# Alt text and captions on UGC are an accessibility issue, not just a legal checkbox Most UGC galleries are the least accessible part of a storefront: missing alt text, uncaptioned video testimonials. What a blind or deaf shopper actually experiences, and a workflow that scales. By Rohin Aggarwal · 2026-08-17 **Quick answer** - A UGC gallery is frequently the least accessible section of an otherwise-compliant storefront, because the theme and core PDP get accessibility attention and imported customer content gets none. - Auto-generated or missing alt text on customer photos, and uncaptioned video testimonials, mean a blind or deaf shopper gets none of the social proof a sighted, hearing shopper is meant to be persuaded by. - Hand-writing alt text for thousands of imported posts doesn’t scale. A practical workflow mixes structured templates, AI-assisted drafts a human spot-checks, and captions burned into the gallery’s video player rather than relying on the source platform. - This is worth treating as a real gap to close, not a checkbox for a lawsuit: the whole point of a UGC gallery is persuasion, and persuasion that only reaches part of your audience is a self-inflicted conversion loss, not just a compliance risk. A storefront can pass a general accessibility audit, WCAG-conformant theme, correctly labeled buttons, sensible heading structure, and still have a UGC gallery sitting in the middle of the product page that a screen reader announces as “image, image, image, image” twelve times in a row. Nobody set out to build it that way. It happens because the gallery is imported content, sourced from Instagram, TikTok and review platforms that themselves rarely carry usable accessibility metadata, and nobody in the pipeline owns adding it back in. ## What this actually sounds like A blind shopper using a screen reader relies on alt text to know what an image contains at all. When a UGC photo has no alt attribute, most screen readers fall back to reading the filename: something like “IMG_4821.jpg” or a long randomized string from the CDN. That’s not a degraded experience, it’s a null one. The photo is functionally absent from that shopper’s version of the page, even though it’s doing real persuasive work for every sighted visitor next to them. A deaf or hard-of-hearing shopper hits the same wall with video testimonials. A customer describing why a product changed their skincare routine, or walking through how a piece of furniture assembled, is exactly the kind of specific, credible content that makes UGC convert. Without captions, none of that reaches a viewer who can’t hear the audio. They see motion and a product, with the actual argument stripped out. Auto-generated alt text is a partial fix and a real trap. A generic image-recognition caption like “person, indoors, clothing” technically fills the attribute and satisfies an automated scanner, but conveys none of the specific claim the photo is making (a jacket in a stated size on a stated body type, a product being used a specific way). It passes a checklist without doing the actual job, which is the gap that gets a program flagged as accessibility theater rather than accessibility. ## Why this gap is so common Source platforms are part of the problem. A photo pulled in from Instagram carries a caption written for engagement, not description, and rarely any alt text at all; TikTok and reviews sites are the same. Whoever builds the gallery inherits content with zero accessibility metadata attached, and unlike a theme’s hero image or a product photo, there’s no single design file to add alt text to once. There are potentially thousands of individual pieces, arriving continuously, from sources that don’t supply the information in the first place. The honest reason this stays unfixed longer than it should is that broken alt text and missing captions don’t show up in a normal QA pass. A sighted reviewer scrolling the gallery on a laptop sees photos and reads them fine. The gap is invisible unless someone specifically turns on a screen reader or mutes the video and watches with captions off, which almost nobody does as part of routine review. ## A workflow that actually scales - **Structured, templated alt text for the common cases.** Most UGC photos fall into a small number of shapes (product in use, product on a body, product in a setting). A template that pulls in the product name and category from tagging data (“Customer photo of [product name] in [context]”) beats no alt text and beats a generic auto-caption, and it can be generated at import time with zero manual effort. - **AI-assisted drafts with a human spot-check, not full automation.** A vision model can draft a more specific description than a template alone, but should be reviewed in the same moderation pass content already goes through for rights and brand safety, catching the cases where the model gets the product, color, or context wrong. - **Captions generated once, burned into the gallery’s own player.** Don’t depend on the source platform’s captions carrying over, they usually don’t survive the import. Auto-transcribe on ingest, let a human correct obvious errors on the pieces getting the most display time, and render captions as part of your own video component rather than hoping a viewer turns on a platform feature that no longer exists in your context. - **Prioritize by exposure, not by volume.** A gallery of thousands of imported posts doesn’t need every single one perfected before you ship anything. Fix alt text and captions on whatever is actually being displayed in the live rotation first (see the case for rotating that content at all in [the piece on refreshing a gallery monthly](/blog/fresh-monthly-ugc-rotation-vs-evergreen-gallery)), and let the backlog catch up as content cycles through. **Why this is worth the effort:** A UGC gallery exists to persuade a hesitant shopper with real evidence from real customers. Missing alt text and uncaptioned video quietly cut a portion of your audience out of that persuasion entirely, not as an edge case but as a predictable, fixable gap in the pipeline. Treat it as a conversion problem with a legal upside, not the other way around. --- Canonical: https://idukki.io/blog/alt-text-and-captions-on-ugc-accessibility Tags: accessibility, ugc, alt-text, video-captions