At Goldman Sachs Communicopia on September 9, Neil Vogel said the number out loud: twenty percent. Google's share of People Inc's business, down from more than seventy percent five years ago. He said it the way you say something you've been waiting to say.
He'd said a version of it before. At a Press Gazette interview, it was "below thirty percent of traffic." In a crawler-blocking discussion, "high-twenties percent." Three venues, three numbers, three units. Traffic. Business. Revenue. Each definition produces a different figure. But they all point in the same direction: down, and fast.
The question is whether the industry has any context for what "twenty percent" means.
The Invoice
It does. Similarweb's July 2026 data provides the receipt.
The industry is spending $1.36 billion a year buying back traffic it used to get for free. Forbes alone accounts for nearly two-thirds of the top-100 total — $72.2 million in a single month, up eightfold in three years — while its organic traffic fell 26.7%. The New York Times more than doubled its paid search budget. CNN and USA Today lost a quarter of their organic reach. Spending is growing faster than the traffic it buys, which means the auction is getting more expensive even as the inventory shrinks.
People Inc isn't in the dataset. Not as a large spender. Not as a small one. Absent.
Three Readings
The first reading is escape velocity. People Inc genuinely diversified. Non-session revenue — Apple News, AI licensing, social, events, D/Cipher — grew 16% in Q2 and now represents more than 40% of the business. The October subscription bundle (People premium: exclusive content, live celebrity chats, games, video) will push that further. Southern Living Insiders launched at $89/year. MyRecipes has 4.5 million registered users. The Netflix content partnership spans six markets. None of these revenue streams need Google to send a single click. When Vogel says twenty percent, the math checks out.
The second reading is accounting. "Twenty percent of our business" and "below thirty percent of our traffic" and "high-twenties percent" are three statements that can all be true simultaneously while describing different things. If Google-referred sessions monetize worse than direct or social traffic — plausible, since brand-loyal readers engage longer — then twenty percent of revenue can coexist with twenty-eight percent of traffic. The number is real, but it's the most flattering version of a family of numbers.
The third reading is substitution. People Inc has successfully reduced its dependency on the platform that was destroying it. This is genuinely impressive. But sixty-eight percent of Google searches now end without a click. The AI summary ate the traffic. And the next layer — agentic commerce, where software agents buy ads and transact without human sessions — has zero People Inc presence. D/Cipher isn't in the AAMP registry. Wasn't at Cannes. Didn't ship an MCP server. Celebrating twenty percent Google dependency while maintaining zero percent agentic presence is swapping one structural risk for a newer one.
The Divergence
What makes this data genuinely unusual isn't that People Inc claims low Google dependency. Every publisher says that. It's that the organic traffic numbers confirm it — people.com is up 27% while the industry bleeds.
The reason is content type. Celebrity news, entertainment coverage, human-interest stories — these are queries that AI Overviews handle poorly. You can summarize a recipe. You can extract a fact from a how-to guide. You can't summarize whether Jennifer Aniston and Adam Sandler are actually friends. The answer requires voice, judgment, sourcing. It requires being People.
Forbes built a business on the kind of content that AI can reproduce. People Inc built a business on the kind of content that AI can't. The search tax falls heaviest on the substitutable.
"We are producing fifty percent more content at the same cost."
— Neil Vogel, Goldman Sachs Communicopia, September 9, 2026
Fifty percent more content at the same cost. With AI. Which means People Inc is using AI to produce content while being protected from AI's consumption of its traffic — because the content it produces is the kind AI can't replace. There's a word for that kind of positioning, and it's the one investors keep circling without quite landing on: moat.
PPLI closed last week at $38.05. Analyst consensus is $57.30. The gap between what the stock says and what the business says has never been wider. But the stock isn't pricing the publishing business. It's pricing the MGM deal at 78 cents on the dollar, and treating everything else as free.
Twenty percent might be the most important number People Inc has produced since the Dotdash Meredith merger. Whether it's an achievement or an alibi depends on what you think the next tax will be.