How many choices belong on an empty screen?

In 1952 a psychologist named William Hick sat people in front of a set of lamps, gave each lamp a key, and measured how long it took them to hit the right key when a lamp lit up. The finding was clean enough to become a law. The time it took to choose didn’t rise in a straight line as he added lamps — it rose with the logarithm of the number of options. Two choices to four felt like a small step. Four to eight, another small step. Every doubling of the options cost the same fixed slice of time, which is another way of saying the mind isn’t scanning the list one item at a time; it’s narrowing, halving, homing in. More options, more halving, more delay.

That’s Hick’s law, and on its own it sounds like an ergonomics footnote. The reason it belongs at the very top of a builder’s mind is what it turns into when the stakes are real. In Hick’s lamp room the person had to press a key. On your first screen, they don’t. When choosing gets hard enough, a brain has a fourth option the lab never offered: choose nothing, and leave.

Two other bodies of work sit underneath Hick’s and make the same point from different angles. Cognitive load theory, from John Sweller, says working memory is small and easily swamped; a screen that asks you to hold six unfamiliar options in mind at once is spending a budget the newcomer barely has. And the famous choice-overload study — Iyengar and Lepper’s supermarket display of 24 jams versus 6 — found that the larger, more tempting-looking table drew more lookers but produced far fewer buyers. More to consider, less actually chosen. (Worth flagging honestly: choice overload has had a bumpy time in replication, and its size depends heavily on context. But the direction — that piling on options can suppress action rather than encourage it — is well supported and matches Hick’s mechanism.)

Now put a fresh user on the first screen of your product. They have no model of what this thing is yet, no muscle memory, no sense of which button matters. This is the exact moment their cognitive budget is smallest and their willingness to bail is highest. And it is the moment builders reliably do the worst thing: they show everything. Every feature they’re proud of, every entry point, a navigation bar with nine items, three call-to-action buttons of equal weight, a tour, a checklist, and a dismiss-able banner. Each addition felt harmless in isolation. Together they hand a brand-new brain a lamp room with forty lamps and no instructions.

The reframe Hick’s law gives you is almost aggressive in its simplicity: the empty state is not a canvas to fill — it’s a decision you’re making on the user’s behalf about what they do first. The strongest onboarding screens I’ve seen are close to insulting in how little they offer. One obvious next action. Maybe a second, quiet one for the person who wants to poke around. Everything else deferred until the user has enough of a model to want it. You are not hiding your product’s richness; you’re sequencing it, so that the first choice is trivially easy and each later choice arrives only once the user has the context to make it without strain.

A few practical corollaries fall straight out of the law. Prefer sequence over simultaneity — a wizard that asks one thing per screen beats a single form with fifteen fields, because Hick’s cost is about options in view at once. Give options a visual hierarchy so the eye doesn’t treat them as an equal-weight list to be scanned linearly — one primary button and three faint links is a smaller decision than four identical buttons. And treat “add a setting for it” as the tax it really is: every toggle you surface early is another lamp in the room. Push it into an advanced pane the beginner never has to see.

The deeper point is that a choice is never free, even when it costs no code and no pixels to spare. It costs the user a small, measurable act of thinking, at the exact moment they have the least to spend and the easiest exit. Design the first screen as though every option you add is borrowing against the one decision you actually need them to make.