The Peak-End Rule: Why You Misremember Your Experiences

You have two selves. One lives the experience. The other remembers it. They don’t agree — and the one that remembers is the one that decides what you do next.

Here’s the finding. Your memory of an experience isn’t the average of every moment in it. It’s driven almost entirely by two data points: the most intense moment (the peak) and how it ended. Everything in between — including how long it lasted — barely registers. Psychologists call this the peak-end rule, and it’s not a metaphor. It’s been demonstrated in a randomized clinical trial where making a painful procedure longer made patients remember it as better.

What is the peak-end rule?

The peak-end rule says that when you look back on an episode, your brain doesn’t tally up every second and compute a fair total. Instead it takes a shortcut: it averages the worst (or best) moment with the final moment, and calls that the memory.

Its companion is duration neglect — the length of an experience has surprisingly little effect on how you remember it. A two-hour dinner and a four-hour dinner are recalled on roughly the same terms: how good the peak was and how it ended. This is why “remembered utility” (what you think an experience was like) can diverge sharply from “experienced utility” (what it was actually like, moment to moment). The two selves keep different books.

The study that proved it: a longer procedure that felt better

The cleanest evidence comes from an unlikely place — a colonoscopy clinic. In a 2003 randomized trial in Pain (Redelmeier, Katz & Kahneman), 682 patients undergoing colonoscopy were randomly assigned to either a standard procedure or one with a small twist: at the end, the physician left the (stationary) scope tip in place for a short extra interval, adding a stretch of mild, low-level discomfort onto the tail of the procedure.

Objectively, the extended group got more total discomfort — the same procedure plus extra time. But their memory was better on every measure. They rated the final moments as far less painful (1.7 versus 2.5 on a ten-point scale, P < 0.001), rated the whole experience as less unpleasant (4.4 versus 4.9, P = 0.006), and ranked it as less aversive against seven other bad experiences (4.1 versus 4.6, P = 0.002). A gentler ending overwrote the memory of the rougher middle.

And it changed behavior. Over a median follow-up of 5.3 years, patients who got the longer, better-ending procedure were more likely to come back for a repeat colonoscopy (odds ratio 1.41, P = 0.038). The remembering self, not the experiencing self, decided whether they’d return for a procedure that could save their life.

When more pain is genuinely preferred to less

That colonoscopy result wasn’t a fluke of the clinic. It was predicted by a lab experiment a decade earlier.

In Kahneman, Fredrickson, Schreiber & Redelmeier (1993), Psychological Science — bluntly titled “When More Pain Is Preferred to Less” — participants held a hand in painfully cold 14°C water for 60 seconds (the short trial). In the long trial, they endured the same 60 seconds, then kept the hand in for 30 seconds more while the water was gradually warmed to a slightly-less-painful 15°C. The long trial contains strictly more total discomfort — every second of the short trial, plus thirty more.

Yet when asked which they’d rather repeat, a significant majority chose the long trial. The mildly-better ending rewrote the memory, and duration was neglected. People volunteered for more pain because the tail end of it felt kinder. That’s the peak-end rule making a choice against a person’s own physical interest.

The honest caveat

Two things keep this from being a magic wand. First, duration neglect doesn’t mean duration never matters — you’d still rather have a short root canal than a long one in the moment, and over big stretches of life, real time obviously counts. The rule is about how episodes get encoded into memory, not a license to ignore how long you make people suffer. Second, it’s not a hack for manufacturing fake happiness: you can’t bolt a good ending onto a genuinely bad experience and expect it to erase everything. What you can do is stop assuming your memory is an honest ledger.

Why this matters if you run anything

Once you see it, the peak-end rule is everywhere you design an experience — a product onboarding, an offsite, a customer-support call, a workout, a relationship. People won’t remember the average. They’ll remember the spike and the finish.

The practical moves follow directly: engineer a strong peak and protect the ending, because a mediocre middle with a great close beats a great middle with a limp one. It also explains why your own recall of a hard week is unreliable — you’re rating it on its worst hour and its last hour, which is close kin to how rumination fixates on the sharpest, ugliest moment and calls it the whole story. The same gap between the living self and the judging self is why the arrival fallacy leaves a goal feeling empty the moment you reach it — the ending you imagined isn’t the ending you get. Knowing the bias is the start of stepping outside it, the same move as coaching yourself in the third person instead of drowning in the moment.

Your experiencing self lives every second. Your remembering self writes the history and casts the votes. The peak-end rule is the reason the two disagree — and the reason it’s worth designing the peak and the ending on purpose. For more on the cognitive biases that quietly run behavior, see our behavioral change work.