Design the Last Moment First

TL;DR:
Users remember the peak and the ending, not the journey in between. A flat, forgettable final screen can erase three weeks of good work. Design the last thing they see first. That's the one they carry with them.

You spent three weeks on that onboarding flow. The copy is sharp. The animations are smooth. Each step earns the next. And then, at the end, the user hits a confirmation screen that says “You’re all set!” in gray text on a white background, and that’s it. That’s the last thing they see.

That moment is the one they will remember.

Not the smart microcopy in step two. Not the progress bar that felt honest. The gray text on white. The flat, empty ending. That’s what sticks.

Memory does not keep a running total

In 1993, Barbara Fredrickson and Daniel Kahneman ran a series of experiments with short film clips, some pleasant and some aversive, that varied in length and intensity. They asked people to rate how they felt moment to moment, then asked for a global evaluation when each clip ended. Duration had almost no effect. What drove how people remembered an experience was a weighted average of its most intense moments and its final moments. Their paper called it “the peak and end rule,” and it predicted evaluations with striking accuracy. A rule combining peak discomfort and final discomfort alone accounted for 94% of the variance in how people rated what they had just been through.

What this means in plain terms: people are not taking notes. They are not running a tally of every screen, every interaction, every moment of friction or pleasure. They collect snapshots, the most intense one and the last one, and average those two together into a verdict.

“Retrospective evaluations appear to be determined by a weighted average of ‘snapshots’ of the actual affective experience, as if duration did not matter.”

Fredrickson & Kahneman

The 1993 cold water study made this concrete in a way that’s hard to shake. Participants submerged one hand in 14°C water for 60 seconds, which was unpleasant. Then they submerged the other hand for 90 seconds: 60 seconds at the same temperature, followed by 30 more seconds as a researcher raised the temperature to 15°C, still cold but less so. Participants then chose which trial to repeat. A clear majority chose the longer one. More total pain, but a better ending. They chose the experience that cost more because the memory of how it ended outweighed the arithmetic of what it contained.

The math users never run

The implications for design are not subtle. Your users do not judge your product by adding up every moment and dividing by the number of screens. They remember two things, and one of them is the last one.

The more you look into this, the stranger it gets. An experience with a rough start and a strong finish can score better in memory than a smooth experience that ends with nothing. The flow you built through steps three to seven may go unnoticed if step eight is a dead-end error page, a confusing redirect, or a success message so inert it might as well say nothing. The ending does not just close the experience. It colors everything before it.

The peak matters just as much. Intensity spikes, good or bad, get disproportionate weight in memory. A sharp moment of confusion becomes the anchor users use to re-order the whole experience in their heads. A moment where the product does something the user did not expect but needed, that can lift the memory of an otherwise flat flow. Most design teams think about the beginning and the middle: getting users through the door, reducing drop-off, optimizing conversion at each step. Very few think about what the experience feels like at its end, or where its sharpest emotional spike lands.

Disney figured this out without the research

Disney had been engineering memory long before Kahneman named the mechanism. The FastPass system, first introduced in 1999, was not just about cutting wait times. It was about controlling what waiting felt like and where the experience peaked and resolved. Disney parks treat queue design as part of the product: visual storytelling in the line, theming that builds anticipation, a layout shaped so the sharpest moment of excitement lands as close to the attraction as possible. Guest satisfaction surveys at Disney parks have shown that queue management shapes overall park satisfaction scores more than the rides themselves, not because queuing matters more, but because it controls the emotional peaks and the remembered endpoint of the visit.

The product did not change. The memory of it did.

Designers miss this when they focus only on the functional layer. A flow that works fine but ends with a form confirmation page and peaks at an error message will stick in memory as unpleasant or forgettable, even if every screen in between was good. The emotional architecture of a product is not a layer you add after the work is done. It is the work.

Engineer the end on purpose

The practical version of this is not complicated: before you ship any significant flow, ask what the last thing a user sees feels like, and where the sharpest emotional moment in the sequence lands.

The ending does not have to be big. Mailchimp’s send confirmation screen put a high-five illustration right at the moment of peak anxiety, that instant after pressing send on a campaign going to thousands of people. The product worked the same either way. The memory of using it did not.

For negative peaks, the question shifts: where does your flow hurt most? Error states, loading failures, dead ends, form resets are all candidates for the memory users take away. They do not average out. They anchor.

Map the emotional shape of your flow, not just the functional shape. Mark where intensity spikes and where the experience ends. Then ask whether those two moments, peak and end, were ones you designed or ones that just happened. Most of the time, they just happened.

The screens between the ones that count

Every hour spent refining a screen that sits between a bad peak and a bad ending is an hour that does not change what users remember. The work in the middle can be extraordinary. If the last screen is gray text on white, that is what users walk away with.

Your best work gets forgotten. The question is whether you designed that.