Warren Buffett made a fortune from The Washington Post because it earned money on quiet days. Half a century later, the Post built a business that only worked in a storm.
In brief
- The internet broke the old newspaper toll road. That alone does not explain why the Post is shrinking.
- The real error was budgeting for the 2020-21 news boom as if it were the new normal.
- Booms are bad teachers. The businesses that survive them build habits, not audiences.
On a January morning in 2021, somewhere far from Washington, a retired schoolteacher refreshes The Washington Post for the fourth time before breakfast. The Capitol has just been stormed, a pandemic is raging, and she is one of roughly 22.5 million people opening the paper that day. In a newsroom that has nearly doubled in eight years, every chart on every screen is pointing up.
Nobody in that room is asking the only question that matters: how many of these readers will still be here when nothing is happening?
A toll road on a slow day
The Post had once been very good at the answer. When Warren Buffett bought into The Washington Post Company in 1973 for about $10.6 million, he was buying something unusual. If you wanted to sell a house, hire a clerk or shift a sofa in Washington, the paper’s classified pages were the road to your buyer, and you paid the toll whether the front page carried a constitutional crisis or a pothole. Twenty years on, his stake was worth more than $400 million.
That is the real definition of a great business. It does not need anything to happen. Its customers turn up on dull Tuesdays because the product is woven into their routine, not because the world has caught fire.
The internet took the toll road away, and nobody should pretend the Post could have stopped Craigslist or Google. But that explains a slow decline across an entire industry. It does not explain why one paper, rescued and well funded by Jeff Bezos, was profitable for years and then lost more than $100 million in a single year.
Hiring for wedding season
Think of a caterer who hires ten extra cooks for wedding season and keeps all ten through the monsoon. The cooks are good. The food is excellent. The arithmetic is fatal.
That is roughly what happened. The Trump years and the pandemic produced the most compulsive news audience in living memory, and the Post built for it, adding journalists, desks and ambitions. Then the cycle cooled. Readers drifted back to their lives while the payroll stayed exactly where it was. By February 2026, a third of the staff had been let go.
Revenue is a forecast. Salaries are a promise. When the forecast misses, the promise still falls due on the first of the month. Economists call this operating leverage, and it works beautifully on the way up and brutally on the way down.
Why clever people keep doing this
In 1886, the statistician Francis Galton noticed that unusually tall parents tended to have children closer to average height. Extremes, he found, rarely repeat. We now call it regression to the mean, and it is the most ignored idea in corporate planning.
A boom hides it. In good times every decision gets an A: hire more and traffic rises, expand abroad and subscriptions rise. Nothing separates the wins that came from strategy from the wins that came from the weather, so managers take credit for the weather and budget for it again next year.
This is not a newspaper story. Peloton announced a new American factory during the pandemic boom and scrapped it within a year. India’s edtech companies fought talent wars for a classroom that lived on laptops, then shed thousands of jobs in 2022 when children went back to school. Different industries, same mistake: a spike treated as a floor.
The cross-current
The New York Times also roughly doubled its staff over the past decade, and it is thriving. The difference lies in what the hiring was attached to. People subscribe to the Times for the news and stay for the crossword, the recipes and the sport, none of which depends on Washington being on fire. The Post built a bigger audience for the fire.
Attention and habit look identical on a dashboard. They behave very differently under stress. When the Post dropped its presidential endorsement in October 2024, at least 250,000 readers cancelled within days. Nobody cancels a crossword in protest. An audience can leave in an afternoon; a habit has to be broken.
The Post now says it is on track to break even. That is real progress, but cutting your way to break-even is shelter from a storm, not a new climate. The harder job is to become a paper people open on a quiet morning.
Somewhere, that retired schoolteacher still reads the news. She simply stopped needing it every hour. The storm was the outlier. Her ordinary Tuesday was always the mean.