Tremoloo logo
English
English
Tremoloo logo
English
brown brain decor in selective-focus photography

Cognitive Interviews: How to Learn What Users Really Think

UX Research

cognitive interview, UX research methods, think aloud, contextual inquiry, user research techniques, what users think, Design for How People Think

Summary

Cognitive interviews are research conversations designed to surface the automatic thinking users cannot report directly.

Drawing from Design for How People Think, the method combines observation with targeted probes: watch users attempt real tasks, ask them to narrate their reasoning, and dig into hesitations and surprises.

The goal is not opinions but cognition: what they noticed, remembered, assumed, and decided.

For designers, the outcome is insight into why users fail, which is the only kind that reliably tells you how to fix it.

Users will happily tell you what they think about your product. The trouble is that their explanations are mostly fiction: rationalizations constructed after the fact by a mind that did not witness its own automatic processes. Cognitive interviews, a core method in John Whalen’s Design for How People Think, are how you get closer to the truth.

What is a cognitive interview?

It is a research session built around real behavior rather than reported opinion. Instead of asking “would you use this feature,” you give the user a real task and observe, then probe the thinking behind what you saw. The method borrows from cognitive psychology and, yes, from how detectives interview witnesses: the goal is reconstructing what actually happened in the mind, not collecting judgments.

How do you run one?

  1. Start with a real task. Behavior first, talk second. “Show me how you would renew your subscription” beats “what do you think of our subscription page.”

  2. Watch for the tells. Hesitations, hovering, rereading, backtracking. Each is a visible trace of invisible cognition.

  3. Probe in the moment. “What are you looking for right now?” “What did you expect to happen?” Short, neutral, immediate.

  4. Chase the assumptions. When users misread a label, the interesting question is what they thought it meant and why.

  5. Separate observation from inference. Write down what you saw before you decide what it means.

What do designers gain?

The difference between knowing that users fail and knowing why. “Thirty percent abandoned checkout” is a metric. “They thought the CTA was an ad because it sat below a banner” is a fix. Cognitive interviews produce the second kind.

Why does Tremoloo share this book?

Because the quality of research findings depends on the quality of the probing, and this book teaches probing better than anything else we have found. It has raised the bar for how our own researchers run sessions.

Want research that explains the why behind your metrics? Book a research session with us.

Ready to Launch
Your Next Project?

If you’re tired of iterating in circles, we can help. We research,
design, and refine products with focused teams.