More choices, longer decisions.
The more choices you present, the longer every decision takes. Hick's Law shapes menus, settings, and onboarding — along with the techniques that work around it.
Reaction time grows with the logarithm of the number of equally probable choices. Double the number of options, and decision time goes up by a fixed amount — not double, because the brain searches the set, it doesn't read it linearly.
The clearest single application of the law: a menu of 4 is faster to navigate than a menu of 12 — even though the 12 might cover more ground. Reduce what's not essential and the rest gets faster.
A flat list of 16 is a Hick's Law nightmare. The same 16 split into 4 groups of 4 is two faster decisions in a row — a 4-way pick, then another 4-way pick. log₂(4)+log₂(4) ≪ log₂(16).
Hide advanced or rarely-used options behind an expander. Most users see only the common 5; power users open the full set when they need it. Same UI, two different decision trees.
A well-chosen default removes the decision entirely. If 90% of users pick the same value, make it the default — they get to skip the choice altogether, and the law no longer applies.
At small n, a flat list is fastest. Past ~7 items, breaking into a 2-level hierarchy beats scrolling. Past ~50, hierarchy alone isn't enough — you need search.
Browsing forces every user through the full Hick's curve. Search collapses it to one decision: type the thing you want. For catalogs of hundreds, a search input is the most accessibility win you can ship.
Hick's Law assumes equally probable choices. In practice, making options visually distinct (icon, color, size) cuts the search space — users discard wrong options before reading them.
Hick's Law applies once. Decision fatigue is the cumulative cost — every choice draws down a finite pool of mental energy. A wizard of 12 short steps is harder than the math predicts.
Hick's Law doesn't say "show one option." Too few choices feels arbitrary or constraining. The right number is the smallest set that still covers the user's real intent — often 3 to 7 for primary actions.
The prompt and the agent both attack decision cost from a new angle — replacing the menu with language, or removing the choice entirely.
Instead of choosing from N options, the user describes what they want. Hick’s Law is sidestepped — there’s a single input box, not a branching list to scan.
Recommenders and agents collapse the option set to a single suggested action — removing the decision entirely. Powerful when right, frustrating when wrong.