A customer photographs a distinctive lamp or ceramic jug, then uses Google to look for something similar. Google has begun giving website owners a clearer view of searches like that. A new Web: multimodal filter in Search Console separates web searches that included an image from those made with words alone.
For a Bath retailer, a Somerset gallery or a Wiltshire visitor attraction, that could reveal a useful route to their website. For many other businesses, it may show very little. The sensible first step is to understand what the new figures actually measure before changing a page or commissioning more photography.
What Google has added
Google announced the reporting change on 24 September 2026 and said it was beginning a global rollout. The new search-type filter appears in the familiar Performance on Search results report and in the separate Generative AI performance report. Google says it covers web searches where an image was used, including Google Lens, Circle to Search on Android, image uploads to Google Search, and Chrome’s “Search this image” option.
This is different from choosing the existing Image search type, which concerns image-search results. The new filter is about web results reached through a search that used an image as part of the query. That distinction matters when comparing reports: a change in one view should not be presented as a change in all visual discovery.
Google says the figures will appear as the rollout reaches a property and the site receives activity from these searches. An empty report on the first day is not evidence that the website has a technical problem.
Why this could matter locally
People often recognise a product, style or place before they know the words to describe it. An independent homeware shop might be discovered through a photograph of a vase; a garden centre through a plant; or a heritage attraction through an image of a building. Those are examples of possible customer journeys, not claims that any particular South West business is already receiving this traffic.
The new view may help an organisation see which pages are being surfaced when someone searches visually. If a product page or useful guide appears repeatedly, the business can check whether that page still describes the item accurately and gives the visitor a clear next step. If a service firm sees little data, it need not invent a visual-search campaign to fill a chart. Its customers may still mainly search by name, service or location.
This fits a broader SEO approach grounded in real customer searches: look at how people find the business, then improve the pages that help them make a decision.
A practical check in Search Console
- Find the right filter. Open the website’s Search Console property, go to Performance on Search results, and choose Web: multimodal under Search type. If the property has the Generative AI performance report, the same search-type choice is available there.
- Look at pages and trends. Check which pages appear, then compare a sensible period once enough data has built up. Country and device views can show whether the audience looks relevant. Export the data if you need to compare it with other reports.
- Keep the measures separate. The ordinary Search results report can show clicks and impressions. Google’s Generative AI report shows impressions for its AI features; an impression does not prove a visit, enquiry or sale. Its data also sits within web-search reporting, so avoid adding figures from the two reports together as though they were independent audiences. The site has already explained why AI search impressions and position need careful interpretation.
- Check the destination page. If a relevant page surfaces, ask whether its image and surrounding words genuinely explain the product, place or service. Is the information current? Can a visitor understand availability, price or how to enquire where those details matter?
Google’s image guidance supports basic, useful housekeeping: place relevant images on relevant pages, use clear pictures that load efficiently, and write descriptive alternative text that also helps people using screen readers. A web developer can check that important images are available to Google in ordinary image elements rather than existing only as decorative backgrounds. There is no need to stuff image descriptions with place names or promise that an alt-text edit will produce visual-search traffic. For many businesses, improving the clarity of the page around the image will help visitors more than chasing a new label in a report.
Use the new view as a clue
This is a reporting change, not an announcement of a new ranking shortcut. It gives website owners another way to ask whether their pages are being found when people search with pictures. A local business should use it alongside ordinary search clicks, website analytics and genuine outcomes such as bookings, sales or worthwhile enquiries.
For now, check whether the filter is available, record what it shows, and give the data time to settle. If a visual customer journey is relevant to the business, improve the pages and images that serve that journey. If it is not, the empty space on a dashboard is not a reason to force one.
