An expert inspection method where evaluators simulate the steps a first-time user would take to complete a task, asking at each point whether the user would know what to do next and whether the interface confirms progress. It targets learnability specifically without requiring real participant recruitment.
Common contexts
- Evaluating a new account setup flow for a first-time user with no prior knowledge of the product
- Reviewing a redesigned checkout before launch to catch steps where progress signals are ambiguous
- Using a cognitive walkthrough to pre-screen a prototype before investing in moderated user sessions
Use when
Run a cognitive walkthrough when you need a fast, low-cost evaluation of learnability — particularly before usability testing, to catch obvious issues that would waste participant time. It's most effective for task-critical flows where a first-time user's failure has real consequences.
Avoid when
A cognitive walkthrough reflects what expert evaluators imagine a novice will experience, not what a real user actually does — it's a useful filter, not a substitute. If the task involves complex real-world context or emotional motivation, user testing will always surface things an expert walkthrough misses.
The two questions of a cognitive walkthrough — 'will the user know what to do?' and 'will they know they did it correctly?' — are deceptively simple, but most interface failures can be traced back to a 'no' answer at one of those two points.
Real-world examples
- Microsoft's usability team performs cognitive walkthroughs of new Office features by simulating a first-time user's thought process step-by-step before any live user testing occurs.
- The UK's Government Digital Service conducts cognitive walkthroughs of citizen-facing forms to identify steps where users would not know what action to take next.
- PayPal used cognitive walkthroughs during a major checkout redesign to identify points where new users would lack the knowledge to proceed, leading to inline guidance copy being added.