Peak–end rule

Based on Wikipedia: Peak–end rule

The Strange Math of Memory

Here's a disturbing experiment. Researchers asked people to plunge their hand into painfully cold water—fourteen degrees Celsius, which feels like holding ice—for sixty seconds. Then they asked the same people to do it again with their other hand, but this time for ninety seconds. The catch? During the final thirty seconds of the longer trial, the water was secretly warmed by just one degree.

When given the choice of which trial to repeat, people overwhelmingly chose the longer one.

They voluntarily signed up for more pain.

This isn't masochism. It's something far more fundamental about how human memory works. The psychologists Daniel Kahneman and Barbara Fredrickson discovered that we don't remember experiences the way a video camera would—recording every moment and averaging them together. Instead, our brains take a brutal shortcut. We remember two things: the most intense moment, and how it ended.

Everything else gets thrown away.

The Snapshot Model

Think of your memory as a photographer with only two shots. One photograph captures the peak—the moment of greatest intensity, whether that's the best part of a vacation or the worst moment of a dental procedure. The other captures the ending. Your brain then averages these two snapshots and calls it the whole experience.

This is the peak-end rule, and it explains an enormous amount of seemingly irrational human behavior.

The rule emerges from what Kahneman and Fredrickson called the "snapshot model of remembered utility." Utility is economics-speak for satisfaction or happiness. The key insight is that there are two different versions of every experience: the one you live through moment by moment, and the one you remember afterward. These two versions often disagree violently.

Consider duration. You might think that a two-week vacation would be remembered as twice as good as a one-week vacation, all else being equal. But the research shows something remarkable: duration barely matters. A week in Paris that ends on a high note will be remembered more fondly than two weeks that peter out with flight delays and lost luggage. Kahneman called this "duration neglect," and it's one of the most counterintuitive findings in all of psychology.

Cold Water and Colonoscopies

The cold water experiment from 1993 was just the beginning. Kahneman and his colleague Donald Redelmeier took their research into hospitals, studying patients undergoing colonoscopies—a procedure that, before modern sedation, was genuinely unpleasant.

They found the same pattern. Patients didn't remember the total amount of discomfort they experienced. They didn't even remember the average discomfort. Their memory of the procedure was almost entirely predicted by two numbers: the worst moment, and the final moment.

This led to a counterintuitive medical intervention. In a follow-up study, doctors randomly assigned some patients to an extended procedure. At the end of a normal colonoscopy, they left the scope in place for an extra three minutes. The scope wasn't moved—the sensation was uncomfortable but not painful. Objectively, these patients experienced more total discomfort than the control group.

But their memories told a different story. The extended-procedure patients rated their experience as significantly less unpleasant. More importantly, they were far more likely to return for follow-up colonoscopies—screenings that could save their lives.

By adding a little more discomfort, doctors had paradoxically created a better memory.

Why Does This Happen?

The peak-end rule isn't a bug in human cognition. It's a feature—one that evolved because it's efficient. Storing a complete record of every experience would require enormous mental resources. Instead, our brains compress experiences into highlights, much like a movie trailer condenses a two-hour film into two minutes.

The peak gets remembered because emotionally intense moments are biologically significant. They're the moments that taught our ancestors something important about the world—where the predators lurk, which foods are poisonous, who can be trusted in a crisis. Evolution tuned our memory systems to prioritize intensity.

The ending matters for a different reason. It's the most recent information, still fresh and accessible. Psychologists call this the "recency effect," and it shows up everywhere from memory tests to performance reviews. What happened last is easiest to recall, so it gets weighted heavily in our judgments.

There's also something deeper going on. Endings feel like conclusions. They carry narrative weight. A story that ends badly is a bad story, even if most of the pages were wonderful. Our minds are built for narrative, and narratives are judged by their resolutions.

The Beginning Matters Too (Sometimes)

Memory researchers have long known about two competing effects: primacy (we remember what came first) and recency (we remember what came last). How do these interact with the peak-end rule?

A clever experiment using potato chips helped untangle this. Researchers had people eat varying numbers of chips while rating their enjoyment. For shorter eating sessions—just five or fifteen chips—the recency effect dominated. People's memory of how much they enjoyed the snack was mostly predicted by how the last few chips tasted.

But here's the interesting part: people had genuinely worse memory for the first chip than the last chip. Each new chip seemed to interfere with the memory of previous chips, like writing over a recording. When researchers showed participants their own ratings of that first chip, suddenly it mattered as much as the final chip in their overall evaluation.

This suggests the peak-end rule isn't about the beginning being unimportant—it's about the beginning being forgotten. The information is lost to interference before it can influence the final judgment.

Red Sox Games and Rose-Tinted Glasses

The peak-end rule helps explain why nostalgia is so powerful and so unreliable.

Researchers asked Boston Red Sox fans to recall a game they'd attended when the Red Sox won. Then they asked how representative that game was of typical Red Sox victories. The fans confidently reported that yes, the game they recalled was pretty typical.

It wasn't. They'd recalled the best game they could remember—the peak experience—and mistaken it for average. Only when explicitly asked to recall "the best game" did fans show any awareness that their first memory was an outlier.

The same pattern appears with television shows, movies, and almost any category of experience. Ask someone about sitcoms from the 1990s, and they'll recall Friends or Seinfeld—the peaks—and assume that's what TV was like. They won't spontaneously recall the forgettable shows that filled most of the schedule unless prompted.

This is why "they don't make 'em like they used to" is almost always false. They made plenty of bad ones back then too. We just don't remember them.

The Business of Better Endings

Once you understand the peak-end rule, you start seeing its applications everywhere.

Consider retail. Most shopping experiences have a clear beginning (entering the store) and end (checkout). According to the peak-end rule, a store could recover from a mediocre experience by nailing the finale. Play music customers enjoy. Offer a small gift or sample at checkout. Have a friendly employee hold the door as people leave.

As the marketing consultant Scott Stratten put it: "A really great salesperson who helps with an exchange can erase negative experiences along the way. The long wait in line and the bad music in the changing room are forgotten."

Hotels have figured this out. The mint on your pillow, the warm goodbye at checkout, the offer to call you a cab—these closing touches are strategic. They're not just hospitality. They're memory engineering.

There are limits, though. Research by Talya Miron-Shatz suggests that day-long experiences don't follow the peak-end rule as cleanly as shorter episodes. A terrible hotel stay with a nice checkout is still remembered as a terrible stay. The rule seems most powerful for experiences that feel like discrete events—a single procedure, a meal, a transaction—rather than extended periods.

The Pricing Paradox

The peak-end rule has infiltrated pricing strategy in subtle ways.

When consumers form a "reference price"—their mental benchmark for what something should cost—they don't simply average all the prices they've seen. Instead, their reference price appears to be a weighted combination of the highest price they remember (the peak) and the most recent price (the end).

This creates a strategic dilemma. If a company offers deep discounts, those low prices can anchor in consumers' memories and permanently erode their willingness to pay full price. The discount becomes the new peak—just in the wrong direction.

Some researchers have suggested that occasionally raising prices above the desired level can reset consumer expectations. But this is a dangerous game. Do it too often and you seem unreliable. Do it for too long and customers leave. The tactic works best for frequently purchased, relatively inexpensive items—things like snacks or streaming music—where the lost sales during the high-price period are quickly recouped.

Streaks and the Workplace

Not all experiences involve continuous sensations like pain or pleasure. Some are sequences of discrete events: did the project succeed or fail? Was the task hard or easy? Was the customer satisfied or angry?

Researchers Kang, Daniels, and Schweitzer wondered whether the peak-end rule applies to these binary sequences. They proposed what they called the "streak-end rule." In sequences of yes-or-no outcomes, the psychological equivalent of a peak isn't a single extreme moment—it's a streak, a run of the same outcome in a row.

They tested this with volunteer workers who were repeatedly assigned tasks that were either hard or easy. Sure enough, workers were disproportionately likely to quit after experiencing a streak of hard tasks, or after their most recent task was hard. A single difficult assignment buried in a sequence of easy ones didn't matter much. But four hard tasks in a row? That was a peak, and it shaped how people evaluated the entire job.

Vacations: Shorter Might Be Better

In 2006, researchers at the University of Canterbury in New Zealand applied the peak-end rule to vacations. They found that travelers' memories of their trips were well-predicted by the peak and end—and barely influenced by duration.

Actually, they found something even more interesting. The best predictor of remembered happiness wasn't quite the peak and end. It was the "most memorable or most unusual 24-hour period." If your vacation included one extraordinary day—a perfect hike, an unexpected encounter, a magical evening—that day would dominate your memory of the entire trip.

The implication is provocative: a five-day vacation with one incredible day might be remembered just as fondly as a three-week trip without such a highlight. If you're optimizing for memory rather than experience, shorter and more intense may beat longer and more relaxed.

Of course, you might reasonably ask whether optimizing for memory is the right goal. More on that in a moment.

The Ethics of Medical Memory

The colonoscopy studies raised an uncomfortable question. If adding a few minutes of mild discomfort at the end creates a better memory, shouldn't doctors routinely do this? After all, patients with better memories are more likely to return for life-saving follow-up screenings.

But there's a catch. Those extra minutes are still discomfort. The patient is still experiencing unpleasant sensations they wouldn't otherwise endure. The memory is better, but the experience is objectively worse.

Doctors face a genuine ethical tension. They can minimize the actual pain their patient feels, or they can minimize the pain their patient remembers. These are not the same thing, and it's not obvious which one medicine should prioritize.

Kahneman himself was skeptical that patients would volunteer for this tradeoff. "It is safe to assume," he wrote, "that few patients will agree to expose themselves to pain for the sole purpose of improving a future memory."

But is that actually true? People often choose present discomfort for future benefit—that's what exercise is. If patients understood that a slightly longer procedure would make them more willing to return for essential screenings, some might accept the deal. The key is informed consent: the patient should be the one to decide whether the memory is worth the extra discomfort.

Teaching to the End

The peak-end rule has found its way into education, particularly in how students receive feedback.

A study by Vincent Hoogerheide and his colleagues looked at how children responded to peer assessments—evaluations from their classmates. They created two versions of the same assessment: a short version and an extended version that added one more comment at the end.

When the overall assessment was negative, the extended version added a moderately negative (but not harsh) final comment. When the assessment was positive, the extended version added a moderately positive (but not glowing) final comment. In both cases, the additional ending was milder than the average of what came before.

The results were striking. Students who received the extended assessments—even when they contained more total criticism—remembered the experience as more pleasant and less difficult to deal with. A gentler landing cushioned the blow.

The practical advice is simple: structure feedback to end on the most palatable note. If you must deliver bad news, save the most acceptable criticism for last. If the news is good, don't bury it in the middle—end on your highest note.

The Pizza Paradox

What about food? Surely the peak-end rule applies to eating—an experience with clear beginning, middle, and end.

It does, but with a twist. Researchers set up a pizza buffet where some customers paid four dollars and others paid eight. Then they tracked which slices predicted overall satisfaction.

For the four-dollar customers, the peak-end rule held beautifully. Their enjoyment of the last slice and the best slice predicted how they rated the meal overall. But for the eight-dollar customers, something different happened. Their first slice was the best predictor of satisfaction.

The explanation seems to be about expectations. People who pay more expect more. If that first bite disappoints, the initial let-down colors everything that follows. But people who pay less don't have as much riding on the experience. They're open to being pleasantly surprised, which makes peaks and endings more influential.

The practical takeaway for restaurants: if you're expensive, lead with your best dish. Make that first impression impeccable. If you're cheap, you have more flexibility—but don't neglect dessert.

Two Selves

The peak-end rule points to something profound about human identity. We don't have one self—we have at least two.

There's the "experiencing self," the one who lives through each moment in real time. This self knows the difference between a sixty-second ice bath and a ninety-second ice bath. Every additional moment of cold is felt, fully.

Then there's the "remembering self," the one who constructs the story afterward. This self doesn't care about duration. It cares about narrative—peaks, valleys, and endings. When you ask someone how their vacation was, the remembering self answers. The experiencing self is already gone, dissolved into the irretrievable past.

Which self is the "real" you? Kahneman spent decades thinking about this question and never reached a satisfying answer. Both selves are real. They just want different things.

The experiencing self wants every moment to be as good as possible. It would choose the shorter cold-water trial, because less pain is obviously better. The remembering self wants a good story, and a story that ends well is a good story—even if it contains more total suffering.

Most of life's big decisions are made by the remembering self. We choose vacations, jobs, and relationships based on how we think we'll remember them, not on a moment-by-moment calculation of pleasure. This isn't irrational—memory is how we learn from the past and plan for the future. But it does mean we systematically ignore duration and overweight intensity and endings.

Living with the Peak-End Rule

Once you understand how memory works, you can start engineering better endings for yourself and others.

Giving a presentation? Save something good for the close, even if it means restructuring your argument. Hosting a dinner party? Don't let the evening peter out—create a definite, positive ending. Having a difficult conversation? Think about how to land it gently.

You might also become more skeptical of your own memories. That amazing vacation you keep talking about—was it really amazing throughout, or just at its peak and end? That terrible job you left—was it truly miserable every day, or did it just end badly? Memory is not a faithful recorder. It's an editor with a particular style.

There's also a case for seeking intensity. If duration doesn't matter to memory, but peaks do, then one transcendent experience may be worth more than many mediocre ones—at least to your remembering self. This might justify occasionally splurging on something extraordinary instead of spreading your budget across many ordinary pleasures.

But don't forget your experiencing self entirely. That self exists in the moment, feeling every sensation that the remembering self will later discard. A life optimized purely for memory might look strange—full of carefully constructed peaks and endings, with everything in between treated as filler to be endured.

Perhaps the best approach is awareness. Know that your memory will distort your experiences in predictable ways. Plan for good endings. But also remember that you're living your life in real time, not just collecting highlights for a future retrospective.

The peak and the ending matter. But so does everything in between, even if you won't remember it.