// AI NEWS · PLATFORM

Ziff Davis and People Inc. Quantify the AI Search Traffic Hit in Q3 Earnings — Google Referrals Down 37%, Core Sessions Down 26%

In early-November 2026 earnings calls, two of the largest US digital publishers put hard numbers on AI search: Ziff Davis (CNET, IGN, PCMag) said organic Google referrals fell about 37% year over year and People Inc. (formerly Dotdash Meredith) said core web sessions fell 26% — and both told investors they are no longer planning on a search recovery.

2026-11-09 · by Moe Ameen

What happened

Two of the largest US digital publishers used their third-quarter 2026 earnings calls to put concrete numbers on something creators have felt for two years: Google's AI search features are shrinking the referral traffic that content businesses were built on. On its call reported November 6, 2026, Ziff Davis CEO Vivek Shah said organic Google referrals to CNET, IGN, PCMag, and the company's other properties fell about 37% year over year in the quarter — an acceleration from a roughly 23% decline a year earlier. Shah told analysts the company is "not planning on a search recovery." Q3 revenue, reported November 5, fell 5.6% to $289.3 million.

People Inc. — the publisher formerly known as Dotdash Meredith, owner of People, Allrecipes, Investopedia, and Verywell — reported results November 2 for the quarter ended September 30, 2026. Core web sessions fell 26% year over year, to 1.61 billion from 2.17 billion, which the company attributed primarily to AI-oriented search and content-discovery experiences. On the call, CEO Neil Vogel said "we've lost nearly 70% of our Google search traffic over the last few years" and that AI Overviews are "now over 90% of the queries that used to send us traffic from search." The company's earnings presentation put AI Overviews penetration at 82% for the quarter.

The disclosures also introduced a reframing worth watching. Ziff Davis reported a metric it calls AI Impressions — drawn from Google Search Console, it counts how many times a link to one of its pages appeared inside a generative AI feature — and said its content was cited roughly 2.7 billion times in the quarter, which Shah described as nearly ten times the clicks those same sites received from Google Search. The comparison sets AI citations against all Google Search clicks, not just AI-driven ones, so it measures visibility, not visits: being cited inside an answer is not the same as a reader loading your page. Treat the specific percentages as company-reported figures for these publishers' own networks, which skew toward news and reference content; they are directional signals of a broad shift, not universal rates. Confirm current numbers in each company's investor materials before planning around them.

Why it matters for creators

  • The biggest SEO operations in media have stopped trying to win search back. When a publisher with a dedicated optimization team tells investors it is "not planning on a search recovery," out-optimizing AI Overviews is not a strategy a smaller creator should bet on either.
  • Citations are not clicks. Ziff Davis being cited 2.7 billion times while its referral clicks fell 37% is the whole shift in one data point — appearing inside an AI answer is brand exposure, not traffic, so being the name a reader already recognizes is what converts a citation into a visit.
  • The damage lands on informational and reference content first. The "what is / how do I / best for" queries that AI Overviews answer in place are exactly the top-of-funnel searches many creators built their reach on; branded and transactional queries hold up better.
  • The audience did not vanish, the front door did. People Inc. and Ziff Davis are not out of readers; they are losing the discovery channel that used to introduce readers to them — which makes owned audience and in-feed presence the parts of the funnel still under your control.
  • Scale is no longer the moat it was. Billion-session catalogs are shrinking faster each year, so the edge shifts to a recognizable voice and consistent presence across many surfaces rather than a large library ranking on one that is giving way.

How to act on this with Kompozy

The tell in these calls is not the percentages — it is that two publishers with real SEO teams publicly conceded the search game and reframed success around being *cited* rather than *visited*. For an independent creator, that reframing is the instruction: stop competing for a click an AI summary now intercepts, and start building the one asset the summary cannot take — a name people recognize and seek out, backed by channels they reach without a results page in the middle. Ziff Davis's own numbers make the point: 2.7 billion citations next to collapsing referrals means visibility is cheap and recognition is the scarce thing. The reader who already knows your face and voice is the one who clicks through an AI answer, forwards your post, or opens your email — and that recognition is a product of showing up consistently, everywhere, in a style that is unmistakably yours.

That is the job [Kompozy](/) is built around. A single [Persona Brief](/glossary/persona-brief) fixes your voice, phrasing, and a recurring face across everything, so each piece reinforces the same recognizable identity instead of reading like anonymous SEO filler — the exact kind of commodity content these AI features summarize and discard. From one source it builds the two halves of a post-search strategy at once: the owned, citable depth AI engines pull from (a Blog Article and an Email Newsletter that lands in an inbox no summary can reroute) and the in-feed presence that builds recognition directly (captioned [Persona Shorts](/glossary/persona-shorts), [Clipped Shorts](/glossary/clipped-short) cut from long-form video, and brand-exact [Carousels](/glossary/hyperframes)). [Autopilot](/glossary/autopilot) then schedules and publishes the set across the eight social platforms plus blog and email behind a per-post review gate. The honest boundary: Kompozy will not reverse an AI Overview or restore a Google click — nothing a creator does will. What it changes is whether, when your name shows up in that answer, there is already an audience that recognizes it.

Quick takeaways

  • Ziff Davis (CNET, IGN, PCMag) told analysts on November 6, 2026 that organic Google referrals fell about 37% year over year in Q3, up from a ~23% drop a year earlier, and that it is "not planning on a search recovery."
  • People Inc. (formerly Dotdash Meredith) reported Q3 core web sessions down 26% year over year — 1.61 billion from 2.17 billion — with CEO Neil Vogel saying the company has lost nearly 70% of its Google search traffic over recent years.
  • Ziff Davis reported a new AI Impressions metric: its content was cited inside Google generative-AI features roughly 2.7 billion times in the quarter, which it framed as nearly 10x its Google Search clicks — visibility, not visits.
  • The through-line: citations are rising while referral clicks fall, so recognition and owned audience, not search rankings, are the assets still under a creator's control.

Frequently asked questions

How much did AI search cut publisher traffic in Q3 2026?

By their own Q3 2026 disclosures, Ziff Davis said organic Google referrals to CNET, IGN, PCMag and its other sites fell about 37% year over year (after a ~23% drop the prior year), and People Inc., formerly Dotdash Meredith, said core web sessions fell 26% year over year to 1.61 billion from 2.17 billion. Both attributed the declines largely to AI Overviews and AI-driven search. These are company-reported figures for their own networks, which skew toward news and reference content.

What is the "AI Impressions" metric Ziff Davis reported?

AI Impressions, drawn from Google Search Console, counts how many times a link to one of a site's pages appeared inside a generative AI feature such as AI Overviews or AI Mode. Ziff Davis said its content was cited roughly 2.7 billion times in Q3 2026 — which it described as nearly ten times the clicks its sites received from Google Search. An impression is an appearance, not a click, so the metric measures visibility rather than traffic to the site.

Why do publishers say they are not planning on a search recovery?

Because the decline has accelerated rather than stabilized — Ziff Davis's ~37% drop followed a ~23% drop the year before, and People Inc. said AI Overviews now cover over 90% of the queries that used to send it traffic. With Google answering more questions on its own results page, these publishers are treating reduced search referrals as the new baseline and shifting toward owned audience and other channels instead of expecting Google clicks to return.

What should a creator do about falling AI search traffic?

Build distribution that does not depend on a search click: a recognizable brand voice and consistent presence in the social feeds, plus owned channels like email and a blog that an AI summary cannot intercept. A content engine like Kompozy turns one source into on-brand shorts, carousels, a blog, and a newsletter under a single Persona Brief, then schedules and publishes them across the eight social platforms plus blog and email.

Related news

← All AI news · Get started →