You can spend weeks polishing a design and still watch someone fail to find the “Pay now” button in ten seconds. That is why usability testing matters: you give real people real tasks, watch where they struggle, and fix the design before it ships. The good news is that your first test needs no lab, no budget and no special software. This guide to usability testing for beginners takes you from planning the test to sharing what you found.
This guide walks through a complete small study, from choosing participants to sharing what you found, with a checklist you can copy. It is written for students and freshers, and it works just as well for a freelance client project.
What usability testing is (and the main types)
In a test, a participant tries to complete tasks with your product or prototype while you observe. You are testing the design, not the person. The output is a list of problems, ranked by how badly they hurt users, plus evidence you can show your team.
It is different from asking people what they think of a design. Opinions are cheap; behaviour is what counts. A participant may say “looks clean” and then get lost in the menu.
| Type | How it works | Good for |
|---|---|---|
| Moderated | A facilitator runs the session live and can ask follow-up questions | Early prototypes, complex flows, learning why people struggle |
| Unmoderated | Participants complete tasks alone, usually through a testing platform that records them | Quick checks of simple tasks, larger numbers of people |
| In-person | You sit beside the participant | Mobile apps, physical context, reading body language |
| Remote | You meet over a video call with screen sharing | Reaching people in other cities, lower cost |
These combine: “moderated remote” over a video call is the most practical starting point for most beginners in India.
How many participants do you need?
The best-known guidance is Jakob Nielsen’s 2000 article Why You Only Need to Test with 5 Users. Based on a model built with Tom Landauer, it estimates that five users uncover about 85% of the usability problems in a design. Each extra user finds fewer new issues than the one before.
The caveats are just as important, and people often skip them:
- The 85% figure is an average across projects, not a promise for your product.
- It assumes your users are fairly similar. If you have distinct groups (say, shop owners and their customers), test 3–4 people from each group.
- The real advice is to test iteratively: rather than one study with 15 people, run three rounds of five, fixing problems between rounds.
- It applies to qualitative testing, where you look for problems. If you need reliable numbers such as average task time, NN/g suggests around 20 users for quantitative studies.
Plan the test: tasks, scenarios and a script
Write realistic tasks
Start from your goals. If you created user personas, pick the two or three jobs those people most need to do. Turn each into a short scenario that gives context without giving away the answer.
- Weak: “Click the Bookings tab and select a date.” (It names the button, so you learn nothing.)
- Better: “You want a bike for this Saturday morning. Find out whether one is available and book it.”
- Give each task a clear end point, so you can judge success.
- Keep sessions to roughly 5–7 tasks so the participant does not tire.
Write a script and use think-aloud
A script keeps every session consistent. It covers your welcome, the consent request, a reminder that you are testing the design, the tasks, and a few closing questions such as “What was the most confusing part?”
Ask participants to think aloud: to say what they are looking at, expecting and feeling as they go. NN/g calls this perhaps the single most valuable usability tool. Your job is mostly to stay quiet. If someone goes silent, a neutral prompt like “What are you thinking now?” works. Avoid hints such as “Did you see the menu at the top?”, which change what they would have done.
Tip: Run a pilot session with a friend first. It will expose unclear tasks and broken prototype links before they waste a real participant’s time.
Recording and consent
Recordings let you review moments you missed and share clips with the team. Before you record anything, tell participants what you will record (screen, voice, face), who will see it, how long you will keep it, and that they can stop at any time. Get their agreement in writing, even if it is a short form.
Do not test with real passwords, bank details or personal data. Prepare dummy accounts and sample data, and delete recordings when the project no longer needs them.
What to measure
For a first study, keep metrics simple and record them in a spreadsheet, one row per participant per task:
| Metric | What you record | Why it helps |
|---|---|---|
| Task success | Completed, completed with difficulty, or failed | The clearest signal of whether a design works |
| Time on task | Seconds from start to finish | Shows friction; treat it as indicative with five users |
| Errors | Wrong paths, mis-taps, back-button use | Points to the exact screen that confuses people |
| Comments | Notable quotes from think-aloud | Explains the why behind the numbers |
You can also hand out the System Usability Scale (SUS) at the end. Created by John Brooke in 1986, it is a 10-item questionnaire with 5-point answers. To score it, subtract 1 from each odd-numbered answer, subtract each even-numbered answer from 5, add the results and multiply by 2.5 to get a score from 0 to 100. MeasuringU reports an average of about 68 across around 500 studies. Note that a SUS score is not a percentage.
Analyse findings and share them
Soon after each session, write down every problem you saw. Then group similar issues across participants (a quick affinity map on sticky notes or in FigJam works well) and count how many people hit each one.
Rate each issue using Nielsen’s 0–4 severity scale, which weighs how often it happens, how hard it is to overcome, and whether it keeps bothering users:
- 0 – not a usability problem
- 1 – cosmetic; fix if time allows
- 2 – minor; low priority
- 3 – major; high priority
- 4 – usability catastrophe; fix before release
Many issues map neatly onto Nielsen’s 10 usability heuristics, which helps you explain the cause. Share a short report: the top five problems, a screenshot or clip of each, how many participants were affected, and a suggested fix. Busy teams read one page; they rarely read twenty.
Finally, fix and test again. If you tested a low-fidelity design, my post on wireframes, mockups and prototypes explains how much fidelity each round needs.
Usability testing for beginners: checklist and budget tips for India
Copy-ready test plan
- Goal: what decision will this test help you make?
- Product and version: link to the prototype or build being tested.
- Participants: who they are, how many (start with 5), how you will recruit them.
- Format: moderated or unmoderated, remote or in-person, session length.
- Tasks: 5–7 scenarios, each with a clear success criterion.
- Script: welcome, consent, think-aloud instructions, tasks, closing questions.
- Consent form and dummy test data ready.
- Metrics sheet: success, time, errors, notes, optional SUS.
- Pilot session done and tasks revised.
- Analysis plan: severity rating, report format, date to share with the team.
Testing on a student or freelancer budget
- Recruit people close to your users. Classmates, relatives and neighbours are fine if they match the audience. Your designer friends are not; they think like designers.
- Ask for help, not favours. A clear 30-minute request and a thank-you, such as a coffee or a snack, goes a long way.
- Use tools you already have. Google Meet or Zoom for remote sessions, OBS Studio to record your screen, Google Forms for consent and SUS, and Google Sheets for notes. Check each tool’s current plan limits before you rely on them.
- Test prototypes, not code. A clickable Figma prototype is enough; see the Prototyping in Figma lesson.
- Turn it into a case study. Before-and-after screens backed by test findings make a strong portfolio case study.
Want the theory in a shorter form? The Usability Testing lesson on this site covers the basics. Then book five people for next week and run your first round of usability testing.
Frequently asked questions
Is 5 users always enough?
No. Nielsen’s estimate that 5 users find about 85% of problems is an average for qualitative tests with similar users. Test 3–4 per group if you have distinct user groups, run several rounds, and use larger samples (around 20) when you need reliable numbers.
Can I test with friends and family?
Yes, if they resemble your real users and you avoid helping them. Do not test with other designers or with anyone who worked on the product.
What is the difference between usability testing and a survey?
A survey records what people say they think. A usability test shows what they actually do when they try to complete tasks, which is far more reliable for finding design problems.
Do I need a finished product to test?
No. Paper sketches, wireframes and clickable prototypes can all be tested. Earlier tests are cheaper because problems are easier to fix.
What is a good SUS score?
MeasuringU reports an average of about 68 across roughly 500 studies, so scores above 68 are above average. Treat SUS from five people as a rough indicator only.
