Search Console’s new image search filter separates image-based search clicks from text queries. Here’s what it shows an online store and what to fix first.
Table of Contents
A shopper points their phone camera at a friend’s sneakers, taps Google Lens, and lands on a product page they never typed a single word to find. Until this week, the store that owns that page had no way to see that visit happening. Google Search Console’s Performance report only ever showed search as one undifferentiated pile of clicks, whether someone typed “waterproof hiking boots size 10” or pointed a camera at a picture.
That changed on September 24, 2026, when Google rolled out a multimodal filter inside the Performance report’s Web search type. It splits traffic into two buckets: text-based, the normal typed queries, and multimodal, anything that started from an image rather than a keyword, whether that’s Google Lens, Circle to Search on Android, a direct image upload to Search, or Chrome’s right-click “Search this image” option. Google product managers Harsh Kharbanda and Moshe Samet described the goal plainly: giving site owners insight into how their content gets surfaced when a search starts with a picture instead of text. (Source: Search Engine Journal)
What the New Multimodal Filter Shows You
Once multimodal traffic reaches a property, the filter breaks it down by page, country, and device, same as any other Performance segment, and the whole thing exports the way the rest of the report does. That’s genuinely useful for a store with a large product catalog, because it answers a question no other free tool answers directly: which specific product pages are earning visits from people who searched with a photo rather than a phrase.
There’s a real limit worth knowing about before you go looking for it. Google isn’t surfacing query text for multimodal searches, because most of them never had one. Someone circling a lamp on their screen didn’t type “mid-century brass floor lamp,” so there’s no keyword string to report. You get the page, the country, the device, and the click and impression counts. You don’t get the search term that led there, because most of the time none existed.
Why This Matters More for a Store Than a Blog
A blog post ranking for a typed phrase and a product photo ranking for a Lens search behave differently enough that lumping them into one number never made sense. Visual search skews hard toward physical products: furniture, clothing, shoes, home decor, anything a shopper can point a camera at and immediately recognize a category for, even without knowing the brand or the right words to search it. A store selling recognizable physical goods has more to gain from this data split than almost any other type of site, because it’s the segment of search traffic where a great product photo does work that no amount of keyword optimization ever could.
That also means the sites that show up with meaningful data in this new filter are the ones that already had decent product photography and clean image delivery before Google ever announced this feature. This isn’t a switch you flip after the fact; it’s a report that finally makes visible work that was already paying off, or already not happening.
Turning the Report Into a Practical Checklist
Seeing that a page gets multimodal clicks doesn’t do much without knowing what to check next. A few things determine whether a product photo is even eligible to surface from an image search in the first place:
- Alt text that actually describes the product, not the file name. “Blue canvas sneaker, size 9, side profile” gives Google’s image understanding something to match against; “IMG_4021.jpg” or a keyword-stuffed string gives it nothing.
- Structured data on the product page (schema.org Product markup: price, availability, brand) so a matched image can resolve to a page Google trusts enough to serve.
- An image sitemap, separate from the standard XML sitemap, listing product image URLs directly. Many store platforms don’t generate one by default, and it’s easy to assume the regular sitemap covers this when it usually doesn’t.
- Fast-loading, properly sized images. A visually matched photo that sits behind a slow page load loses the click before the shopper ever sees the product; this is a case where hosting performance and SEO performance are the same problem wearing different names.
The Part Most Coverage of This Update Skips
Most of what’s been written about this filter treats it as a Search Console feature announcement and stops there: here’s the toggle, here’s the rollout date. What that misses is that OpenCart and PrestaShop stores, specifically, tend to have a structural gap in exactly the areas that determine multimodal visibility. Default OpenCart product image handling uses auto-generated cache filenames rather than descriptive alt text unless a store owner sets it manually per product, and neither platform ships an image sitemap out of the box the way some hosted platforms do. That’s a real, checkable gap: open any OpenCart product edit screen and the image alt-text field is blank unless someone filled it in.
For a store on either platform, the practical order of operations is: audit a sample of product pages for missing or generic alt text first (usually the biggest single gap), confirm whether an image sitemap extension is installed and actually submitted in Search Console, and only then start watching the new multimodal filter to see whether any of it moved the needle. Checking the report before fixing the underlying gaps just confirms there’s nothing to see yet.
Where This Fits Into the Bigger AI-Search Shift
Image search isn’t the only way discovery is moving away from typed keywords. Chatbot-style AI search tools are doing something similar on the text side, answering a shopper’s question directly instead of returning a list of links for them to click through. The two trends share a root cause: search increasingly starts from something other than a precise typed phrase, whether that’s a photo, a voice question, or a conversational prompt to an AI assistant.
A store that’s already disciplined about alt text, structured data, and clean product information is better positioned for all of these at once, because the same underlying signals (clear entity names, accurate specs, machine-readable structured data) are what both an image-matching algorithm and an AI answer engine rely on to identify a product correctly. Treating this multimodal filter as an isolated checkbox misses that it’s one visible piece of a broader shift in how products get found, not a one-off feature to check and forget.
What a Realistic Timeline Looks Like
Google’s own rollout note is honest that data won’t appear for every property immediately, and that’s not a bug to wait out, it’s a function of whether the property already gets meaningful multimodal traffic. A niche B2B software page has little reason to show anything in this filter ever; a home goods or apparel store with decent product photography should start seeing a nonzero multimodal number in the weeks after the September 24 rollout finishes, assuming Google is already crawling and indexing those product images, which itself depends on the image sitemap and crawl budget being in reasonable shape.
The realistic expectation for most small stores is a data trickle at first, not a dashboard full of numbers on day one. That’s consistent with how Google has rolled out other segmented Performance report splits before: the filter existing and the data being dense enough to act on are two different milestones, and only the first one landed this week.
If you’re running a PrestaShop or OpenCart store and haven’t looked at your product pages’ alt text and image delivery speed recently, this update is a good forcing function to do that regardless of whether the multimodal filter shows meaningful numbers yet. The underlying fixes help ordinary image search and page experience either way; the new report is just the first tool that will eventually let you measure whether they’re working for the image-search half of the traffic specifically.
Key Takeaways on the Image Search Filter
- The image search filter separates clicks that came from image-based searches from ordinary text queries.
- Use the image search filter to see which product pages attract visual searches, then check their images and alt text.
- A store gets more from the image search filter than a blog, because product photos are what shoppers match against.
- Open the image search filter in the Performance report to compare image-based clicks with text clicks.
- A image search filter result is only as useful as the images behind it, so check alt text and file names first.
- Revisit the image search filter monthly; image-based search traffic changes as more people use visual search.