Measuring Accessibility at Design Review: Catching Barriers Before They Are Built

2026-10-19

Every accessibility program knows prevention is cheaper than remediation, and most programs still measure almost only remediation: audits of built pages, fixes to shipped code. The design stage, where a barrier costs a conversation to remove, runs largely unmeasured and therefore unmanaged.

Design review metrics are not about policing designers. They are about proving that early review happens, learning what it catches, and feeding production failures back into the checklist so the same barrier is never built twice.

Track coverage before counting catches

The foundational metric is coverage: what proportion of new screens and flows received a documented accessibility check before development started. A program reviewing 20 percent of designs cannot claim a prevention strategy, whatever its catch rate is on that sample.

Coverage is also the easiest number to raise. It responds to scheduling, a standing review slot, a checklist embedded in the design tool, an accessibility annotation step in the handoff. Measure it monthly and report it alongside remediation metrics so prevention gets equal standing.

Log what reviews catch, by category

A simple log of issues found in review, categorized as contrast, missing labels or names, focus order, target size, state communication, motion, or content structure, builds the evidence for where design systems help most. Recurring categories are component library problems wearing a project disguise.

That log also calibrates effort. If most catches are contrast pairs in one component, a design system fix ends the category. If catches are mostly focus order in complex flows, the fix is guidance and review time for interactive work specifically.

Escaped defects close the loop

An escaped defect is a production accessibility issue whose origin was a design decision that a review should have caught. Tagging production issues this way, honestly and without blame, measures review quality rather than review activity.

The escape list is the curriculum for the next quarter's checklist. Every escape names something the current review missed: a state that was never designed, an error message with no text equivalent, a gesture with no alternative. Add it to the checklist and the escape rate for that category should fall measurably.

Keep the review checklist short and specific

Design accessibility review fails when it tries to be an audit. A one page checklist tied to design decisions works: text and UI contrast pairs listed, interactive elements visibly focused in the mock, touch targets at size, information never carried by color or position alone, error and success states designed, motion and auto playing content specified.

Short lists get used. A review that takes 20 minutes and happens for every flow will catch more real barriers in a year than a heroic two hour review that happens twice and is then quietly dropped.

Frequently asked questions

What should design review accessibility metrics include?

Coverage, the share of designs reviewed before build; catches per review by category; and escaped defects, production issues that trace to design decisions review should have caught.

Why track escaped defects?

They measure whether reviews actually prevent barriers, not just whether reviews happen. Each escape identifies a gap to add to the review checklist.

What belongs on a design accessibility checklist?

Contrast pairs, visible focus intent, target sizes, information not carried by color alone, designed error and success states, and motion specifications. Short and specific beats exhaustive.

How long should a design accessibility review take?

About 20 minutes for a typical flow with a focused checklist. Frequency and coverage matter more than depth on any single review.