Usability Testing Basics: How to Find UX Problems Without a Research Lab

You can't find usability problems by testing your own app. You built it. You know how it works. Your mental model is the product's mental model.

Users don't share your mental model. Watching them try to use your app for the first time is humbling and necessary.

---

Why Five Users Is Enough

Research from Jakob Nielsen shows that five users uncover 85% of usability problems. Beyond five, you're mostly discovering the same problems again.

The first user reveals the biggest problems. The second confirms them and finds a few more. By the fifth user, you've identified the priority issues and have enough signal to act.

This is why expensive, large-scale usability testing isn't necessary for most apps. Five people, two hours, and you know what to fix.

---

What Usability Testing Is (and Isn't)

Is: Observing users attempt to complete tasks with your app, noting where they struggle, get confused, or take unexpected paths.

Is not: Asking users what they think. Opinion surveys are user research. Usability testing is behavioral observation.

The gap between what users say and what they do is significant. "I would totally use that feature" from a user who never opens the settings menu where you put it.

---

Setting Up a Lightweight Test

Participants: People who match your target user profile. Ideally not your friends or colleagues — they know too much about your work. 5 participants.

Tasks: 3-5 specific tasks that represent core user journeys:

• "You've been using the app for a week. Create a task for next Monday's meeting with John."
• "You completed the grocery shopping. Mark that task as done."
• "Find all tasks you haven't completed that were due last week."

Tasks should be realistic scenarios, not step-by-step instructions. "Click the + button and type a title" is not a usability test — it's reading the manual.

Facilitation: You observe. You don't help. Even when participants are confused.

The hardest part: staying silent when participants struggle. The moment you help is the moment you invalidate the observation.

---

The Think-Aloud Protocol

Ask participants to narrate their experience as they use the app:

"As you're trying to complete each task, please tell me what you're thinking and what you're looking for. There are no wrong answers — we're testing the app, not you."

Think-aloud reveals:

• What users expect to find and where they look for it
• When they're confused but not giving up
• Incorrect mental models they have about how the app works
• Moments where they succeed by luck rather than understanding

---

Analyzing and Prioritizing Findings

After five sessions, categorize each observed issue:

Issue	Frequency	Severity	Priority
Users can't find the completed tasks filter	4/5	Major	P1
Users expect to swipe to delete, but find the menu confusing	3/5	Minor	P2
Date picker format is unclear to one user	1/5	Minor	P3

Focus on issues that were frequent (3+ users) and severe (major functionality affected). Minor issues affecting one user are signals to monitor, not immediate action items.

---

Remote Usability Testing

In-person testing isn't always possible. Remote alternatives:

Moderated remote: Video call with screen share. You observe in real time. Ask participants to share their screen and use the app while talking you through it.

Unmoderated remote: Tools like Maze, UserTesting, or Lookback record task completion on real users without moderation. You review videos afterward.

Moderated is better for complex tasks where nuance matters. Unmoderated is faster and cheaper for well-defined task flows.

---

What to Do With the Findings

Write a brief report with:

• Top 3-5 findings with video clips or screenshots showing the behavior
• Severity and frequency for each
• Suggested improvements (but not final decisions — design iteration comes next)

Share with the team or, as a solo developer, use it as your prioritized UX fix list for the next sprint.

---

Testing Your Android App Specifically

For Android apps, specific things to watch during usability testing:

• Do users find their way through navigation without guidance?
• Do users understand bottom navigation vs top navigation affordances?
• Does the back button behavior match expectations?
• Do swipe gestures work intuitively or do users look for buttons?
• Does the keyboard dismissal behave as expected?

These are Android-specific conventions. Violating them creates friction even if users can technically accomplish their goals.

---

Takeaways

• Five users reveal 85% of usability problems — you don't need a big study
• Tasks should be realistic scenarios, not step-by-step instructions
• Stay silent during the test — your intervention invalidates the observation
• Think-aloud protocol reveals mental models and confusion that task completion alone misses
• Prioritize by frequency × severity — fix what affects most users most severely first