We specialize in usability testing of mobile app prototypes, offering both moderated and unmoderated options. Our certified UX researchers guarantee actionable insights. Tests start from $800 for a basic unmoderated session.
How to Conduct Usability Testing on Mobile App Prototypes
We run usability tests where a real person performs a task in a prototype, and we silently observe — no hints. That "silent" part is the hardest. The urge to say "no, tap here" kills the entire value of the test. The moment a user freezes for three seconds in front of a button that seems obvious to the team — that's the data. Our experience: 10+ years in mobile development and 50+ tests for fintech, e-commerce, and B2B apps. Every test will uncover critical issues before launch. Typically, 20% of problems account for 80% of user frustration.
How testing works
Test formats: moderated and unmoderated
Moderated usability test — a facilitator is present (online or in person), gives tasks, asks follow-up questions. Suitable for Figma/XD prototypes: the facilitator controls transitions between states that the prototype doesn't cover. Record via Zoom + device screen (iOS ReplayKit over AirPlay, Android via scrcpy or USB recording).
Unmoderated remote test — the participant runs the test independently, recording is sent automatically. Tools: Maze (integrates directly with Figma), Useberry, UserZoom. Maze provides automatic metrics: misclick rate, time on task, path analysis. For quantitative hypothesis validation — faster and cheaper than moderated.
The process involves:
- Define objectives and tasks.
- Prepare prototype and screening questionnaire.
- Recruit target participants.
- Run moderated or unmoderated sessions.
- Analyze data and prioritize issues.
- Report findings with video clips.
| Feature | Moderated | Unmoderated |
|---|---|---|
| Depth of insights | High (can clarify) | Low (recording only) |
| Number of participants | 5–8 | 20+ |
| Analysis time | More | Less |
| Cost | Higher | Lower |
For mobile app prototypes, we prefer testing on real devices, not desktop. Figma Mirror or Maze on-device — the prototype opens on the participant's phone. Touch targets that work on a 1440p monitor break on a 360x780dp screen.
Why testing on real devices matters
Testing on a simulator does not replicate app behavior under real conditions: sensor sensitivity, network speed, overlays like notifications. We recorded a case where 4 out of 6 users could not tap the "buy" button because its bottom was covered by the iPhone X home bar. This was only revealed on a physical device. We use Lookback.io for moderated sessions and Maze for unmoderated.
How many participants do you need?
The classic Nielsen formula: 5 participants uncover 85% of usability issues for one user group (Jakob Nielsen, NN/g). In practice, 5–8 participants for a moderated test of one prototype. For unmoderated with quantitative metrics — at least 20 participants to achieve statistical significance.
Crucial: all participants must match the target profile. Testing with the wrong audience yields false insights.
How do we recruit participants?
We use a screening questionnaire covering demographics and usage experience. We filter out "professional testers" who know the scenarios. For B2B apps, we recruit via LinkedIn; for B2C, via targeted ads.
What we test and how we build tasks
Tasks are scenario-based, not directive. Not "click the registration button," but "imagine you want to create an account — do it." The task should describe the goal, not the path.
Typical task set for an e-commerce prototype:
- Find a product in category [X] and add it to the cart
- Place an order with delivery to a new address
- Find the status of your last order
After each task, we ask a Single Ease Question (SEQ): "How easy was it to complete this task?" on a 7-point scale. After the entire test, we use the System Usability Scale (SUS), 10 questions. SUS benchmark: above 68 is acceptable, above 80 is good.
Analysis of results
After each session — brief debriefing notes while fresh. After all sessions — affinity mapping of observations: which problems occurred with several participants, which are isolated.
We prioritise results using a frequency × severity matrix:
| Frequency | Severity | Priority |
|---|---|---|
| 3+ participants | Blocks task | Critical — fix before release |
| 3+ participants | Slows but not blocks | High |
| 1–2 participants | Any | Medium/Low |
Deliverables
A report with video clips of key moments (timestamp + problem description), a prioritised list of UX problems, and fix recommendations with alternative solutions. Not just "button not obvious" — but "3 out of 6 participants tapped on the cart icon in the header instead of the floating button — consider replacing the FAB with an inline button on the product card."
What is included in the work?
- Development of scenarios and screening questionnaire
- Participant recruitment according to the target profile
- Conducting sessions (moderated/unmoderated)
- Data analysis with expert evaluation
- Preparation of a report with video, metrics, and recommendations
- Consultation on fixing identified issues
Timeline: from 2 to 5 business days turnkey. Write to us — we will evaluate your project for free. Contact us to discuss details.
Common questions: How many participants? For moderated, 5–8; for unmoderated, 20+. What metrics? SEQ, SUS, task time, misclick rate, path analysis. The report includes video clips and prioritized recommendations.