The babysitter is alone in the house. The kids are asleep. The phone rings.
“Have you checked the children?”
She calls the police, and they trace it. Then the detective calls back with the scariest line in horror-movie history:
“The call is coming from inside the house.”
I’ve been thinking about that babysitter this week, mostly because of Netflix.
To be clear, Netflix is not a company in crisis. It has more than 325 million paid memberships. Last quarter it grew revenue 12% to $12.9 billion and posted a 34.6% operating margin. It’s the most successful subscription business in entertainment history.
Netflix reported third-quarter results on October 20, and the core retention numbers were good. Churn improved year over year, both globally and in the United States, and came in better than the company’s own forecast. Co-CEO Greg Peters pointed to retention as the clearest evidence that members are getting value.
But something strange is happening.
Netflix’s share of US TV viewing time slipped to 7.8% this summer. The stock fell 14% in September alone. Wells Fargo expects hours watched for Netflix’s top 100 originals to drop 21%. And on the call, management conceded what the numbers already show. US viewing hours are growing modestly, not as fast as the company wants, and US content execution needs to improve. Co-CEO Ted Sarandos, to his credit, said it plainly: the company isn’t growing as fast as he wants.
The suspects outside are easy to spot. YouTube now has almost double Netflix’s share of TV viewing. The hits dried up, and Netflix had its weakest Emmys in a decade. Live sports bring in signups but, as Sarandos admits, not much viewing.
But I think there’s one more problem, and it’s coming from inside the house.
Netflix may be the most data-driven media company in history. It knows exactly how many people watched, for how long, and where they stopped. And it acts on that data with brutal consistency. Perhaps too consistently.
The Boroughs, a sci-fi series from the Duffer Brothers, lost 61% of its audience in its third week. Cancelled, even with a Season 2 writers’ room already open. Ransom Canyon dropped 43% in first-week views from Season 1 to Season 2. Cancelled.
Now, each of those decisions is defensible. Seasons usually get more expensive over time, and paying more for a shrinking audience is nobody’s idea of discipline.
But what happens when you make that decision a hundred times?
Call it the cancellation tax. Viewers have noticed that Netflix shows often end after a season or two, and plenty of them now say they wait to see whether a show survives before starting it. If enough people do that, fewer start, the week-three numbers drop, and the show gets cancelled. Which proves the viewers right.
I can’t prove how big that tax is. But I’d bet it makes every other problem a little worse.
Judge every show on its first few weeks, and you’re running your service like a Windows 97 launch, or a Marvel movie premiere. But a subscription isn’t a series of product releases over time. Imagine trying to run a software company that kept every failed feature iteration you’ve made over the past decade a) actively online and b) accessible through a train wreck of a central dashboard! Talk about a haunted house!
Maybe this is simply a UI issue. Maybe it’s the inevitable consequence of a broad portfolio approach to entertainment. But the answer might lie in making fewer crazy expensive swings, focusing on a discrete number of audiences, and actively developing those cohorts by investing in stories that may not explode out of the gates, but grow steadily over time.
Look, it’s entertainment. It’s always been a tricky business. In the words of the screenwriter William Goldman: “Nobody knows anything. Not one person in the entire motion picture field knows for a certainty what’s going to work. Every time out it’s a guess and, if you’re lucky, an educated one.”
But right now it feels like Netflix is applying the worst aspects of Hollywood economics (boom or bust, yea or nay) to a subscription business (steady growth over time based on winning customer trust).
If I were in that house, I wouldn’t just stare at the phone.
I’d go check on the children.
