Image SEO & Visual Search

Brass-SEO treats this as the authoritative checklist for image discoverability. Google's documentation states it derives image understanding from five signals — filename, alt text, surrounding page text and captions, page title and meta tags, and structured data — and lists supported formats as BMP, GIF, JPEG, PNG, WebP, SVG, and AVIF. It also states directly that "Google doesn't index CSS images," meaning a background image set via CSS is invisible to Google Images no matter how well it's described elsewhere on the page; only images placed in an actual img element are eligible.

Examines:

Google's first-party guide to how Google Images discovers, crawls, and ranks images, and how to structure a page to optimize for image search.

Brass-SEO draws on:

The five documented ranking signals and the CSS-image exclusion — the exact checklist Brass-SEO's page audit checks an image against.

Primary source

developers.google.com/search/docs/appearance/google-images

Brass-SEO draws on this for the specific numeric threshold article images need to hit. Google's structured-data documentation recommends "multiple high-resolution images... with the following aspect ratios: 16x9, 4x3, and 1x1" at a minimum of 50,000 total pixels, meaning width multiplied by height. That's a hard, checkable production spec, not a rule of thumb — an image below that pixel count can carry perfect alt text and still fail to qualify for the enhanced visual placement.

Examines:

Google's official structured-data specification for Article rich results, including the required properties and image size constraints for eligibility.

Brass-SEO draws on:

The 50,000-pixel minimum and the 16x9/4x3/1x1 aspect ratio set — cited when Brass-SEO flags a blog post's featured image as too small for rich-result eligibility.

Primary source

developers.google.com/search/docs/appearance/structured-data/article

Brass-SEO uses this as the vocabulary reference underlying every image-related schema property it recommends. ImageObject sits in the type hierarchy Thing > CreativeWork > MediaObject > ImageObject, inheriting contentUrl, encodingFormat, and width/height from MediaObject and author, datePublished, and license from CreativeWork, with caption and representativeOfPage as ImageObject-specific properties. Understanding the inheritance chain explains why a property like license is valid on an image's markup even though it's not defined directly on ImageObject itself.

Examines:

The canonical schema.org structured-data type definition for images, including its property inheritance from MediaObject and CreativeWork.

Brass-SEO draws on:

The MediaObject/CreativeWork inheritance chain — cited when Brass-SEO explains which properties are valid on an ImageObject markup and where each one actually comes from.

Primary source

schema.org/ImageObject

Brass-SEO monitors this for real-world image format and loading benchmarks. The 2024 Web Almanac found WebP reached 12% of all mobile image requests (up 3 percentage points since 2022) and AVIF reached 1.0% — nearly four times its 2022 share — while native lazy-loading is used on 33% of sites, up from roughly 25% two years earlier. It also found 9.5% of images responsible for a page's Largest Contentful Paint were incorrectly marked lazy-loaded, a direct performance regression on the exact image the browser needs to paint first.

Examines:

Annual independent measurement of image format adoption, sizing, and lazy-loading practices across millions of real pages.

Brass-SEO draws on:

The 12%/1.0% WebP/AVIF adoption figures and the 9.5% incorrectly-lazy-loaded LCP image stat — used to benchmark a site's format and loading choices and flag a specific, common mistake.

Primary source

almanac.httparchive.org/en/2024/media

Brass-SEO cites this to establish that image optimization is usually the primary Core Web Vitals lever, not a secondary one. The 2024 Web Almanac found images are the Largest Contentful Paint element on 83.3% of desktop pages and 73.3% of mobile pages. For the large majority of real sites, the single biggest factor in a page's LCP score is an image — format, compression, and load timing decisions on that one element move the metric more than almost anything else on the page.

Examines:

Annual independent measurement of which element type is responsible for Largest Contentful Paint across the crawled web.

Brass-SEO draws on:

The 83.3%-desktop / 73.3%-mobile image-as-LCP-element figures — cited when Brass-SEO prioritizes image fixes over other performance work in a page audit.

Primary source

almanac.httparchive.org/en/2024/performance

Brass-SEO draws on this to identify exactly which page elements can even become the LCP bottleneck. Google's documentation limits LCP element candidates to img elements, the image inside an svg, video elements (via their poster frame), elements with a CSS background-image set through url(), and block-level elements containing text nodes. A decorative CSS background image and a hero img tag are both in scope for LCP optimization; most other image uses on a page are not, which narrows exactly where to look first.

Examines:

Google's official technical specification of the LCP Core Web Vital metric, including the exact DOM element types eligible to be measured as the LCP element.

Brass-SEO draws on:

The specific list of LCP-eligible element types — cited when Brass-SEO identifies which image on a page is actually driving its LCP score versus which ones are irrelevant to it.

Primary source

web.dev/articles/lcp

Brass-SEO monitors this as first-party evidence of visual search's real scale. Google's own announcement states Google Lens "is now used for nearly 20 billion visual searches every month." That's a company disclosure, not a marketing estimate, and it establishes that optimizing images for camera- and screenshot-based discovery is a real-volume channel, not a speculative one worth deprioritizing behind classic text search.

Examines:

Google's first-party blog disclosure of Google Lens visual search usage scale, alongside new search capabilities announced in the same post.

Brass-SEO draws on:

The nearly-20-billion-monthly-searches figure — cited when Brass-SEO makes the case for treating image optimization as a search-visibility channel in its own right.

Primary source

blog.google — Lens October 2024 updates