Established by William Edmund Hick (On the Rate of Gain of Information, 1952) and Ray Hyman (Stimulus Information as a Determinant of Reaction Time, 1953).
States that decision time rises with the logarithm of the number of choices, so every added option slows people down and degrades their choices.
Hick's law on a ticket kiosk: when a short first screen beats one long one
The curve is Hick's law for one screen: the time to choose grows with the logarithm of the number of options, so each added option costs a little less than the last, but they keep adding. The square is a two-screen design: the most used options plus More on the first screen, everything else behind it. Below the curve the short screen is faster on average; it wins only when enough visitors find their option on the first screen. Illustrative.
The design callKeep one screen: too many buyers pay for the second step.Use the short first screen: it saves time on average.
An example
A regional rail operator's ticket kiosks put all twelve ticket types on one screen, and at the morning peak the queues backed into the concourse. The design team's first idea was a two-level menu: four category buttons, each opening its own list. The usability lead ran the numbers on Hick's law before anyone drew a screen. Two choices among small lists take longer than one choice among twelve, because the time per choice grows only with the logarithm of the count, so splitting a list into categories adds a whole extra decision for every buyer.
The sales data offered a better split. Four ticket types made up 85 percent of purchases. A first screen with those four and a More button would make the common case quick and the rare case slower. At 85 percent, the average buyer would save about half a second; below about 75 percent, the design would cost time. The team built the short screen, retired two ticket types nobody bought, and timed the peak queue before and after.
Why it works
Choosing is a search through the options, and the time it takes grows with how much the chooser has to tell apart. When options are equally likely and familiar, that grows with the logarithm of their number: doubling the list adds a fixed slice of time, so each new option costs a little less than the last, but they never cost nothing. Every option added to a screen, a menu, or a form slows everyone, including people who never wanted it.
The law also says what does not help. Splitting one list into categories turns one choice into two, and two logarithms add up to more than one. What does help is putting the likely options where most people stop, removing options nobody picks, and making the list familiar so people skip the search altogether.
The thinking behind it
William Edmund Hick William Edmund Hick, a British experimental psychologist at the Medical Research Council's Applied Psychology Unit in Cambridge, reported in On the Rate of Gain of Information (1952), in the Quarterly Journal of Experimental Psychology, that the time needed to choose among alternatives rises with the logarithm of their number. The title states his framing: a choice is information taken in, and more alternatives mean more of it to take in before the hand can move.
Ray Hyman Ray Hyman, a University of Oregon psychologist, followed Hick's work with Stimulus Information as a Determinant of Reaction Time (1953), in the Journal of Experimental Psychology. His experiments showed that choice reaction time rises with the information in the stimulus, which for equally likely choices is roughly the logarithm of their number. Measuring the stimulus as information is what makes the law usable on real screens, where some options are picked far more than others. The combined result is known as the Hick-Hyman law.
How to run it
- Count the options at each decision. List every choice a user faces on the screen, form, or menu, including ones that are rarely used.
- Get the real usage. Pull how often each option is chosen; the simple form of Hick's law assumes equal odds, and real use is almost never equal.
- Cut what nobody picks. Remove or hide the options with negligible use, since each one slows every user.
- Put the common choices first. Give the few options most people want their own short first screen or top position.
- Price the second step. Estimate the time a More or a submenu adds for the rest, and find the share at which the design breaks even.
- Avoid splitting for its own sake. Categories that every user must pass through add a decision; use them only when the list is unfamiliar and hard to scan.
- Time it with real users. Measure choice time and errors before and after, at peak load, before rolling it out.
Where it goes wrong
The law is applied to a search it does not describe. Hick's result holds for familiar options; a long list of unfamiliar labels is read one by one, and time grows with the count itself. Spot it when users scan with a finger. Group and label for scanning first.
Options are cut by guesswork. A team removes what it assumes is rarely used and strands a small group that needs it. Cut on usage data, and keep a path to everything.
The decision time is the wrong target. A screen that saves half a second but doubles wrong purchases is a loss. Measure errors and abandonment beside time.
When to use something else
- Queueing Theory. When the question is how long the line gets at peak, given how long each transaction takes.
- Pareto Principle. When the first need is finding the few options that account for most of the use.
- Forcing Function. When the aim is to make the wrong choice impossible, beyond making the right one fast.
Where it sits in SOLVE eX
Go to the source
- On the Rate of Gain of Information, Quarterly Journal of Experimental Psychology, 1952
- Stimulus Information as a Determinant of Reaction Time, Journal of Experimental Psychology, 1953