Picture two identical afternoons. In both, you do four hours of honest, difficult work. You write the hard section, untangle the problem you'd been dodging, and answer the email that scared you. In the first, you close the laptop at 5:30, tidy your desk, and write down where you'll start tomorrow. In the second, the work trails off into a vague drift. You open a tab "for a second", then another, and at 6:15 you surface from a feed with a faint sourness you can't name.

If someone asked you a week later which afternoon was better, you wouldn't average the hours. You'd pick the first, even though the four hours were identical. That isn't sloppy thinking. It's how human memory is built.

Memory doesn't add up your day

In the early 1990s, Daniel Kahneman and colleagues ran an experiment that sounds mildly cruel. Participants held a hand in painfully cold water for 60 seconds. In a second trial, they held the other hand in the same cold water for 60 seconds, then kept it there for another 30 while the water warmed slightly, still uncomfortable but a little less so. Offered a third trial, a majority chose to repeat the longer one. They picked 90 seconds of pain over 60, because it ended better.

A later study of colonoscopy patients (Redelmeier and Kahneman, 1996) found a similar pattern. Patients' retrospective ratings of the whole procedure tracked the worst moment and the final moments, not how long it lasted. In a follow-up trial, patients whose procedure was deliberately extended with a milder final stretch later remembered the experience as less unpleasant than those who got the shorter one.

This is the peak-end rule, and its companion, duration neglect. When we look back, we don't replay the whole recording. We consult a summary built from the most intense moment and the ending. The length matters surprisingly little.

The rule isn't perfect or universal. It describes a strong tendency in how experiences are remembered, and later work has shown it bends under some conditions. But it's robust enough to be worth designing around.

What this means for a workday

Most of us treat the end of the workday as nothing: the leftover edge where the real work stopped. The research suggests the opposite. The ending is half of what you'll keep.

That explains a feeling many people recognize and can't explain. You had a productive day. You can list what you did. Yet by evening you feel scattered, vaguely guilty, as if you'd wasted it. The hours were fine. The ending was a slow leak into a screen, and the ending is what your memory grabbed.

It also explains why the evening scroll costs more than its minutes. A feed is almost perfectly designed to produce a bad ending. It has no natural stopping point, it tapers off into low-grade fatigue, and it rarely leaves you feeling anything in particular. You don't remember what you saw. You remember the flat, slightly disappointed feeling of having arrived nowhere. That feeling then colours the whole day retroactively.

The same logic applies to individual focus sessions. If every stretch of deep work ends with you hitting a wall, getting interrupted, and sliding into your phone, your brain files focus as something that ends badly. Next time, starting feels heavier, and you don't know why.

You can't always get a peak, but you can always choose an ending

The "peak" half of the rule is mostly out of your hands. You can't schedule a breakthrough. The ending is different, because you usually decide when and how a session closes, and that makes it the cheapest lever you have.

Two features make an ending good in this sense. First, it's deliberate. A stop you chose feels different from a stop that happened to you. Second, it closes the loop. It leaves a sense of completion, or at least a clear handoff to tomorrow, instead of a trailing mess.

There is solid research behind the second feature, too. Work on psychological detachment from work (Sonnentag and colleagues, among others) links the ability to mentally step away in the evening with better recovery. Closing a day cleanly, with tomorrow's starting point already written down, is one of the more practical ways to make that step.

A good ending is small and repeatable

Nothing here requires a ceremony. The best endings are brief enough that you'll still do them on a bad day. What they share is a few minutes of attention, aimed at the right thing.

Name one thing that moved. Not a full review, just one real thing that went forward today. This hands your memory a concrete "peak" to attach to, instead of leaving it to find whatever was loudest.

Write the first line of tomorrow. Leave a note stating the exact next action, not a project name. It closes the open loop so your mind stops rehearsing it, and it makes the next morning's start nearly frictionless.

Make a physical or visual stop. Close the laptop. Stand up. Put the phone in another room. Your brain reads a clear boundary as an ending, and a boundary you can feel is stronger than one you only intend.

Put something good after it. Plan the thing that follows work: a walk, a meal, a call. If the ending is followed by something you look forward to, the stop has somewhere to land instead of a vacuum your phone will fill.

Your next moves

  • Tonight, pick a hard stop time and set an alarm labelled "close it out". Choose a time you can actually keep, even if it's not your ideal.
  • Write a three-line shutdown note before you stop: one thing that moved, one thing that's still open, and the exact first action for tomorrow.
  • Do one physical closing act every day for a week. Close the laptop lid, clear the desk, or put your phone in a drawer or another room for the first twenty minutes after work.
  • Schedule something specific right after your stop time, such as a ten-minute walk or starting dinner, so the gap isn't left for a feed to fill.
  • At the end of the week, notice how your days feel compared with how they felt before. Don't audit your hours. Just ask whether the days seem to have held together better in memory.

A small note on where this gets hard

The trouble with endings is that they happen at your lowest-willpower moment. You're tired, the work has thinned out, and the phone is right there offering something easy. That's why intention alone often fails here. The stop you planned at 3 p.m. loses to the drift you fall into at 5:30.

That's the one place where a little outside structure earns its keep. Reclaim is built around that moment: a way to set the boundary while you're clear-headed so the version of you at the end of the day doesn't have to win an argument from scratch. If you're curious, you can see how it works at reclaim.lumenlabs.works. But the idea stands without it. Your day will be remembered mostly by how it ended, so decide, ahead of time, how yours does.