Write image alt text to convey the image’s purpose in its specific context. A useful description helps someone understand information they cannot see; a decorative image may need an empty alternative. Do not turn alt attributes into lists of search keywords.
Audit the purpose before the wording
Create a short list of images from an important page and describe what each one contributes. Separate informative screenshots, data charts, linked graphics and decoration. The same image can serve different purposes on different pages, so a filename-based default is rarely enough for editorial review.
W3C’s image accessibility tutorials distinguish image purposes and appropriate text alternatives. Google’s image search guidance covers image context and descriptive alt text. Use these references to check implementation, while treating the examples below as editorial suggestions.
Example: explain a screenshot’s relevant evidence
Suppose a fictional tutorial includes a screenshot of a sitemap report showing two unavailable URLs. “SEO tool best free sitemap checker” says nothing about the evidence. “Sitemap report listing two URLs with 404 responses” communicates the relevant observation.
The article should also explain which URLs matter and what the reader should do. Do not make a screenshot the only location of essential instructions. If it contains account identifiers or private addresses, create a sanitized example before publication and label illustrative data clearly.
For a step-by-step interface tutorial, avoid transcribing every navigation item into one long alt attribute. Describe the meaningful state and put the complete procedure in nearby text where everyone can use it.
Handle charts with a written explanation
A chart comparing crawler requests across several dates needs more than “traffic graph.” State the main observation in a nearby paragraph and provide the values or an accessible table when the detail is necessary to understand the argument.
Keep the interpretation modest. A chart of requests does not establish crawler identity, and an increase in visits does not prove a content change caused it. Link to the measurement method so readers can inspect what the visualization actually counts.
When preparing an article, ask someone to read it without the image. If they cannot understand the conclusion, improve the surrounding explanation before polishing the alt text. See AI crawler log analysis for how to separate observations from attribution.
Review linked and decorative images deliberately
For an image that functions as a link or control, the accessible name should communicate the destination or action. Check the complete link, including any adjacent text, rather than reviewing the image attribute in isolation. Avoid making a user hear the same label twice through redundant descriptions.
For purely decorative graphics, document that purpose in the design review and use the appropriate empty alternative. Do not invent a verbose description of an abstract flourish because an automated audit says every image needs some nonempty prose.
Check the published page
Inspect the final HTML after the CMS transforms the image. Confirm the intended alternative survives, the image loads, and the caption still matches the final asset. Review small-screen layout and make sure a detailed chart remains readable or has a usable text equivalent.
Keep descriptive asset names and relevant nearby copy, but do not treat any single attribute as a promise of image rankings or AI citation. The priority is a page that communicates its evidence clearly.
For broader document structure, use the semantic HTML audit. The agent readiness checker offers limited raw control observations and does not replace manual accessibility testing.
Frequently asked questions
Should alt text contain the main keyword?
Only when that wording naturally describes the image’s purpose. Choose the information a reader needs, rather than adding a phrase solely because it appears in the article’s keyword brief.
Is an image caption a replacement for all alt text?
Review them together. A caption provides visible context, while the appropriate alternative depends on what the image contributes and whether surrounding text already conveys that information.