Your Next Customer Won't Type a Word: SEO Is Going Visual

The short version

Search stopped being a text box a while ago

We have spent years talking about keywords, rankings, clicks and content. All of that still matters. But the way people start a search has quietly changed, and a lot of teams are going to overlook it.

Google Images itself exists because text was not enough. In its 25th anniversary post in July 2026, Google traced the product back to 2001, when people searching for a famous dress wanted to see it rather than read about it. Twenty five years later, the camera is a search box.

Source for the timeline: Google's 25th anniversary post on Google Images.

The volume is no longer niche. In October 2024, Google reported nearly 20 billion visual searches a month and said younger users aged 18 to 24 engage most with Lens. By December 2025, Google's own marketing research site put the figure at more than 25 billion.

Lens queries are one of the fastest growing types of Search, with more than 25 billion queries per month.

That line comes from Think with Google. If a fraction of your audience searched that way and you had no idea how your site performed for them, you would treat it as a blind spot. Until this week, that is exactly what it was.

What Google just gave you in Search Console

On September 24, Google announced multimodal performance reporting in Search Console. According to Search Engine Journal, the Web search type in the Performance report now splits into text-based traffic and a multimodal option, which Google describes as web results "where an image was used as part of the search."

You get impressions, clicks, click-through rate and average position, broken down by page, country and device. Google's stated goal, as Search Engine Land reported, is to help site owners "better understand your site's performance in the modern search landscape and optimize your visual content." The global rollout started on the day of the announcement, so some properties will see it later than others. We covered the launch alongside the rest of that week's news in our September 21 to 25 roundup.

Before you build a dashboard on it, know what the report cannot tell you.

Without queries, the page list is the insight. The pages earning multimodal clicks tell you which images Google already trusts, and the pages that should be earning them and are not become your to-do list. Our guide to building strategy from Search Console data covers how to fold a new report like this into a regular review.

Visual discovery is a buying signal

Visual search is not people idly identifying flowers. The same Think with Google research found that 1 in 5 Lens searches shows commercial intent. Someone photographing a lamp in a hotel lobby or a jacket on the street is often one step from buying it.

Google keeps building for that moment. In February 2026, it upgraded Circle to Search to identify several objects in one image, so a shopper can circle an entire outfit and get results for each piece. Google said Circle to Search now handles billions of queries a month and that shopping is among its top uses.

In AI Mode, the shopping layer is even deeper. Google's September 2025 update added visual results and a "visual search fan-out" technique that looks at secondary objects and subtle details, not just the main subject. Those results draw on a Shopping Graph of more than 50 billion product listings, with more than 2 billion refreshed every hour.

What this means. If your product photography is generic, the same stock shot every competitor uses, or missing from your structured data, visual search has nothing distinctive to match to you. Original images of your actual products, shown from the angles people photograph them in real life, are now a ranking asset.

AI search reads images too

The shift is not limited to Lens. Google said in April 2025 that AI Mode can understand the entire scene in a photo, including how objects relate to each other and their materials, colors, shapes and arrangements.

With Gemini's multimodal capabilities, AI Mode can understand the entire scene in an image, including the context of how objects relate to one another and their unique materials, colors, shapes and arrangements.

AI Mode then uses query fan-out, running multiple searches about the image and the objects in it. One photo becomes several searches, and every one of them is a chance for your page to be cited or skipped. In March 2026, Google expanded Search Live to more than 200 countries and territories, letting people point their camera at something and talk to Search about it in real time.

Visual content is also winning citations in AI answers. Comscore data reported by Search Engine Land found YouTube was the most cited source in AI answers for retail searches in June 2026, at 23%. That is a strong hint that AI systems value content they can see, not only content they can read. For the bigger picture on how discovery now spans many surfaces, see SEO is now a multi-surface system.

The questions every SEO team should be asking

The new report changes the questions worth asking in your next planning meeting.

If the honest answer to most of these is "not really," you are in good company. That is also why it is an opportunity. Many competitors will be slow to act on this.

How to make your images discoverable

None of this requires a new playbook. It requires applying the playbook Google already publishes, consistently. Google's image SEO best practices, last updated March 2, 2026, cover most of it.

Two lines in that documentation deserve a place on every developer's wall: "Google doesn't index CSS images," and "Google uses alt text along with computer vision algorithms and the contents of the page to understand the subject matter of the image." If your hero product shot is a CSS background, it does not exist to Google Images.

Google's alt text example is a useful gut check: "Dalmatian puppy playing fetch" beats "puppy," and a list of keywords stuffed into the attribute is worse than both. The same documentation says you can include images hosted on other domains, such as a CDN, in an image sitemap, provided you verify that domain in Search Console.

Two further signals are easy wins. The image metadata structured data guide explains how to add creator, credit and license details, which can earn a Licensable badge in Google Images and travels with the file when you use IPTC metadata. And the image SEO guide notes that the primaryImageOfPage property, the image property on the main entity, and the og:image tag all influence which image represents your page.

What to do this week

The fundamentals of SEO have not changed, but the front door has. When more than 25 billion searches a month can start with a camera, images stop being decoration and start being content. Treat them that way now, while most of your competitors are still only reading their keyword reports.