Field-tested guides on alt text, EXIF metadata, GPS geo-tagging, filename structure and image sitemaps — written for the people who actually have to ship the images.

You set a hero, you set an Open Graph tag, and the search result shows your footer logo anyway. How the displayed image gets selected, the four reasons the wrong one wins, and what you can actually influence.
Read the guide
Image SEO advice assumes photographs, but the images that carry the most information on a technical site are screenshots, charts and diagrams. They need different formats, different alt text and a different quality setting, and almost nobody gives them any.

Content refreshes rewrite the text and leave the images untouched, which is how a 2021 hero ends up on a 2026 article. When to replace, when to keep the URL, when to redirect, and the caching trap that serves the old file anyway.

Alt text is content and needs translating. Filenames mostly do not. The image itself sometimes should be a different image entirely. A practical split for sites running two or twenty locales, plus the sitemap rule that trips everyone.

Picking a compression quality is treated as a global setting and it should not be. Where the visible threshold actually sits, why it moves depending on the subject, and a practical file-size budget that survives a real page.

An honest capability map of DALL·E 3 for commercial website imagery. The subjects it renders convincingly, the four failure modes that recur, and the practical rules for getting usable output at volume instead of one lucky image.

You embed EXIF, IPTC and GPS data, upload the file, and the served version has none of it. Which platforms and pipeline steps discard metadata by default, how to test yours end to end in ten minutes, and what to do when the answer is bad.

Every location page needs its own image, and the same stock photo on forty city pages is the cheapest-looking shortcut there is. Real per-image economics for AI generation, stock libraries and a photographer, plus the cost that never appears on the invoice.

Image work is the easiest SEO to do and the hardest to prove. Where the image numbers hide in Search Console, what they do and do not include, and a reporting format that survives a sceptical client.

Shopify renames your files on upload, serves everything from its own CDN, and generates an image sitemap you cannot edit. Here is what that means in practice, which levers you still control, and where the real product-image wins are.

Most keyword research ignores whether a query produces image results at all, which means image effort routinely lands on pages that can never earn image visibility. A method for finding the queries where images are the opportunity.

Google's filename guidance is one sentence. Turning it into a convention that survives twelve people, three CMSes and four years is the real work. A token-based spec, the rules that make it enforceable, and how to retrofit a site that never had one.

Discovery, fetch, deduplication, understanding, then serving into Images, Lens and AI answers. Knowing the pipeline turns most image SEO problems from mysteries into a question of which stage the file stopped at.

Geo-tagging a storefront photo is good practice. Publishing a customer's house at rooftop precision because a phone recorded it is a data-protection incident. Which fields are safe to publish, which are not, and how to decide per image rather than per site.

Individually good images that do not belong together read as worse than mediocre images that do. How to build a reusable style anchor, which variables to lock, and the small set of rules that keeps a large batch visually coherent.

Machine-written alt text is now good enough for most images and wrong for the ones that matter most. Which tools write it, how to tell an acceptable description from a useless one, and the review step that makes bulk generation defensible.

Responsive images are the largest single saving available on most sites, and the sizes attribute is wrong on most of the ones that have implemented them. A practical guide to srcset, sizes, picture and the breakpoint maths.

Image URLs accumulate rankings, links and history the same way pages do — and a redesign discards all of it by default. The mapping, redirect and verification work that keeps image traffic through a rebuild.

Serving images from a CDN, a subdomain or a third-party host changes what gets crawled, what gets indexed and what metadata survives. The hosting decisions that quietly determine whether your images appear in search at all.

Alt text is one requirement among several. Text baked into images, complex charts, missing captions and low-contrast overlays are the accessibility failures that alt text cannot fix — and the ones auditors actually cite.

Which photos to upload to a Business Profile, how many, how often, and which widely repeated tactics do nothing. Plus the rules on what Google removes and why the cover photo is not what you think it is.

Google does not penalise duplicate images, but reusing one photo across forty pages costs you in ways that look like a penalty. What actually happens, when it matters, and how to decide which pages deserve original imagery.

Most AI image generators hand back a file stripped of metadata. Here is which categories of tool write EXIF, IPTC and GPS data on your behalf, what automatic actually means, and how to check the file you were given.

Generated images fail in predictable, specific ways — hands, garbled text, letterboxing, uncanny repetition. A review pass that catches them in seconds each, plus the batch-level checks that matter more than any single image.

A crawl-based audit that finds the six image defects worth fixing, sorts them by the pages that actually matter, and produces a list someone can work through — rather than a 400-row report nobody opens.

Service businesses run out of image ideas by page three. A working inventory of what to show, what to generate, what must be a real photograph, and where the honesty line sits.

The image that wins a scroll-past in a feed and the image that earns traffic from Google Images have almost nothing in common. Sizes, formats, markup, caching, and why one of them should contain text.

AI Overviews and assistant answers do show images — but there is no markup to target them and no report to measure them. What actually makes an image eligible, and which controls genuinely exist.

Which image structured data Google genuinely uses, which is decorative, and how the licence badge in Google Images is earned. Includes the field-level requirements and the errors that silently void the whole block.

Hosted site builders make most image decisions for you. Here is how to find out what your platform is really serving, which levers still exist, and which battles are not worth fighting.

IPTC digital source type, C2PA Content Credentials and invisible watermarks — what each one records, what survives a CMS upload, and what search engines currently do with any of it.

How agencies actually supply images across many client accounts without a photographer, a shared asset mess, or a licensing question they cannot answer. Structure, ownership, approvals and cost.

A practical format decision for real websites: where AVIF earns its encode time, why WebP is still the safe default, and the three cases where plain JPEG is the right answer.

Where stock still wins, where generation wins outright, and the hidden cost neither vendor mentions — the metadata work that decides whether either one ever appears in search.

The difference between a prompt that returns a plastic 3D render and one that returns a believable photograph is structure, not adjectives. A repeatable prompt formula with worked examples.

Nine causes in the order they actually occur, with the check for each. Most 'my images don't rank' problems turn out to be discovery failures, not quality ones.

Most failing LCP scores are an image problem. How to identify the element, why preloading beats compressing, and the lazy-loading mistake that quietly costs a second.

WordPress solves the delivery half of image SEO almost by itself. The half it cannot solve is the reason most WordPress media libraries are full of IMG_4471.jpg with empty alt text.

Visual search matches pixels, not keywords — but the things that get you found are mostly the image SEO fundamentals plus two that are specific to Lens. Where the hype ends.

Product images are where image SEO has a direct revenue line. Structured data requirements, the variant-image mistake that splits your equity, and where generated imagery belongs — and doesn't.

Yes — for the major paid generators, commercial use is permitted by their terms. Copyright ownership is the harder question, and the answer as of July 2026 is more limited than most teams assume.

Hyphens vs underscores, how many words, handling variants and near-duplicates, and a naming convention that stays consistent across hundreds of files and multiple locations.

Most sites do not need an image sitemap. Here is how to tell whether yours is one of the exceptions, the exact XML syntax, and the diagnostics to run before you build anything.

A practical pre-publish checklist for image SEO — filenames, alt text, dimensions, compression, lazy loading, structured data and metadata — with the reasoning behind each item.

Alt text rules that hold up in practice: what to describe, when to leave it empty, how long it should be, and the keyword-stuffing habits that make pages worse for everyone.

What GPS coordinates embedded in an image do and don't achieve for local search, how to add them correctly per location, and the privacy rule that matters more than any ranking claim.

An honest look at EXIF and IPTC image metadata: which fields search engines demonstrably use, which are ignored, what gets stripped in your pipeline, and why it is still worth writing.

A production workflow for image SEO at agency volume: where the hours actually go, what to standardise, what to automate, and an honest cost comparison against stock and freelancers.

DALL-E 3, Midjourney, Stable Diffusion, Firefly and Ideogram compared for content and local SEO production — output quality, licensing, batch workflow and what each one leaves you to do by hand.
SEOpix writes the filename, alt text, EXIF fields and GPS coordinates while it generates the image. Ten images a month are free — no credit card.
Start Generating Free