You get one "look here" every half second

Here’s a strange thing your attention does. Flash a stream of letters at someone one at a time, fast — a dozen a second — and ask them to spot two particular ones. They catch the first easily. But if the second target shows up somewhere between about a fifth of a second and half a second after the first, they often simply don’t see it. Not “see it and forget it” — don’t register it at all. Push that second target a little later, past the half-second mark, and it snaps back into view like nothing happened. The eyes never closed. The letter was right there, at full brightness, dead center. The brain was just briefly, involuntarily elsewhere.

This is the attentional blink, and the name is honest: it’s a blink of attention, not of the eye. The going explanation is that catching and processing the first target isn’t instantaneous — it ties up the machinery of attention for a beat while it consolidates what it just saw. During that beat, a second thing can land on the retina in perfect focus and never get promoted from “light hitting the eye” to “thing I noticed.” Attention isn’t a floodlight that’s always on; it’s more like a single pair of hands. Hand it something to catch, and for a short window afterward it can’t catch the next thing, no matter how much you’d like it to. The refractory period is small — a few hundred milliseconds — but it is remarkably reliable, and you cannot train it away by trying harder.

Now watch what products do to a fresh user. They land, and in the first second we fire everything: a success toast slides in and a coach-mark points at the sidebar and a “you earned a badge” animation pops and a red dot appears on the menu. We think we’re being informative. What we’ve actually done is hand attention its first catch — whichever of those grabbed it — and then throw the other three straight into the blink. The user leaves genuinely never having seen the one that mattered, and we conclude “people don’t read our tooltips” when the truth is closer to “we made it neurologically impossible to read the second one.” The same trap sits inside notifications, multi-step confirmations, and any moment where two things worth noticing arrive on top of each other.

So the build decision is about sequence and spacing, not just content: when two things both deserve attention, don’t fire them together — give the first one its half second before you ask for the second. If a user’s action produces both a confirmation and a next-step prompt, let the confirmation land, breathe, and then bring the prompt in a beat later rather than in the same frame. When you must show several things at onboarding, stage them — one clear beat at a time — instead of decorating the whole screen at once. And decide, honestly, which single element gets the first catch, because whatever wins that race buys itself a short monopoly on the user’s attention and briefly starves everything else.

Two honest cautions. First, half a second is not a rule you can set a timer by. The blink is a real, replicated tendency, but the exact window shifts with the person, the task, and how demanding that first catch was — treat it as “back-to-back important cues fight each other,” not as a precise 500-millisecond law you can engineer around to the millisecond. Second, this is an argument for staggering, not for slowness. The fix isn’t to make your product feel sluggish with artificial delays between everything; it’s to stop stacking your two or three genuinely important signals into a single instant. Most of what’s on a screen isn’t competing for that first catch anyway — the discipline is only needed at the moments where two things both truly need to be noticed, and there, a beat of spacing is worth more than any amount of making them bigger.

The science, to look up: the attentional blink, first characterized by Jane Raymond, Kimron Shapiro and Karen Arnell (1992) in a rapid-serial-visual-presentation task, showing a reliable dip in detecting a second target roughly 200–500 ms after the first; and the large downstream literature debating exactly why the blink happens. Treat the phenomenon as well-established and the precise timing and mechanism as still-refined — the effect is robust; the millisecond boundaries and the theory behind them are not settled.