Theory
Watch, and be horrified
You interviewed students; they said FestConnect seemed fine. Confident, you sit ONE first-year in front of the prototype and say: "Register for Garba Night."
And you watch, silently, as they tap the wrong thing, hunt for the button, panic at the date field, and after 3 minutes ask, "is this right?"
Everything the interviews missed, this one silent observation revealed. That is usability testing: watching real users attempt real tasks, and it is the most honest feedback in design, because behaviour cannot lie the way words can.
Theory
Usability testing, formally
Usability testing is observing representative users as they attempt realistic tasks with a design or prototype, to find usability problems.
The defining principle: you WATCH behaviour, you do not just collect opinions: this is the antidote to the say-do gap. Users said it was fine; watching them proves otherwise.
One liberating fact (from usability research widely associated with Jakob Nielsen): testing with only about 5 users uncovers the majority of usability problems. You do not need hundreds: a handful of the RIGHT users, watched carefully, exposes most of what is broken. Small, cheap, devastatingly effective.
At a glance
Usability testing methods (the overview)
| Method / axis | What it is | When to use |
|---|---|---|
| Moderated | A facilitator guides live, can probe | Deep insight, understand WHY they struggle |
| Unmoderated | Users do tasks alone, often remote | Fast, cheap, larger numbers |
| In-person vs remote | Same room vs online tools | Context vs reach |
| Think-aloud | Users voice thoughts while doing tasks | Hear their reasoning in real time |
| Qualitative vs quantitative | Observe problems vs measure metrics | Find issues vs prove them numerically |
Theory
Think-aloud, and what to measure
The signature technique is think-aloud: you ask users to say their thoughts OUT LOUD while working: "I'm looking for a register button... is this it? No... where's the date?" It turns silent confusion into a narrated map of where the design fails.
And usability has 3 measurable qualities:
- effectiveness: can they complete the task at all (success rate)?
- efficiency: how much time and effort does it take?
- satisfaction: how did it feel?
Related expert methods worth naming: heuristic evaluation (experts check the design against usability principles) and A/B testing (compare 2 versions with real users). But watching real users remains the gold standard.
Quiz
Roughly how many users does usability testing typically need to uncover MOST usability problems?
- Hundreds, for statistical significance
- Around 5 representative users
- Exactly one is always enough
- Zero: expert opinion replaces user testing
Show the answer
Around 5 representative users
A well-known usability finding (associated with Jakob Nielsen) is that testing with about 5 representative users surfaces the MAJORITY of usability problems: because the same big issues recur quickly across users, extra testers mostly repeat what you have already seen. This makes testing cheap and practical, which is the point. Option A confuses usability testing (finding PROBLEMS, qualitative) with quantitative studies that DO need large samples. Option C undersells it: one user misses issues others would hit. Option D ignores that watching real users catches problems experts overlook. The takeaway: you do not need many users, you need to actually WATCH a few.
Think first
The hardest rule: do not help
During the test, the first-year is stuck, visibly struggling to find the register button. Every instinct says 'just point them to it'. Why must you NOT? Then tap.
Show the answer
Because the struggle IS the data. The moment you help, you have hidden the exact problem you came to find: in real life, no designer will sit beside every user whispering hints. If they cannot find the button alone, the BUTTON is the problem, not the user, and you must let the difficulty play out to see its full shape. Helping feels kind but destroys the test. The discipline: give TASKS, not instructions; then go quiet and OBSERVE, however uncomfortable. (You reassure them it is the DESIGN being tested, not them: but you do not rescue them.) Silent observation of genuine struggle is the whole method.
Watch out
Usability testing slips
Helping or leading: rescuing a stuck user erases the finding; stay quiet, give tasks not steps.
Asking instead of watching: 'do you like it?' returns to the say-do gap; OBSERVE behaviour.
Testing yourself or friends: use representative REAL users.
Waiting until it is built: test PROTOTYPES early, when fixing is cheap (the cost-of-change curve); you can even test a paper wireframe.
Blaming the user: if they struggle, the DESIGN failed, not them.
Theory
Unit 2 closes: you know your users
You now hold the research toolkit: why research matters, personas to model users, interviews to hear them, usability testing to watch them. FestConnect's users are no longer a guess. Unit 3 turns understanding into DESIGN: interaction design (how the product behaves), information architecture (how content is organised), wireframes (the blueprint), and navigation. And usability testing returns in Unit 5 with its full step-by-step process. Understanding done; designing begins.
Summary
Key takeaways
- Usability testing observes representative users attempting realistic tasks to find usability problems.
- It watches BEHAVIOUR, not opinions: the antidote to the say-do gap.
- About 5 representative users uncover most usability problems: testing is cheap and effective.
- Methods: moderated vs unmoderated, in-person vs remote, think-aloud, qualitative vs quantitative.
- Measure effectiveness (task success), efficiency (time/effort), satisfaction; related methods: heuristic evaluation, A/B testing.
- Give tasks not instructions, stay silent, do not help; if users struggle, the DESIGN failed, not them.
- Memory hook: five users, real tasks, quiet observer, honest truth.