Search Console Now Reports Traffic From Image Searches

Google has added a multimodal filter to Search Console’s Performance report, separating searches that began with an image from searches that began with typed words. The change was set out in a Google Search Central post on 24 September, which says the data covers searches made through Google Lens, Circle to Search on Android, image uploads to Google Search, and Chrome’s right-click "Search this image" option. The rollout is global from that date, and the metrics appear only for sites that already receive traffic from those searches.

The filter sits inside the Web search type rather than beside it. Google’s Performance report help page now divides Web into two lines. Text-based covers traditional traffic from queries entered into the standard search bar. Multimodal covers web results where an image was used as part of the search. Site owners choose between them in the search type filter at the top of the report, and the Export button pulls the numbers into other tools.

The Queries Tab Goes Empty

The dimension SEOs reach for first is the one Google cannot supply. Google told Search Engine Journal that because multimodal searches mostly use images rather than text, specific text query data is not available for this traffic. The Queries tab is unavailable once the multimodal filter is on, which leaves Pages as the working dimension.

That changes what the report is good for. A page with multimodal clicks tells its owner the page was reached through a camera, not what the camera was pointed at. The work it points to is visual rather than lexical: which product shots, diagrams and photographs sit on the pages earning those clicks. There is also no documented way to automate the pull, because Google’s Search Analytics API reference lists web, image, video, news, Discover and Google News as report types and carries no multimodal value.

Google’s Reporting Layer Is Catching Up

The pattern this year is a reporting layer chasing a search box that has stopped being a box for text. Google added AI Mode traffic data to Search Console in June 2025, gave generative AI features their own report in June and extended it to every site on 31 August, then opened Search Console to social and video posts in July. Multimodal continues that line, and it continues another one too: each of these reports has arrived with less query-level detail than the report before it.

At Google Marketing Live in May, Google called its new Search box, which accepts video, images, files and longer queries, the biggest upgrade to the box in more than 25 years. A search box built for cameras and files was always going to break a reporting model built on strings of words, and multimodal reporting is the first measurement Google has shipped that admits it. The bylines make the same point: the post is signed by the product lead for Google Lens as well as the product lead for Search Console.

What To Check First

Neither Google’s post nor its updated help page says whether the multimodal line carries history from before 24 September, or whether pulling multimodal out changes the Web totals site owners have reported for years. Until that is answered, the first job is a like-for-like reading: take a full day of Web traffic, then read the text-based and multimodal lines against it and see whether they reconcile.

Sites with catalogues, recipes, plants, parts and anything else people point a phone at are the most likely to find a number worth acting on. For everyone else the value is negative information. If image-led search sends almost nothing, the visual optimisation work that has been sold on the back of Lens can wait.