Fitts's law
Time to hit a target grows with distance and shrinks with target size.
Movement time to acquire a target is predicted by the ratio of distance to target width, on a logarithmic scale. Doubling the distance costs less than halving the width. The law was derived from physical tapping and transfers well to pointing devices and touch.
Pointing cost by target size and distance
How it shows up in software
Button size, hit area, menu placement and thumb reach on phones are all Fitts problems. Screen edges and corners behave as targets of effectively infinite width, because the pointer stops there, which is why the macOS menu bar sits flush with the top edge. On touch devices the relevant width is the finger contact area, not the drawn pixel size.
Using it well
- Give the primary action a hit area at least as large as a fingertip, and extend the tappable region past the visible edge where layout allows.
- Put frequent actions where the pointer or thumb already is: near the current selection, in a context menu, or in the lower reach zone on a phone.
- Use screen edges and corners for actions that need to be fast, since the pointer cannot overshoot them.
- Put destructive actions far from frequent ones and give them a smaller target, using the law in reverse.
Where it turns manipulative
- Making an accept button enormous while the decline control is a small link, so the law is used to steer rather than to speed.
- Placing an ad or upsell in the exact spot a frequent action used to occupy, harvesting clicks from motor memory.
- Expanding hit areas until they overlap, which produces mis-taps that cost far more time than the law saves.
Where you have seen it
macOS
The menu bar is flush with the top screen edge, so the pointer cannot overshoot the menu targets.
iOS
Reachability and bottom-anchored tab bars place frequent destinations inside the thumb arc on large phones.
Figma
Right-click context menus open at the pointer, so the distance to the next action is near zero.
What the research says
- Fitts, 1954Well evidenced
Time to tap between two plates followed a logarithmic function of the ratio of distance to target width, across stylus tapping, disc transfer and pin transfer tasks.
- Card, English and Burr, 1978Well evidenced
Applied the model to text selection with a mouse, joystick and keys, and found mouse pointing times fit the Fitts formulation.
- MacKenzie, 1992Well evidenced
Reviewed formulations of the law for human-computer interaction and established the version used for pointing device evaluation.
Two-dimensional targets need a width definition along the movement axis, so the model is less clean for irregular shapes than the one-dimensional case suggests.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.