A second person can take notes, recording where participants hesitate, make mistakes, succeed or give up. Sessions can be held in person or remotely using video calls and screen sharing.

• Stay neutral: Avoid reacting to success or failure.
• Let silence work: Give participants time to think.
• Ask open questions: "What are you looking for?" rather than "Is it this button?"
• Record with consent: Video helps analysis and convinces stakeholders.

6. Analyse what you saw

After the sessions, review notes and recordings to find patterns. List each problem, how many participants experienced it and how serious it was. Rate severity, for example from minor annoyance to task failure. Problems that affected several participants and blocked important tasks should be prioritised. Short video clips of key moments are very effective when sharing results.

Common testing mistakes

• Testing with colleagues instead of real users
• Writing tasks that give away the answer
• Helping participants when they struggle
• Asking whether people like the design instead of watching them use it
• Running one big test instead of several small ones

Avoiding these mistakes keeps results honest and useful. Even an imperfect test teaches more than no test at all, so start small and improve the process each time.

7. Act on the findings and share them

Turn the most important problems into concrete design changes and add them to the team’s backlog. Share a short summary with stakeholders, focusing on the top issues, their impact and recommended fixes. Then test again after changes are made, to confirm they worked. Our guide to sharing research insights offers tips for presenting results so they lead to action.

Frequently asked questions

How many participants do I need for a usability test?

For qualitative testing, five to eight per user group usually reveals most major problems. Run several small rounds rather than one large test.

How long should a usability test session be?

Typically 30 to 60 minutes, including introduction, three to five tasks and follow-up questions.

Can usability tests be done remotely?

Yes. Moderated remote tests use video calls; unmoderated tools let participants complete tasks on their own.

When should we run usability tests?

Throughout the process: on early sketches, prototypes, before launch and on live products.

What is the think-aloud method?

Participants say what they are thinking as they use the product, which helps observers understand their expectations and confusion.

Do we need a lab for usability testing?

No. Most tests can be run in a quiet room or remotely over a video call with screen sharing.

Related articles

• Designing for usability: principles every team should know
• How to conduct your first UX design audit
• 15 user research methods beyond usability testing