The Peak-End Rule: Why Your Customer Remembers Only Two Moments

The Magic Castle Hotel is a converted 1950s apartment building in Hollywood, painted canary yellow, with forty-three rooms. The furniture is modest. The lobby is nothing. There's no spa, no rooftop bar, no marble anything. If you're imagining a place that competes with the Four Seasons Beverly Hills, stop. The Four Seasons charges three times the nightly rate and has every physical advantage a hotel can have.

The Magic Castle Hotel has consistently ranked among the highest-rated hotels in all of Los Angeles on TripAdvisor, beating five-star properties that outspend it on every measurable dimension of hospitality. It has received TripAdvisor's Certificate of Excellence for nine consecutive years.

The rooms don't explain it. The location doesn't explain it. What explains it is a red phone mounted on a wall by the pool. Guests pick it up and someone answers: "Hello, Popsicle Hotline." Minutes later, a staffer wearing white gloves delivers cherry, grape, or orange popsicles on a silver tray. Free of charge. That phone, along with a breakfast magician who performs at your table and snacks delivered on silver platters throughout the day, is by far the most mentioned detail in the hotel's reviews. Not the rooms. Not the beds. Not the check-in process. The popsicle phone.

The Magic Castle doesn't try to be the best hotel in Los Angeles. It doesn't try to fix every flaw. The carpet is dated. The bathrooms are converted apartment bathrooms. A hundred consultants would walk through that building and generate a hundred punch lists of things to improve. The hotel ignores most of them. Instead, like the elevator mirror that solved a complaint without fixing the actual problem, it pours its budget into a handful of moments so unexpected and so specific that they become the only thing guests talk about when they get home. The peak-end rule is the psychological principle that explains why this works: people judge an experience not by its average quality or its duration, but by its most intense moment and by how it ends. The Magic Castle doesn't deliver a uniformly excellent experience. It delivers two or three peaks so vivid that the brain discards everything else.

The hotel's management team probably never read a Kahneman paper. They didn't need to. They figured out what the research would later confirm: your customer's memory of the experience is not a recording. It's an edit. And the edit keeps two things: the peak and the end.

The Experiment That Proved More Pain Can Feel Better

In 1993, Daniel Kahneman and three colleagues at Princeton brought participants into a lab and asked them to do something unpleasant. Each person immersed one hand in painfully cold water, 14 degrees Celsius, cold enough that the body registers it as genuine pain.

Every participant did two trials. In the short trial, they kept their hand in 14-degree water for 60 seconds. In the long trial, they kept their hand in 14-degree water for 60 seconds, then stayed in the water for an additional 30 seconds while the temperature gradually rose to 15 degrees. Still painful, but distinctly less so.

The long trial contained objectively more total pain. Ninety seconds of suffering versus sixty. Then Kahneman asked: which trial would you like to repeat?

The majority chose the long trial. They voluntarily signed up for more pain because the last thirty seconds had been slightly less painful than the first sixty. Their experiencing self, the one whose hand was in the water, had registered more total discomfort in the long trial. Their remembering self, the one that made the decision about which to repeat, evaluated the experience by its worst moment and its ending. The ending of the long trial was warmer. So the remembering self called it the better experience.

The experiencing self and the remembering self are running different accounts. One integrates moment by moment, summing up pain in real time. The other takes two snapshots — the peak and the end — and throws the rest away. Guess which one decides what to do next.

Why Does Your Brain Edit This Way?

The neuroscience underneath the peak-end rule comes down to two systems operating on different timescales.

Start with what happens during a peak. Kevin LaBar and Roberto Cabeza's review in Nature Reviews Neuroscience documented the circuit: when an experience reaches high emotional intensity, the amygdala detects the spike and signals the hippocampus to bind contextual details to that moment with extra precision. Higher amygdala activation during encoding corresponds with improved later recall. Intensity functions as a marker of importance, and the brain allocates its best memory storage accordingly. The mundane middle of an experience, the twenty minutes where nothing remarkable happens, generates weaker emotional arousal and weaker memory traces. The brain doesn't waste premium encoding on the forgettable parts.

Now consider why endings carry so much weight. Hermann Ebbinghaus first demonstrated the serial position effect in 1885, and it's been replicated so many times it barely qualifies as controversial: items at the end of a sequence are better remembered than items in the middle. The final moments of an experience are still in working memory when the remembering self sits down to construct its narrative. They get the last word.

Put these together and the peak-end rule stops looking like a quirk. It's a predictable consequence of how memory encoding works. The amygdala flags the most intense moment for deep storage, the same neural coupling that makes stories stick when they hit an emotional peak. The most recent moment is still sitting in working memory when the evaluation happens. Everything between those two anchors gets compressed, averaged, largely discarded. Your customer's brain doesn't have the architecture to remember the whole experience. It was never designed to.

What Happened When Doctors Added More Pain to a Procedure?

Kahneman took the cold-water finding and tested it somewhere the stakes were real.

In 2003, Donald Redelmeier and Joel Katz, working with Kahneman, ran a randomized controlled trial with 682 patients undergoing colonoscopy at Sunnybrook Health Science Centre in Toronto. Half received the standard procedure. For the other half, the colonoscope tip remained stationary in the rectum for an extra period at the end, producing mild but reduced discomfort. The procedure was objectively longer and contained more total discomfort.

The group that endured the extended procedure rated the final moments as significantly less painful: 1.7 versus 2.5 on a 10-point scale. They rated the entire experience as less unpleasant. They ranked it as less aversive compared to seven other unpleasant experiences.

Then the researchers waited. The median follow-up was 5.3 years. Patients in the extended-procedure group were 41 percent more likely to return for a repeat colonoscopy. Adding discomfort to a medical procedure, by making the ending less painful, improved the memory of the experience so much that it changed actual health behavior half a decade later.

This is the peak-end rule with consequences that matter beyond satisfaction surveys. The way an experience ends doesn't just color how your customer feels about it. It changes what they do next, the same mechanism that makes the first 48 hours after purchase so critical for retention and referrals.

Filling Potholes vs. Building Peaks

Chip and Dan Heath's book The Power of Moments, published in 2017, turned the peak-end rule into a practical framework. Their central argument is one most businesses will find uncomfortable: the standard approach to customer experience is wrong.

The standard approach is pothole-filling. You identify your worst moments, your longest wait times, your clunkiest processes, your most common complaints, and you fix them. You bring bad experiences up to acceptable. The result is a smooth, competent, forgettable experience. Nobody complains. Nobody talks about it either.

The alternative is building peaks. Instead of distributing your improvement budget across fifty small fixes, you concentrate it on two or three moments designed to be extraordinary. A popsicle hotline. A breakfast magician. White-glove snack delivery. These moments become the story your customer tells. The Magic Castle Hotel has plenty of potholes. Dated furniture, converted apartment bathrooms, no luxury amenities. None of that matters because the peaks are so far above the baseline that the remembering self discards the rest.

The 2022 meta-analysis by Alaybek and colleagues, analyzing 174 independent samples, confirmed this asymmetry: peak intensity predicted retrospective evaluations more strongly than the trough. Positive peaks drive memory more powerfully than the absence of negatives, which means one extraordinary moment delivers more return than a hundred incremental improvements spread across the journey. This is the flip side of the 9X problem: customers overvalue what they already have, but they also disproportionately remember what surprised them.

This runs directly counter to how most founders think about product and service quality. The instinct is to bring everything up to a consistent standard. Consistency feels responsible. But the research says consistency is invisible. Peaks are what get encoded.

How Does the Peak-End Rule Apply to Customer Experience?

Consider what happens at checkout.

For most businesses, checkout is the worst moment of the customer experience and the last moment before the customer leaves. The customer hands over money. They feel the pain of paying. They haven't received full value yet. They're not sure the purchase was right. The emotional low point and the ending are the same moment. From a peak-end perspective, this is a disaster. The most negative moment is also the most recently encoded, which means the remembering self constructs the evaluation of the entire experience around a feeling of loss.

Costco's $1.50 hot dog combo has been the same price since 1984, a price anchor so powerful it reframes the entire shopping trip. The food court is positioned at the exit of most Costco stores. After the sensory overload of bulk shopping, after the cognitive load of evaluating twenty-seven varieties of paper towel, the last thing the customer encounters is an absurdly good deal. Whether intentional or not, it functions as a peak-end strategy: it replaces the payment pain with a feeling of unexpected generosity, and it does it at the moment the remembering self is taking its final snapshot.

IKEA follows a similar logic. Their food must offer the cheapest meals within a 30-mile radius of each store. The 50-cent soft-serve cone, positioned after checkout, ensures the last taste of the IKEA experience is sweetness, not the memory of assembling a mental budget for flat-pack furniture. Gerd Diewald, who oversaw IKEA's food operations in the U.S., put it plainly: "It's hard to do business with hungry customers. When you feed them, they stay longer, they talk about their purchases, and they make a decision without leaving the store."

Apple engineered the same principle into a box. According to Jony Ive, "Steve and I spend a lot of time on the packaging. I love the process of unpacking something. You design a ritual of unpacking to make the product feel special. Packaging can be theater." The iPhone box is designed so that lifting the top creates a slight vacuum, the bottom descends slowly, and the reveal happens in a controlled, almost ceremonial moment. The unboxing is the first product experience, and it's designed to be a peak so vivid that it anchors the customer's entire relationship with the device.

None of these companies are filling potholes. They're engineering the two moments the brain actually keeps.