Ziff Davis, one of the largest digital publishers in the world, has told investors that organic Google referrals to its properties fell about 37% year over year — and that it is not planning on a search recovery. The statement, from chief executive Vivek Shah on the company's third-quarter earnings call, is the bluntest admission yet from a major publisher that the old search traffic model is not coming back.
On the third-quarter earnings call, Shah said organic Google referrals to Ziff Davis properties fell about 37% year over year. The figure covers the company's portfolio of publishing properties, and it is a measure of the clicks Google actually sends — not impressions, not visibility, but people arriving on Ziff Davis sites from Google search. A drop of that size, across a portfolio that large, is not a blip or a single site's problem. It is a structural shift in how much traffic Google sends to publishers.
Just as striking as the number was the framing. Shah said the company is "not planning on a search recovery" — a deliberate choice of words that tells investors the company has stopped waiting for Google traffic to return to its old levels. That is a planning decision as much as a prediction: Ziff Davis is building its business on the assumption that the referrals it has lost are gone for good.
Shah also gave a number that explains where the traffic went. He put about 2.7 billion AI Overview and AI Mode impressions in Search Console for the quarter — the number of times Ziff Davis content appeared inside Google's AI answers — at nearly 10 times "the clicks those same sites received from Google Search" over the same period. In other words, Google's AI surfaces showed Ziff Davis content billions of times, and the clicks that came back were roughly a tenth of that exposure.
The gap between impressions and clicks is the defining fact of the AI search era. Being seen inside an AI answer is not the same as being visited, and the ratio Shah described — about ten impressions for every click — makes the point in hard numbers. For a publisher, an impression inside an AI Overview is exposure without a visitor: the brand may be visible, but the reader never arrives, never sees an ad, and never becomes a subscriber.
Ziff Davis is not a small niche site. It is a large, diversified publisher with the resources to adapt, and its experience is a leading indicator for the rest of the industry. When a company of that size reports a 37% drop in organic referrals and publicly abandons the expectation of recovery, it is telling the market that the change is permanent and industry-wide, not a temporary dip that better SEO will fix.
The statement also reframes the debate about AI Overviews. For months, the question has been whether AI answers hurt publisher traffic, and by how much. Shah's numbers answer it directly: billions of impressions, a fraction of that in clicks, and a 37% year-over-year fall in the referrals that used to sustain the business. The AI answer is not an addition to search traffic; for Ziff Davis, it is a replacement for it, and a much less generous one.
For SEOs, the message is that the metric that matters has changed. Ranking and impressions are no longer a reliable proxy for traffic, and a page can be highly visible inside AI answers while sending almost nobody to the site. The work now is to optimise for the click that still happens — the queries where users still visit, the answers where a citation is prominent enough to earn a click, and the content that gives a reader a reason to leave the answer and arrive on the page.
For link builders, the shift cuts both ways. On one hand, the value of a citation inside an AI answer has risen, because the citation is now often the only link a user ever sees — and being the cited source is the closest thing to a click that many queries still offer. On the other hand, the old model — rank, get the click, earn the link — is weaker when the click itself is rarer. The response is the same work, aimed more precisely: build pages that are genuinely the best answer, earn links that make them authoritative enough to be cited, and measure success by citations and brand visibility as well as by sessions.
Watch whether other large publishers report similar numbers on their own earnings calls, and whether any of them push back on Shah's framing by arguing a recovery is still possible. Also watch what Ziff Davis does next: a company that has stopped planning for a search recovery must be planning for something else, and its next moves — toward direct traffic, subscriptions, licensing, or new distribution — will show the rest of the industry what the post-search publisher looks like. For now, the number to remember is 37%, and the sentence to remember is that no recovery is being planned.
Sources: Search Engine Journal, SEO Pulse, November 14, 2026.
Related reading on this site: November Search Console Click Spikes May Be Bots, SEJ Reports, SEJ Pulse: Core Update, ChatGPT Business Hub and AI Mode's New Model, November Core Update Collides With Black Friday, and a Clean Read Will Not Come Until December, Core Update Volatility Started Days Before Google Announced It, AI Overviews Are Replacing the Brand Knowledge Panel, and Black Friday SEO Starts Eight Weeks Out, and November Is Already Late.
Linkbaits helps you create content that earns citations and backlinks. Start with the toolkit and build your first linkable asset this week.
Start free trial