1. What Google Has Added

In the Search Console performance report, the search type can now be set to "multimodal", just as it was previously possible to switch between web, image, video and news. The change was announced on 24 September 2026 on the Search Central Blog, and the rollout is global. Anyone who receives such traffic sees the figures in their existing report without having to set anything up.

Search Console Now Shows Lens and Image Search Separately – Grünberg.Digital. AI Blog

What matters is the note Google gives on the origin of these data: the multimodal visits were not recorded before. The existing web figures are not supposed to fall as a result, so something is being added and not redistributed. That is the most important practical piece of information in the whole update, because it removes the obvious worry that your own history suddenly looks different. As of September 2026 this stands as a statement by Google, not as a measurement of our own.

The second point is a gap, and it is large enough to shape the analysis. There are no queries for this search type — the dimension is switched off. Anyone who opens the report expecting to see now what people search for with the camera will be disappointed.

2. The Four Entry Points

The search type brings together four routes by which an image rather than a text triggers the search. The first is Google Lens, that is, camera search in the Google app and on the smartphone. The second is Circle to Search on Android, where a section of the screen is circled and sent straight to search. The third is image upload in Google Search, the classic route via an existing photo.

The fourth entry point is the Chrome feature "Search this image", which runs through the context menu of an image on a web page. These four are combined in the report and cannot be separated individually. For the analysis that means: you see the volume of the category, not the behaviour of a single feature.

This grouping is a simplification, but a sensible one. The four routes differ technically but lead to the same kind of request — someone has seen something and wants to know what it is or where it comes from. For the question of whether the channel is worthwhile for a business, the category is the right level.

3. Why the Queries Are Missing

Google's explanation is simple and plausible: because multimodal searches predominantly use images rather than text, specific text queries are not available. There is simply no query that could be displayed, because the user did not enter one. The dimension is therefore deactivated for this search type, not for data protection reasons or out of reticence, but because the data do not exist.

That is a different case from the familiar gaps in queries, where Google withholds existing data for threshold reasons. Here, then, there is also no hope that the information will be supplied later. Anyone who builds their analysis on queries has nothing to connect to at this point and has to change the question.

The consequence in practice is uncomfortable but clear. The usual chain — see the query, align the content with it, measure the effect — does not work here. What remains is the reverse direction: starting from the page and considering which image or which object on it might have brought someone here.

Without a query, the direction reverses

The usual chain does not work for the multimodal search type; what remains is the route starting from the page.

Usual chain Multimodal search type Starting point the query the page Approach align content with… consider which… Measurement effect on the… pages, countries…

4. What Can Be Analysed Without Queries

The remaining dimensions and metrics stay available: pages, countries, devices and date, as well as clicks, impressions, click-through rate and position. That sounds like little, but it is sufficient for the opening question. The page dimension tells you which content is found by this route at all, and that is the information from which a decision can be derived.

In concrete terms, three analyses are worth a look. First, the list of pages by impressions: if product pages, detail photos or object pages are at the top there, the channel really is being used. Second, the comparison of the click-through rate with the web search type, which shows whether visibility also leads to visits or whether the user already gets their answer in the results view. Third, the device dimension, which with a camera-based channel can be expected to be mobile — and if it is not, it is worth asking where that comes from.

Analysing the multimodal report in a few minutes

Three analyses are enough to answer whether the channel counts for your own business.

Switch search type in the performance… Pages by… shows which content… CTR and devices compared with… Result treat the channel…

The country dimension is usually the least interesting of the four for small and medium-sized businesses, but it has one use: it shows whether the multimodal traffic comes from the sales territory at all. For manufacturers with an export share that can be a pointer; for a regional business it is a control figure and nothing more. Anyone short of time looks at pages and click-through rate and leaves the rest.

What cannot be answered with this should be stated just as clearly. Which specific object was photographed, whether it was your own product or someone else's on the same shelf, and whether the user intended to buy — none of that is in the report, nor can it be inferred from it. The report says that this route exists and how large it is, not why it is used.

5. Which Businesses Should Take a Look

There is a workable rule of thumb: the channel is worthwhile wherever someone is standing in front of a thing and wants to know what it is. That concerns manufacturers and retailers with visible products, the spare parts sector, where someone photographs a component instead of typing out a number, and everything with signs, labels and type designations. Buildings, installations and place references fall under it too, as do menus and notices in hospitality.

For other businesses the report is a footnote. Anyone who sells consultancy services, offers software or is found predominantly through technical terms simply has no object that someone could photograph. There the search type will be close to zero, and that is not an omission but the right number.

The in-between case worth watching lies between the two. Some service providers have more of a connection to objects than they think — a vehicle with lettering, a site board, a shop sign. If impressions show up there, that says something about everyday visibility that no keyword research shows. Checking takes five minutes, and the result is usable in either direction.

6. Export Instead of API

One practical point slows the analysis: the Search Analytics API does not yet include the multimodal search type. The API reference lists web, image, video, news, Discover and Google News, but not multimodal. Anyone who pulls Search Console data automatically into a dashboard or a spreadsheet will therefore not get these figures there for the time being.

That leaves the export button in the interface, and for getting started that is enough. One export a month, filed next to the other reports, fully answers the question of the size and development of the channel. An automation project is not worthwhile at this point as long as it is unclear whether any figures worth mentioning accumulate at all.

Whether and when the API will catch up has not been announced as of September 2026. Anyone now planning to rebuild their own reports should factor that in and plan for the manual work instead of waiting for an extension. There is little to suggest it will not come, but no date has been given.

7. Checklist

  • Open Search Console and set the search type in the performance report to "multimodal"
  • Check whether there are any impressions at all — at zero the matter ends here
  • Sort the pages by impressions and look at which content is found by this route
  • Compare the click-through rate with that of the web search type to judge whether visibility leads to visits
  • Cross-check the device dimension — a predominantly mobile share is to be expected with this channel
  • Limit the period to the past few weeks, since the data collection is new and has no long history
  • Export the report and file it, because the API does not yet carry the search type
  • Set a follow-up date in three months to see a development rather than a snapshot

8. Conclusion

The new search type answers a question that until now could only be guessed at: whether your own content is found via the camera, and on what scale. For businesses tied to products, objects or places that is real information, because it makes visible a behaviour that has long been taking place in everyday life but did not show up in the reports. The fact that the figures come on top and do not reduce the existing web values takes the nervousness out of the whole thing.

Even so, expectations should be kept low. Without queries the report is a statement of quantity with pages assigned to it, not a tool for content planning — it says that something is happening, not what follows from it. Anyone who wants to derive a separate optimisation discipline from it is building on four metrics without context, and experience shows that this does not lead far.

My assessment would therefore be: look once, put the result in context, and then either treat the channel as relevant or file it away. For most small and medium-sized businesses the latter will be the outcome, and that is a good result, because it closes an open question. For the others, the consequence that applies anyway is worth acting on: decent product photos, unambiguous labelling and pages that match the object someone has just photographed.

9. Frequently Asked Questions

Will my existing figures fall because part of them is now counted as multimodal?

No. Google states that these data were not recorded before and that the existing web figures are not supposed to fall as a result. It is therefore an additional data collection, not a redistribution of existing clicks. As of September 2026 that is Google's statement; comparing your own monthly values before and after 24 September costs little and provides certainty.

Why do I see no queries in this report?

Because there are none. Multimodal searches predominantly use images rather than text, so no text query exists that could be displayed. The dimension is switched off for this search type, and it is not to be expected that it will be supplied later.

Can I pull the data into my dashboard automatically?

Not for the time being. The Search Analytics API does not yet carry the multimodal search type — the reference lists web, image, video, news, Discover and Google News. Anyone who needs the figures regularly works with the export from the interface until further notice.

Which four routes count as multimodal in the first place?

Google Lens, Circle to Search on Android, image upload in Google Search and the Chrome feature "Search this image". These four are combined in the report and cannot be analysed individually. You therefore see the total volume of the category, not the distribution across the individual features.

What do I do if the report shows zero?

Nothing. With a camera-based channel, a value of zero generally means that there is no object anyone would photograph — that applies to many service providers and software vendors. Instead of turning it into a task, you note the result and look again in a few months.

Is it worth optimising images differently now?

No optimisation instruction can be derived from the report itself, because it lacks the queries. What applies anyway continues to apply: clean product photos, unambiguous labelling and pages that also name the object shown. Whether that has an effect on multimodal searches cannot be proven with the available data.

Stephan Michalik
About the Author
Founder Grünberg.Digital. · CEO Grünberg.Digital. GmbH

Maximum performance through the synergy of experience and innovation: As Founder of Grünberg.Digital. and CEO of Grünberg.Digital. GmbH – a leading business incubator and enabler – Stephan Michalik designs holistic online marketing strategies. Whether precise SEA, high-revenue email marketing, or high-converting landing pages: He seamlessly combines these core disciplines with cutting-edge AI. The result: highly efficient, AI-powered marketing ecosystems for maximum digital advantage.

LinkedIn