The peak-end rule describes how people summarize a past experience: memory weights the most intense moment (the peak) and the final moment (the end) far more heavily than the average of the experience or its duration. Kahneman and colleagues established the effect across studies — Redelmeier and Kahneman (1996) found that patients’ remembered pain from a medical procedure tracked its peak and final moments rather than its total or duration, and Kahneman, Fredrickson, Schreiber, and Redelmeier (1993) showed people would even choose to repeat a longer, objectively worse ordeal if it ended more gently. “Duration neglect” is the companion finding: how long something lasted barely registers in recall.
As a manipulation, the rule is exploited by designing the peak and the ending while letting the mediocre or unpleasant middle stand. A service encounter that was slow and error-filled can be remembered warmly if it closes with a sincere apology and a small gift; a subscription that delivered little can end on a generous gesture that colors the whole. The most pointed version is the manufactured rescue — a problem allowed to build, then dramatically resolved at the finish, so that recovery, not the underlying failure, becomes the story you carry.
Harm here is usually indirect: the tactic rarely costs you money in the moment, but it corrupts the memory you will later use to rate, renew, or recommend — steering future decisions on a distorted record. It is also genuinely hard to catch, because the distortion happens in recall rather than perception. The defense is to refuse to let the finale stand in for the whole: evaluate against contemporaneous notes, weigh the full experience including its duration, and delay any verdict until the emotional ending has cooled.