Mobile App Usability Testing
Usability test is when real person tries to complete real task in prototype while you watch silently and don't hint. Exactly "silently" — this is hardest part. Urge to say "no, click here" kills all test value. Moment when user freezes for three seconds before obvious (in team's opinion) button — this is the data.
Testing formats
Moderated usability test — facilitator present on session (online or in-person), gives tasks, asks follow-up questions. Good for Figma/XD prototypes: facilitator controls transitions between states prototype doesn't cover. Recording via Zoom + device screen (iOS ReplayKit via AirPlay, Android via scrcpy or USB recording).
Unmoderated remote test — participant passes test independently, recording sent automatically. Tools: Maze (integrates directly with Figma), Useberry, UserZoom. Maze gives automatic metrics: misclick rate, time on task, path analysis. For quantitative hypothesis checking — faster and cheaper than moderated test.
For mobile app prototypes prefer testing on real devices, not desktop. Figma Mirror or Maze on-device — prototype opens on participant's phone. Touch target sizes that work on 1440p monitor break on 360×780dp screen.
How many participants needed
Nielsen's classic formula: 5 participants reveal 85% usability problems in one user group. Practice — 5–8 people for moderated test of one prototype. For unmoderated with quantitative metrics — from 20 participants for statistically significant data.
Important: all participants must match target profile. Test with "wrong" audience gives false insights.
What we test and how to structure tasks
Tasks — scenariomatic, not directive. Not "click registration button", but "imagine you want to create account — do it". Task should describe goal, not path.
Typical task set for e-commerce prototype:
- Find product in category [X] and add to cart
- Place order with delivery to new address
- Find status of your last order
After each task — Single Ease Question (SEQ): "How easy was this task?" on 7-point scale. After full test — SUS (System Usability Scale), 10 questions. Benchmark SUS: above 68 — acceptable, above 80 — good.
Results analysis
After each session — quick debrief notes while fresh. After all sessions — affinity mapping observations: which problems occurred with several participants, which — single.
Prioritize results by matrix frequency × criticality:
| Frequency | Criticality | Priority |
|---|---|---|
| 3+ participants | Blocks task | Critical — fix before release |
| 3+ participants | Slows but doesn't block | High |
| 1–2 participants | Any | Medium/low |
Output
Report with video clips of key moments (timestamp + problem description), prioritized UX problem list, fix recommendations with alternative solutions. Not just "button unclear" — "3 of 6 participants tapped cart icon in header instead of floating button — consider replacing FAB with inline button in product card".
Timeline: 2–3 working days on testing and analysis (with ready participant pre-screening). Cost — individually.