Lost users are lost money
Every conversion rate is the share of people who got through. The people who didn't are what the experience costs you. Here is the evidence, step by step, and a way to estimate your own number.
- Poor UX loses users. Most people who start a purchase don't finish, and a large share leave because of design they could have been spared.
- Accessibility barriers lose more of them. One in six people has a significant disability, almost every site fails basic checks, and most of those users leave a site that blocks them.
- Every lost user is lost revenue. The losses run to billions, they compound through loyalty, and since 2025 they carry legal cost too.
Poor UX loses users
70%
of online shopping carts are abandoned. Some of that is browsing, but much of it is the experience.
Baymard Institute, 2025 [1]
22%
of US shoppers have abandoned an order only because checkout was too long or complicated. The average checkout has about 23 form elements; Baymard finds 12 to 14 is enough.
Baymard Institute, 2025 [2]
+35%
conversion is available to the average large online store from checkout design changes alone. Across US and EU ecommerce that is about $260 billion in orders recoverable through better design.
Baymard Institute, 2025 [2]
+8.4%
retail conversions from making a mobile site 0.1 seconds faster, with shoppers spending 9.2% more. Small frictions move money.
Deloitte Digital for Google, 2020 [3]
+32 pts
more revenue growth over five years for companies in the top quarter of McKinsey's design index than their industry peers, and 56 points more shareholder return.
McKinsey & Company, 2018 [4]
Accessibility barriers lose more of them
1 in 6
57%
of working-age computer users were likely to benefit from accessible technology because of mild to severe difficulties with vision, hearing, dexterity or thinking. Zoom, large text, keyboard use and reduced motion are everyday settings, not a niche.
Forrester Research for Microsoft, 2003 [7]
94.8%
of the top million home pages have WCAG failures a scanner can find, 51 on average. Low-contrast text, form fields without labels and empty buttons lead the list, and those are what checkouts are made of.
WebAIM, 2025 [8]
69%
of disabled online shoppers leave a site they find hard to use. Most don't complain, so the loss never reaches a support queue.
Click-Away Pound, 2019 [9]
83%
limit their shopping to sites they already know are accessible, and 86% have paid more for something from an accessible site rather than buy it for less from one that was harder to use.
Click-Away Pound, 2019 [9]
The second step makes the first worse. A checkout that annoys most people can be impossible for someone using a keyboard, 200% zoom or a screen reader. Many of them don't convert less; they don't convert at all.
Every lost user is lost revenue
£17.1bn
a year in spending by 4.9 million UK disabled shoppers who leave sites that block them, up from £11.75bn in 2016.
Click-Away Pound, 2019 [9]
$6.9bn
in potential North American online sales missed because sites were inaccessible to people with vision impairments alone.
Nucleus Research for Deque Systems [10]
$18tn
in global spending power held by disabled people together with the family and friends who choose where to shop with them.
Return on Disability Group, 2024 [11]
5,000+
digital accessibility lawsuits were filed in the US in 2025, about 70% against online stores. 36% of the companies sued made more than $25M a year, and 1,427 had been sued before.
UsableNet [12]
€1M
Estimate what it costs you
Industry figures show the pattern. They can't tell you your number. This estimate needs five figures you can read from your own analytics.
This counts only people using an accessibility setting. The same design problems slow everyone else down too, so the full cost of poor UX is larger.
How strong is each piece of evidence?
If you take these figures to a finance team, expect questions. This is which ones hold up, and why.
| Evidence | Strength | Why |
|---|---|---|
| Baymard abandonment and checkout research | Strong | Large, repeated usability testing over 16 years. The recoverable total is a model built on it. |
| WebAIM Million failure rates | Strong | An automated scan of a million home pages, repeated every year. It undercounts, because automated checks find only some WCAG failures. |
| WHO disability and ageing figures | Strong | Global health data. It counts people, not how many of them visit your site. |
| Lawsuit counts and European Accessibility Act fines | Strong | Court filings and law. Fines vary by country and the figures are maximums, not typical amounts. |
| Deloitte speed study | Moderate | Real site data, but commissioned by Google, and about speed rather than design in general. |
| Click-Away Pound | Moderate | A self-reported survey of UK disabled shoppers. The direction is clear; the pound figure is an estimate. |
| McKinsey design index | Moderate | A correlation across 300 companies. Good design and growth may share a cause. |
| Spending power totals | Use with care | Market sizing, not lost sales. Good for showing scale, not for a forecast. |
| Forrester's 57% | Use with care | From 2003. Still widely quoted, so give the year when you use it. |
Every source has the same gap: none of them measured your users.
Your own number, measured
Metrickle records which accessibility settings people use, alongside every funnel, task and conversion, so the estimate above becomes a figure from your own data.
Every flow, split by setting
Funnels, tasks and conversion rates for keyboard, zoom, large text and other users (and screen reader users in mobile apps), each against people using none.
The release that broke it
Protected flows catch the release that made a flow worse for one setting, browser or device, with the attempts and money lost.
The element behind it
Accessibility scans from CI are matched to the elements people actually get stuck on, so you fix the failures that cost conversions first.
Sources
Figures are as each source published them, checked in October 2026. Numbers match the citations under each figure.
- Cart abandonment rate statisticsBaymard Institute · 2025 · Research summary
- The current state of checkout UXBaymard Institute · 2025 · Usability benchmark
- Milliseconds Make MillionsDeloitte Digital for Google · 2020 · Study
- The business value of designMcKinsey & Company · 2018 · McKinsey Quarterly
- DisabilityWorld Health Organization · Fact sheet
- Ageing and healthWorld Health Organization · Fact sheet
- The wide range of abilities and its impact on computer technologyForrester Research for Microsoft · 2003 · Study, as cited by Microsoft
- The WebAIM MillionWebAIM · 2025 · Annual accessibility analysis
- Click-Away Pound Survey 2019Click-Away Pound · 2019 · Survey results (PDF)
- Research shows how the internet is unavailable to blind usersNucleus Research for Deque Systems · Study summary
- Annual report 2024Return on Disability Group · 2024 · Market report
- Digital accessibility lawsuits surge in 2025UsableNet · Year-end lawsuit report
- Directive (EU) 2019/882, the European Accessibility ActEuropean Union · 2019 · Law
- Penalties for European Accessibility Act non-complianceLevel Access · Summary of national fines
Common questions
How many of my visitors use an accessibility setting?
It depends on your audience, and few analytics tools can tell you. Metrickle records pinch zoom, large text, keyboard navigation, reduced motion, high contrast and forced colours on the web, and screen readers and text size on mobile, so you see the real share for each flow.
What gap should I use in the calculator?
If you don't know yet, try a range. Teams usually find the gap varies by flow: small on content pages, much larger on forms, checkouts and anything with custom controls. Measuring it for one flow is the fastest way to replace the guess.
Is this mainly about avoiding lawsuits?
Legal risk is real and growing, but the bigger cost for most products is the conversions lost every day. Fixing the barriers that cost the most money usually fixes the ones most likely to draw a complaint too.
Find out what your experience is costing you
Free to start. See every flow split by accessibility setting, and the release that made it worse.
Start free