Cognitive Walkthroughs

A cognitive walkthrough is a structured usability inspection method in which one or more evaluators work through a defined task step by step, asking at each action whether a first-time user would know what to do, whether they would recognise the correct action, whether they would know they have made progress, and whether they could recover from errors — producing a systematic record of where the interface fails to communicate its own operation to someone encountering it for the first time. Unlike heuristic evaluation, which assesses an interface against a set of general principles, the cognitive walkthrough is specifically focused on learnability — whether a new user can figure out the interface through exploration without prior training or documentation.

The method was developed by Clayton Lewis, Peter Polson, and colleagues at the University of Colorado in the early 1990s, building on cognitive science research into how users form action plans when interacting with unfamiliar systems. What makes it specifically valuable for product teams is its cost and speed: a cognitive walkthrough can be conducted by a designer or PM in a few hours without recruiting participants, producing actionable learnability findings on any interface that can be documented as a sequence of steps — including prototypes, competitor products, and flows that have not yet been built.

A cognitive walkthrough works by forcing the evaluator to adopt the perspective of a first-time user at each step of a task sequence. Instead of reviewing the interface as a whole, the evaluator moves through it action by action, asking four specific questions at each step that probe the learnability failure modes identified in cognitive science research. The discipline of answering these questions at each individual step forces the evaluator to confront moments of likely user confusion that are invisible at the level of overall impression.

When Monzo's growth team investigated why a meaningful proportion of users who downloaded the app were not completing account setup, a cognitive walkthrough of the sign-up flow was conducted before any user research was commissioned. The walkthrough defined the user profile as a smartphone user in their late twenties with moderate tech literacy, no prior experience with app-based banking, and a goal of opening a current account in under ten minutes.

At the step where users were asked to verify their identity by uploading a photo of their passport, the cognitive walkthrough identified a failure on question three: users would see the upload button and know what to press, but the label "Upload document" did not communicate clearly enough that the document needed to be the specific identity document previously selected, and that a blurry or partial photo would fail verification and require the step to be repeated. The feedback on a failed upload — a generic error message with no specific guidance — was a failure on question four. Both failures were identified without recruiting a single participant, in a session that took ninety minutes. Both were confirmed in the subsequent usability test.

Cognitive walkthroughs sit within the expert inspection toolkit alongside heuristic evaluation — both assess interfaces without recruiting participants, but each catches different categories of problems. Understanding how they differ helps teams deploy them at the right moment for the right question.

The most important pairing is cognitive walkthroughs with Usability Testing. Cognitive walkthroughs predict learnability failures from the evaluator's simulation of novice cognition; usability testing confirms whether real novice users actually experience those failures and surfaces failures the simulation missed. Neither alone produces a complete learnability picture — but the cognitive walkthrough before the usability test makes the test session more efficient by removing the predictable failures, allowing the test to surface the unpredictable ones.