Target Size Failure Rate: Measuring the Buttons People Cannot Tap

Nobody sets out to build a button too small to tap. Target size failures accumulate from density decisions: a pagination strip that must fit ten numbers in a row, a table row with four icon actions, a social icon set sized to match a footer line of text. Each decision looks reasonable on a desktop monitor and fails on a phone, in a hand with a tremor.

WCAG 2.2 made this a formal requirement, Level AA, in Criterion 2.5.8. Most organisations have never counted their exposure, because their tooling predates the criterion and their audits still test mainly for contrast and alt text. Counting it is the first step, and it is more mechanical, and more fixable, than most accessibility work.

The thresholds, stated plainly

Criterion 2.5.8 Target Size Minimum, Level AA, requires interactive targets to be at least 24 by 24 CSS pixels, with exceptions including inline text links, user agent controls, and cases where a larger target would be essential or where spacing offsets apply. Criterion 2.5.5 Target Size Enhanced, Level AAA, sets 44 by 44. Apple and Google platform guidelines have recommended roughly that larger size for years, which is why 44 is a sensible internal standard even though AA asks for less.

The metric is the count of failing targets divided by total interactive targets, per template or component type. A site wide single percentage hides the story. A rate by template shows the header team, the shop grid team and the legal team with its dense footer exactly where their failures live.

Where failures concentrate

Pagination, calendar pickers, star ratings, table row actions, carousel dots, and map markers are the usual suspects, because all of them pack competing targets into a fixed strip. Icon only toolbars in account and admin areas are a second cluster, often skipped in audits because the pages sit behind a login. Logged in pages are part of the service and part of the risk.

Spacing can save a small target under the criterion's offset logic, where a target under 24 pixels can still pass if a 24 pixel circle centred on it does not intersect another target. That means the measurement is not just size, it is size plus neighbours. Automated measurement helps here, followed by manual review of the exceptions, because inline links in paragraphs are legitimately exempt and should not inflate the failure count.

Motor disability is the user story behind the pixels

Small targets are talked about as a mobile convenience issue, which understates them. For users with tremor, limited dexterity, or switch and head pointer input, a 16 pixel icon in a row of icons is not inconvenient, it is a wall. Adjacent target errors, deleting instead of editing, logging out instead of saving, are the predictable outcome, and they rarely appear in analytics as accessibility failures.

Include target size in the set of metrics you show alongside contrast and keyboard trap rate. Together they describe operability in practice. A page can be perceivable in perfect contrast and still be unusable by the people the operable principle exists for, and a metric set that misses that is measuring the easy half of the standard.

Fix by component, measure the drop

Remediation is efficient because failures are component driven. Enlarge the pagination component once and hundreds of pages pass. Add padding around icon actions, increase row height in data tables for the action column, and space carousel dots. Each fix should be followed by a re count of that template, so the rate shows the improvement and the remaining failures get a named owner.

Set a target rate and a date by template priority, checkout and account templates first. Target size is one of the most objectively measurable criteria in WCAG 2.2, which makes it a good place for an organisation to prove its metric programme works, before applying the same discipline to the harder, judgment based criteria.

Related reading

These earlier guides go deeper on connected questions: Keyboard Trap Rate: The Metric That Finds What Your Scan Score Hides, Alt Text Coverage Lies When Decorative Images Count as Covered, and Form Label Coverage: A Metric That Counts Labels and Misses Meaning.

Frequently asked questions

What is the minimum target size in WCAG 2.2?

Criterion 2.5.8 Target Size Minimum, Level AA, requires targets of at least 24 by 24 CSS pixels, with defined exceptions and a spacing offset provision. The AAA Criterion 2.5.5 requires 44 by 44. Platform guidelines from Apple and Google have long recommended around the larger size.

How is target size failure rate calculated?

Count interactive targets that fall below the applicable threshold, accounting for spacing and legitimate exceptions such as inline text links, and divide by the total interactive targets measured. Calculating it per template or component type is more useful than a single site figure, because failures cluster in specific components.

Which elements most often fail target size?

Pagination controls, calendar pickers, star ratings, table row icon actions, carousel dots, map markers and dense icon toolbars. These components pack many targets into a small area, often on logged in or transactional pages that receive less audit attention.

Does spacing count, or only the size of the target itself?

Spacing counts under 2.5.8. A target smaller than 24 pixels can still conform if a 24 pixel circle centred on its bounding box does not intersect another target or its circle. Measurement therefore needs to consider neighbouring targets, not just each element in isolation.