The Map Nobody Had
Until recently, no brand could confidently state how much visibility they occupied within an answer generated by a language model. You could track how many clicks you received on Google, how many impressions on Meta, and how much you were spending per conversion in an affiliate channel. But when a consumer asks ChatGPT which running shoes to buy for asphalt, that trail disappears. This blindness is not a minor issue; it signals that the entire performance marketing measurement framework has been built on ground that no longer exists.
On March 25, 2026, Rakuten Advertising and Similarweb announced an integration connecting Similarweb's user behavior data directly with Rakuten's analytics platform. The stated goal is singular: to enable brands to measure their representation within AI-driven channels, particularly large language models, and connect that visibility to measurable performance outcomes. Nick Stamos, CEO of Rakuten Advertising, described it as "a new level of transparency that gives brands an edge in a rapidly changing landscape for brand discovery and engagement." Baruch Toledano, VP and General Manager of Digital Marketing Solutions at Similarweb, added that his company has been on the frontier of helping businesses understand their presence in these environments and the synergies between performance channels.
The announcement did not include investment figures, revenue projections, or prior impact metrics. However, the absence of numbers does not diminish the revelation of this move; it does the opposite.
Why This Agreement Matters Beyond the Press Release
What Rakuten and Similarweb are building is not a brand-new product from scratch. It is the repurposing of an existing capability towards a territory where conventional measurement instruments simply do not reach. Similarweb was already working with Rakuten for publisher verification, competitive analysis, and seasonal trend planning. What changes now is the direction of the telescope: instead of looking back to understand what happened in known channels, the instrument is aimed at a space where visibility rules are written in real-time.
There is concrete financial logic here. Brands that currently invest in content to establish themselves in traditional search engines are beginning to discover that an increasing fraction of their potential customers never clicks on anything; they receive an AI-generated response and make decisions based on it. If that response doesn't mention them, or mentions them unfavorably, their spending on content and affiliates does not compensate for the loss of relevance at the first point of contact. The cost structure of performance marketing remains the same, but the return begins to migrate toward a channel that no one had budgeted for.
Rakuten understands this. Its Innovation Labs program, previously launched within its Insights & Analytics portal, aimed to compress decision cycles under the group's commitment to achieve a 20% increase in efficiency for employees, teams, and clients. This integration with Similarweb is not an isolated bet: it's the next layer of an architecture that has been years in the making with data, artificial intelligence, and selective partnerships.
The Trap for Those Who Only Measure
This is where the analysis becomes uncomfortable for many marketing executives. Having visibility into how a brand appears within a language model is a necessary condition, but it is far from sufficient. The history of digital marketing is full of companies that invested in sophisticated measurement tools but continued to lose market share because they confused the ability to observe with the ability to act.
The operational question this agreement still does not answer is: once a brand knows it is represented poorly or insignificantly in the responses of a language model, what concrete lever does it pull to change that? The answer isn’t in the dashboard. It lies in the content architecture, in the quality of the sources that the models use to train their responses, and in the thematic authority that a brand builds over time in formats that models consider reliable. No measurement tool, no matter how sophisticated, can solve that problem on its own.
This is not a critique of Rakuten or Similarweb. It's a diagnosis of the market both companies serve. Brands that use this tool as an end in itself — to report in a committee that "they have visibility in AI" — will waste its value. Those that use it as a starting point to redesign their content strategy and positioning in sources that the models cite will be the ones to see a measurable return.
The initial launch to a select group of clients, with additional reporting functionalities coming soon, suggests that Rakuten is being deliberately careful with the learning curve. This is wise. Controlled scarcity in early access allows for iteration on real use cases before scaling, something that many analytics platforms have historically sacrificed in favor of speed to market.
The New Battleground Doesn’t Have the Same Rules
What this agreement strongly signals is not a technical solution. It is the confirmation that the field where brand visibility is gained or lost has shifted, and the tools that dominated the last fifteen years of performance marketing were not designed for that terrain.
Brands still optimizing click, impression, and conversion metrics in traditional channels are not doing something wrong. They are doing something incomplete. The consumer who reaches a point of sale after receiving a recommendation from a language model has already made a consideration decision before entering the funnel that brands know. Acting only within the funnel means losing half the conversation.
Rakuten and Similarweb are betting that brands will pay to understand that first half. Historically, every time marketing gained visibility over a new consumer touchpoint — from banners to tracking pixels, from SEO to multi-touch attribution — a marketplace of tools, agencies, and services was generated around that new layer. This will be no exception.
The leadership that matters right now does not involve replicating the measurement capabilities that all competitors already have. It consists of having the discipline to eliminate investment in metrics that no longer drive real decisions and to build intelligence about the space where the consumer forms their preferences before any ad intercepts them. The one who understands that first will not need to struggle for attention in saturated channels.
