Why You Remember the End of a Task More Than the Middle

When you remember a vacation, you do not replay the whole trip second by second. You replay a handful of moments — the best meal, the worst delay, the way it ended. Your "memory of the vacation" is mostly those few frames standing in for the whole.

Psychologists call this the peak-end rule. It is one of the more counterintuitive findings in memory research because it implies something that does not feel true: that how long a good or bad experience lasts barely affects how you remember it. Only the peak intensity and the ending do.

This was not obvious until Daniel Kahneman, Barbara Fredrickson, and their collaborators ran the experiments that pinned it down — and the way they pinned it down involves a procedure most readers will recognize and many would prefer not to think about.

The colonoscopy study

In a 1996 paper led by Donald Redelmeier and Daniel Kahneman, researchers studied 154 patients undergoing colonoscopies — at the time, painful, often without sedation. Patients reported their pain level every 60 seconds during the procedure, and afterward rated their overall memory of how unpleasant it had been.

Half the patients had a standard colonoscopy. The other half had an "extended" version: the scope was left in place an extra minute or so at the end, with the tip in a less sensitive area. The total amount of pain experienced by the extended group was, by definition, greater — they had more painful minutes. But because the ending was less painful, their memories of the procedure were significantly less negative. Many were more willing to come back for a follow-up.

This sounds like a small effect described from a distance. Reading the paper makes clear it is not small. The extended-procedure patients had more total pain and remembered less of it as bad. The duration of the experience barely mattered. Only the peak and the end did.

Why memory works this way

The likely explanation is that your brain does not have the storage budget to keep continuous experience in full resolution. So it compresses. The compression algorithm — to use the metaphor that will obviously occur to a software person — keeps a handful of high-information samples and drops the rest.

Peaks are high-information because they tell you something extreme happened. Endings are high-information because they tell you the final state, which is what you most need to know to decide whether to do this again. Middle minutes that look like other middle minutes get summarized away.

It is the same reason you can vividly remember a five-minute argument from a relationship but not which morning you ate cereal in May. The cereal was forgettable in the engineering sense — there was no signal to extract.

What this changes about how you design things

Once you know about the peak-end rule, you start noticing where it has already been quietly built in by people who pay attention to user experience:

Disney's queue design. The waits for Disney attractions are long and the middle of the wait is, by intention, the least interesting part — you barely remember it. The peak is whatever entertainment they hand you partway through. The ending is the ride itself. The trick is that you remember the visit by the ride, not by the wait, even though the wait was most of the time.

Hotel checkout vs check-in. Hotels that invest in a smooth, slightly delightful checkout — a real goodbye from the front desk, a small piece of fruit, a courteous offer to store luggage — get better reviews than hotels with great check-ins and average checkouts, even when the room and service were otherwise identical. The ending sticks.

Meetings. The last two minutes of a meeting carry disproportionate weight in how people remember whether the meeting was useful. A meeting that wraps with a clear list of next steps feels productive even if decision quality drifted in the middle. A meeting that runs long, then ends in confusion about who is doing what, feels unproductive even if the middle was sharp.

Workouts. This is the most personal one for me. A run that finishes with a short downhill or a satisfying stretch is remembered as easier than the same run finishing on a steep climb to the door. Knowing this, I deliberately route my last block as a flat or downhill section. The total distance is the same. The memory is different.

Where the rule breaks down

The peak-end rule is not a universal law of memory — it is a heuristic that explains some of the variance, particularly for short, contained experiences. Several caveats are worth knowing:

• It works less well for very long experiences. A two-week vacation has multiple peaks and multiple endings (each day has its own arc). The rule applies more cleanly to bounded experiences like a procedure, a meal, or a meeting.
• The peak does not have to be the most positive moment. It is the most intense. A very bad peak in an otherwise fine experience still gets disproportionate weight. One ten-minute fight on a good vacation can dominate the memory of the trip.
• "How it ended" is sensitive to ambiguity. If an experience ends with unresolved questions, the brain seems to encode that uncertainty rather than a clean valence. This is why open-ended endings (movies, projects, conversations) can produce stronger emotional aftereffects than tidy ones.

How to use this without manipulating yourself

You cannot fool your memory into liking something you hated — and trying to will mostly produce cynicism about your own enjoyment. But you can shape the parts of an experience that matter most:

• Build a small, deliberate ending into anything you want to remember well. A quiet minute at the end of a workday. A short call with someone you like at the end of a long deep-work session. A photo at the end of a hike.
• For experiences you do not control, recognize that your judgment will be skewed by the ending and adjust accordingly. Do not write the post-trip review on the way home from a delayed flight.
• When designing things for other people — products, services, classes, meetings — invest disproportionately in the last two minutes. That is where most of the memory is made.

The strange truth the colonoscopy study revealed is that "how good was that?" is not a question about the whole thing. It is a question about a few high-resolution snapshots your brain decided to keep. Knowing which moments those are, and shaping them where you can, is one of the more useful applications of psychology to ordinary life.