Take your best product photo. The one you hired a photographer for, or spent an evening getting the lighting right. Beautiful shot. Real craftsmanship in the frame.
To Google, that photo is a file called `IMG_20260214_093412.jpg` containing nothing.
Not "nothing interesting." Nothing at all. Because here's the fact that reshapes how you think about every image on your website: search engines don't see pictures, they read the text attached to them. And most business websites attach no text worth reading.
The blindness problem, explained properly
When Google crawls your product page, the image itself is close to opaque. Machine vision helps at the margins, but the ranking decision leans overwhelmingly on textual signals:
- The filename, `IMG_4471.jpg` says nothing; `handmade-leather-tote-brown.jpg` says everything
- The alt text, the written description in your image's code, blank on most business sites
- The surrounding page, whether the text near the image confirms what it shows
- Metadata and structured data, captions, titles, product schema wrapping the image in meaning
Every one of those blank means your photo doesn't compete in image search. It's not losing, it was never entered.
Why this costs more than you'd guess
Image search isn't a novelty channel. For anything visual, furniture, fashion, food, crafts, interiors, machinery, a serious share of buyers start with images. They search "brown leather office chair," scan the grid of photos, and click into whichever ones match. Every image in that grid is a doorway into a business.
Your competitors with named, described photos are in the grid. Your better product, wearing the filename `WhatsApp Image 2026-01-12.jpeg`, is not. The customer never chose against you. They never saw you.
And it stacks: image results now appear inside regular search results, inside shopping surfaces, and, increasingly, inside AI answers that pull visuals. Invisible images are locked out of all of it simultaneously.
The fix, in the order that matters
1. Rename before you upload. Filenames are read as content. Describe the thing in plain words, hyphen-separated: `ankara-print-cushion-cover-blue.jpg`. Thirty extra seconds per image; permanent signal. (Already-uploaded images can be renamed too, it's just more tedious, which is why doing it at upload is the habit worth building.)
2. Write alt text like you're describing the photo to someone on the phone. Specific, natural, honest. Not a keyword pile, a description. This deserves its own discussion, because how image descriptions win you customers goes further than SEO. Short version: describe what's actually in the frame, include the detail a buyer would care about, stop.
3. Put images near text that agrees with them. A photo of a product, on a page describing that product, with a caption naming it, three signals confirming each other. Google trusts corroborated images the way you trust corroborated stories.
4. Compress without destroying. Heavy images load slowly, and slow pages sink in rankings, taking their images down with them. Modern formats (WebP) keep quality at a fraction of the weight. Speed is an image-SEO signal wearing a disguise.
5. Add product schema where it applies. Structured data tells search engines "this image shows a product, here's its name and price," which is how photos earn rich results instead of plain listings.
The habit beats the project
Here's the practical truth: fixing 300 old images is a slog. Naming and describing every new image properly takes seconds and compounds forever. Start the habit today, chip at the backlog weekly, and in a quarter your image library goes from invisible to indexed.
We bake image optimization into every SEO and AI search optimization engagement for exactly this reason, it's some of the highest-leverage, lowest-glamour work in search. Nobody brags about filenames. The filenames don't mind. They just quietly bring customers.