12 · Deciding
Hick's Law
The more choices people have, and the harder they are, the longer it takes to decide.
Evidence: Holds, with limitsWhat we do with it
- 01
Count decisions, not buttons. Three choices that each need research are slower than eight obvious ones.
- 02
Make the order predictable: alphabetical, grouped, or most-used first. People who know where to look don't pay the full cost of a long list.
- 03
When people stall at a choice, find out what they're missing before you remove options. It's often one fact they need to compare.
Examples
Getting it wrong, getting it right.
Don't
A settings page with 40 toggles in the order they were built.
Do
The same 40 grouped under six plain headings, with the five people change most at the top.
Don't
A homepage that opens on twelve links of equal weight.
Do
One main action at the top, and everything else grouped under a few clear headings below it.
The evidence
What the research says.
Strong for simple, practiced choices: press the button that matches the light. Menus are messier. When people don't know the list, they read it item by item, so time grows in a straight line. The gentler curve shows up once people know the order, as in Thomas Landauer and D. W. Nachbar's 1985 menu study. Andy Cockburn, Carl Gutwin and Saul Greenberg put both into one model in 2007, visual search for a newcomer and Hick's law for an expert, and it matched menu data closely.
How our view changed
For a long menu we care more about whether people can predict the order (alphabetical, grouped, most-used first) than about the raw count.
Sources
- Hick, W. E. (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology, 4(1).
- Hyman, R. (1953). Stimulus information as a determinant of reaction time. Journal of Experimental Psychology, 45(3).
- Landauer, T. K., & Nachbar, D. W. (1985). Selection from alphabetic and numeric menu trees using a touch screen. CHI '85.
- Cockburn, A., Gutwin, C., & Greenberg, S. (2007). A predictive model of menu performance. CHI '07.
Last reviewed November 7, 2026. How we keep this current