10 Image SEO Optimization Techniques

Images can bring traffic from Google Images, and they are also among the elements that affect page speed most. Google's own documentation says images are often the largest contributor to overall page size. Image SEO is therefore about both visibility and performance.
This article covers 10 techniques based on Google's image SEO best practices documentation. The speed side is covered in detail in the Core Web Vitals article, and the wider technical foundation in the technical SEO guide.
Short answer: Add images with the HTML img element, give them descriptive alt text and meaningful file names, and serve the right size in efficient formats such as WebP or AVIF with srcset. Lazy-load images below the fold but never the LCP image. Image sitemaps and structured data help Google find images and display them correctly.
How does Google find and understand images?
Google first finds an image in the page HTML, then works out what it shows from the alt text, file name, surrounding text, page title and structured data. That is why the first techniques below deal with discovery and meaning, and the later ones with speed and extra signals.
- Discovery: Techniques 1, 7 and 9 (img element, lazy loading, sitemap)
- Meaning: Techniques 2, 3 and 8 (alt text, file name, context)
- Speed: Techniques 4, 5 and 6 (format, size, srcset)
- Extra signals: Technique 10 (structured data and image metadata)
Three core techniques for discovery and meaning
1. Add images with the HTML img element
Google's documentation states plainly that it does not index CSS images. Product photos, infographics and any image you want found in search should use an img element with a src attribute. Keep CSS backgrounds for decorative images only.
2. Write descriptive alt text
Google calls alt text the most important metadata for an image; it helps Google understand the subject and is essential for screen reader users. Describe the image briefly and in context, and do not stuff keywords. I show the role of alt text for portfolio driven businesses in SEO tactics for architecture firms.
- Brown leather backpack with a front pocket
- Entrance of our Kadikoy, Istanbul branch
- Chart showing the 2.5 second LCP threshold
- image1
- bag bag cheap bag leather bag best bag
- Empty alt text, or alt text identical to the file name
3. Use meaningful file names
Google says a file name can give clues about an image's subject and recommends avoiding generic names like IMG00023.JPG. Choose short, hyphenated names such as "brown-leather-backpack.webp". For Turkish sites I prefer writing Turkish characters as ASCII in file names, because some CMSs and CDNs encode them inconsistently.
Three techniques for speed and size
4. Choose efficient formats like WebP or AVIF
Google Search supports BMP, GIF, JPEG, PNG, WebP, SVG and AVIF. For photos, WebP or AVIF usually give a smaller file at the same visual quality, and the picture element lets you offer a fallback for older browsers. Make sure the file extension matches the actual format.
| Format | Best for | Note |
|---|---|---|
| AVIF | Photos and large images | A picture fallback is advisable |
| WebP | Photos, product images, transparent images | Broad browser support |
| JPEG | Photos as a fallback format | No transparency |
| PNG | Screenshots, sharp-edged graphics | Large files for photos |
| SVG | Logos, icons, simple illustrations | Vector, sharp at any size |
5. Serve images at display size and compress them
Uploading a 4000-pixel photo into a slot displayed at 800 pixels is the speed mistake I see most. Scale images to their display size, compress them at a sensible quality setting and give the img element width and height to prevent layout shift (CLS) while loading. Fast but sharp images are what Google recommends too.
6. Use responsive images with srcset and sizes
Google supports the srcset and sizes attributes, or the picture element, for responsive images, and asks you to always provide a fallback URL in src. That way phones download a small file and desktops a large one.
<img src="/img/leather-backpack-800.webp"
srcset="/img/leather-backpack-400.webp 400w,
/img/leather-backpack-800.webp 800w,
/img/leather-backpack-1600.webp 1600w"
sizes="(max-width: 600px) 100vw, 800px"
width="800" height="600"
alt="Brown leather backpack with a front pocket"
loading="lazy">Two techniques for lazy loading and context
7. Use lazy loading, except for the LCP image
For images below the fold, the browser's built-in loading="lazy" attribute is enough, and because the src is in the HTML Google still sees the image. The large image in the first screen is often the LCP element, and lazy-loading it delays LCP. Load it normally and prioritise it with fetchpriority="high". On mobile, do not hide primary content behind lazy loading that waits for user interaction either; see mobile SEO.
8. Place images next to relevant text
Google recommends using images on pages about the same topic and near related text, and the page title and description may be used in image results. That is why supporting product images with a proper product description matters; see writing ecommerce product descriptions. Reusing the same image URL across pages also makes caching easier for Google. I explain how original photos and descriptions set listing pages apart in SEO tactics for real estate websites.
Two techniques for sitemaps and structured data
9. Add an image sitemap
Google recommends image sitemaps for images it might not otherwise find, and images hosted on a CDN can be listed too. Since 2022 Google supports only the image:image and image:loc tags, so there is no need for the old title or caption tags.
<url>
<loc>https://example.com/leather-backpack</loc>
<image:image>
<image:loc>https://cdn.example.com/img/leather-backpack-1600.webp</image:loc>
</image:image>
</url>10. Add structured data and image metadata
In markup such as Product, Article or Recipe, the image property is required for rich result eligibility. If you license images, ImageObject lets you provide license, acquireLicensePage, creator, creditText and copyrightNotice; Google says markup that includes license can make an image eligible for the Licensable badge, though display is not assured. See the schema and structured data guide for types and the schema generator for code.
Which sites should prioritise image SEO?
Traffic from image search is not equally valuable in every sector. Ecommerce, furniture, fashion, travel and recipe sites, where appearance drives the decision, benefit most. On these sites, plan image optimisation together with the catalogue structure in the ecommerce SEO guide; if you also use video, the video SEO article complements this one.
For stores with thousands of product images I handle this at template level; see ecommerce SEO consulting and, for the speed side, performance and growth, or simply book an intro call.
Frequently Asked Questions
Does every image need alt text?
Every image that carries information does. For purely decorative images, an empty alt attribute (alt="") is considered correct for accessibility; screen readers skip those images.
Does the image title attribute help SEO?
Google's image documentation highlights alt text as the most important metadata. The title attribute shows a tooltip on hover and is not a substitute for alt text.
Can I stop my images from being shown inline in Google results?
Google's documentation describes responding with 200 or 204 and no content when the request comes from a Google domain if you do not want inline display. Removing an image from search entirely uses separate methods such as robots.txt or noindex.
Is it a problem to use AI-generated images?
Google's image documentation has no rule against AI images. You can state how an image was created with the IPTC Digital Source Type field in image metadata; on product pages, images that do not reflect the real product damage trust.
Does using a CDN affect image SEO?
Not negatively; images can be hosted on another domain and listed in an image sitemap. Verifying the CDN domain in Search Console helps you see crawl errors.

8+ years of hands-on SEO and GEO experience across 150+ projects in e-commerce, SaaS, healthcare and local business sectors. Writes based on real client data, not theory.





