A real accessibility audit for a small business site runs somewhere around $3,000. A complex application with custom components can push past $25,000. Neither number is the one worth remembering. What matters more: a free scanner and a $50,000 enterprise audit often turn up roughly the same first 80% of problems, missing alt text, low contrast, unlabeled buttons. The cost sits entirely in the last 20%, the part no scanner can see and most DIY passes never reach.
Automated tools catch somewhere between 30% and 50% of real issues. The rest needs a person: a screen reader, a keyboard, no mouse, working through the site the way someone using assistive technology actually would. That's where checklists run out of road. Alt text is easy to add. A custom date picker's keyboard focus order isn't. A modal that traps focus correctly, a data table that reads sensibly to a screen reader: none of that shows up on a scan, and none of it gets fixed by working through a punch list once. A site can tick every box on a popular accessibility checklist and still fail for a user who can't use a mouse, because the checklist tests for presence (is there alt text) instead of function (does the keyboard navigation actually work).
Professional audits price the way they do because the testing is mostly manual. A thorough ecommerce audit runs 150 hours of hands on testing across a handful of core page types. Not a script running overnight, a person, clicking through everything by keyboard. And the audit is the cheap half. Remediation, actually fixing what gets found, commonly costs more than the audit itself, especially once the fixes touch shared components used across dozens of pages. Skip a round of retesting after the fixes ship and you pay for the audit twice: once to find the problems, again later when the same class of issue resurfaces because nobody verified the fix in every state it needed to work in. Most sites that fail an audit fail on the same handful of things. Images. Links. Form fields. Keyboard navigation. Exactly the areas a DIY pass tends to skip.
Doing nothing has a price too, and it isn't only a legal one. Settlement figures for smaller businesses commonly land between $5,000 and $50,000, on top of legal defense fees that can run $10,000 to $30,000 on their own. Some states allow damages per visit, so one determined repeat visitor can generate real numbers fast. Underneath the legal exposure sits a quieter cost: a checkout or a navigation menu a meaningful share of visitors can't use is lost revenue every day it stays broken, whether or not anyone ever files a complaint.
None of this is an argument against the free scanner. Run it. Fix what it finds, that part genuinely is DIY work, and there's no reason to pay someone for an afternoon of alt text. The honest line sits right where the scanner stops: custom components, assistive tech testing, retesting after fixes ship. If your site has forms, filters, or anything more interactive than static pages, get a real quote for that manual gap before deciding it isn't worth closing. Knowing where your own pass ends is worth more than pretending it doesn't.