Definition: Target size is the area you have to hit with a mouse, finger or stylus to activate a control. Under WCAG 2.5.8 (Level AA, new in WCAG 2.2), it must be at least 24 × 24 CSS pixels – or the target needs enough space around it [1].
German: Zielgröße · Also: touch target, tap target · Last reviewed: 27 September 2026
Who it helps
Small targets are hard to hit for many people: people with a tremor or limited fine motor control, people using a mouth stick or head pointer, older people, and anyone using a phone one-handed on a bus. A missed tap can trigger the wrong control – "Delete" instead of "Edit".
What WCAG 2.5.8 requires
A target passes if it is at least 24 × 24 CSS pixels. If it is smaller, spacing is enough: draw a 24-pixel circle centred on each undersized target, and those circles must not overlap another target or another undersized target's circle [1][2].
Exceptions:
- Equivalent: the same function is available on the page through a target that is large enough.
- Inline links in a sentence, whose size is set by the line height.
- User-agent controls whose appearance the page has not changed.
- Essential: the size is legally required or essential to the information, such as pins on a map.
The stricter Level AAA criterion (2.5.5) asks for 44 × 44 CSS pixels, and many design systems use 44 pixels as their baseline anyway.
Common mistakes
- Icon buttons with a 16-pixel icon and no padding, so the clickable area is only as large as the icon.
- Tightly packed pagination ("1 2 3 … 10") or star ratings.
- Close buttons on banners and dialogs.
- Forgetting the mobile view: big enough on desktop, squeezed together at 360 pixels wide.
In code
/* Clickable area = icon: 16 × 16 px */
.icon-button { padding: 0; }
/* The target grows, the icon stays small */
.icon-button {
min-width: 44px;
min-height: 44px;
display: inline-grid;
place-items: center;
}
How Reviseberg handles it
Reviseberg checks target size with the axe-core rule target-size on every page – at 1280 pixels wide and again at 360 pixels, because targets often crowd together on narrow screens. Menus that open on hover or focus are closed when axe runs, so on one page per template Reviseberg opens them and measures the links inside. Neither check can always judge the exceptions, the equivalent-target one in particular, so 2.5.8 shows as "partially" testable by automation in the overview.
Related terms
Voice control · Reflow · Screen magnifier · Assistive technology