ALIVE LIBRARY
MIND & BEHAVIOR

The peak-end rule

Last updated: June 2026

The peak-end rule states that people judge an experience primarily by its most intense moment (the peak) and how it ends, rather than the total sum or average of the entire duration.

01

The Principle

Daniel Kahneman and Barbara Fredrickson’s research in the 1990s demonstrated this through studies like the colonoscopy experiment, where patients remembered the procedure based on the worst moment and the ending far more than its total length. When the end was made slightly less uncomfortable (even if it prolonged the overall experience), remembered pain decreased significantly. The mechanism is rooted in how memory works: our brains don’t store continuous footage but rather representative snapshots, with emotional intensity and recency carrying disproportionate weight.

This isn’t just about pleasure or pain. Positive peaks (moments of delight, mastery, or relief) and a strong closing shape retrospective evaluation. The lived experience — the full sequence of moments — often fades, while these two anchors dominate what we recall and how we feel about the product afterward.

In my own building, I’ve seen this play out repeatedly. A feature that felt solid during development could still leave users with a lukewarm memory if the final interaction landed flat. Understanding the rule shifted my focus from optimizing every micro-step equally to deliberately engineering memorable peaks and clean, satisfying endings.

02

Why It Matters for Design & Building

For designers and builders, the peak-end rule highlights a fundamental mismatch: we optimize for the lived experience (smooth flows, efficiency metrics), but users remember and decide based on the remembered one. A long but well-peaked and well-ended session can feel better in memory than a shorter one that ends abruptly or on frustration.

This has practical implications across onboarding, task completion, error recovery, and even logout. As a Design Engineer, I now audit flows by asking: where is the emotional peak, and how does this end? In one client project involving a reporting dashboard, the core analysis was strong, but the export process ended with a generic “success” toast that disappeared instantly. Users remembered the tool as “okay but forgettable.” Adding a simple summary preview with shareable highlights at the end created a positive closing peak that noticeably improved perceived value in follow-up feedback.

In AI interfaces this matters even more. The generation process can be unpredictable; a brilliant output can be undermined by a clunky dismissal or save experience. Respecting the rule means designing not just functional interactions but memorable ones that leave users with a positive retrospective judgment.

03

Real-World Examples

Duolingo applies the rule effectively during language lessons. Completing a streak or hitting a milestone triggers celebratory animations, confetti, and the owl mascot’s enthusiastic feedback — clear positive peaks. Sessions often end with a calm progress summary and gentle encouragement to continue tomorrow, leaving users with a sense of accomplishment rather than exhaustion. This helps explain why many stick with the app despite its repetitive nature.

A major European banking app illustrates the downside during fund transfers. The core flow is efficient and secure, but the final confirmation screen presents dense legal text, multiple required checkboxes, and a delayed “success” state that sometimes hangs. Even users who complete the task successfully often recall the experience as stressful or untrustworthy because the ending lands on friction rather than relief.

A healthcare appointment booking portal offers a mixed case. The search and selection steps can feel tedious, but a thoughtful end — a clear calendar integration prompt, immediate confirmation email preview, and reassuring “we’ve got you covered” message — elevates the overall memory. Patients remember the convenience of the close more than the earlier friction, improving satisfaction scores despite average lived efficiency.

Related entries

References

  1. Kahneman, D., Fredrickson, B.L., et al. (1993). "When More Pain Is Preferred to Less: Adding a Better End." Psychological Science.
  2. Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
  3. NN/g: The Peak–End Rule. nngroup.com
  4. The Decision Lab: Peak-End Rule. thedecisionlab.com
  5. Laws of UX: Peak-End Rule. lawsofux.com