Multimodal SEO: how do you measure visual search?
Photographing a product, circling an object on a screen or uploading an image to Google are increasingly common ways to start a search.
With Google Lens and Circle to Search, users can find information, compare products and reach websites based on what they see.
For SEO professionals, however, this behaviour has long been difficult to analyse using traditional reporting tools.
On 24 September 2026, Google introduced a dedicated multimodal search filter in Search Console, allowing website owners to distinguish searches made using images from those based on text.
This is an important development for understanding which pages are discovered through Google visual search and how that traffic contributes to an SEO strategy.
What is multimodal search on Google?
Multimodal search allows users to find information by combining different types of input, such as images and text.
Google Lens is a familiar example. Users can point their camera at an object or upload a photograph to find similar products, identify features or discover related content.
With Circle to Search, available on compatible Android devices, they can select an element directly on their screen to start searching.
These experiences change how search intent is expressed.
Someone searching for a jacket using a photograph might want to identify the model, find a retailer or compare colours and prices.
The search begins with a visual element, while the information needed to answer it may be found across images, written content and product pages.
We have already explored the evolution of search intent in our article on SEO and AI-powered search: what impact on keywords?.
Multimodal search adds another consideration: some users no longer begin their searches with words at all.
What's changing in Google Search Console?
On 24 September 2026, Google announced a new multimodal search filter in Search Console's Performance reports.
The update allows website owners to analyse web searches involving images, including those originating from:
- Google Lens: searches using photographs or a smartphone camera.
- Circle to Search: elements selected directly on compatible Android screens.
- Google Search: images uploaded to the search engine.
- Google Chrome: the Search this image feature available through the right-click menu.
The feature began rolling out globally, with data appearing when a property receives traffic from supported multimodal experiences.
The update also extends to the Generative AI performance report, although the available metrics and reporting purposes differ.
The main advantage is the ability to isolate a type of organic search visibility that previously lacked dedicated reporting.
How can you track Google Lens traffic in Search Console?
Accessing the new data is relatively straightforward.
- Open Google Search Console and select the relevant property.
- Navigate to Performance > Search results.
- Open the Search type filter.
- Select Web: multimodal, if available.
- Choose a date range and review the results.
The filter separates multimodal searches from Web: text-based, which covers searches initiated through traditional text queries.
You can also use 'Export' to download the data for analysis in other tools.
If the option is missing, the feature may not yet be available for the property, or the website may not have received relevant multimodal traffic.
It is also worth remembering that dedicated reporting was introduced in September 2026.
The new filter does not provide a complete historical record covering previous years, so long-term comparisons are currently limited.
Which metrics are available for multimodal SEO?
Within the Search results Performance report, the multimodal filter provides the familiar metrics used to analyse organic search.
These include:
- Clicks: how often users select a website result following a multimodal search.
- Impressions: how often website content appears in relevant search results.
- CTR: the percentage of impressions that generate a click.
- Average position: the average position of search results, calculated according to Search Console's reporting rules.
Data can also be examined by page, country, device and date range.
This makes it possible to identify which URLs receive visibility from image-based searches and compare their performance with traffic generated by text queries.
Why do pages matter more than keywords?
The new filter has one significant limitation: it does not display the queries used by searchers.
Because multimodal searches rely primarily on images, the usual query dimension is unavailable.
Website owners cannot see which photograph was submitted or which object triggered a particular impression.
They can, however, identify the pages appearing in these searches.
Imagine a furniture retailer discovering increasing multimodal visibility for product pages featuring table lamps.
That suggests those pages are appearing in searches initiated through images.
The retailer can then review the photographs, product descriptions, specifications and supporting information to identify potential improvements.
The report reveals which URLs are gaining visibility, without identifying the individual image responsible for a particular result.
Multimodal search vs Google Images: what's the difference?
An important distinction concerns the two search types available in Search Console.
Google Images measures how content performs in results displayed within Google's Images tab.
Web: multimodal measures web results generated when users include an image as part of their search input.
They therefore represent different search experiences.
One person might type a query and open Google Images to browse photographs. Another could use Google Lens to identify a product from a screenshot.
In the second scenario, the new filter helps analyse how web pages perform in the resulting visual search experience.
Combining these datasets without distinguishing between them could lead to misleading conclusions, particularly when comparing impressions, clicks and individual page performance.
For a comprehensive visual search SEO strategy, both reports can provide useful information, but they should be interpreted separately.
Multimodal SEO and AI Overviews: what can you measure?
The multimodal filter is also available in Search Console's Generative AI performance report, which covers AI Overviews and AI Mode.
Here, however, the measurement approach changes.
The report provides impression data for links to website content appearing within Google's generative AI experiences. It also distinguishes multimodal searches from those initiated through text.
It does not provide the same traffic and conversion information as the standard Search results report.
This distinction matters.
A page appearing in an AI-generated response indicates visibility within that experience, but does not automatically mean it has generated a website visit, enquiry or sale.
As discussed in our analysis of Search Console in the age of AI, these signals need to be interpreted alongside organic traffic and commercial performance.
How can you optimise images for Google Lens and visual search?
Better reporting helps identify opportunities, but no optimisation technique guarantees visibility in Google Lens.
Google's official image SEO best practices remain the most useful starting point.
For ecommerce and business websites, I would focus on four areas.
1. Image quality and relevance
Images should represent the content or product clearly.
For ecommerce, photographs showing product details, materials, variants and realistic usage scenarios can be particularly helpful.
Clear, representative photography makes it easier for users to recognise what they are looking for.
2. Alt text, filenames and surrounding content
Google uses several signals to understand visual content, including alternative text, filenames and surrounding page content.
Useful alt text describes what an image actually shows, avoiding unnecessary keyword repetition.
Headings, descriptions and the wider page content also provide important context.
3. Accessibility and indexing
Images must be accessible to Google's crawling and indexing systems.
This means checking supported HTML image elements, potential crawling restrictions, resource loading and mobile usability.
Performance matters too, particularly when large image files slow down product or content pages.
4. Product information and structured data
For ecommerce businesses, images should be supported by accurate product information covering prices, availability, variants and specifications.
Consistency between page content, structured data and Merchant Centre feeds helps Google interpret product information correctly.
We have explored these principles in our article on Google Merchant Centre and AI Shopping for AI Mode.
How should you use multimodal search data in your SEO strategy?
The real value of multimodal reporting emerges when the data informs practical decisions.
Start by identifying pages receiving impressions and clicks from visual searches, distinguishing between product categories, editorial articles and other content types.
Then compare their performance over time to identify opportunities for improvement.
For example, a page generating substantial multimodal impressions but relatively few clicks may deserve closer examination of its content relevance and presentation in search results.
However, CTR alone cannot explain the underlying cause.
The next step is connecting Search Console with analytics data.
Search Console reveals visibility and clicks, while GA4 can help evaluate what users do after reaching the website, including whether they complete meaningful conversions.
These figures should be assessed with their respective measurement and attribution methods in mind.
Finally, identify priority pages, make targeted improvements and monitor subsequent performance, accounting for available data volumes and changes in search demand.
Google visual search: what does it mean for SEO?
With the new multimodal filter, Google Search Console provides a dedicated way to measure searches that begin with images.
For ecommerce businesses, this can improve understanding of product discovery.
For publishers and corporate websites, it creates another opportunity to assess visibility through visual content.
The main limitation remains the absence of query data and information about the image submitted by the user.
Performance analysis should therefore begin with individual pages and their contribution to business objectives.
At RichClicks, we combine technical SEO, content strategy and performance analysis to help brands understand changing search behaviour.
Explore our SEO consultancy services and discover how we can help optimise your content for traditional, visual and multimodal search.
Giovanni Coppola
