Google is experimenting with how many shopping products fit across a row in Search. The reported variations of the “More products” module show three larger cards or a denser arrangement of four or five products. For ecommerce teams, the interesting question is how presentation affects the opportunity to be noticed, even when a product’s position in an ordered list stays the same.
What the reported test shows
Search Engine Roundtable reported the experiment on October 1, 2026, citing screenshots and a video posted by Khushal Bherwani. The material documents different row densities and a grid that expands or contracts. Bherwani’s accompanying description associates one three-card variation with clicking a product.
This is an observed interface test. The report does not establish a general rollout, eligibility rules, affected markets or a change to Google’s ranking systems. It also provides no measurement of merchant clicks or conversions. Three to five refers to products across a row, rather than the total number of products available in the module.
Product position and screen exposure are different measures
Consider a hypothetical product in fourth position. In a five-column arrangement, it could appear in the first row. In a three-column arrangement, it would move into the second row while retaining fourth position. Whether that puts it outside the visible screen depends on the device, the surrounding page and the height of the cards. This example illustrates a possible layout effect; it is not a measured outcome of the reported experiment.
Larger cards can also give each product more room for its image and supporting information, while reducing the number of alternatives presented side by side. A denser row creates a different comparison environment. Neither arrangement is automatically better for every merchant: prominence, competing offers and shopper intent all matter.
How ecommerce teams can investigate a visibility change
When performance changes, record the interface context alongside the query and product position. Useful observations include device, viewport size, products per row, card dimensions and whether the shopper has opened a product. A screenshot taken before interaction may describe a different exposure opportunity from the screen shown afterwards.
Use these observations to form hypotheses, then compare them with available performance data. A fall in clicks alone cannot identify a layout test as its cause. Demand, pricing, availability, competing listings and other changes can produce similar patterns. Repeated observations under comparable conditions are more useful than a single screenshot.
The implication for SEO and AI Search
For shopping SEO, this experiment makes a practical measurement question visible: where does an eligible product actually appear on the screen? Tracking position remains useful, but an ordered result and a rendered shopping card describe different aspects of discovery.
The connection to AI Search is an editorial inference. This report does not document an AI feature. However, teams evaluating any interface that assembles product recommendations can apply the same distinction between selection, ordering and presentation. Being included in a recommendation set does not explain how much attention an individual item receives.
For now, merchants should treat dynamic grid density as an interface variable to monitor. The reported test is a reason to examine exposure more carefully, without assuming a ranking update or a proven impact on sales.