Amazon Image SEO: File Names, Alt Text, and Backend Terms

Here's a confession. For years I named image files whatever came out of Photoshop. IMG_4821_final_FINAL2.jpg. You know the file. Everyone has one. It didn't matter back when I was just handing a designer a folder. It matters a lot when that folder goes to Amazon, because the file name is one of the few pieces of text Amazon actually attaches to the image itself.

I've seen listings where the photography was beautiful and the search performance was limp, and the fix wasn't a reshoot. It was naming, order, and backend terms. Boring stuff. The kind of stuff nobody posts about. That's exactly why I'm writing about it.

What "image SEO" actually means on Amazon

The direct answer first: on Amazon, image SEO means making your images findable, readable, and consistent with the words shoppers search for. It's not a ranking hack. It's alignment between what the image shows, what the file is called, and what the listing says.

Here's where sellers get confused. On a normal website, you upload an image, type alt text into a box, and Google reads it. Amazon doesn't work like that. There is no public alt-text box for each gallery image in the listing editor, the way a WordPress site gives you one. What you actually control is different: the file name at upload, the image order, the title and bullets around the images, the backend Search Terms field, and the text inside A+ modules where it applies.

So when someone sells you "Amazon image alt text optimization" as a standalone service, ask them which field they mean. If they can't name it, they're selling you a website concept with an Amazon label stuck on it. I've had that conversation more than once, and it never ends well for the seller who paid.

File names: the system I actually use

The direct answer: name files before you export, with the product keyword first, then the slot, then a short descriptor, all lowercase, separated by hyphens. Never ship IMG_4821 anything.

Why does this matter if shoppers never see file names? Two reasons. First, file names are how you stay sane across a catalog. When a listing needs a swap six months later, you want to find the right file in ten seconds, not open twenty mystery images. Second, in bulk upload workflows Amazon uses a naming convention of its own, and sloppy names are where uploads break.

For bulk image uploads, Amazon's convention uses the product identifier, a variant code, and the extension, separated by periods. The main image uses MAIN, extra images use PT01 through PT99, and angle shots use codes like FRNT, BACK, LEFT and RGHT. A clean example looks like this: B000123456.MAIN.jpg for the main image, B000123456.PT01.jpg for the first extra image. Spaces, dashes, or extra characters in that convention stop the file from uploading cleanly, so this is the one place where you follow Amazon's format exactly instead of inventing your own.

For everyday work before upload, I use a simpler descriptive system:

A few rules I don't bend on. One idea per file name, not a keyword salad. No brand names that aren't yours. No claims like best or number-one in a file name, because that text has a way of surfacing somewhere you didn't intend. And keep the original high-resolution export, because Amazon's own image rules are strict about quality: images must be at least 500 pixels on the longest side, and zoom only switches on at 1,000 pixels or more. I export at 2,000 pixels on the longest side as a working standard, which keeps zoom working and the image sharp on large screens. That 1,000-pixel zoom threshold and the 500-pixel minimum come straight from Amazon's product image guidance, and they're worth memorizing because a beautiful image uploaded too small is just a beautiful thumbnail.

If you want the full compliance side of this, my Amazon main image requirements checklist covers the white-background, 85% frame fill, and banned-elements rules that sit next to the naming work. Naming gets you organized. Compliance keeps you live. You need both.

Alt text: where it still matters, and where it doesn't

The direct answer: don't chase a gallery alt-text field that Amazon doesn't give you. Put your descriptive words where Amazon and Google can actually read them: the title, the bullets, the A+ module text, and the file name.

This is the part most guides get wrong, so let me be blunt. If you spend an afternoon writing alt text for gallery images and pasting it nowhere Amazon reads, you've done website SEO on an Amazon listing. The effort feels productive. It isn't.

What does work? Descriptive, honest text in the places that exist. Your main image should match your title's core product words, not fight them. Your second image's job is usually a different angle or a detail, and the words in your bullets should say what that detail is. If your A+ Content has image modules, the text in those modules is real, readable copy, so write it like a human explaining the product, not like a keyword dump. And on your own site or blog, where alt text genuinely exists, write it properly. That's where Google Images actually meets your work. I've written before about how many listing images you need and what each slot should do, and the same logic applies here: every image earns its place by answering one shopper question, and the words around it should say which question that is.

One more honest note. Amazon's own systems increasingly read the image itself, not just the text around it. That means a clear, unambiguous photo beats a clever file name every time. A close-up that actually shows the material, a scale shot with real dimensions, a lifestyle shot where the product is obviously the hero. No file name rescues a confusing photo. Naming supports good images. It doesn't replace them.

Amazon recommends at least six additional images and one video per product, on top of a compliant main image. That recommendation is about giving shoppers enough to decide, and it's the simplest image SEO advice there is: more complete galleries answer more questions.

Backend search terms: the field most sellers waste

The direct answer: backend search terms are the hidden Search Terms field in Seller Central, and their only job is covering relevant words your title, bullets, and description don't already use. Fill them deliberately, stay under the limit, and never repeat what's already visible.

Here's the number that matters. In the US marketplace, the Search Terms field is limited to 250 bytes, not 250 characters. Bytes, not characters. Spaces count. Accented characters can take two bytes or more. Go over the limit and Amazon may ignore the whole field, not just the extra words. I've seen listings where someone pasted a paragraph of keywords into that box and wondered why nothing changed. The field was over the limit the entire time. It looked fine in Seller Central. It was doing nothing.

Amazon's own Seller Central guidance for the Search Terms field sets out that 250-byte limit, and it's the one number in this post I'd ask you to actually check before you trust anyone else's version of it, including mine. The rules around it are simple once you see them: separate terms with spaces, not commas. Don't repeat words already in your title or bullets, because repetition buys nothing. Don't put competitor brand names in there. Don't put ASINs, subjective claims like best, or temporary words like sale in there.

So what belongs in the field? The words a shopper might type that would sound awkward in your visible copy. Synonyms and alternate names. Common misspellings. A use case or occasion, like camping or dorm. An audience word, like toddler or senior, if it genuinely fits. A material or fit detail that reads clumsily in a bullet. That's it. If a term already appears in your title, leave it out and spend that byte on something new.

Write the field, count the bytes, then stop. A backend field stuffed with 250 bytes of near-duplicates is worse than a clean field with 120 bytes of genuinely new coverage, because the clean one actually gets read.

The 20-minute image SEO checklist I run before upload

The direct answer: run this checklist in order, and don't upload until every box is ticked. It takes twenty minutes on a new listing and saves you the reshoot conversation later.

1. Export at 2,000 pixels on the longest side, JPEG, RGB, sharp at full size. Zoom needs 1,000 pixels minimum, so don't flirt with the floor.
2. Rename every file with the product keyword, slot, and one descriptor. No IMG files, no FINAL2 files, no spaces in bulk-upload names.
3. Check the main image against compliance: pure white background, product filling about 85% of the frame, no text, badges, or props that aren't included.
4. Order the gallery by shopper question, not by how pretty the shots are. Main, angle, detail, scale, in use, what's in the box. If you haven't mapped that order yet, start with my guide to designing listing images that convert and steal the slot order.
5. Write backend terms last, only words not already visible, spaces not commas, under 250 bytes. Count bytes, not characters.
6. Read the A+ text out loud. If it sounds like a robot wrote it, rewrite it. A+ earns its keep through conversion, and conversion is what feeds everything downstream.

That's the whole system. It's not exciting, and that's the point. The sellers who win on Amazon images aren't the ones with a secret trick. They're the ones whose files are named, whose galleries are complete, and whose backend fields aren't silently over the limit. I know which camp I'd rather be in. After eight years of doing this work, I'm Mirza Rizwan, and I build Amazon listing images at trendingbar.com around exactly this kind of boring, checkable craft, because boring and checkable is what holds up when a listing scales.

If your images are already shot and you're not sure where the gaps are, that's what my free listing audit is for. Send me the listing, and I'll tell you the three fixes I'd make first. No pitch attached, just the checklist applied to your product.