Hick's Law

There was a long stretch where showing more felt like being generous. From the inside it makes sense: one customer needs this feature and another needs that one, so putting everything on the surface means nobody gets stranded. Menus grew by a row every quarter, and each new feature landed on the home screen next to the last one.

The argument that followed was always the same argument. A designer would say the list had too many items. A product manager would say some customer uses every one of them, so which one are you proposing to take away. One side called the screen cluttered and the other side called the removal wasteful, and since both were describing a feeling, neither could move the other.

Hick's Law puts something measurable into that room. Measure how long people take to pick from two options, then four, then eight, and the times line up in a consistent pattern. Once that pattern exists, "this looks cluttered" becomes "this screen costs people time before they can act," and whether to cut an option turns into a comparison between that cost and what the option is worth.

Adding one more option doesn't add a fixed amount of time. The first few options cost a lot; after that, each additional one costs less than the one before it.

The shape is this: every time the number of options doubles, roughly the same amount of time gets added. Going from two items to four adds about as much as going from four to eight. That has a practical consequence — trimming three items off a list of twenty is nearly invisible to the user, while adding a third option to a screen that had two is very noticeable.

Three techniques fall out of that shape.

1. Grouping speeds things up. Twenty-four items in a flat list is one twenty-four-way decision. Split into four groups of six and it becomes a four-way decision followed by a six-way one. Even added together, the two small decisions land faster than scanning twenty-four at once.
2. A default removes the decision entirely. If the value most people want is already selected, those people effectively face no choice at all.
3. Deferring beats deleting. Progressive disclosure — revealing options only when they become relevant — works because of this law. Move a setting that nine out of ten users never touch behind an "advanced" link and the count on the first screen drops.

There's a precondition. The pattern only holds when a person can tell the options apart at a glance. Put Settings, Preferences, and Configuration next to each other and the count is three, but nobody can distinguish them, so they open each one in turn. The time that costs blows past anything the law predicts. This is why renaming and reordering usually comes before deleting.

Pricing pages call on it most often. List seven plans and a prospective customer stops choosing and starts comparing, and comparing frequently ends with a closed tab. Most products settle on three and mark one of them as recommended. That badge is worth noticing: it shortens the decision without removing a single option.

Navigation and information architecture are the second home. The judgment that a menu past a dozen items needs grouping comes from here. But the grouping isn't free-form — it has to follow boundaries users already carry in their heads, and finding those boundaries is half the work of structuring information.

Third is the first screen a new user sees. An empty state offering five possible starting actions usually produces none of them. The fix is to make the single most common action large and let the rest sit quietly underneath, or leave them out until there's something to work with.

Some screens use the law in reverse. Deleting an account or exporting everything is hard to undo, so those flows are built so they can't be finished in one motion. An extra confirmation step, a box where you type the workspace name — each one lengthens the decision, and here the delay is the point. Knowing the law is not the same as always spending it on speed.

"Five plans is a Hick's Law problem. Let's feature one and push the rest into a comparison table."

We can't cut the count right now, so cut it only where the eye lands first.

"Twelve menu items and no grouping. Three clusters alone would change how this feels."

You can shorten the decision without removing anything.

"We're running Hick's Law backwards here. Deleting an account shouldn't be one click."

The action is irreversible, so slowing it down is deliberate.

At Pagedeck the phrase came up over the document sharing screen. Noa's mockup had six permission levels for anyone holding a share link. View, comment, and edit were understood by everyone; the other three couldn't be told apart from their names alone. Mia pulled the usage numbers and found that the top three accounted for 94 percent of every share ever created.

Noa didn't delete the other three. She moved them. The main screen kept three options and the rest went under advanced settings. The same argument that used to stall — cut it versus keep it — closed in half an hour once two things sat on the table together: a 6 percent usage figure, and a law saying the count itself charges everyone who opens the screen.

The most common misreading turns it into a rule that fewer is always better. The conclusion is short and easy to remember, so it hardens that way. But the law describes one relationship, between option count and decision time, and says nothing about whether a feature deserves to exist. Cut on time alone and the customers who relied on that option file a ticket or quietly leave. The speed gain shows up in next week's numbers and the lost accounts show up at the end of the quarter, which is why this mistake tends to bill late.

Treating the count as the whole story is the same error wearing different clothes. Eight items is eight items, so any two eight-item screens should feel equally slow. In practice, eight options with sharply distinct names arranged in sensible clusters and eight options with overlapping names in arbitrary order differ by several times over. Looking at the labels and the grouping usually pays more than looking at the number.