Google is removing the product carousel – what now decides whether products get seen

Google has almost completely removed organic product carousels in the EEA. From now on, visibility belongs to products whose data works everywhere: on Google, on comparison sites and in AI search.

What happens to a shop window when the street in front of it is moved overnight?

In mid-September, online retailers in Europe found out. Organic product carousels have almost completely disappeared from Google results across the European Economic Area. These were the popular-products blocks at the top of purchase-intent searches. They vanished within a few days, with no transition period.

For many retailers, this carousel was the most important free shop window on the web. The model was simple: feed products into Merchant Center, keep the data clean, collect the clicks. In Europe, that no longer works for now.

The obvious question is: how do we get back into the carousel? The more important question is a different one.

How do products stay visible when that one shop window is gone?

What exactly happened

The background is the Digital Markets Act. In July, the European Commission fined Google 460 million euros for favoring its own shopping, hotel and travel results over third-party services. On September 8, Google rolled out redesigned search results for the EEA. The company itself called them the biggest drop in quality in the 29-year history of its search engine.

Around September 15 and 16, the change reached product search. The analytics platform Productrise measured the share of queries showing a product carousel. In Germany, France, the Netherlands, Belgium and Sweden, it fell by 90 to 100 percent. In the UK, the US and Australia, nothing changed.

The carousel has been replaced by a block called “Comparison Sites”. It lists several price comparison services one below the other, with one of them expanded. Google itself appears there as one provider among several. Opening Google's entry brings back something resembling the old carousel. But the retailer names are missing from the tiles, and everything sits one click further away. Google has also removed the Shopping tab from the navigation.

From shop window to a chain of intermediaries

For the individual retailer, this means an extra layer now sits between the search and the product page. The click no longer goes straight to the shop but first to a comparison site. And that site in turn pulls product data from its own feeds.

An example: a mid-sized garden equipment retailer has placed its cordless lawnmowers in the carousel for years. The titles are optimized, the GTINs maintained, the prices current. One Monday morning in September, Merchant Center shows organic clicks collapsing. The products haven't got worse. They simply no longer appear in a spot where they always used to be. Whether they are now found via a comparison site or not at all depends on how well the data arrives there.

That is the real shift. Visibility no longer comes from a single place you can optimize. It is spread across many systems, and each one has its own requirements for product data.

Why this directly affects AI search

It would be easy to dismiss this as a pure SEO issue. That would be a mistake. The development shows where product search as a whole is heading: away from a list of many options, toward a few preselected answers.

That is exactly how AI search works. ChatGPT, Perplexity, Copilot or Google's AI Mode don't show ten carousel slots. They recommend a handful of products, often just one. Productrise measured in the US and the UK how many products AI Mode shows for the same queries. It is around 95 percent fewer than in classic search.

Fewer slots mean a higher bar. An AI agent doesn't pick the product with the strongest ranking signal. It picks the product whose attributes it understands and can match against the request. Say someone asks for a “quiet cordless mower for 400 square meters”. If our retailer's lawnmower lacks information on noise level, battery runtime or area coverage, it simply drops out.

One detail deserves attention: according to Productrise, Google's Shopping Graph currently appears not to be connected to AI Mode in the EEA. European AI product search therefore relies even more on what machines find directly on websites, in feeds and with third parties. Anyone supplying incomplete data there won't be recommended.

Data resilience instead of channel optimization

What remains when the surfaces keep changing? The data behind them.

We call this data resilience: the ability of product data to arrive complete and understandable in every system. It doesn't matter whether Google displays it, a comparison site lists it or an AI agent recommends it.

Resilient product data meets three conditions:

  • It is complete. Every attribute that matters for the purchase decision is there, not just the mandatory fields of a single channel.
  • It is structured. The data is machine-readable, typed and consistent across all products.
  • It is channel-ready. Every target platform gets the data in the format it actually processes.

Google itself points in this direction. In the EEA, websites can earn their own carousels through a new beta feature. The prerequisite is structured data, specifically ItemList and Product markup on category pages. Visibility thus moves from the Merchant Center feed back to your own website. And it rewards exactly those retailers whose product data is cleanly structured.

What ainavio takes care of

This is where ainavio comes in as an agentic PIM. Each AI agent handles its own task:

  • The Research Agent finds missing manufacturer and product information.
  • The Enrichment Agent adds attributes, titles and descriptions.
  • The Classification Agent assigns products to the right categories and proposes the attribute structure.
  • The Quality Agent detects gaps, duplicates and inconsistencies before a product goes out anywhere.
  • The Mapping Agent and Publishing Agent translate the golden record into the requirements of shops, marketplaces and other sales channels.

The result isn't a better spot in one particular carousel. It is a dataset that works wherever products are searched for, today and after the next change.

The shop window isn't the merchandise

Google will probably keep adjusting the EEA layout. The Commission's deadline is running, and much is still in flux. Maybe the carousel will return in another form. Maybe not.

For retailers, that changes little. Anyone who built their business around a single shop window is now facing a locked pane of glass. Anyone who invested in the merchandise behind it can put it in every new window. That means complete, structured, channel-ready product data.

Shop windows come and go. Good product data stays.

Björn Thomsen

Head of Marketing, ainavio

Björn Thomsen is Head of Marketing at ainavio, specializing in B2B SaaS, demand generation, marketing automation, and leveraging AI to scale modern marketing processes.

contact@ainavio.com
+49 (0) 2842 - 929987-3

Künstliche Intelligenz (KI)

From raw material to enriched content: The art and science of AI-based product onboarding

Erfahre mehr!
Künstliche Intelligenz (KI)

Behind the scenes of the AI workflow: How multiple LLMs are working together to deliver future-oriented product data

Erfahre mehr!
Künstliche Intelligenz (KI)

PIM and AI: Why artificial intelligence is essential in your product data strategy

Erfahre mehr!

Fangen Sie noch heute an!

Ihre Strategie. Ihr Tempo. Ihr Wachstum. Unsere Unterstützung – auf ganzer Linie.