Peak-End Rule: How the Climax and Finale Rewrite the Story

Imagine two versions of the same medical procedure. In Version A, you endure eight minutes of moderate pain. In Version B, you endure the same eight minutes of moderate pain — but then the procedure continues for an additional three minutes at slightly reduced, though still uncomfortable, intensity. Version B is objectively worse: it lasts longer and contains all of Version A's discomfort plus more. Yet when participants in Daniel Kahneman's landmark study were asked which procedure they would choose to repeat, a substantial majority chose Version B. They preferred the longer, more painful experience. The explanation is the peak-end rule: memory of an experience is governed not by its total duration or integrated discomfort, but by two data points — the most intense moment (the peak) and the final moment (the end).

The Cold Water Experiments

The empirical foundation for the peak-end rule came from Kahneman, Fredrickson, Schreiber, and Redelmeier in a pair of studies conducted in the early 1990s. In the cold pressor paradigm, participants submerged their hand in painfully cold water (14°C) for 60 seconds — then, without warning, experienced the same initial trial but extended to 90 seconds, with the water temperature rising slightly (to 15°C) for the final 30 seconds. The extended trial was objectively worse in total painful-second terms. Yet when asked which they would repeat, 69% of participants chose the longer trial. They preferred more pain because the ending was marginally less intense.

Kahneman described this as a conflict between two selves: the experiencing self, which lives through events moment-to-moment and accumulates actual suffering or pleasure, and the remembering self, which narrates the story afterward and bases its verdict on peak and end. Decisions about future experiences are made by the remembering self — which means we systematically choose options that our experiencing self will regret, while passing on options our experiencing self would have preferred.

Duration Neglect

A corollary finding that makes the peak-end rule even more counterintuitive is duration neglect: the total length of an experience has remarkably little effect on how it is remembered. Whether a painful procedure lasts 60 seconds or 60 minutes, the remembered unpleasantness is shaped far more by how bad the worst moment was and how it felt at the end. Similarly, a vacation's enjoyment in memory is poorly predicted by how many good days it contained — a three-week trip with a disastrous finale will be remembered less fondly than a four-day trip that ended perfectly.

This is particularly unsettling from a welfare perspective. If we are trying to maximise the wellbeing of people undergoing medical procedures, educational programmes, or any extended experience, optimising for the experience as it is lived leads to different design choices than optimising for how it will be remembered. The two selves have different utility functions — and we live our forward-looking lives by the remembering self's accountancy.

The Colonoscopy Study

Kahneman, Redelmeier, and Katz (2003) extended the laboratory findings to a genuine clinical context with important practical stakes. In a randomised trial, patients undergoing colonoscopies were assigned to either a standard procedure or one modified to end more gently: at the conclusion, the scope was left in place for an extra minute with minimal movement, reducing discomfort at the moment of termination. The modified procedure was technically longer. But patients who experienced the gentler ending reported the procedure as significantly less unpleasant in retrospect, and — crucially — were more willing to return for future screening.

This has direct implications for preventive medicine: colon cancer screenings save lives, but compliance is a chronic problem. Patients who remember the procedure as agonising avoid coming back. A minor modification that makes the ending less abrupt, with no change to the medically necessary portion, shifts the memory and increases the likelihood of life-saving return visits. The peak-end rule is not just a laboratory curiosity — it has measurable effects on health behaviour.

Vacations, Films, and Customer Experience

Kahneman's framework has been applied extensively beyond medicine. Some notable domains:

Vacations

Travel research consistently shows that memorable vacation moments — a stunning sunset, a lucky encounter, a day things went unexpectedly right — dominate retrospective assessments, while the accumulated ordinary days contribute little to remembered enjoyment. This creates a gap between experienced and remembered quality that travel marketers have learned to exploit: package holidays are sold on peak-generating highlights (the "moments you'll never forget") rather than on consistent comfort throughout.

One practical upshot: if you want a vacation to be remembered as wonderful, spend your limited budget creating one genuinely extraordinary peak and ensuring the journey home is pleasant — rather than distributing expenditure evenly across the trip. The mathematical logic of experience optimisation is different from the logic of memory optimisation.

Films and Stories

Narrative structure has always instinctively followed peak-end logic. Stories climax before they conclude; the climax is the peak, the denouement is the end. Audiences forgive a slow middle if the climax is powerful and the ending satisfying. Research on film evaluation confirms that audiences' overall ratings are disproportionately affected by the final act and the most emotionally intense scene — consistent with the peak-end model.

UX and Service Design

Customer experience design has absorbed peak-end thinking substantially. The principle that "customers will forget 90% of what happened and remember the worst moment and the last moment" has become a design heuristic in service industries. Airlines invest in improving the final boarding and landing experience; restaurants train staff to ensure goodbyes are warm. Disney's famous queuing designs are partly motivated by ending the wait experience on a high note rather than minimising average wait time.

The Nielsen Norman Group (2018) applied peak-end thinking systematically to UX design, noting that interfaces should be engineered not just for efficiency across a session but for strong peaks (delightful moments, successful task completions) and clean, positive endings (confirmation screens, success messages). The same total interaction quality, restructured to front-load positivity and end cleanly, will be remembered more favourably than one that distributes quality evenly but ends neutrally.

The Two-Selves Problem

The deeper philosophical question that the peak-end rule raises is which self's welfare should count. If the experiencing self suffers, does it matter if the remembering self edits the story to a happier conclusion? Kahneman himself has argued that welfare policy and medical ethics have generally failed to distinguish between the two — conflating "this person had a good experience" (experiencing self) with "this person remembers having a good experience" (remembering self).

For decisions about repeated experiences, the remembering self clearly governs: you will return to the colonoscopy based on your memory of it, not your moment-to-moment felt experience during it. But for experiences that are lived once and never repeated, optimising only for memory at the expense of experience seems ethically questionable. The person who endures additional pain so that the memory is more favourable has not been made better off — their experiencing self has been made worse off to manufacture a better story.

These are not purely theoretical puzzles. They arise in end-of-life care, in education design, in the management of chronic pain, and in any setting where we must choose between making an experience feel better as it happens versus making it remembered as better afterward.

Related Biases

The peak-end rule connects to several other memory and evaluation distortions. Availability Heuristic also shows how ease of recall — rather than statistical accuracy — governs judgment; vivid peaks are highly retrievable and thus dominate. The general asymmetry between experiences and their remembered versions relates to the broader pattern of how emotions, rather than objective outcomes, drive evaluations — which connects to concepts like Loss Aversion, where negative peaks (losses) loom larger than positive ones of equivalent magnitude.

Sources & Further Reading

• Kahneman, D., Fredrickson, B. L., Schreiber, C. A., & Redelmeier, D. A. "When More Pain Is Preferred to Less." Psychological Science 4, no. 6 (1993): 401–405.
• Redelmeier, D. A., Katz, J., & Kahneman, D. "Memories of Colonoscopy: A Randomized Trial." Pain 104, no. 1–2 (2003): 187–194.
• Kahneman, D. Thinking, Fast and Slow. Farrar, Straus and Giroux, 2011. Chapter 35: "Two Selves."
• Kane, L. "The Peak-End Rule: How Impressions Become Memories." Nielsen Norman Group, 2018. nngroup.com
• Wikipedia: Peak–end rule