Image Optimization

Image optimization is the technical work of making images load quickly, display at the right size without shifting the layout, and carry enough text (alt text, file names, captions) for browsers, search engines and AI systems to understand them. It is mainly about performance and accessibility. Getting images found in Google Images, Google Lens and AI answers is a related but separate job, covered under image search optimization.

Images are usually the heaviest part of a page. Google's image SEO guidelines note that images are often the largest contributor to overall page size, which makes pages slow and expensive to load. A large hero image is frequently the Largest Contentful Paint (LCP) element in Core Web Vitals, and images without reserved dimensions cause Cumulative Layout Shift (CLS) as they load. Heavy pages also cost more to crawl, which matters for large catalogs and for AI crawlers that request many URLs in a short burst (crawl budget).

The standard techniques are well established:

• Serve modern formats such as WebP or AVIF, with a fallback inside a picture element where older clients matter. Google Search indexes BMP, GIF, JPEG, PNG, WebP, SVG and AVIF.
• Use srcset and sizes so each device downloads an appropriately sized file, and always keep a plain src attribute as a fallback, which Google recommends because some crawlers ignore srcset.
• Set width and height attributes (or a CSS aspect ratio) so the browser reserves space and the layout does not jump.
• Lazy-load images below the fold with loading='lazy', but never the LCP image; giving that image fetchpriority='high' helps it load first.
• Put content images in img elements rather than CSS backgrounds, because Google does not index CSS images.
• Reference a shared image with the same URL everywhere it appears, so it can be cached and crawled once.
• Compress to the lowest quality setting that still looks sharp, and serve images from a CDN close to users.

The text attached to an image is the other half of the job. Alt text is the most important image metadata for accessibility, and Google says it combines alt text with computer vision and page content to understand what an image shows. Write it as a short description of what the image shows and why it is on the page, not as a list of keywords. A descriptive file name such as blue-ceramic-mug-350ml.webp gives a small additional hint. Captions and nearby copy matter most for charts, diagrams and screenshots, because the data in them is otherwise locked inside pixels.

This text matters more in AI search than it did in classic SEO. Answer engines like ChatGPT, Perplexity and Google AI Overviews mostly retrieve and quote text. A chart with no caption or summary contributes nothing to a passage a retriever can quote, even if a multimodal model could read the chart when shown it directly. Restating the key number from an image in the surrounding text makes it citable.

To measure progress, check LCP and CLS for real users in PageSpeed Insights and the Core Web Vitals report in Search Console, review total image bytes per template, and audit alt text coverage in a crawler during a technical SEO audit. Improvements usually show up first in page speed metrics and only indirectly in rankings or citations.