Updated 6 October 2026. Website accessibility means people can use your content and complete tasks with different ways of seeing, hearing, moving and processing information. For a small business, the practical question is which barriers prevent a customer from reading an article, finding an answer or completing an enquiry—and who will fix them.
This guide covers ownership, priorities and budgeting. For a repeatable browser test, use our separate keyboard, screen-reader and regression testing workflow.
Start with the customer task
Choose three important journeys rather than beginning with a site-wide score. A publisher might select finding an article, reading it with enlarged text and moving to a related guide. A service business might choose understanding an offer, completing a form and receiving confirmation. Write down the expected result before testing.
Accessibility work removes barriers to these tasks. It does not guarantee higher search rankings, more sales or compliance with every applicable law. A useful business case connects a specific problem with a task people need to finish, then measures whether the change helps.
Use a standard, then define the review scope
The W3C WCAG 2.2 Quick Reference organizes technical success criteria into perceivable, operable, understandable and robust content. WCAG 2.2 Level AA can be a documented engineering target; it is not a substitute for determining which obligations apply to a particular business.
Specify the pages, templates, devices and journeys included in your review. Shared navigation, consent interfaces, embedded media and error states matter as much as the default page. Keep documents and third-party components in the scope when customers need them to complete the task.
Prioritize blockers before cosmetic improvements
| Observed barrier | Business task affected | First action |
|---|---|---|
| A keyboard user cannot reach Submit | Sending an enquiry | Fix the control and focus order; repeat the complete form journey |
| An error is shown only by a red border | Correcting a failed submission | Provide a clear message associated with the field |
| Text becomes clipped when enlarged | Reading the offer or article | Repair the layout and verify the affected template |
| Several articles use confusing headings | Finding the relevant section | Improve the shared structure and check representative articles |
A blocked task generally deserves attention before a minor defect that has a workable alternative. Record the affected users, frequency, workaround and number of templates involved. One shared component fix can help many pages, but confirm the result in each important context.
Make ownership explicit
Assign one person to coordinate the backlog. Content editors own meaningful headings, link wording and image descriptions; developers own controls, semantics and interaction; designers own readable presentation and visible states. The coordinator checks that a ticket has an owner, acceptance evidence and a retest date.
The W3C’s planning and management guidance provides a framework for integrating accessibility into ongoing work. Translate that into your own publishing process: include a short content check before publication and test shared templates when they change.
Budget for diagnosis, repair and verification
As an illustrative planning exercise, a team with 20 available hours could reserve five for investigation, ten for repairs and five for retesting. Those numbers are an example allocation, not a cost estimate or a promise that a website can be made accessible in 20 hours.
Ask a supplier to separate the scope of evaluation, implementation, manual testing and follow-up support. A quote described only as an “accessibility score improvement” does not explain which customer journeys will work afterwards. Request examples of the evidence delivered for a resolved issue.
Measure task completion and keep the evidence
Track unresolved blockers, repeated defects and successful completion of the chosen journeys. Compare support requests and abandonment only when the data can answer the question; small samples can fluctuate. Include people who use assistive technology in suitable research and record the actual feedback received.
Finish each repair with reproducible steps, a supported browser or assistive-technology combination and a clear result. Accessibility becomes maintainable when new content and design changes pass through the same checks, rather than relying on a one-time badge.
