How to Track Visual Search Traffic in Search Console

Google Search Console now splits out multimodal traffic from Lens and Circle to Search. Here is how to find it, export it and decide what to do with it.

On 24 September 2026, Google split the "Web" search type in the Search Console Performance report into two: text-based and multimodal. If someone found your page by pointing a phone camera at a product or circling something on their screen, that visit now has its own line in your data.

Until now, that traffic was mixed into your general web numbers, so nobody could say how big it was. This guide walks through where to find the new filter, what it does and does not tell you, and a simple routine for tracking it without much effort.

What changed and who it affects

Google's own help page now lists "Web: multimodal" as a search type, described as search traffic when users search using images, such as with a smartphone camera. It sits next to Web: text-based, Image, Video and News in the filter at the top of the Performance report.

Four entry points feed the multimodal bucket:

• Google Lens
• Circle to Search on Android
• Image uploads to Google Search
• Chrome's right-click "Search this image" option

According to a summary of the announcement, the filter appears in both the Search results Performance report and the Generative AI features report. Clicks from the Images tab still count as "image" type, not multimodal.

Who should care? Any business whose customers might photograph something and then look for it: retail and ecommerce, furniture and home goods, fashion, food, travel, local services with a visual product, and anyone with a lot of product or location imagery. If you sell something people can point a camera at, this is your data.

Step 1: Find the multimodal filter

1. Open Search Console and choose your property.
2. Go to Performance, then Search results.
3. Click the search type filter at the top, currently showing Web.
4. Select Web: multimodal and apply it.
5. Repeat with Web: text-based so you have a comparison.

If you do not see the option yet, check again in a few days. Google said the rollout is global, but features like this rarely arrive everywhere at once.

Step 2: Know what you are looking at

Google says this is previously unreported traffic, not a reshuffle of existing numbers, so your text-based figures should stay as they were, according to the same write-up. That matters for reporting. Do not tell your boss that text search dropped when you see the split.

The big limitation is queries. Because these searches mostly start with an image rather than words, the Queries dimension is not available for multimodal traffic. You can see pages, countries, devices and dates, along with clicks, impressions, CTR and position. You cannot see what the image showed.

That means you will be working backwards. If a product page suddenly collects multimodal clicks, the page itself is your clue to what people photographed.

Step 3: Pull the pages, countries and devices

With the multimodal filter on, switch between the Pages, Countries and Devices tabs and look for patterns:

• Pages: Which URLs get multimodal clicks? Are they product pages, category pages, blog posts, or something you did not expect?
• Devices: Expect a heavy mobile skew, since cameras and Circle to Search live on phones.
• Countries: Does the mix match your normal audience, or does visual search reach markets text search does not?

Compare click-through rate and average position against text-based traffic for the same pages. A page that ranks well in multimodal results but converts badly on mobile is a fixable problem, and now you can see it.

Step 4: Export a baseline now

Do this part early. The Search Analytics API did not yet list a multimodal search type when the first reports came out, with its reference page last updated on 11 August. That means any automated pull through the API, including most reporting dashboards, may not capture it yet.

The reliable route for now is the Export button in the Performance report. Export pages, countries and devices with the multimodal filter applied, and save the file with the date in the name. Note which report the data came from, since the Generative AI features report has its own view.

One caution on history. One write-up says the data starts on 24 September, so do not expect a long look-back. Treat your first export as day one of your own trend line.

Step 5: Build a light weekly routine

You do not need a new dashboard for this. A short routine is enough to spot a trend:

1. Every Monday, export the previous week's multimodal data (pages, countries, devices).
2. Keep text-based and multimodal numbers side by side in one sheet.
3. Once a month, join multimodal landing pages to your analytics revenue or lead data to see whether this traffic actually converts.
4. Flag any page that gains multimodal clicks quickly and check what image, product or location it relates to.

The monthly join is where the value is. Clicks are interesting, but a business owner cares whether a Lens visitor buys. If you do not have clean landing-page attribution yet, a proper web analytics setup is the first job, and SEO reporting can then fold this new segment into your regular numbers.

What to do with what you find

Once you have a few weeks of data, the pages that attract multimodal traffic become your priority list. Some practical checks:

• Image quality and variety: Product pages should have sharp, well-lit photos from several angles, because Google is matching what the camera sees.
• Descriptive alt text and file names: Plain, accurate descriptions help both accessibility and image understanding.
• Mobile page speed: Visual searchers are on phones. If the page takes five seconds to load, the click is wasted. Our technical SEO work covers this kind of fix.
• Clear next step: Someone who photographed a lamp wants the price, stock status and a buy button above the fold.
• Structured data: Product, price and availability markup keeps your listing accurate wherever it appears.

None of this is new advice. The difference is that you can now measure whether it moves a traffic source you could not see last month. Visual search also overlaps with how AI answers pick sources, so it is worth reading the multimodal numbers next to your Generative AI features report. If AI visibility is a priority, our GEO and AIO optimisation service looks at that side of things.

Common mistakes to avoid

• Expecting keywords. There is no Queries data here, so do not build a keyword strategy around it.
• Reading too much into one week. Volumes may be small for many sites. Wait for a few weeks before changing anything.
• Relying on the API. If your dashboard pulls from the API, it may show nothing for multimodal. Check before you assume there is no traffic.
• Forgetting the Images tab. Image tab clicks are reported separately, so a low multimodal figure does not mean your images are not working.

Key takeaways

• Search Console now shows Web: multimodal traffic from Lens, Circle to Search, image uploads and Chrome's Search this image.
• Queries are not available, so work from pages, countries and devices.
• Export a baseline now, because API support was missing at launch and history appears to start on 24 September.
• Compare against text-based traffic and tie multimodal landing pages to revenue or leads.
• Improve images, mobile speed and product details on the pages that attract this traffic.

If you would rather have someone set up the tracking and tell you what the numbers mean for your site, our team is happy to look at it. You can get a free quote and we will start with a look at your current Search Console data.

Sources

• Search Console Multimodal Filter: Lens Data, No Queries (Relevant Audience)
• Search Console Now Separates Visual Search Traffic (Digital Applied)
• Performance report (Search Console Help)