A descriptive psychological heuristic: when people form a retrospective evaluation of a single episodic experience, their global judgment is heavily influenced by the intensity of the episode's most extreme moment (the peak) and by the quality of its final segment (the end), rather than by the arithmetic average or total duration of the experience.

Principle

Retrospective evaluation of an episode is disproportionately determined by its peak intensity and its ending; therefore modifying those moments produces larger shifts in remembered evaluation than equivalent changes distributed elsewhere.

Demonstration

Illustrative scenario → A museum curator tests two guided tours of equal length and content. Tour A includes a moderately engaging sequence and a strong, memorable finale; Tour B has consistently good segments but a neutral ending. Surveyed visitors rate Tour A higher despite similar cumulative engagement metrics because the strong peak and positive ending dominate memory-based judgments.

Misapplication

Mistaken application treats the rule as implying overall experience quality is unimportant; the error is assuming that optimizing only peak and end always yields better real outcomes, ignoring that some objectives (e.g., safety, learning retention) depend on sustained or cumulative properties.

Consequence

Designers who optimize peak and end can increase remembered satisfaction and choice probability without proportionally changing average moment-to-moment experience; this can reallocate resources toward crafting salient moments but may create trade-offs with steady-state performance or long-term outcomes.

Reversal

When the target metric is real-time behavior, objective performance, cumulative outcomes, or repeated exposures (habitual interactions), peak–end weighting may be weak or overridden; likewise, when memory is probed for multiple specific moments rather than an overall judgment, different summary rules apply.

Boundary

Clearly within: single, bounded episodic experiences evaluated retrospectively (e.g., a single concert, a single service encounter). Boundary case: a multi-session program where participants form both session-level and program-level judgments. Clearly outside: continuous lifetime well‑being measures or aggregate metrics computed from logged data rather than memory-based reports.

Semantic Tension

Remembered satisfaction (peak–end weighting) ↔ Aggregate/real-time measures (duration, average experience, cumulative safety): design choices that improve remembered evaluation can conflict with choices that improve objective or distributed quality.

Synthesis

The rule reframes experience design as management of memory anchors (high-intensity moments and endings) rather than as uniform improvement across time; effective application requires aligning those memory-focused interventions with the objective goals that remain sensitive to cumulative or repeated effects.