Google Search Console Multimodal Report: A D2C Guide

The traffic you've been paying for but never counted

If someone points their phone camera at a moisturiser bottle, or long-presses a screenshot of a kurta and taps Circle to Search, and then lands on your product page — where did that visit show up in your reports? Until last week, nowhere useful. It got swept into your Web totals with no label, so you couldn't tell an image-led visit from a typed-query visit.

That changed on Sep 24, 2026. Google started rolling out multimodal search reporting inside Search Console's Performance report, and it's a bigger deal for D2C stores than the announcement made it sound. For the first time you can isolate the traffic that starts with a picture instead of a phrase — and picture-first discovery is exactly how a lot of Indian shoppers find products they can't name.

What actually shipped

The new filter lives under the Web search type in the Performance report. When it reaches your property, you can break out "multimodal" results and see them apart from ordinary text search. Google's own framing is that this covers "web search results where an image was used as part of the search."

In practice, that bucket includes four things:

• Google Lens searches
• Circle to Search on Android
• Images uploaded directly to Google Search
• Chrome's right-click "Search this image"

All four share one trait: the query didn't begin with words. That single fact shapes everything about how the report behaves, including its biggest limitation.

What it shows, and the one thing it hides

Here's the useful part. For multimodal traffic you still get the four metrics you already trust — impressions, clicks, CTR and position — plus the country and device breakdowns. So you can answer real questions: how many people saw us through an image search, how many clicked, and where they were.

Here's the catch. There are no query terms. Google says it plainly: "Because multimodal searches mostly use images rather than text, specific text query data isn't available for this traffic." That makes sense — there was no phrase typed — but it means you can't reverse-engineer intent the way you do with text keywords.

There's a second place this data appears: the generative AI performance report is also getting multimodal reporting, but that one counts impressions only, not clicks. Keep the two straight, because it's easy to double-read the same visit and think you're doing twice as well as you are.

Multimodal is not "AI search" — untangle the three reports

Plenty of store owners already blur AI Overviews, AI Mode and image search into one worry. They're different behaviours and now they sit in different places.

The quick version

The AI reports answer "am I showing up when Google generates an answer?" The multimodal filter answers "am I showing up when someone searches with a camera or a screenshot?" A shopper using Lens to identify a serum isn't asking an AI question — they're doing visual lookup. If you conflate the two, you'll optimise for the wrong thing: writing more FAQ-style copy when your real gap is that your product images are unreadable to Google.

Why this lands harder for Indian D2C

Two habits make image-first search unusually common here. First, a lot of discovery happens on Android, where Circle to Search is a swipe away and Lens is baked into the camera and Photos. Second, shoppers often can't name what they want — "that printed cotton kurta from the reel," "the amber glass serum bottle" — so they screenshot and search the picture instead of typing.

For a catalogue-heavy store, that's a channel you were flying blind on. If your category is visual — apparel, jewellery, home decor, beauty, footwear — a meaningful slice of your Web impressions may have been multimodal all along. Now you can size it.

If you'd rather have someone map this against your paid and organic numbers together instead of squinting at one report in isolation, that's the kind of cross-channel read our growth marketing team does for D2C brands — but you can get most of the value yourself with the steps below.

How to actually earn multimodal visibility

The report only rewards you if Google can understand your images. Most Shopify stores lose here on basics, not on anything clever.

Fix file names and alt text

"IMG_4821.jpg" tells Google nothing. Rename product images to describe the product — "amber-glass-vitamin-c-serum-30ml.jpg" — before upload, and write alt text a human would recognise, not a keyword dump. On Shopify, alt text is editable per image in the product media panel; make it part of your product-upload checklist so it never gets skipped.

Give images structure and a path to be crawled

Product structured data helps Google tie an image to a price, availability and reviews, which is what makes a listing worth surfacing. Make sure your theme outputs valid product markup, and that your image sitemap (or your sitemap's image entries) actually lists your product images so they get discovered in the first place. An image Google never crawls can't win an image search.

Serve images that render fast and sharp

Compressed-to-mush images and slow-loading galleries hurt twice — once on Core Web Vitals, once on whether your visual is legible enough to match. Use appropriately sized images, modern formats, and keep the primary product shot clean and uncluttered so the object is obvious.

A quick example of what you'll see

Picture a jewellery store. A shopper screenshots an oxidised silver jhumka from an Instagram reel and runs it through Lens. Google matches the image to your product page and shows your listing. If they tap through, that visit now lands in the multimodal bucket — one impression, maybe one click, on that product's URL, with no query attached. Multiply that across a festive month and you get a real number for a channel you used to log as anonymous Web traffic. What you won't get is "oxidised silver jhumka" as a keyword, because nobody typed it. That's the trade: you learn the scale and the winning pages, not the exact words.

Where this fits with your merchant listings

If you sell through Google's free product listings or Shopping, strong product imagery already does double duty — it feeds Shopping surfaces and now this multimodal channel too. The same clean, well-named, well-structured images help both. So this isn't a separate SEO project bolted on for one report; it's the image discipline your whole Google presence already runs on, finally with a number attached. Prioritise the products that carry your revenue, not your entire catalogue in one weekend.

Reading the report without fooling yourself

A few guardrails, because this is week one for the data:

• Don't celebrate a spike yet. Google hasn't said whether the split backfills history. If numbers appear from a standing start, that's the reporting turning on, not demand arriving.
• Watch clicks, not just impressions. High multimodal impressions with near-zero clicks usually means your thumbnail or title in the result looks nothing like what the shopper photographed.
• Segment by device. Since this skews Android and mobile, check that those product pages load and convert on a mid-range phone, not just your desktop preview.
• Don't merge it with the AI report. Different behaviour, different metrics — the AI report has no clicks to give you.

What to do in the next two weeks

A tight starting plan:

1. Check whether the multimodal filter has reached your property yet; if not, it's still rolling out.
2. Once it's there, note your baseline — impressions, clicks, CTR — for the whole property.
3. Pull your top 20 revenue products and audit their image file names, alt text and structured data.
4. Confirm your product images are in your sitemap and returning 200s.
5. Recheck the report in two weeks and compare click trends on the products you fixed versus the ones you didn't.

None of this is exotic SEO. It's the same image hygiene that's mattered for years — the difference is that Google finally hands you a scoreboard for it. Set the baseline now, clean up your top sellers, and let the next month tell you whether picture-first shoppers were finding you all along.