AI Surfacing Is Messy: Publisher Visibility and Traffic Often Misalign

The Murky Waters of AI Visibility and Traffic

For instance, Digiday reports that Buzzstream, using the Xofu tool, analyzed 4 million AI citations across ChatGPT, Google's Gemini, AI Overviews, and AI Mode in January. Their findings showed news publications accounted for 14% of all citations, primarily in evaluative or decision-making queries. Conversely, Similarweb identified a different mix of top publishers, focusing on how often a brand name or source was mentioned, not just linked.

This lack of standardization means "nobody knows anything yet," according to Vince Nero, director of content marketing at Buzzstream. He draws a parallel to the early days of SEO, emphasizing that expertise in this new domain is still nascent. The immediate implication is that publishers cannot rely on a single metric to gauge their influence in the AI era.

One crucial distinction highlighted by Digiday is between "citations" (a direct link to a publisher's site) and "mentions" (referencing a brand name or source in an AI summary). Reuters, a wire service, was the only publisher to appear in analyses from Buzzstream, Similarweb, and Profound, suggesting a unique advantage for rapid, factual reporting in AI search. However, even for top publishers, AI visibility has been volatile over the past nine months, a fluctuation attributed by Adelle Kehoe, Similarweb's director of product marketing, to the news cycle and ongoing adjustments to large language models (LLMs).

Navigating the AI Paradox: Visibility vs. Value

The most striking revelation in the evolving AI landscape is the disconnect between high visibility in AI responses and actual referral traffic. While some publishers are frequently cited, the number of clicks they receive from LLMs can be surprisingly low. For example, Similarweb data indicates The New York Times receives the most AI referral traffic, despite ranking eighth in Similarweb’s AI visibility index. Conversely, Reuters, which ranks first in AI visibility, is ninth in terms of referral traffic. This demonstrates that being seen doesn't always translate into reader engagement on a publisher's site.

Publishers are, however, adapting. Alphonse Hardel, head of Reuters’ news agency, attributes their frequent AI mentions to the "speed, accuracy, and breadth" of their reporting. Reuters has also secured content licensing deals with major LLMs, including Microsoft and Meta, acknowledging that while AI citations are growing at a "very low base," they align with their mission of disseminating factual information. Hardel also highlighted that AI search helps them build business relationships with AI companies.

The Washington Post has taken a proactive approach, signing deals with OpenAI, Google, Amazon, and Perplexity, and hiring a head of AI discovery. According to Karl Wells, The Washington Post’s chief revenue officer, people referred from LLMs spend more time on their site and convert to subscriptions at a rate four to five times higher than traditional search referrals. This highly engaged audience is crucial for their advertising and subscription businesses. Meanwhile, Business Insider has reported year-over-year growth in visits and return visits, even as Google's AI Overviews have reduced traffic for many other publishers, underscoring the potential for strategic engagement.

The broader market trend suggests a shift away from traditional search as a primary traffic driver. Condé Nast CEO Roger Lynch anticipates a future where Google search is "no longer a meaningful driver" for their business, despite increased revenue. This perspective, along with the fact that LLMs currently drive only about 1% of overall web traffic, suggests that AI visibility is more about brand authority and direct content licensing opportunities than a new channel for mass referral traffic.