Publishers Shift to Licensing Models
Stuart Kearney, co-founder of Vetted, on AI agents in shopping
The center of gravity is shifting from referral economics to data economics. Review publishers that used to win by ranking first on Google are becoming upstream inputs for AI shopping agents, which means their durable asset is not pageviews alone but proprietary testing data, trusted editorial judgment, and a brand strong enough that platforms will pay to ingest and cite it. This is the same basic move Reddit made when it turned a widely used public corpus into a licensed data product.
-
The immediate pressure comes from distribution. In the interview, incidental SEO traffic is described as the part most likely to fade, while direct traffic remains with a small set of brands that actually buy products, test them, and build trust over time. That creates a split between audience businesses and data licensing businesses.
-
Reddit showed what a licensing path looks like in practice. Google announced an expanded Reddit partnership in February 2024, and OpenAI announced a Reddit partnership in May 2024 to bring Reddit content into ChatGPT and other products. The precedent is that high signal user generated or editorial content can be sold as structured input for AI systems.
-
The model is already spreading across publishers. OpenAI signed content partnerships with outlets like TIME in June 2024, and The New York Times reached an AI licensing deal with Amazon in May 2025, even as broader copyright fights continued elsewhere. That points to a future where premium review catalogs are monetized through a mix of subscriptions, commerce, and platform licensing.
The next step is that the best review brands and category experts become infrastructure for shopping agents. Instead of waiting for shoppers to land on a webpage, they will package test results, rankings, and category expertise into feeds, APIs, and licensing deals that power answers inside chatbots, assistants, search results, and merchant experiences.