Google has added a way to see how your pages perform when people search with a picture instead of words. This guide is for bloggers, marketers, agency owners and small business owners who already use Search Console and want to know what the new multimodal data means, where to find it and what to do about it.
What Google announced
On 24 September 2026, Google announced web multimodal Search performance reporting in Search Console. It is rolling out globally from that date. The new data appears in the Performance report for Search results and in the Generative AI features report, and it shows how your content is surfaced when users search using images. Google says metrics appear only if your site actually receives traffic from these kinds of queries, so an empty view is not necessarily a fault.
The announcement does not say how far back the data goes or whether it changes how pages rank. Treat it as a measurement tool, not a ranking signal.
What counts as multimodal search
“Multimodal” here means a search that starts from an image rather than typed text. Google’s announcement and the Search Console help page list these entry points:
- Google Lens, for example pointing a smartphone camera at an object.
- Circle to Search on Android.
- Image uploads to Google Search.
- Chrome’s right-click “Search this image” option.
This is different from the existing Image search type, which covers traffic from the Google Images tab. Multimodal traffic starts with a picture the searcher already has, and the results can lead to ordinary web pages.
Where to find the data in Search Console
The Performance report search type filter
Open Search Console, choose your property and open the Performance report for Search results. Use the search type filter at the top of the report. According to Google’s help page, the options are:
| Search type | What it covers |
|---|---|
| Web: text-based | Text queries typed into the standard Google search bar |
| Web: multimodal | Image-based searches, including Lens, Circle to Search, uploaded images and Chrome image search |
| Image | Traffic from the Google Images tab |
| Video | Traffic from the Google Videos tab |
| News | Traffic from the Google News tab and news results |
Pick “Web: multimodal” and the chart and table update to show only that traffic. If you are new to the report itself, our guide to finding quick SEO wins in the Search Console Performance report covers the basics of clicks, impressions, CTR and position.
The Generative AI features report
The Generative AI features report tracks how your site appears in AI Overviews and AI Mode. Google’s help page says it covers only text-based and multimodal web searches, so multimodal activity can show up here too. It shows impression changes over time, your pages with the most and fewest impressions, and where impressions come from by device or country. The help page says it rolled out to all sites worldwide as of 31 August 2026.
Limits to know about, all from the help page: the report may not appear if your site has not had enough impressions in these features; sites excluded from Search generative AI features show no data; Search Labs experiments are not included; the standard 1,000-row limit applies; the newest data may be preliminary; and dates use Pacific Time. For background on controlling how your content is used in these features, see our post on what you can control in Google AI Overviews and AI Mode.
How to analyse and export the data
The help page says the Performance table can be grouped by Queries, Pages, Countries, Devices, Search appearance or Dates. A simple routine:
- Set the search type to “Web: multimodal” and note the date range. The default view covers the past three months, and you can change it.
- Switch to the Pages grouping and sort by impressions. These are the pages Google is surfacing for image-based searches.
- Switch to Devices and Countries to see whether this traffic is mostly mobile, which is plausible for camera-based searching but is something to confirm in your own data rather than assume.
- Compare the same pages under “Web: text-based” and “Image” to see which entry points matter for each page.
- Click Export to download the chart and table data for a spreadsheet or a reporting tool such as Looker Studio.
One export quirk from the help page: values shown as “~” or “-” download as zeros. Do not read those zeros as real zero traffic.
Because the feature is new, expect small numbers at first and a short history. Look for patterns across weeks rather than reacting to a single day, and remember that the newest data can be preliminary.
Practical image SEO steps
Google has not published a separate checklist for multimodal search, and it does not say that any of these steps raises your position in those results. What it does publish is its image SEO best practices for Google Images. They are sensible foundations for any image-based discovery, so use them as the starting point.
Make images easy to find and index
- Use HTML img elements. Google finds images in the src attribute, including inside a picture element. It does not index CSS background images, so important product or editorial images should not live only in CSS.
- Use a supported format. Google lists BMP, GIF, JPEG, PNG, WebP, SVG and AVIF, and says the file extension should match the real file type.
- Submit an image sitemap for images Google might not otherwise discover. Google notes image sitemaps can list images hosted on other domains, such as a CDN.
- Reference images consistently. Using the same URL for an image across pages lets Google cache and reuse it.
Give Google context about each image
- Descriptive file names. Google’s own example is that “my-new-black-kitten.jpg” is better than “IMG00023.JPG”. Avoid generic names such as “image1.jpg”.
- Useful alt text. Google calls alt text the most important metadata for images. Describe what the image shows in the context of the page and do not stuff keywords into it.
- Place images near relevant text. Put an image on a page about its subject, close to the text that discusses it. Captions and page titles also help Google understand the subject.
- Name a preferred image. Google says you can do this with schema.org markup or the og:image meta tag, and recommends one that is relevant, representative, high resolution and not an extreme aspect ratio. It advises against generic images such as logos and images that contain text.
Keep the landing page itself in good shape
Google’s guidance also covers the page around the image: write clear titles and descriptions, add structured data where it applies, keep images sharp, and watch page weight, since images are often the largest part of a page. Our walkthrough on adding and testing structured data is a good companion if you have not used markup before. Google also suggests testing accessibility and checking speed with PageSpeed Insights.
Common mistakes
- Leaving camera-style file names such as IMG_0042.jpg on every upload.
- Repeating the same keyword phrase in every alt attribute instead of describing each image.
- Using stock images or logos as the preferred image when a real, representative one exists.
- Treating one week of small multimodal numbers as proof that an optimisation worked or failed.
- Blocking images from Google by accident. Google notes you can block images entirely, which would remove them from results.
Key takeaways
- Search Console now reports “Web: multimodal” traffic, covering Lens, Circle to Search, image uploads and Chrome’s “Search this image”.
- Find it through the search type filter in the Performance report; multimodal activity can also appear in the Generative AI features report.
- Group by Pages first, then compare with text-based and Image traffic, and export for deeper analysis.
- Apply Google’s published image SEO basics: descriptive file names, honest alt text, images near relevant text, an image sitemap and a healthy landing page.
- Google does not state that these steps improve multimodal rankings, so use them as good practice and measure the result.
FAQ
Why can’t I see the multimodal option or any data?
Google says the rollout began globally on 24 September 2026 and that metrics appear only if your site receives traffic from these queries. The Generative AI features report may also be missing if your site has had too few impressions in those features.
Is multimodal search the same as Google Images?
No. In the Performance report they are separate search types. Image covers traffic from the Google Images tab, while Web: multimodal covers searches started from an image through tools such as Lens or Circle to Search.
Will optimising images improve my multimodal rankings?
Google has not said so. Its image best practices help Google discover and understand images, which is worthwhile, but there is no stated guarantee of more multimodal traffic.
How far back does the multimodal data go?
The announcement does not state a history length, so this is unconfirmed. The Performance report’s default view covers the past three months, but check the date range available on your own property.