Multimodal Search: Google Lens, Circle to Search, and Visual Queries

Nearly 20 billion visual searches happen through Google Lens every month. That number, disclosed directly by Google, is easy to skip past because it’s not framed as dramatically as an AI Overviews headline — but it represents a genuinely enormous, mostly overlooked search surface that most SEO strategies don’t account for at all.

The actual scale of visual search

Google disclosed 12 billion monthly visual searches through Lens at I/O 2023. By its own disclosure around October 2024, that number had grown to nearly 20 billion monthly visual searches. In the same update, Google stated Lens has more than 1.5 billion monthly users, and that roughly 20% of Lens searches are shopping-related — someone pointing their camera at a product, or a photo of one, specifically to find it, compare it, or buy it.

That 20% shopping figure is worth sitting with on its own: at nearly 20 billion monthly searches, that implies roughly 4 billion monthly visual shopping searches happening through a single feature — a volume that’s easy to underestimate if your mental model of “search” is still built entirely around a text box.

Circle to Search: a different entry point into the same behavior

Circle to Search launched January 31, 2024, alongside Samsung’s Galaxy S24 series at their Unpacked event, and via a January 25, 2024 Pixel Feature Drop for Google Pixel 8 and 8 Pro devices. It lets someone circle, tap, or scribble on literally anything on their screen — a photo in an app, a video frame, an image in another app entirely — and search visually without switching apps. It’s since expanded beyond the original Samsung and Pixel launch to other Android manufacturers including OnePlus, Motorola, Xiaomi, and Tecno, as well as tablets.

Circle to Search removes the step of opening a dedicated search app entirely — the search starts from whatever’s already on screen, in whatever app the person happens to be using. That’s a meaningfully lower-friction entry point than Lens itself required.

Why this matters for a site that’s never thought about visual search

Most SEO effort goes toward text — titles, headings, body copy. Visual search surfaces reward a completely different set of signals, and most sites haven’t optimized for any of them:

Image quality and clarity, directly. A Lens or Circle to Search query is often matching against the actual visual content of your images — product photography, diagrams, screenshots — not just the alt text or filename around them. A blurry, low-resolution, or heavily stylized image is genuinely harder to match visually, independent of how well it’s tagged.
Multiple images per product, from different angles. Given how much of Lens’s volume is shopping-related, product pages with a single low-effort image are missing an entire class of query — someone photographing a similar item and searching for a match.
Alt text and surrounding context still matter, just differently. Even in a visual-first search, the text context around an image helps confirm and refine what the visual match actually represents — a well-written alt attribute isn’t obsolete, it’s a supporting signal rather than the primary one.
Structured data on product images (accurate Product schema, as covered by Merchant Center’s own requirements) reinforces what a visual match is actually looking at, connecting the image to specific, verifiable product facts rather than leaving Google to infer them purely from pixels.

A category of query text search simply can’t handle

The genuinely distinct value of visual search is that it captures an entire category of intent that never gets typed into a search box at all — someone who sees a chair in a café and wants to find it has no efficient way to describe it precisely enough in words to get a useful text-search result, but can photograph it and get a visual match instantly. That’s not a marginal improvement on text search; it’s a different entry point into search behavior entirely, which is part of why the volume has grown as quickly as Google’s own disclosed numbers suggest. Sites that only think about being found through typed queries are, by definition, invisible to this entire category of search intent, regardless of how well their text content is optimized.

What’s realistic to actually do about this

Visual search optimization isn’t a separate discipline requiring new tools most sites don’t have — it’s mostly a matter of taking image quality as seriously as text content, which a lot of sites still don’t:

Audit your primary images for genuine visual clarity — resolution, lighting, background — not just file size and page-speed impact. These are different quality dimensions and both matter, for different reasons.
For any product or reference content, include multiple images covering different angles or contexts, giving a visual-match system more surface area to work with.
Keep alt text accurate and specific rather than generic or keyword-stuffed — it’s a supporting signal for visual search, not the primary one, but it still needs to describe what’s actually in the image.
If you sell physical products, make sure Merchant Center feed images meet the actual quality specifications (covered in Google’s own Merchant Center image guidelines) — this is the same foundation that supports both traditional Shopping results and visual-match features like Lens.

Visual search doesn’t get the attention text-based AI search does, but the volume numbers Google has actually disclosed suggest it deserves more attention than it currently gets from most site owners. Almost 20 billion monthly searches, a fifth of them shopping-related, running through a feature most SEO checklists don’t even mention — that’s a real gap between where search behavior actually is and where most optimization effort is still pointed.