Google Search Console Finally Measures Visual Searches: How to Analyze Your Lens Traffic?

Since September 24, 2026, Google Search Console has finally distinguished some of the traffic coming from visual and multimodal searches. The new Web: Multimodal filter lets you isolate performance from Google Lens, Circle to Search on Android, images uploaded to Google Search, and Chrome’s “Search this image” feature.

For an e-commerce site, portfolio, project gallery, product documentation, or editorial content rich in screenshots, this development changes one thing above all: it is now possible to measure separately a type of visibility that was previously difficult to isolate. That does not mean image optimization will automatically generate more traffic. However, Search Console now provides concrete data to help you decide which pages and visuals deserve priority in your audit.

What’s changing in Search Console since September 24, 2026

On September 24, 2026, Google announced a dedicated report for multimodal web searches. This data is available in the Search results performance report, as well as in reporting related to Google Search’s generative AI features.

The new search type groups together several situations in which users do not start solely from a query typed on a keyboard:

• Google Lens, for example when a user photographs an object, product, building, or document;
• Circle to Search on Android, when an item visible on screen is circled to launch a search;
• images uploaded to Google Search;
• Chrome’s “Search this image” feature.

In Search Console, the filter appears as Web: Multimodal. It should be distinguished from Image, which corresponds to traffic from the Google Images tab. This distinction matters: a page may receive few visits from Google Images while still being discovered through Lens or Circle to Search.

What Search Console can actually measure

The standard performance report retains its usual metrics: clicks, impressions, CTR, and average position. Once the “Web: Multimodal” search type is selected, these metrics apply only to the visual searches concerned.

The available dimensions remain particularly useful for an audit:

• Pages: which URLs appear or receive clicks in multimodal search;
• Queries: which terms are associated with impressions or clicks, when available;
• Devices: useful for seeing the role of mobile, which makes sense for journeys such as Lens or Circle to Search;
• Countries and dates: to identify geographic differences or changes over time.

This new feature does not provide an automatic diagnosis of your images, however. Search Console will not necessarily tell you, “this specific product photo triggered this Lens search.” The report mainly shows you which pages are discovered in a multimodal context. From there, you need to revisit the page, identify its main visuals, and determine which ones may have helped Google understand or present the content.

How to audit Google Lens and multimodal traffic in Search Console

The most useful method is to start with page data, then work back to the images. This avoids reworking an entire catalog or every illustration on a site without knowing whether they actually contribute to visual discovery.

This approach turns the question “Should we optimize our images for Lens?” into a more practical one: which pages does Google already associate with visual searches, and which have the most commercial or editorial value?

Compare products, projects, screenshots, and editorial content

The new filter is most useful when you look not only at individual URLs but also at groups of pages. A site may discover that its product pages perform well in multimodal search while its articles almost never appear—or vice versa.

A good practice is to add a “page type” column in a spreadsheet or analytics tool, then aggregate multimodal impressions and clicks by group. This gives you a much more useful view for deciding between a new product photography campaign, a revamp of article visuals, or portfolio improvements.

How to interpret the main signals

Average position also needs to be interpreted with caution. Visual and multimodal interfaces may present results differently from a conventional text-based SERP. It remains a useful metric to monitor, but should not be used on its own to conclude that an image “ranks better.”

Which images should you improve first?

Once you have identified priority pages, the goal is not to stuff images with keywords. Google mainly recommends making images easy to discover, relevant to the page, and supported by sufficient context.

Google Search Central’s official recommendations still apply to this new report:

• use actual HTML <img> elements with a src attribute that search crawlers can access;
• provide clear, high-quality images without compromising loading performance;
• place the image near text that accurately describes its subject;
• use short, descriptive filenames rather than generic ones;
• write useful, contextual alt text without artificially stuffing it with keywords;
• provide responsive images and a src fallback;
• where appropriate, specify a representative image using structured data or og:image.

For an e-commerce site, it may be more worthwhile to start with products that already receive multimodal impressions and generate a strong margin. For an agency or tradesperson, projects appearing in Lens may justify a new series of photos, more precise captions, or better-structured dedicated pages. For a software company, illustrated documentation may deserve a different approach if screenshots become a real entry point.

The multimodal filter is also available in generative AI reporting

Google also says the Web: Multimodal search type is available in the performance report for Search generative AI features. This report covers impressions in AI Overviews and AI Mode, among other things.

It is not interpreted exactly like the standard search report: generative AI reporting focuses mainly on impressions, with dimensions such as pages, countries, devices, and dates. This lets you see whether certain URLs appear in AI experiences following a multimodal search, without concluding that a particular visual triggered the citation or exposure.

For an advanced audit, it is therefore useful to compare three sets of data:

• Web: Text-based, for conventional text searches;
• Web: Multimodal, for searches initiated from an image;
• the generative AI report, also filtered to multimodal where available.

This comparison may reveal that a page performs well in text-based SEO but poorly in visual discovery, or, conversely, gains unexpected visibility through its images, diagrams, or screenshots.

A 30-day visual SEO action plan

Search Console’s main contribution is therefore not a new SEO formula, but a new level of visibility. Until now, teams could work on their images based on Google Images, manual Lens tests, or hypotheses. They can now isolate some multimodal searches in Google’s SEO measurement tool.

For a sound audit, the right sequence is simple: measure the pages, understand which content groups stand out, inspect their images, improve a sample, and then check whether visibility increases. It is less dramatic than a promise of “Lens SEO,” but far more useful when deciding where to invest your time.