Skill Maps

Stage 5: Test — Validate with Real People

Testing is the moment of truth. Here you find out if your solution works. The main rule: Don't defend your idea. Listen.

Why Test?

1. You Are Not the User

What seems obvious to you might be confusing to others. You know how the system works, the user doesn't.

2. Cheaper to Find Errors Now

Fixing a problem at the prototype stage costs $1. After launch—$100.

3. Insights You Wouldn't Think Of

Users do things you didn't predict. They click the wrong places, read the wrong things, understand differently.

How to Conduct Testing?

1. Usability Testing

This is observing how a user completes a task with your prototype.

Steps:

  1. Prepare a task scenario:

    • ❌ Bad: " Try using the app. "
    • ✅ Good: " Imagine you need to pay an electricity bill for $150. Try to do this. "

  2. Find participants (5-8 people from target audience).

    • Why 5? Jakob Nielsen's research showed: 5 users find 85% of problems. More is redundant.

  3. Conduct one-on-one tests:

    • Sit nearby but don't help.
    • Ask them to think aloud: " Tell me what you see now and what you're going to do. "

  4. Observe and record:

    • Where did the user stumble?
    • What did they say?
    • What emotions did they show? (Frustration, joy, confusion)

  5. Don't explain or defend:

    • ❌ " But you can see the button is right there! "
    • ✅ " Interesting, why didn't you notice the button. Tell me what you were looking for? "

2. A/B Testing

You show two groups of users different versions and see which works better.

Example:

  • Version A: Green " Buy " button.
  • Version B: Red " Add to Cart " button.

Watch metrics: which version gets more clicks?

When to use:

  • When you already have traffic (live product).
  • For testing details (color, text, placement).

3. Guerrilla Testing

You go to a cafe, park, or shopping center and ask random people to test your prototype for 5 minutes.

Pros:

  • Fast and free.
  • Diverse audience.

Cons:

  • People might not be your target audience.
  • Little time for deep testing.

4. Remote Testing

User tests at home, and you watch screen recording or video call.

Tools:

  • Maze, UserTesting, Lookback: Platforms for remote testing.
  • Zoom + Figma: Simple way for startups.

Pros:

  • Can test people from other cities and countries.
  • User in familiar environment (more natural behavior).

Cons:

  • Harder to read emotions and body language.

What to Measure?

Qualitative Metrics

  • Where did the user stumble? (Pain Points)
  • What did they say? (Quotes)
  • What emotions did they show? (Joy, frustration, confusion)

Quantitative Metrics

  • Task Success Rate: How many users completed the task?
  • Time on Task: How long did it take?
  • Error Rate: How many mistakes did they make?
  • SUS (System Usability Scale): A 10-question survey giving usability score from 0 to 100.

Analyzing Results

1. Affinity Diagram

Write all observations on sticky notes. Group similar problems.

Example:

  • Group 1: " Couldn't find payment button " (5 people).
  • Group 2: " Didn't understand how much to pay " (3 people).
  • Group 3: " Were afraid to press 'Confirm' " (4 people).

2. Prioritizing Problems

Not all problems are equally important. Use a matrix:

  • Critical: User can't complete the task.
  • Serious: User completes the task but with difficulty.
  • Cosmetic: Minor inconveniences.

Fix critical ones first.

Iteration: Going Back

Design Thinking is a cycle, not a straight line. After testing, you can:

  • Return to Prototype (fix details).
  • Return to Ideate (come up with a new solution).
  • Return to Define (reformulate the problem).
  • Return to Empathize (turns out you didn't understand the user).

This is normal! Failing at the test stage is success because you learned the truth before launch.

Practical Example: Instagram

Originally Instagram was called Burbn and was a complex check-in app with lots of features. Tests showed: people only use photo filters. The team removed everything else, left photos—and the app exploded.