Builder's Brain
Builder's Brain · the neuroscience of building · ◉ Evergreen

Your users can't tell which screen the feeling came from.

by Shreyansh Ojha·4 min·Working Theory

Two groups of men cross a river to meet the same interviewer. One crossing is a low, solid footbridge; the other is a swaying rope bridge strung over a gorge. Afterward, the men from the scary bridge were far more likely to call the interviewer back. Their bodies were still humming from the height, and somewhere in the handoff the hum got relabeled — not I’m frightened, but I’m interested. (The classic study behind this — Dutton & Aron, 1974 — has been argued over for fifty years; treat it as the vivid illustration, not the proof. The idea underneath is sturdier than the anecdote.)

That idea is really two, and both hold up. First: an emotion is arousal plus a story about where it came from — Schachter and Singer’s two-factor account. Your heart doesn’t announce its reasons; your mind reaches for the nearest plausible cause. Second: arousal drains slower than events change. The spike from one moment is still leaving your bloodstream when the next moment arrives, and that leftover charge gets added to the new thing’s account — credited to whatever you happen to be looking at now. That’s the core of Zillmann’s excitation-transfer work.

Here’s why a builder should care: your product is a sequence of moments, and the feeling does not reset at the screen boundary. Whatever the user felt during step N is still in the room during step N+1, looking for something to be about.

The negative case first, because it’s the expensive one. A payment fails. An upload spins for eleven seconds. An error line quietly blames the user. And right there, in the wake of it, you show them your best feature, or ask for a review, or surface the upgrade. The residual arousal is negative and unattached — and your shiny new thing is the nearest available cause. It gets relabeled as I don’t like this. You didn’t get a fair hearing; you got charged for the previous screen’s crime.

The move is to put distance — a beat, a clean confirmation, a moment of visible resolution — between a high-arousal bad event and any judgment you actually care about. Let the charge drain before you ask the question.

The positive case is the one you can spend. A real hit of delight — the thing worked, the result beat expectations, the involuntary oh, nice — leaves a warm residue that lasts a few seconds. That’s the moment for the ask that needs goodwill: invite a teammate, grant the permission, start the trial. Not cold, three screens later. In the wake.

time → arousal event 1 event 2 leftover from event 1 event 2's own — what gets blamed or credited
Arousal drains slower than events change, so a judgment made at event 2 is paid partly with charge left over from event 1. Original diagram · Working Theory

One caution, because the wedge is honest or it’s nothing: you can manufacture arousal. Countdowns, fake scarcity, a little engineered panic — they do raise the charge, and the charge does transfer. But it transfers into an association with your product, and anxiety is a bad thing to be remembered by. Borrow arousal from real delight and real stakes. Don’t mint it.

The through-line: your users are excellent at feeling and unreliable at sourcing. They will tell you, sincerely, that they liked or disliked this — when half of what they felt was still leaking from the last thing. So design the order, not just the screens.

The science, to look up: Schachter & Singer's two-factor theory of emotion (1962); Dutton & Aron's suspension-bridge study (1974), and the long debate over it; and Dolf Zillmann's excitation-transfer theory.

Liked this? Get the next one in Working Theory.

Going weekly in August (it's in beta now). One genuinely interesting read on building, the brain, and the science most people missed.

Got a reaction, a counter-example, or something I missed? Reply by email — I read everything.
◉ join in

Where have you hit this — in a product you use, or one you're building?

Threads open here soon. For now, the conversation lives two clicks away — discuss on GitHub, or just reply by email. I read and answer everything.