Personalize for user goals.
Adapt what helps the user’s task, not only what increases engagement.
Contents 8. Adaptation & Evolution 5. Personalization
How the system adapts to the individual.
First, ask yourself
“Does personalization in your system help each user without making the system unpredictable?”
Mission statement
Ensure users benefit from adaptation by making it visible, controllable, and consistent with the shared model of the system.
Key heuristics Personalization
- Personalization must serve the user’s goals, not only engagement metrics.
- Adapted elements must be identifiable as adapted.
- Users must be able to see why something was personalized.
- Users must be able to adjust or turn off personalization.
- Core structure and interaction rules must not change per user.
- Adaptation must change gradually, not abruptly.
- Users must be able to return to the default experience.
- Personalization must not narrow what users can discover.
- Explicit settings must take precedence over inferred ones.
Executive brief
The system must adapt to individuals without breaking what all users share.
Core questions Personalization
“What does the system change for this user, and does the user know it?”
“Can the user see, adjust, or reset it?”
Focus areas
Visibility-first
“Which parts of the interface are personalized?”
“Can users tell why?”
Control-aware
“Can users tune or disable adaptation?”
“Is there a way back to default?”
Stability-aware
“Do core layout and behavior stay the same for everyone?”
“Does adaptation change too fast to learn?”
AI-aware
“Is adaptation driven by learning the user did not consent to?”
“Does personalization create a filter bubble?”
UX directives Personalization
Mark adapted content.
Indicate which elements are personalized.
Explain personalization.
Show the reason an item was adapted or recommended.
Let users tune adaptation.
Provide controls to adjust or turn off personalization.
Keep the core model shared.
Do not personalize structure or interaction rules that all users rely on.
Adapt gradually.
Avoid abrupt changes that invalidate what users have learned.
Provide a way back to default.
Let users reset personalization at any time.
Preserve breadth of discovery.
Prevent personalization from hiding content users have not yet encountered.
Rank explicit settings above inference.
Let stated preferences override inferred ones.
Executive summary
- Personalization is adaptation to the individual within a shared model, not a separate product per user.
- It increases relevance without breaking predictability.
- The system must make adaptation visible, explainable, and controllable.
- Core structure and rules must remain the same for everyone.
- Adaptation must be gradual, reversible, and subordinate to explicit settings.
- Personalization succeeds when each user gets a better fit without losing the ability to predict or reset the system.
Success indicators
- Personalized elements are identifiable.
- Users can see why something was adapted.
- Users can adjust, disable, or reset personalization.
- Core layout and behavior are the same for all users.
- Users still encounter content outside their history.
One-line summary
Before a system can fit the individual, it must remain the same system for everyone.