Zero-Click Search Is Here. What Retailers Should Do Next.

Pull up your analytics platform and look at year-on-year organic search traffic for informational and navigational queries. For most UK retailers, it will not make pleasant reading.

Between 58% and 68% of Google searches now end without a click to any website. Google AI Overviews now appear on 48% of searches as of February 2026 (up 58% year-on-year), and when they do, the zero-click rate rises to 83%. Organic CTR falls from 1.62% to 0.61% when an AI Overview is present, per Seer Interactive's Q3 2025 study across 42 organisations and 25.1 million impressions.

This is not a temporary dip. This is the search interface completing a transformation that began with featured snippets around 2015, accelerated with voice search, and is now reaching its logical conclusion. The model where Google's job was to send users to websites is being replaced by a model where Google's job is to answer questions. Websites are one input into that answer, not the destination.

The Retail-Specific Picture

The consoling data point you will hear is that transactional queries ("buy Nike Air Max size 10", "order pizza delivery Manchester") are relatively protected from AI Overview encroachment because the AI cannot complete those transactions itself yet. That is broadly true. In January 2025, 91.3% of queries triggering AI Overviews were informational; by October, that had dropped to 57.1%, with commercial queries rising to around 19%, per Semrush's analysis of ten million keywords.

The trend is clearly toward AI Overviews expanding into commercial territory. The Universal Commerce Protocol that Google announced at NRF in January is explicitly designed to enable agentic completion of commercial queries within the Google interface. It is a US-led initiative, though Shopify's role as a co-developer makes it directly relevant to UK retailers on the platform. The transactional safety zone is also eroding, just more slowly.

For retailers whose content strategy has been built on dominating informational queries in their category (buying guides, comparison articles, "best X for Y" content that drives top-of-funnel awareness), the 2025 data is a direct challenge to that strategy's return. The traffic that informational content was meant to generate is increasingly not arriving.

The Silver Lining Worth Taking Seriously

There is one finding in the zero-click data that is genuinely useful: brands cited inside AI Overviews get 35% higher organic CTR and 91% higher paid CTR compared to brands present in AI Overview queries but not cited. (The caveat matters: Seer Interactive cannot prove citation causes the uplift. It may be that brands with stronger authority signals get both cited and clicked more. The direction of the relationship is real; the causation is not confirmed.)

What this means practically is that the game is not purely zero-sum. Being the source that the AI cites and attributes is meaningfully better than not being cited, even in a world where fewer people click through.

This shifts the strategic goal from "rank first on this query" to "be the authoritative source that the AI draws from when answering this category of question." Ranking required keyword density, backlink acquisition, and the other paraphernalia of traditional SEO. Being cited by AI systems requires genuinely authoritative, well-structured, factually accurate content. That is closer to old-fashioned editorial quality than to SEO mechanics.

There is a machine-readability angle to this as well. Content that AI systems can parse and confidently attribute gets cited more. Content that is vague, poorly structured, or contradicts other sources on the page gets cited less.

The Zero-Click Commerce Angle

Zero-click commerce is a slightly different but related problem: transactions completing inside AI interfaces without the user ever loading a merchant site. If the Universal Commerce Protocol achieves its vision and AI agents can complete purchases without visiting a product page, then the product page as a conversion mechanism is partially deprecated. The moment of influence shifts earlier in the journey, inside the AI interface.

This is worth thinking about as a medium-term structural shift rather than an immediate crisis. Agentic purchases are still a very small fraction of total ecommerce. The trajectory is consistent, though, and planning as if organic search traffic will recover to 2022 levels seems unwise.

What Actually Helps

A few things that are genuinely useful for retailers navigating this:

Stop measuring zero-click as a binary problem. The question is not "do we get the click or not?" but "what is our brand presence in the zero-click answer?" Being prominently and accurately cited in an AI Overview that does not generate a click is still brand exposure, still authority building, and still a pathway to the eventual transaction.

Audit your informational content for citation quality. Not keyword density. Actual quality, accuracy, and structure. Would an AI system confidently cite this as authoritative? Is the factual content clearly attributed? Is it free of the vague marketing language that AI systems tend to discount?

Invest in structured data. Adobe's analysis of retail site machine readability found that retail homepages average a machine-readability score of 75%, and individual product pages 66%. Roughly a quarter to a third of typical retail content is currently invisible to LLMs. Schema.org markup for product pages, articles, and FAQs is part of what closes that gap.

Think about direct relationship building as a counterweight: email, SMS, app push notifications, loyalty programmes. Channels that do not depend on Google sending you traffic. Tesco Clubcard, Boots Advantage, and M&S Sparks look considerably more strategically valuable in a zero-click world than they did when search was reliable.

The traffic crisis is real. It is also not uniformly distributed. Retailers with strong direct relationships and high brand recall are less exposed than those whose customer acquisition model has been primarily search-dependent. Whether you are in the first group or the second is worth checking.