Peak-End Rule
Retrospective evaluation of an experience can be disproportionately influenced by its most intense moment and its ending rather than its full duration.
Direct explanation
What it means
Memory compresses an episode. A painful breakdown or supportive recovery and the final state may shape later judgment, but average experience, duration, expectations, and domain also matter.
Designers should protect critical moments and endings without neglecting ordinary steps or manufacturing a theatrical finish around a poor service.
Research anchor: Daniel Kahneman, Barbara L. Fredrickson, Charles A. Schreiber, Donald A. Redelmeier, 1993
Research and interpretation
What the research found
Participants’ retrospective ratings of painful episodes were strongly associated with the most intense pain and the final moments, with duration comparatively neglected in some studies.
Peak and end effects appear in several domains, but a 2022 meta-analysis found both mean and peak/end summaries can predict retrospective judgments.
Popular advice versus careful use
Common shorthand: Only the best moment and ending matter; everything between is forgotten.
Treat peaks and endings as influential features of memory while measuring average, cumulative, and outcome quality too.
How designers apply it
- 01
Design humane recovery for a failed identity check.
- 02
End onboarding with a real completed outcome and next step.
- 03
Close a cancellation flow with confirmation and accessible records.
Original UI examples
mobile travel
Flight disruption
After a stressful rebooking, the final screen confirms itinerary, assistance, and refund status.
The ending reduces unresolved uncertainty without pretending the disruption was pleasant.
desktop admin
Data export
A long task ends with verified file location, contents, and deletion schedule.
The final state is useful and trustworthy rather than a confetti-only celebration.
Subscription cancellation
Less effective
Add a cheerful animation after a coercive maze of retention screens.
Better fit
Make the entire path direct and end with durable confirmation plus reactivation terms.
A pleasant ending cannot ethically compensate for prior obstruction.
Limits and failure modes
Effect size and which summary wins vary with episode structure, delay, measurement, and personal meaning.
Common mistakes
- Ignoring average and cumulative experience because two moments are memorable.
- Engineering emotional peaks to distract from service failure.
When it should not dominate
- To justify leaving routine friction unfixed.
- To intensify emotion in vulnerable or consequential contexts.
Responsible use
Accessibility
Completion states must communicate outcome without depending on animation, sound, or color and should remain available long enough to review.
On mobile: Restore context after interruptions so the true ending is not an unexplained dismissed screen.
Ethics and context
Do not manipulate remembered evaluation to mask harm; prioritize honest resolution and the quality of the full experience.
On desktop: Use persistent receipts and summaries rather than transient celebratory overlays.
Designer checklist
- Map the likely high and low moments.
- Verify the final state answers what happened and what is next.
- Measure the whole journey as well as retrospective rating.
Check your understanding
Which statement applies Peak-End Rule most carefully?
Practice and continue learning
See ideas in real interfaces
These external interfaces are editorial references, not examples created to demonstrate this principle.
Sources and further reading
- 01Open source
When More Pain Is Preferred to Less: Adding a Better End
Daniel Kahneman, Barbara L. Fredrickson, Charles A. Schreiber, Donald A. Redelmeier · 1993 · Psychological Science
Primary research · primary or original source
- 02Open source
Patients' Memories of Painful Medical Treatments: Real-Time and Retrospective Evaluations of Two Minimally Invasive Procedures
Donald A. Redelmeier, Daniel Kahneman · 1996 · Pain
Primary research · primary or original source
- 03Open source
What Do We Remember About Employee Performance? A Meta-Analysis of Peak-End Effects
Banu Alaybek, Yitzhak Fried, Jennifer L. Ford · 2022 · Organizational Behavior and Human Decision Processes
Peer reviewed review
Continue the graph
Related principles
effect
Serial-Position Effect
Recall often varies by an item’s position in a sequence, with advantages for early and recent items under many test conditions.
effect
Recency Effect
Recently presented items often receive a recall advantage when they remain accessible at the time of test.
effect
Goal-Gradient Effect
Motivated effort can increase as a salient, valued goal appears closer, although progress framing and context shape the pattern.