Width is the effective size along the movement direction

What it is

In Fitts's law, effective width is neither a control's area nor always its visually horizontal width. It is the tolerable extent of the target along the actual direction of approach. A long horizontal button can be wide when approached head-on yet narrow when approached from the side. Padding, adjacent controls, and hit-testing rules can further make the usable target smaller than its visible outline.

Why it happens

The final correction controls where movement stops along its axis. Extra room perpendicular to that axis cannot compensate for stopping too early or too late. Effective width is therefore the target's projected tolerance band in the movement direction. Change the approach angle and the tolerance changes even if the shape does not.

Studying it

Record endpoints and project them onto the movement axis to estimate effective width from the endpoint distribution rather than from a design bounding box. Manipulate target shape, approach angle, spacing, and hit criterion; measure movement time, misses, and endpoint variability. When trajectories arrive from several directions, analyse them separately—an averaged width can hide a severe directional problem.

Where it stops holding

This definition is for aimed pointing. Large drag-and-drop actions, two-finger gestures, or targets that change while moving may not have one sufficient movement axis. On touchscreens, contact area, finger occlusion, and tolerance algorithms make the effective target differ from its geometric projection. Visual whitespace counts only when it accepts input without colliding with a neighbouring action.

Applying it

• Inspect hit areas from the directions users actually come from, especially the narrowest projection along common approaches.
• For elongated controls, increase clickable extent along the main travel direction or change orientation; decorative whitespace in the perpendicular direction may not help.
• Aggregate missed taps by approach direction in testing or telemetry, then adjust hit areas and spacing where failures cluster.

Related

• Same group: B1.01.1 Pointing time is jointly determined by target distance and width · B1.01.2 Distance and width affect time logarithmically, not linearly · B1.01.4 The law describes an average trend, not one operation · B1.01.5 Enlarging a target and shortening travel have different costs
• Nearby: B1.02 Fitts's law on touchscreens · B1.03 Infinite width at screen edges
• Search terms: effective width · movement axis · endpoint variability