Friday May 8, 2026, The New York Times published a guest essay by investigative journalist Julia Angwin with a title that deserves attention: “Meta is dying.” She emphasizes that Meta lost daily active users in Q1 2026going from 3.58 billion in the fourth quarter of 2025 to 3.56 billion.
Angwin sees it as the start of a long, slow decline, comparing the company’s trajectory to that of AOL in 2003 and Yahoo in 2015: technically alive, still profitable, but entering what she bluntly calls “the zombie era.”
Maybe she’s right. And if so, Theodore Levitt told us exactly why this would happen, 66 years ago.
The meta lesson never learned
In 1960, Theodore Levitt, professor at Harvard Business School, published “Marketing myopia“in the Harvard Business Review. His central argument was that businesses fail not because demand disappears, but because they define their business too narrowly. The railroads collapsed because they thought they were in the rail business rather than the transportation business. The tram manufacturers were replaced by automobiles of which they could have been the pioneers. “People don’t want a quarter-inch drill,” Levitt wrote. “They want a quarter-inch hole.”
Now let’s look at Meta’s six major pivots over 22 years and ask yourself: what industry did Mark Zuckerberg actually think he was working in?
In 2021, he said the answer was “the metaverse business” — a gamble whose Reality Labs division has since racked up about $80 billion in operating losses. Users disagreed. In 2023, it turned its attention to generative AI and has since committed more than $100 billion to building models that, as Angwin notes, currently perform worse than the competition. First-quarter 2026 results show record revenue of $56.3 billion, up 33% year-over-year, but also $33.44 billion in total costs, a 35% increase and AI spending prospects that have investors rattled.
Earnings look strong. The trajectory resembles that of a company that continues to move toward new product definitions while its core users quietly disengage.
What Traffic Data Really Shows
This is where opinion meets evidence, and the Similar Web Traffic for March 2026 is informative.
Google is the world leader with 86.9 billion monthly visits. YouTube follows with 29.3 billion. Facebook comes in third place with 11.9 billion and Instagram comes in fourth with 7.1 billion. This gap between Google and Facebook is the data equivalent of what Levitt described. Google has defined itself as being active in the sector of access to information. Facebook defines itself as being active in the social networking sector. One of these definitions evolves indefinitely. The other lacks space.
The AI category data is even more detailed. ChatGPT records 5.7 billion monthly visits worldwide, with year-over-year growth of 28.5%. Gemini is growing strongly at 283.8% year-on-year. Claude.ai jumped 423.7% to 613.7 million visits over one year.
Meta.ai does not appear in the top 100 most visited websites.
Meta spent $100 billion to enter the AI race. It’s not winning.
The compression game described by Angwin
When an aging platform’s user base begins to decline, the immediate response is almost always the same: monetize stronger. Angwin documents this clearly. Meta’s first-quarter ad impressions increased 19% year-over-year, while average ad prices increased 12%. Revenue per user jumped 27%. The company is running more ads on its platforms and charging advertisers more for each one.
This is the decision that maximizes short-term revenue while accelerating long-term decline. More ads means worse user experience. A worse experience means slower growth. Slower growth means that ad inventory eventually stops growing. Levitt described this as the trap companies fall into when they focus on making their current product a harder sell rather than understanding what customers actually need.
For digital marketers and SEO professionals, this creates a short-term concern. The advantage of Meta+ The ad suite delivers truly solid performance data: a return of $4.52 per dollar spent, or 22% higher than comparable manual campaigns, according to Meta’s own revenue reports. But these returns depend on a healthy and engaged user base, generating meaningful behavioral signals. If the user base contracts and advertising load increases simultaneously, signal quality degrades and performance follows.
The counterargument to take seriously
Angwin’s essay is compelling, but she is writing an opinion, not an analysis, and the full picture of the first trimester is more complicated than the term “dying” suggests. Year-over-year, Meta’s daily active user base grew another 4%. The quarter-over-quarter decline has a partially verifiable explanation in Iran’s Internet disruptions and WhatsApp ban in Russia. 33% revenue growth is not the profile of a company in terminal decline.
What this is about is the profile of a company that is spending at a scale that requires continued growth, while its investments in AI have yet to generate significant new revenue streams. Like the Wall Street JournalIt is Asa Fitch observed this week, “spending growth appears increasingly unsustainable.”
Levitt’s lesson is not that myopic companies always die quickly. AOL and Yahoo persisted for years. The lesson to be learned is that once a company loses the notion of its real activity, recovery becomes structurally difficult. Every dollar spent defending a bad definition is a dollar not spent understanding the customer.
The question Levitt would ask is not whether Meta is dying. The question is whether Meta has already clearly understood in which sector of activity it really operates. Through six pivots in 22 years, the answer seems to be: not consistently.
This uncertainty is now visible in traffic data. And traffic data doesn’t lie.
More resources:
Featured image: Roman Samborskyi/Shutterstock
