Search Console's performance report adds a multimodal search filter
Google added a multimodal search filter to the Search Console performance report on 24 September 2026, splitting camera, upload and right-click starts from typed searches. Here is where the filter lives, what the four entry points mean, and what it does not tell you.
RULE CHANGE · 2026-09-24 · Google Search Console
Google added a multimodal search filter to the Search Console performance report on 24 September 2026. The filter separates visits that begin with a camera, an uploaded image, or a right-click on a picture from visits that begin with typed words. For a product site it is the first place you can see how much discovery starts at your images instead of your text.
What the Search Console performance report now shows
The new value sits inside the existing Search type filter at the top of the performance report. Google's announcement of 24 September 2026 names four entry points it counts, and it said the data was rolling out globally from that day. You only see rows if your site already receives traffic from these starts, so an empty filter is normal for a text-only site.
| Entry point | Where the search starts |
|---|---|
| Lens | A phone camera pointed at a real object |
| Circle to Search | A finger circling part of a screen on Android |
| Image upload | A picture uploaded into Google Search |
| Chrome "Search this image" | A right-click on a picture in the browser |
All four are the same shape of interaction: the person did not type words to find you. That is why the filter matters more than it first looks. A typed search is a phrase you can plan around. An image search puts your product photo, your packaging, or a screenshot of your interface to work as the match itself.
Why this traffic was invisible before
Until this change, those four entry points were folded into the general web numbers, so a product that got found through a photograph looked the same as one found through a phrase. Google says the same multimodal data also appears in the report for Generative AI features, not only in the Search results report. That matters because image starts and AI answers now share a supply chain: the picture that wins a Lens match is often the same asset an AI answer can quote.
The filter is a reporting change, not a ranking change. Google did not announce that visual content now ranks differently. What it did is hand you a measurement you did not have, and measurement is the part most small teams skip. If you never look, you cannot tell whether the work you put into product images is pulling anyone in.
What to do with the filter
Three things, in this order. None of them requires new tooling; the data is already in the account you use for the rest of Search Console.
- Open the performance report and switch the Search type filter to Multimodal search. Your done mark is seeing clicks and impressions for at least one date, or confirming the filter is genuinely empty for your site.
- Compare its CTR against the web tab. If multimodal impressions are high and clicks are low, people see the image but do not choose you. That points at the image or the surrounding snippet, not at your copy.
- Fix the images that carry the traffic. Descriptive alt text and a clear main subject help image search, and Google's image guidance says to make the subject of the page and the image match. A cropped or cluttered photo loses the match before text can help.
Step two is the one people skip, and it is where the filter earns its place. A high-impression, low-click multimodal row is a picture problem, and it is invisible in the site-wide average.
What it cannot tell you
We have not measured multimodal impressions on our own sites, so we cannot tell you what share of a typical product site's traffic arrives this way. Treat any percentage you see quoted elsewhere with care until it comes with a sample and a date.
The filter also stays coarse. It reports clicks, impressions, CTR, and position by page and country, the same dimensions as the rest of the performance report. It will not name the single image that won a placement, and it does not separate Lens from Circle to Search from an upload. If your product is image-heavy, you get a direction, not a diagnosis.
Common questions
Is multimodal search the same as the Images report?
No. The Images report is about your pages appearing in Google Images. Multimodal search is about image-initiated searches that land on your pages, including Lens and camera starts, and it lives inside the performance report as a filter. A page can be strong in one and weak in the other.
Why is my multimodal filter empty?
Google only shows these rows when your site already receives traffic from these entry points. An empty filter usually means none of your pages are being matched by camera, upload, or right-click yet. It is not an error and it is not something you can force on.
Does the Generative AI report show it too?
Yes. The 24 September 2026 announcement says the same multimodal search data appears in the report for Generative AI features as well as the Search results report. Check both before you conclude you have no image-driven discovery.
How did you verify this?
We read Google's own announcement and its image documentation on 6 October 2026 and quoted the entry points and the rollout wording directly. We have not run a first-party measurement of multimodal traffic, and we say so above rather than estimate a number.
The takeaway is narrow and firm: one new filter separates image-started discovery from typed discovery, and almost no small site has looked at it yet. If you want to connect that signal back to the product work behind it, that is what QueryWin works on; for reading the rest of the report, see what the Generative AI report can and cannot tell you, and to pull the same numbers programmatically, pulling Search Console data through the API.
