How Many Customers Your Inaccessible Site Quietly Turns Away

If your checkout threw an error for one customer in twenty, you’d know within a day. Someone would email. Someone would tweet. Your revenue graph would dip and you’d drop everything to find the bug. But there’s a class of failure that behaves differently: it turns customers away in complete silence, and the silence is not incidental. It’s built into the nature of the failure.

Accessibility failures work like this. A person using a screen reader lands on your site, hits a wall of unlabeled buttons and images described only as “image,” and leaves. A person who navigates by keyboard, because a tremor or an injury makes a mouse impractical, tabs into your menu and can’t tab out, and leaves. A person with low vision meets gray text on a white background, can’t read your prices, and leaves. None of them files a bug report. Why would they? From their side, this wasn’t an error. It was your site working as designed, and the design said: not for you.

This is the part owners consistently get wrong. They reason from the absence of complaints to the absence of a problem. But complaints require a working feedback channel and a person who believes complaining might help. Someone who couldn’t even read your contact form has neither. The people your site excludes are precisely the people your site prevents from telling you so. The silence isn’t reassurance. It’s the sound of the problem.

It’s worth being concrete about who these people are, because “users with disabilities” conjures too narrow a picture. Yes, it includes blind users with screen readers, software that reads a page’s structure aloud, which is why that structure needs to actually exist in the HTML rather than only in the visual layout. But it also includes the enormous population of people with partial vision, who don’t use assistive software at all and just need reasonable text sizes and reasonable contrast. It includes people with colorblindness, for whom a form that marks errors only by turning fields red marks them not at all. It includes people with motor limitations for whom small, closely packed tap targets are a wall. It includes older customers, who tend to have more money than average and more of these needs than average, a combination worth sitting with. And it includes, part-time, everyone: anyone in bright sunlight, holding a baby, on a cracked screen, or watching a video with the sound off is temporarily relying on the same affordances, captions, contrast, big targets, that disabled users rely on all the time.

So how many customers is your site actually turning away? I won’t pretend to give you a number, and you should be suspicious of anyone who does, because the honest answer is that it depends on your audience and your site’s particular failures. But you don’t need a number to reason about it. You need only two observations. First, the population affected by accessibility barriers is not a rounding error; disability is common, becomes more common with age, and the temporary and situational versions touch everyone. Second, and this is the economically interesting part, the loss concentrates at the worst possible spots. An inaccessible marketing page loses you a reader. An inaccessible checkout loses you a buyer, someone who had already decided to give you money and was physically prevented from doing so. The most expensive accessibility bugs are the ones closest to the purchase, and forms, with their unlabeled fields, mouse-only date pickers, and color-coded errors, are historically the most broken things on the web.

The maddening thing is how detectable most of this is. Whether an image has alternative text is a yes-or-no fact in the HTML. Whether a form input is associated with a label, whether the contrast between your text color and your background meets the thresholds the accessibility guidelines define, whether your page can be operated by keyboard alone, these are checkable properties, not mysteries. When GazeSite audits a site, accessibility is one of its six focus areas precisely because it’s the area where invisible-to-the-owner and mechanically-detectable overlap most. The owner can’t experience the screen-reader version of their site and will never hear from the customers who did. But the page’s HTML confesses everything, if something bothers to read it. Each finding comes with the specific issue and the specific fix, and a surprising share of the fixes are small: an attribute here, a label there, a color nudged darker.

I’d encourage you to run one experiment yourself, though, before or after any audit, because it changes how you feel about the problem in a way a report can’t. Put your mouse out of reach and try to buy your own product using only the Tab, Enter, and arrow keys. If you get stuck, and on most sites you will, sit for a moment with the fact that for some of your would-be customers, that dead end wasn’t an experiment. It was Tuesday, and they’re someone else’s customer now.

Silent losses are the most dangerous kind, because every feedback system you have is calibrated to detect noise. The customers who complain get served. The customers who can’t complain just disappear, and the books never show what was never rung up. The first step isn’t fixing everything. It’s refusing to mistake silence for health.