Google Search Console now counts the searches people make with a camera. Since September 24, 2026, a new multimodal filter in the Performance report shows traffic from Google Lens, Circle to Search, image uploads, and Chrome's right-click image search — searches where a photo, not a typed query, started the search. If your pages appear in those results, you can finally see it.
There is one big blind spot: the report shows no queries. Because the search started with an image rather than text, Google has no search terms to show you — only the pages that appeared, the countries and devices, and the dates. This article explains how to open the report, how to read what it shows, what it cannot tell you, and which image fixes are actually worth doing.
What Google Announced on September 24, 2026
Google announced web multimodal search performance reporting on the Search Central blog on September 24, 2026, with a global rollout starting the same day. The new reporting appears in two places: the Performance report for Search results and the report for Generative AI features. It is a new search-type filter, not a new report: under the Web search type, traffic now splits into "Text-based" and "Multimodal".
Multimodal means an image formed part of the search input. Google lists four routes into it: Google Lens, Circle to Search on Android, image uploads to Google Search, and Chrome's right-click "Search this image" option. In each case, the person pointed a camera at something — a product on a shelf, a plant, a landmark, a piece of clothing — and Google matched that photo to web pages.
The important distinction is between how someone searches and what kind of result they see. A multimodal search does not have to land on an image or a product page. The destination can be an ordinary article or guide that Google decided answers what the photo shows. That is why this report matters beyond image-heavy sites: any page can appear for a photo search.
How to Open the Multimodal Report
Open your property in Search Console, go to Performance, then Search results. Click the search-type filter at the top — it defaults to Web — and under Web, pick "Multimodal". The table below it switches to multimodal traffic only, with the familiar metrics: clicks, impressions, CTR, and average position. You can export the data like any other performance view.
You will only see numbers if your site actually receives traffic from these searches. The filter may not be visible in every property yet, since the rollout is gradual, and a property with no multimodal traffic shows nothing at all. An empty report is not an error — it means nobody has photographed their way to your pages yet, or the rollout has not reached your property.
The same filter also appears in the Generative AI features report, where the label is "Web: multimodal" and the report counts impressions. The two views cover the same underlying traffic in different contexts, so do not add their numbers together — you would be counting the same searches twice.
The Missing Queries: What the Report Cannot Tell You
The defining limitation of the multimodal report is that there is no Queries dimension. Google's own documentation states it plainly: because multimodal searches mostly use images rather than text, specific text query data is not available, and the queries dimension is not available when this search type is selected. You get pages, countries, devices, and dates — never what the person was looking at.
That changes how you read the report. With ordinary search traffic, the query is the starting point: you see the words, then the page that ranked for them. With multimodal traffic, the page is the query. If a product page appears for multimodal searches, someone photographed something that Google decided your page answers — but you cannot see the photo, the object, or the intent behind it.
John Mueller clarified one more thing that matters for anyone comparing numbers: this data was not previously included in the web search counts. It is genuinely new data, not a re-slicing of old traffic. That means there is no baseline to compare against — a line that climbs through October may only be the rollout reaching your property, not your site gaining visual-search traffic.
| What you get | Text search report | Multimodal report |
|---|---|---|
| Queries | Yes — full query list | No — not available |
| Pages | Yes | Yes |
| Clicks and impressions | Yes | Yes |
| CTR and position | Yes | Yes |
| Countries and devices | Yes | Yes |
| Historical baseline | Yes — 16 months | No — data starts Sep 24, 2026 |
What the First Real Numbers Look Like
Early numbers suggest multimodal traffic is real but small for most sites. In a published analysis of his own site, Christian Ott reported 407 multimodal impressions and two clicks for September 10 to 21, 2026 — entirely before the public announcement, which suggests some properties saw the data before the public rollout.
Two clicks from 407 impressions is a CTR under 0.5%, which is normal for this kind of traffic: people searching by photo are browsing, not hunting for a specific page.
That low CTR is not a problem to fix. A photo search is the top of the funnel by nature — someone saw something in the world and wanted to know what it was. Most of those searches end without a click.
The clicks that do happen go to the page Google judges most relevant to the photo. What matters is which of your pages appear at all, because that tells you which of your images Google can match to real-world objects.
Do not treat a rising multimodal line as a ranking win, either. Since the data did not exist in the counts before September 24, any growth in the first months is at least partly the rollout itself. Compare pages against each other — which pages get multimodal impressions and which get none — rather than comparing this month against a past that never had the metric.
Image SEO Tools That Help With Visual Search
No tool can directly optimise for multimodal matching, because Google has not published how an image gets matched to a page — and nobody outside Google knows the weighting. What tools can do is fix the fundamentals that decide whether your images are usable at all: fast loading, real image files, descriptive filenames and alt text, and structured data that tells Google what the image shows.
| Job | Free tool | Why it matters for visual search |
|---|---|---|
| Compress images | FrostRank Image Compressor | Fast pages get crawled and served more; heavy images slow both down |
| Check image files load | Browser DevTools, PageSpeed Insights | An image that fails to load cannot be matched to anything |
| Validate structured data | Google Rich Results Test | Product and ImageObject markup tells Google what the image depicts |
| Check indexing status | Search Console URL Inspection | Confirms Google can see the page the image lives on |
The honest version of this advice is that image SEO tools help with the parts you control, and the matching itself is Google's black box. A tool that promises to "optimise for Lens" is selling certainty Google has not published. Spend your effort on images that load fast, have real URLs, and sit on pages that clearly explain what they show.
What to Fix on Your Pages
Start with the pages the report already shows. Open the multimodal view, switch the dimension to Pages, and look at which pages get impressions. Those pages are already matching real photos — make sure the images on them are doing their job: real image files with descriptive filenames, alt text that says what the image shows, and captions where they help a reader. Then compress your images so the pages load fast on the phones doing the searching.
Next, check for images embedded as Base64 on those pages. A photo embedded as a Base64 data URI has no image URL of its own, so it cannot surface as an image result the way a real image file can.
If a page that matters for visual search carries its key photo as a data URI, move that photo to a real file. The same applies to images loaded through JavaScript that Googlebot cannot see — if the crawler cannot fetch the image, Lens cannot match it either.
Finally, add structured data where it fits. Product pages should carry Product markup with an image property; articles with a defining photo can use ImageObject. Structured data does not guarantee a multimodal appearance, but it is the clearest signal you can send about what an image depicts — and unlike the matching algorithm, it is a signal Google has documented. Validate it with the Rich Results Test before you call the job done.
What Not to Do
Do not rebuild your image strategy around this report. Google has announced no new ranking factor and no optimisation checklist for multimodal traffic, and the report itself is a measurement tool, not a ranking signal. A page that appears for photo searches is a page Google already trusts for the topic — the photo is a new way in, not a new thing to rank for.
Do not chase the missing queries, either. There is no workaround that reveals what was photographed: the Queries dimension is unavailable by design, and third-party tools cannot recover data Google does not expose. Any tool claiming to show your multimodal queries is guessing. Read the Pages dimension instead — it is the only honest view of this traffic.
And do not panic about the numbers being small. Most sites will see a handful of multimodal impressions at first, and some will see none for months. That is not a penalty or a crawling problem — it is a new channel growing from zero. If a page has Crawled - currently not indexed status or sitemap errors, fix those first: a page Google cannot index cannot appear for any search, camera or keyboard.
Frequently Asked Questions
How do I see multimodal searches in Search Console?
Open Performance, then Search results, click the search-type filter, and under Web choose "Multimodal". The table shows clicks, impressions, CTR, and position for searches made with an image. The filter appears only if the rollout has reached your property, and numbers appear only if your site receives traffic from these searches.
Why does the multimodal report show no queries?
Because the search started with an image, not text. Google's documentation states that specific text query data is not available for traffic that mostly uses images, so the Queries dimension is not offered for this search type. You get pages, countries, devices, and dates instead. For this traffic, the page is effectively the query.
Does multimodal search affect my rankings?
Google has announced no new ranking factor with this report. It is a measurement feature: it shows traffic that was already happening but was not previously counted in web search totals. Optimising your images — fast loading, descriptive alt text, structured data — helps the pages that appear, but there is no separate multimodal ranking checklist to follow.
What are the best image SEO tools for visual search?
The tools that matter are the ones that fix the fundamentals: an image compressor for fast loading, the Rich Results Test for validating Product and ImageObject markup, and Search Console's own URL Inspection to confirm Google can see the page. No tool can optimise directly for Lens matching, because Google has not published how an image gets matched to a page.
Which searches count as multimodal?
Google lists four routes: Google Lens, Circle to Search on Android, image uploads to Google Search, and Chrome's right-click "Search this image" option. Google has not said where Search Live sessions or images asked inside AI Mode are counted, so treat the multimodal figure as a floor rather than a complete count of image-led searches.
Should I compare multimodal traffic with my old totals?
No. Google confirmed this data was not previously included in the web search counts, so there is no baseline. A rising line in the first months may only reflect the rollout reaching your property. Compare pages against each other within the multimodal view instead of comparing against a past that never had the metric.
The Bottom Line
Search Console now shows you a kind of search you could not see before: people pointing a camera at the world and landing on your pages. The report is young, the numbers are small, and the missing queries are a genuine frustration — but the pages it lists are a real signal about which of your images Google can connect to real things.
Open the filter, look at which pages appear, and fix the basics on those pages first: real image files, descriptive alt text, fast loading, and structured data where it fits. That is the whole job for now. The matching algorithm is Google's secret; making your images findable, fast, and clearly described is yours.
