What is user testing?
User testing is the practice of watching real people attempt real tasks with your product (interviews, usability sessions, surveys and A/B tests). Design decisions then rest on observed behaviour rather than opinion. It surfaces the problems users actually hit, ranks them by severity, and settles internal debates with evidence.
Why does user testing matter?
Every product team carries assumptions about its users. Most of those assumptions are wrong, and they fail in ways nobody predicts from a meeting room. User testing checks them against real behaviour, and it is at its cheapest while the design is still easy to change. A usability problem caught in a testing session costs an afternoon to fix. After launch, the same problem costs a release cycle, a support queue, and a measure of user trust.
Testing also settles arguments. A team can debate a flow for weeks, then watch two participants fail it the same way in a single session. Once everyone has seen the same failure, the discussion moves to the fix.
- 01User testing observes behaviour. It does not collect opinions, and what people do outranks what they say.
- 02Five participants per round surface the majority of severe usability issues in a single flow.
- 03Test before the design is polished: feedback on a finished-looking screen drifts to aesthetics.
- 04Qualitative sessions find the why. Quantitative methods (surveys, A/B tests) size the how much.
- 05The output is a ranked, evidenced list of problems, not a report that sits in a drawer.
When should you run user testing?
The honest answer is: earlier than feels natural. The strongest signal comes from testing prototypes before a line of production code exists, because at that stage nothing is precious and everything is fixable. We test at three moments in most engagements. The current product baselines where it hurts, prototypes settle the direction, and the built product confirms the fixes landed.
If you can only afford one round, test the riskiest assumption, not the finished interface. A single afternoon with five participants against a clickable prototype answers more than a month of internal review.
- 01
Frame the assumption
We write down our assumptions about who the user is, what they came to do, and what would count as failure. The session exists to prove them wrong.
- 02
Recruit real users
Participants who match the product's real users, in their context where it matters. For South African products that includes testing on the devices and data conditions people really have.
- 03
Run the sessions
A task, not a tour. The participant works while we watch, stay quiet, and let the failures happen. Rescuing someone thirty seconds in throws away the finding.
- 04
Rank and deliver
Findings arrive as a severity-ranked list tied to the screens they occurred on. We walk your team through it, so the next sprint's fixes are already scoped.
Qualitative or quantitative: which do you need?
Both, at different moments. Qualitative methods (moderated sessions, contextual interviews) explain why people struggle, and five participants per round is enough to see the pattern. Quantitative methods (surveys, analytics, A/B tests) measure how many users are affected and which variant performs better, and need volume instead of depth. A mature practice alternates: qualitative to find and frame problems, quantitative to size and confirm them.
What’s included in User Testing?
- Moderated usability testing
- A facilitator watches a participant attempt real tasks, probing the why behind each stumble. The richest signal per session, and the fastest way to understand a failure.
- Unmoderated testing
- Participants complete tasks in their own time while the tool records screens and outcomes. Depth per session drops and the weekly count climbs.
- User interviews
- Structured conversations about needs, workflows and frustrations: the qualitative ground truth that shapes what is worth testing at all.
- Surveys
- Surveys give quantitative reach. They size how widely a finding applies, and track sentiment over time across a whole user base.
- A/B testing
- Two live variants, one measured winner. The method of choice once a product has the traffic to answer questions statistically.
Seen in our work
User Testing is not a line on a rate card. It runs through the engagements below.
- Five participants per round is the working standard for qualitative usability testing, enough to surface the majority of severe issues in a single flow. Run several small rounds: five users, fix what they found, five more.
- Usability testing is one method inside user testing. User testing covers every way of learning from users (interviews, surveys, field observation, A/B tests). Usability testing specifically watches someone attempt tasks with an interface to find where it fails them.
- Before, during and after design, not as a final check. Test the existing product to baseline problems, test prototypes to choose between directions cheaply, and test the built product to confirm the fixes worked. The earlier a round runs, the cheaper its findings are to act on.
- A focused qualitative round (five moderated sessions, analysis and a ranked findings list) fits inside one to two weeks. Recruiting hard-to-reach participants extends it. Unmoderated and survey-based rounds run faster in parallel.
- Yes. That is when it is most valuable. We test prototypes exactly like live products. Participants attempt tasks against clickable screens, and the findings shape the build before development costs accrue.