KENYA · GUIDE 0591 · EDITION 4.7
Website accessibility audit in Kenya: Launch Risks and Decision Checklist
Editorial update: 28 October 2026 · Business planning reference · Review scope
Set the limits before committing
Tie the website accessibility audit trial decision to its actual acceptance gates, unresolved evidence and cash limit.
This is a planning reference for website accessibility audit. The order below is an invented teaching case. It is not a completed customer job, an operating approval or a prediction of demand or earnings.
The example offer and its boundary
Fictional accessibility review A11Y-91 examines five pages and one enquiry journey on a small business site: home, services, process, about and contact. Output is a scoped issue register with URL, reproduction steps, user impact, evidence and proposed remedy. Example finding: 'On contact page, keyboard focus reaches Send but no visible focus indicator appears.' It is a preliminary practical review, not full WCAG conformance certification. Remediation, legal assurance and testing every browser/assistive technology are excluded unless separately agreed.
Risk signals to put in the trial record
| Specific warning | What to inspect before a decision | Decision record |
|---|---|---|
| Automated report says pass but keyboard fails | Reproduce the task manually and inspect focus/control behaviour. | Responsible person; evidence; hold/revise/proceed/stop; review date |
| Form error only red border | Check labels/error text and non-colour cues. | Responsible person; evidence; hold/revise/proceed/stop; review date |
| Issue cannot be reproduced | Compare page state/browser/settings. | Responsible person; evidence; hold/revise/proceed/stop; review date |
Resolve the outstanding questions
- No real site, assistive-technology session or conformance evaluation performed.
- Scope and reviewer competence must be matched to any requested formal audit.
Milestones to define from the scoped order
| Gate | Evidence needed |
|---|---|
| Scope agreed | Assess a website against an agreed accessibility scope and report evidence. |
| First prerequisite addressed | A default home page alone does not cover error messages, menus or form confirmation. |
| Inputs suitable | Use fictional form data and avoid unapproved real submissions. |
| Handover acceptable | Automated scan output is evidence to investigate, not proof of complete compliance or accessibility. |
| Cash collected | Actual receipts, refunds, credits and receivables reconciled to the same order |
Decide from observed results
Proceed within a defined trial limit only after the buyer's acceptance, actual full costs and collections are recorded, and the guide's unresolved checks are addressed. The final business-specific checkpoint is: Automated scan output is evidence to investigate, not proof of complete compliance or accessibility. Otherwise record a hold, revision or stop decision in file 10.
Set a maximum affordable test loss and an unpaid-balance limit from your real funds. Do not invent a launch date where the work depends on growing periods, training, approval, product validation or a customer cycle. A single teaching case and a single real delivery do not establish repeat demand or safe expansion.
References and their limits
Evaluating Web Accessibility Overview — W3C WAI
Human evaluation required; tools alone do not establish conformance. Reference check recorded 2026-10-26.
Forms Tutorial — W3C WAI
Labels, instructions and clear form structure. Reference check recorded 2026-10-26.
Images Tutorial
Purpose-based text alternatives; does not validate marketing performance. Reference check recorded 2026-10-26.
Recorded references support only the scope stated beside them. They do not prove current supplier stock, a local quotation, actual buyer demand, operator competence or the feasibility of this particular business. Later editorial changes do not refresh earlier source-check dates.
Prepared by Business Starter Guides using AI-assisted drafting and desk editorial checks. Independent operator trials and specialist certification are not claimed. See the editorial method for the review scope.
Editorial method and evidence limits