Visual search is now a shopping channel, and most product pages aren't ready

Somewhere in the last couple of years, pointing a phone camera at something stopped being a novelty and became a habit. You see a jacket on someone at a café, a lamp in a video, a pair of shoes in a photo a friend sent you, and instead of typing a description into Google, you just point and tap. Google Lens now handles close to 20 billion of these visual searches every month, and Google itself says around a fifth of them are shopping related. The feature has grown well beyond identifying plants and landmarks. You can now circle or tap an item directly on your screen, in a photo, a screenshot, even a paused video, and Lens will pull up price, reviews, and where to buy it. That's not a gimmick sitting in the corner of the app. It's a genuine browsing behaviour, and it's already at a scale most retailers haven't accounted for.

How the recognition actually works

It helps to understand what's happening technically, because it changes how you think about preparing for it. Lens isn't matching a typed phrase against your page text the way traditional search does. It's running computer vision over the image to identify shape, colour, material, and context, then matching that against an enormous index of product images and the structured data attached to them. The system is essentially asking what does this look like and what do we know about things that look like this, rather than which page contains these words. That distinction matters, because it means a product can be invisible to visual search even if the page it sits on ranks perfectly well in ordinary text search.

The groundwork that's easy to skip

A lot of what makes a product recognisable to Lens is basic, and a lot of retailers still get it wrong, which is exactly the gap that's pushing more Auckland businesses to bring in an award-winning SEO agency in Auckland to get the fundamentals sorted before visual search becomes unavoidable. Product photos need to be clear, well lit, and shot from multiple angles against a clean background, because a cluttered or busy image makes it genuinely harder for the model to isolate shape and colour correctly. Reused supplier stock photos, the same image sitting on twenty other sites, give the system nothing unique to match against your specific listing. Descriptive file names and proper alt text still matter here, not for keyword stuffing but because they're additional signal about what the image actually contains. And critically, accurate Product schema markup needs to be implemented properly, covering price, availability, and review data, because that's what actually populates the overlay Lens shows once it recognises the item. Get the schema wrong or leave it out, and even a correctly recognised product can surface with no useful information attached, which tends to kill the click before it happens.

Why page speed still decides what happens next

Getting recognised and getting tapped through is only the first half of the job. If the product page on the other end is slow or poorly optimised, that visit is largely wasted, and it also quietly damages how much Google trusts the page as a reliable source for that visual match going forward. Image compression still matters here in a very practical sense. Serving properly compressed WebP images rather than oversized JPEGs keeps load times down without sacrificing the image quality Lens needs to recognise the product in the first place. Core Web Vitals haven't gone anywhere just because visual search has arrived. If anything, they matter more, because the visit is starting from a highly specific, high-intent moment, and a slow page burns that intent almost immediately.

A local angle worth paying attention to

There's a local dimension to this too, and it's one Auckland retailers in particular are still catching up on. Someone spots a product in a local store window, in a friend's home, or in a piece of local content, and increasingly the instinct is to point a camera at it rather than search by name. Whether that search resolves to a nearby retailer depends on the same fundamentals: accurate product data, clear imagery, and local listing information that's actually connected to structured data rather than sitting in a separate, disconnected Google Business Profile. This is precisely the kind of technical audit a business rarely thinks to run until it's already losing that traffic to a competitor who has it sorted, and getting the image optimisation, schema implementation, and local data structure right isn't something that happens by accident.

Visual search stops being a novelty

None of this is speculative or far off. The volume is already there, the shopping intent behind it is already proven, and the technical requirements are well understood, they're just widely ignored. The gap between businesses that show up in visual search and those that don't isn't about budget or brand size. It comes down to whether the basics, clean product imagery, correct structured data, and a fast page underneath it, were ever actually put in place. As camera-first search keeps growing, that gap is only going to become more visible.