SEO-37
Image in an outdated format without a modern alternative
What the check measures
On every live page we look for images in an outdated format (jpg, jpeg, png, gif, bmp) that also have no modern alternative offered, that is, webp or avif in a <picture><source>. If at least one is found, the finding is raised with a count and an example.
An empty <img src=""> does not count (it displays nothing), SVG does not count, nor do images without a recognisable extension: for a vector or for data embedded in the page, a webp alternative makes no sense.
What the check does not do, and here it matters especially:
- It does not measure file size. An optimised 40 kB JPEG gets the finding while a badly saved 2 MB WebP passes. The check looks at the format, not the weight.
- It cannot see server-side conversion. Sites and CDNs that send
webpunder the original.jpgextension based on theAcceptheader get the finding undeservedly, and that is a common setup. - It judges neither quality nor dimensions.
- It cannot see images inserted via CSS.
So it is a coarse sieve on format, not a performance measurement. The finding is page-level; severity is warning.
How strong the evidence is
We recommend it because it does no harm or has some other benefit, but we promise nothing about whether it makes language models cite you. Nobody has demonstrated that yet.
For visibility in AI answers we have no documented effect: images in any format change nothing about what a model reads off the page.
The connection to classic search is indirect, and it is fair to describe it as indirect: modern formats are smaller, smaller images download faster, and the loading speed of the hero image is measured by SEO-20 (LCP), a metric Google's own documentation names as used by its ranking systems. The chain exists, but it is a chain of three links, not a direct relationship.
And honestly about what the finding can and cannot do: because it does not measure file size, it cannot tell a site with well-optimised images (just in an older format) from one serving unresized camera photographs. The first gets our finding and need not act; the second has a problem this check does not measure at all.
The practical conclusion: before converting anything because of this finding, look at what your images actually weigh. If the largest image on the page is 80 kB, converting to WebP gains you nothing. If it is 3 MB, you have a problem, but format will not fix it, resizing will.
How to fix it
Measure first, convert second. The network tab in your browser's developer tools lists image sizes sorted, and you will usually find two files are the problem, not fifty.
If conversion does make sense, the right markup offers the modern format and keeps a fallback for older browsers:
<picture>
<source srcset="pump.avif" type="image/avif">
<source srcset="pump.webp" type="image/webp">
<img src="pump.jpg" alt="Heat pump on the north wall of a house"
width="1200" height="800">
</picture>- The browser picks the first format it understands: order from the most efficient down.
alt,widthandheightbelong on<img>, not on<source>. Without them you manufacture aSEO-10and aSEO-21.- Do not convert by hand. Let your CMS or an image service do it on upload; a hand-converted library goes stale with the first new image.
- Automatic conversion at a CDN is a full solution: it sends
webpto whoever understands it, under the original URL. Our finding will persist for it (see above), and that is our limitation, not your problem. - Replace animated GIFs with video. An animated GIF is often ten times larger than the same thing as
mp4orwebm; this is where the saving is biggest of all.
What the report says about it
Finding description
We found [imagesPhrase] on the page in an older format (JPG/
Recommendation
Add a `<picture>` with a `<source type=
Sources
- Google Search Central: Google image SEO best practices (accessed 2026-09-12)
- web.dev: Largest Contentful Paint (LCP) (accessed 2026-09-12)
Text verified 2026-09-19