The Blue Book of UX Directives
1. Human Cognition & Behavior

4. Mental Models

Contents 1. Human Cognition & Behavior 4. Mental Models

How users believe a system works.

First, ask yourself

“Do users form accurate beliefs to predict your system’s behavior correctly?”

Mission statement

Ensure users can predict system behavior by aligning with their existing mental models.

Key heuristics Mental Models

  1. System behavior must intentionally shape the mental model users form.
  2. Correct understanding must emerge from consistent interaction and feedback.
  3. Similar actions must produce similar outcomes across contexts.
  4. Users must be able to trace how actions lead to results.
  5. Behavior-influencing states must be visible.
  6. System behavior must align with user intuition, not technical convenience.
  7. Departures from established patterns must be prominent and justified.
  8. Metaphors must strengthen correct behavioral expectations.
  9. Mental models must support consistent recovery and cross-context application.

Executive brief

The system must actively shape accurate mental models or users will form their own, often incorrect ones.

Core questions Mental Models

“What do users think the system does?”

“Does that understanding let them reason, predict, and recover correctly?”

Focus areas

Responsibility-forward

“What understanding is the interface teaching?”

“What beliefs does it actively reinforce?”

Expectation-aligned

“Can users reliably predict what will happen next?”

“Does the system behave as they expect?”

Structure-aware

“Does the interface reveal how the system works, not just outcomes?”

“Are relationships, rules, and dependencies visible and understandable?”

AI-aware

“What does the user think the system knows, decides, or controls?”

“Does AI encourage accurate understanding or anthropomorphic illusion?”

UX directives Mental Models

Directive14/01

Design explicitly for the mental model users will form.

Do not allow users to infer core system behavior accidentally or inconsistently.

Directive14/02

Design consistent system behavior to teach the correct mental model.

Do not rely on documentation to teach users the correct mental model.

Directive14/03

Maintain consistent behavior to keep users’ mental models stable.

Similar actions must produce similar results across contexts.

Directive14/04

Make causality visible and traceable.

Users must be able to perceive how actions lead to outcomes.

Directive14/05

Make states visible when they influence behavior or outcome.

Avoid hidden conditions, as they corrupt learned mental models.

Directive14/06

Design for behavioral coherence rather than technical convenience.

Prioritize behavior that is intuitive to users over internal system logic.

Directive14/07

Make behavioral deviations explicit and justified.

Do not introduce unannounced changes to established patterns.

Directive14/08

Design metaphors that reinforce correct mental models.

Avoid metaphors that teach incorrect rules about system behavior.

Directive14/09

Design mental models that support recovery and transfer.

Ensure users can use their understanding to correct errors and apply knowledge across contexts.

Executive summary

  • Mental Models are explanatory structures users form about how the system works.
  • They are inevitable and continuously active.
  • Users form beliefs whether the system intends them to or not.
  • Mental Models determine whether users can reason correctly about action and consequence.
  • The system must teach the correct model through consistent behavior, not documentation.
  • Mental Models succeed when users can predict outcomes, recover from errors, and transfer knowledge across contexts.

Success indicators

  • The system behaves in ways users can easily understand.
  • Similar actions produce similar results across the system.
  • Users can see how their actions lead to outcomes.
  • Important system states are visible and easy to recognize.
  • Users can use their understanding to fix mistakes and continue working.

One-line summary

Before users can trust the system, they must be able to reason about how it works.