Google is adding a new multi-modal filter to the Performance report in Search Console. Multimodal searches are searches in which a user used an image as part of the search, including Google Lens and Circle to Search.
Google’s rollout of the change to their properties is global, and it adds a significant caveat in its rollout. Query data isn’t available for them.
Google announced the change to the Search Central blog and is rolling out globally. Search Engine Journal wrote about the change when they published on September 24, 2026, and the rollout was just getting underway.
What The Filter Covers
Google states that the new multimodal data covers these image-based searches:
- Searches made with Google Lens
- Circle to Search on Android
- Image uploads to Google Search
- The Chrome right click option to search an image
The new filter provides a view of the traffic in its own area, so that it’s separated from the text based results that make up the rest of the Web numbers.
Where To Find It
The option is located in the Web category in the Performance report. Google has updated the help page for the Web view to now split it into two parts. Text based traffic covers queries done in the standard search bar. The multimodal section covers web results where an image was part of the search.
Lead for Google Lens, Harsh Kharbanda and Lead for Search Console Moshe Samet described the intent of the change in the announcement. They’re intended to offer a view of how the content is surfaced under searches done through images, such as a smartphone camera.
Metrics will only appear for sites that do see some traffic of this nature. Sites not seeing any may not be able to see much in this new category, and that’s not a signal of error.
Why There’s No Query Data
Anyone working with Search Console knows well the Queries tab is a place to spend a lot of time. You don’t have one here. You can’t see specific queries behind this category of image searches. When selecting the multimodal search type, the Queries tab is gone.
Google’s help page for dimensions provides an explanation. As multimodal searches are largely image driven and not text driven, there are not specific text queries for this traffic in the way one would expect. A photo of a jacket will have no accompanying keywords to report as part of a search, and so Search Console has no way to report that as a query might be done against a text search.
The help page does not list any other dimension as unavailable for multimodal searches. Pages, countries, and devices are the views available and, in Google’s eyes, the ones they suggest you use as a reference for comparisons against text-based traffic.
Getting The Data Out
Google’s post draws attention to the Export button as a way to get the numbers out, providing a good enough guide to manual reporting needs, whether building a simple spreadsheet or sending a quick snapshot to a client.
The API side is not specified clearly. The Search Analytics API reference document (last updated August 11) lists web, image, video, news, Discover, and Google News as report types. It does not note a separate value for the type parameter for multimodal. Google’s announcement does not indicate whether that will change either.
Those that pull Search Console data into dashboards through the API may want to test the waters early, as assuming multimodal numbers will flow into an existing dashboard system may miss an update.
A Connection To The Generative AI Report
Google also states that multimodal reporting will be available to the generative AI performance report. That report reached all domains on August 31 and counts impressions rather than clicks
One will have to take care when making comparisons between the two. The Web filter can have clicks from image-based searches, while the generative AI report uses only impressions as a measurement. The two are not easily compatible as they are presented.
Why It Matters For Sites With Product Photo
Perhaps the most-clear example of the use of the multi-modal filter across image types is in the world of ecommerce. Any sites that contain product photographs can leverage this view of their traffic to provide a page level look at the pages that receive traffic from image-based searches, like a phone photo or a screenshot. It provides an opportunity to compare across the pages, by country and device, the traffic coming from image-based queries against that of text-based queries.
Think about the way real world usage could happen. A user seeing sneakers on the street snaps a photo with it, and does a search. Another takes a screenshot from a post showcasing a chair and runs it through Lens. They both want to find something that’s in a picture, but they may not be sure how to phrase it in plain text. If either of those result in a click on your product page, you can see the visit in the new area.
While it’s a small change on paper, it can provide site owners a new view, and it’s fairly specific enough that it’s worth examining if your site catalog makes heavy use of photography.
The Limits Of A Page Level View
Without queries, you’re left with pages. A product page showing multimodal clicks tells you it was clicked from a search where an image was used. It doesn’t tell you which image the user used, what they were trying to find, or how they would have described in words to get what they wanted.
This leaves some guesswork on the table. If one product page pulls a lot of multimodal clicks, look at what it has in common with the other pages that see such, whether it’s product category, photography style, or visibility in the real world. This is your own guesswork; there’s nothing for the report to tell you.
What Google Hasn’t Clarified
A few questions on this rollout come out as unclear.
Google’s post does not specify whether the multimodal data will include history from before the time of this rollout, and whether the new split will impact existing Web totals.
Both of these are important in the event of you planning to run trends based on this new view. Without history being included, any comparison prior to this time is going to have to start over from zero. And should the split change to reflect what’s seen under the Web view, existing totals reported to clients and managers for Web searches could be shifting.
Practical Steps
Some of these steps are worthy for you to take as soon as the option is available to you in your property.
- Check availability. This rollout is global, but it’s rolling out gradually. If you don’t yet see the multimodal option, it’s likely not yet available to you.
- Compare against a full day of Web totals. One you see the option available, examine the new split alongside the totals for a full day of Web data. This can give you a good sense of the scale of the numbers.
- Review the pages that appear. Look on whether the product details, images, and page content are current on whichever pages that will collect multimodal clicks. A visitor coming in with a photo will want to find the item.
- Test your API setup. If you’re using API pulls, confirm what will be available before developing reports.
Looking Ahead
The rollout is still early on, and there will be some properties that may not see this option for a while. Keep tabs on the coming days, and consider what the first numbers you see are in context of an ongoing picture.
FAQs
Q1: What is the new multimodal filter in Search Console?
It is a new filter in the Performance report for image-based searches as part of the search.
Q2: Which searches does the filter include?
Google says it includes searches made with Google Lens, Circle to Search on Android, image uploads to Google Search and the Chrome right click option to search an image.
Q3: Is the multimodal filter available everywhere?
Google says it is rolling out the change globally. Since the rollout is gradual, some properties may not see the option yet.
Q4: Why is there no query data for multimodal traffic?
Google says these searches are largely image driven, so specific text query data isn’t available. The Queries tab isn’t offered when you select the multimodal search type.
Q5: What data can I still see for multimodal searches?
The help page does not list any other dimension as unavailable for multimodal. You can compare these pages against text-based traffic by country and device.
Q6: Who sees multimodal metrics?
Only sites that get traffic from image-based searches, Google’s post states. If your site has none, there may be little to see.
Q7: Can I export data for multimodal searches?
Yes and no. Google’s guidance points to the Export button in the report. The Search Analytics API reference document (last updated August 11) does not note a separate value for the type parameter for multimodal.
Q8: How is the new multimodal traffic view related to the generative AI performance report?
Google says that multimodal reporting will also be rolled into the generative AI performance report. It counts impressions, and not clicks, so it’s not a fit against the Web filter.
Q9: Does the filter show which image someone searched with?
No. A page with multimodal clicks only tells you an image was used as part of the search, but not which image it was.
Q10: Does the multimodal data include history from before the rollout?
Google’s post does not indicate whether this is true. It also doesn’t specify whether the split will change existing Web totals.
