Usability Testing & Iterative Design
UI/UX Design Principles
Chapter 9 · Usability Testing & Iterative Design
Chapter 2 introduced research methods; Chapter 7 covered expert-based heuristic evaluation. This chapter goes deep on the real thing both of those set up: watching a real person actually attempt real tasks with the prototype from Chapter 8, then feeding what you learn back into another round of design.
Moderated vs. Unmoderated Testing
| Type | How it works | Real tradeoff |
|---|---|---|
| Moderated | A facilitator guides the session live, asking follow-up questions in real time | Richer, deeper insight; slower and more resource-intensive to run |
| Unmoderated | A participant completes tasks alone, often via a remote testing tool | Faster, cheaper, more participants; no chance to ask a real-time follow-up |
The Think-Aloud Protocol
Asking a participant to verbalize their thoughts continuously while using the interface — "I'm looking for a way to change my email... I don't see it, so I'll try this menu" — surfaces the real reasoning behind a click, not just the click itself. It's the single most common technique used inside a moderated usability test.
Real Task Metrics
- Completion rate — the real percentage of participants who finished the task successfully
- Time on task — how long it genuinely took, compared against a reasonable benchmark
- Error rate — how many real mistakes were made along the way, even in a task eventually completed
A Real Correction: Does Being Watched Change Behavior?
It's extremely common to hear that recording or observing a usability test participant meaningfully distorts their real behavior — often citing the Hawthorne effect, the idea that people change behavior simply because they know they're being observed.
The Iterative Design Loop
A single test round is rarely the end of the process — the same non-linear spirit from Chapter 1's own Design Thinking stages applies directly here:
Test
Run the session with the current prototype (Chapter 8's own fidelity level).
Learn
Identify real, specific problems from what was observed.
Redesign
Change the design to address what was actually learned.
Retest
Confirm the fix actually worked — and didn't introduce a new problem.
This loop is exactly why Nielsen's own real research (Chapter 2) recommends three separate 5-user rounds over one large 15-user study — the loop only works if there's a real retest after each redesign.
Hands-On Exercises
A team needs deep, real-time follow-up on why participants hesitate at a specific step, but also needs to test with 50 participants within one week. Explain, in your own words, why this creates a real tension between moderated and unmoderated testing, and how the team might resolve it.
📄 View solutionA stakeholder argues that recording usability test sessions should be avoided entirely, since "the Hawthorne effect proves people behave completely differently on camera." Explain, in your own words, why this overstates what the real research on the Hawthorne effect actually supports.
📄 View solutionA team redesigns a checkout flow based on one usability testing round, ships the change immediately, and moves on without testing again. Explain, in your own words, which real step of the iterative design loop this skips, and what real risk that creates.
📄 View solutionChapter 9 Quick Reference
- Moderated (live, deep, slow) vs. unmoderated (fast, scalable, no real-time follow-up) testing
- Think-aloud protocol — verbalizing reasoning while completing a task, surfacing the "why" behind a click
- Real task metrics: completion rate, time on task, error rate
- Real correction: the original Hawthorne effect studies are now considered weak evidence by later analysis (Levitt & List; Nisbett's "glorified anecdote") — being observed is worth being mindful of, but isn't the large, proven distortion the popular version claims
- The iterative loop: Test → Learn → Redesign → Retest — skipping the retest defeats the point of the loop