Peak-End Rule

People judge a past experience mostly by how it felt at its most intense moment and at its end, not by the sum or the average of every moment in it.

Last reviewed

The law

People judge a past experience mostly by how it felt at its most intense moment and at its end.

How long the experience lasted counts for surprisingly little. Psychologists call this duration neglect.

Where it comes from

Daniel Kahneman, Barbara Fredrickson, Charles Schreiber and Donald Redelmeier reported the key experiment in When more pain is preferred to less: Adding a better end, published in Psychological Science in 1993. Participants held a hand in painfully cold water for 60 seconds. In a second trial they held it in the same water for 60 seconds and then for 30 seconds more while the water was warmed slightly, still uncomfortable but less so.

Asked which trial they would rather repeat, most participants chose the longer one, even though it contained all the pain of the shorter one and more. The better ending improved the memory of the whole.

Related studies by the same group found the same pattern elsewhere: remembered discomfort tracked the worst moment and the final moments, while duration had little effect. Redelmeier and Kahneman's 1996 study of patients undergoing colonoscopy reported this in a medical setting.

What it means for an interface

  • The memory of a product is not the average of its moments. People decide whether to come back based on what they remember, so the peak and the ending of key flows deserve the most care.
  • Endings matter. The last screen of a checkout, a sign-up or a support conversation weighs heavily in how the whole flow is remembered. A clear confirmation and next step is a better ending than a dead end.
  • Remove the worst moment first. A single bad peak, such as a confusing error or a lost form, can define the memory of an otherwise smooth flow. Fixing it does more than polishing the parts that already work.
  • Measure at the right time. A satisfaction question asked straight after a bad moment and one asked at the end of a flow measure different things. Know which one you are collecting.

Examples

Order confirmation. A checkout that ends on a page with the order number, the delivery date and a clear way to track the parcel ends on certainty. One that ends on a generic "Thank you" leaves people unsure whether it worked.

Error recovery. A payment that fails and keeps everything the person typed, explaining what to fix, turns a potential worst moment into a small one. A failure that empties the form becomes the peak people remember.

Cancellation. A subscription that is easy to cancel, with a confirmation that says what happens next, can leave a good last impression. A hard-to-find cancel button often becomes the moment people remember and talk about.

Where it is misapplied

  • Polishing the end to hide the middle. A pleasant ending does not cancel a painful flow in practice: people still drop out before they reach it. Fix the problems first.
  • Manufacturing peaks. Confetti and celebration screens are not peaks if they interrupt people. A peak is a moment that matters to the person, not a decoration.
  • Ignoring duration entirely. Duration neglect describes memory, not behavior during the experience. A long, tedious flow still loses people along the way.
  • Generalizing from one experiment. The finding comes from short, controlled experiences. It is a strong tendency, not a formula for computing how people will rate your product.

Checklist

  • Every key flow ends on a clear confirmation and a next step.
  • The worst moment in each key flow is known and on the list to fix.
  • Errors keep people's input and explain how to recover.
  • Satisfaction questions are timed deliberately and reported with that timing.
  • Celebration moments never block the person's next action.

Sources

  1. Kahneman, D., Fredrickson, B. L., Schreiber, C. A., & Redelmeier, D. A. (1993). When more pain is preferred to less: Adding a better end. Psychological Science, 4(6), 401–405.
  2. Fredrickson, B. L., & Kahneman, D. (1993). Duration neglect in retrospective evaluations of affective episodes. Journal of Personality and Social Psychology, 65(1), 45–55.
  3. Redelmeier, D. A., & Kahneman, D. (1996). Patients' memories of painful medical treatments: Real-time and retrospective evaluations of two minimally invasive procedures. Pain, 66(1), 3–8.