Think about the last time you saw a pair of shoes on someone at a wedding, or a plant on a friend's balcony, and wanted to know what it was. You probably didn't type anything. You opened the camera, pointed, and tapped.
A growing share of searches start that way now. And until very recently, site owners had no clean way to see them in Search Console. They were either hidden inside the regular "Web" numbers or not obvious at all.
That changed on 24 September 2026. Google's Search Central team announced web multimodal Search performance reporting in Search Console. In plain words: searches that begin with an image now get their own row.
Here's what the new report covers, how to open it, how to read it if your audience is in India, and how it differs from the text search data you already know.
What Google actually launched
The Search type filter in the Performance report used to offer one "Web" option. It's now split in two:
- Web: text-based covers traffic from text queries typed into the standard Google search bar.
- Web: multimodal covers traffic when people search using images, such as with a smartphone camera.
Google's Performance report help page lists four sources for the multimodal bucket:
- Google Lens
- Circle to Search on Android
- Image uploads to Google Search
- Chrome's right-click "Search this image"
The filter shows up in two places: the Search results Performance report (clicks, impressions, CTR, position) and the Generative AI features report. Google says it is rolling out globally from 24 September, so if you don't see it yet, give it a few days.
One quiet but important detail: Google's announcement tells you to click Export to take the data into other tools. It does not mention API support. If your reporting runs on the Search Console API or a Looker Studio connector, check that the multimodal split actually comes through before you trust an automated dashboard.
How to open the multimodal report (step by step)
- Sign in to Search Console and pick your property.
- Go to Performance → Search results.
- Click the Search type filter at the top.
- Choose Web: multimodal and apply.
- Set a date range of the last 3 months, then look at where the line actually starts. Don't assume a start date; check your own property.
- Switch to the Pages tab to see which URLs get visual-search clicks and impressions.
- Repeat the same steps in Performance → Generative AI features to see how often your pages appear in AI answers triggered by an image.
- Hit Export and save the file. You'll want a baseline before the numbers start moving.
The big gap: there are no queries
This is the part that will catch most people off guard. When the filter is set to multimodal, you won't get a list of search terms. That makes sense once you think about it: the "query" was a photo, a screenshot or a camera frame. There's no typed text to show.
So you're working with page-level, country-level, device-level and date-level data only. That changes how you analyse it:
- Pages become your proxy for intent. If your saree product pages get multimodal clicks but your blog posts don't, that tells you what people are photographing.
- Look at the images on those pages. The image on the page is often the closest thing you have to a "keyword".
- Don't try to reverse-engineer queries. Any tool that claims to show you the exact images people searched with is guessing.
How to track Google Lens traffic in India
India is an Android-first, camera-first market. Circle to Search lives on Android, and people here already use the camera to shop, translate menus, check plants and solve homework. If your audience is Indian, this report deserves a regular slot in your monthly review.
Here's a simple routine that works:
- Filter to India. Set Search type to Web: multimodal, open the Countries tab and click India. Everything below now applies only to Indian visual searches.
- Split by device. Open the Devices tab. Expect mobile to dominate; if desktop shows up, that's likely Chrome's right-click image search.
- Find your visual winners. On the Pages tab, sort by clicks. Note the top 20 URLs.
- Compare with text. Keep the same India and page filters, then switch Search type to Web: text-based. Pages that are strong in multimodal but weak in text are your hidden visual assets.
- Group by template. Tag the URLs as product, category, recipe, how-to, or article. Patterns usually jump out fast.
- Track it monthly. Export once a month into a sheet. Since there's no query data, the trend line per page group is your main signal.
Who in India should care most? Fashion and ethnic wear, home decor, plants and gardening, auto spare parts, electronics and accessories, food and recipe sites, and education platforms. In all of these, the thing people want is easier to photograph than to describe, especially when they'd otherwise have to type it in English rather than Hindi, Tamil or Marathi. A photo doesn't need a language.
Multimodal vs text-based web traffic: what's different
The two buckets measure the same Google Search, but the behaviour behind them is not the same.
| Web: text-based | Web: multimodal | |
|---|---|---|
| How the search starts | Typed words in the search bar | A photo, screenshot or image (Lens, Circle to Search, upload, Chrome) |
| Query data in GSC | Yes, Queries tab available | No, there's no typed query to show |
| Typical intent | Research, comparison, questions, brand lookups | "What is this?", "Where can I buy this?", "Find similar" |
| Pages that tend to win | Guides, answers, category and brand pages | Product pages, image-rich pages, visual how-tos, identification guides |
| Main lever | Content relevance and keywords | Clear, original images with strong page context |
| Generative AI report | Text-led AI Overviews and AI Mode impressions | Impressions in AI features triggered by an image |
One caution before you compare year on year: Google hasn't spelled out how this traffic was counted before the split. Keep any before-and-after comparison inside the new filters, and avoid mixing old "Web" exports with new "Web: text-based" ones.
How to earn more multimodal visibility
None of this is new advice, but the report finally lets you measure whether it works:
- Use original photos. Stock images look like everyone else's. Your own product shots, from several angles, give visual search something unique to match.
- Put the image next to the words that explain it. Captions, nearby headings and product details give Google context about what's in the picture.
- Write real alt text. Describe the item plainly: "maroon Banarasi silk saree with gold zari border", not "saree1".
- Name files sensibly. banarasi-silk-saree-maroon.jpg beats IMG_4821.jpg.
- Add Product structured data with price and availability on ecommerce pages.
- Don't block images in robots.txt, and serve them at a good resolution and in a fast format like WebP.
- Keep pages fast on mobile. Visual searchers are almost always on a phone.
Related reading on ClicZeo
- Search Console AI Impressions: What They Mean for Indian SEO Teams
- Google Search Console: The Free SEO Tool Guide
- Google AI Performance Report in Search Console (2026)
- Google Discover Dive Deeper: Publisher Playbook for 2026
The bottom line
The multimodal filter doesn't hand you keywords, and it won't explain every click. What it does is make visual search visible for the first time, by page, country and device. That alone is enough to show which of your images are already working and which pages deserve a better photo.
Open Search Console today, export your first multimodal baseline, and come back to it next month. The trend will tell you more than any single day of data.
Sources
- Announcing web multimodal Search performance reporting in Search Console, Google Search Central Blog, 24 Sep 2026
- Performance report (Search results): Overview and basic setup, Search Console Help
