Alt Text for Social Media: Examples and Review Checks
- How do you write useful alt text for social media images?
- What does the image contribute to this post?
- How should you handle words inside an image?
- How long should the description be?
- What should you do with charts and infographics?
- Where does the description go on the platform?
- What should the final review check?
- Sources
How do you write useful alt text for social media images?
Write alt text that conveys the image's relevant meaning in the context of the post. Identify the important subject or action, include essential words shown in the image, and check every detail against the final asset. Keep it concise. For a detailed chart or diagram, provide an accessible longer explanation as well. Use the platform's image-description feature and verify the published result.
As the W3C images tutorial explains, images need text alternatives that communicate their information or function. This guide offers an editorial drafting and review process, with fictional examples; it is not a claim that adding one field makes an entire campaign accessible.
What does the image contribute to this post?
Start by asking what a reader would miss without the image. The W3C informative-images tutorial explains that a text alternative conveys meaning or information, rather than necessarily describing every visible object. Context determines which details matter.
For an imaginary workshop announcement, “people in a room” could omit the action the photograph was chosen to show. “A facilitator demonstrates folding a paper model while participants follow at their tables” communicates more, provided that is what the approved photograph actually depicts.
Do not add an unverified identity, emotion, relationship or result. “Delighted customers celebrate our success” is a promotional interpretation if the image only shows people standing together. Have the factual owner confirm names and roles when those are necessary. The PR brief should identify that owner.
How should you handle words inside an image?
Read the final graphic, not an earlier caption draft. W3C's alt decision tree distinguishes text already available nearby from words supplied only by the image. Essential image text needs an equivalent that readers can access.
Here is a fictional example. A graphic says “Registration closes Friday at 17:00 UTC,” while the post says only “Join our briefing.” The closing information is missing from the post text. Add the verified deadline to the accessible message rather than describing the graphic merely as “blue event poster.” In a real announcement, use the full date as well as the time zone wherever needed to avoid ambiguity.
The decision tree discusses HTML image attributes. Do not assume a social platform exposes every HTML option, including an explicit decorative-image setting. Use the platform's supported controls and assess the complete post. Leaving its description blank is not automatically equivalent to deliberately marking a decorative web image in HTML.
How long should the description be?
Use the shortest wording that preserves the relevant information. The W3C writing tips recommend concise descriptions and a longer-description approach when a short phrase or sentence cannot convey the content. There is no reason to fill a platform's entire allowance merely because it exists.
Use normal punctuation and put the important information first. Read the caption and description together. Remove irrelevant colour, camera or background detail unless it contributes to the message. Conversely, keep colour or placement when it carries meaning, such as distinguishing two labelled series in a chart.
What should you do with charts and infographics?
W3C's complex-image guidance describes a short identification plus a longer account of the essential information. Charts may need values, relationships, scales and trends. A label such as “results infographic” does not communicate those results.
Our suggested social workflow is to put a useful summary in the post and make the detailed explanation available in accessible text at a clearly identified destination. Check that the destination actually contains the information. A link to another image of the same chart has not solved the problem.
For a fictional graphic with two labelled bars, 12 enquiries and 18 enquiries, the difference is six enquiries. That comparison does not prove a campaign caused the change. Preserve definitions and caveats from the source data, as in our PR measurement guide.
Where does the description go on the platform?
Check the current platform help for the format you are publishing. X's image-description instructions, checked on September 13, 2026, allow a description for each image and state a limit of 1000 characters. Their image-description feature is not available for videos. Those are X-specific instructions, not universal social-media specifications.
Prepare the wording alongside the asset, then have the authorized publisher enter it in the correct field. A spreadsheet cell marked “alt text approved” does not establish that the description reached the published image. For video or another unsupported format, arrange the appropriate accessible alternative separately.
What should the final review check?
Use this original handoff table for each image, including each separate image in a multi-image post:
| Check | Evidence to inspect |
|---|---|
| Asset match | The final image and its assigned description |
| Factual accuracy | Approved names, visible action and verified graphic text |
| Missing information | Caption plus description, read together |
| Complex detail | Working destination containing the longer explanation |
| Publication | Description attached to the correct published image |
Give the image, description and platform version a shared identifier. If a crop removes an object or an updated graphic changes a number, reassess the description. Check through the platform's available accessibility features rather than assuming an export preserved everything.
Keep this review inside the social repurposing workflow, alongside the caption and asset. The practical result is a matched, verified description for the image that readers actually receive.