:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/63010155/youmain.0.jpg)
In mid-January, a series of stories swept the entertainment industry press that suggested a seismic shift has happened in the way people watch TV and movies: Netflix started releasing viewership numbers for some of its biggest hits — and they were massive.
The service announced that 40 million accounts were projected to watch the dark stalker drama You and the British teen dramedy Sex Education over the course of the two shows’ first month of streaming. Note that it didn’t say 40 million viewers — it said 40 million accounts, which (Netflix hopes you assume) could translate to double or even triple that number of raw viewers.
The implications were huge: If Netflix’s projections were accurate — if two shows that seemed to be solid doubles for Netflix were, instead, massive home runs — the news was a game changer. What if the long-lamented monoculture, that world where everyone cared about the same stuff, had been hiding right under our noses all along? What if Netflix itself was our new monoculture, and all other pretenders to the throne needed to get out of the way?
But if you possess even an ounce of skepticism, I hope you can see why many TV journalists found Netflix’s numbers suspect. Both You and Sex Education appear to be performing well for Netflix. Both have been renewed for additional seasons, after all (though You’s was technically ordered by Lifetime before Netflix picked it up). But if 40 million people are watching them worldwide, that would put them among the biggest TV hits on earth, on the level of something like Game of Thrones, a legitimate global phenomenon. And just think about the culture around you. Does it feel like either of those shows is that big? Be honest.
As the weeks have passed since Netflix’s January announcement, a variety of sources have thrown cold water on the streaming platform’s self-reported numbers (more on that below). But you don’t even need to hear them out to be skeptical. You only need to take Netflix’s announcement at face value.
If you look closely, you’ll see that Netflix’s “40 million” figures are heavily manipulated
:format(webp):no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13720920/sexeducation.jpg)
Let’s start with something basic: Netflix almost certainly has some data points somewhere that can back up its “40 million accounts” claims. It made those claims in a letter to shareholders, and blatantly lying to shareholders would be asking for all sorts of legal headaches.
But even if we accept Netflix’s numbers completely at face value, the “40 million” number shouldn’t be compared to, say, The Big Bang Theory’s Nielsen ratings. After all, the Nielsen numbers for The Big Bang Theory are only for the United States, while Netflix’s numbers are worldwide. The population of the US is estimated to be just under 329 million, which would mean that if 40 million Americans have watched You, then a little over 12 percent of the country has seen the show. But 40 million viewers out of the entire global population of 7.7 billion doesn’t even hit 1 percent of the world’s population. My point: Cobbling together 40 million out of 7.7 billion is a lot easier than doing so from 329 million.
On top of that, anybody who writes about TV is well aware of how networks tend to poke and nudge at numbers to get them to say much more impressive things than they actually do. The most common way to do this is to count something called the total number of eyeballs that watch any portion of a program — which is to say, everybody who’s watched a couple of minutes of a program, even if they turn it off.
That’s how the Oscars arrive at their highly suspect claim that 1 billion people watch the ceremony annually, which they repeat every year, despite slumping ratings for the show most years. It is entirely possible that 1 billion people watched at least one minute of the Oscars in the past, especially in years when global phenomenons like Titanic or The Lord of the Rings won Best Picture. But there’s no way that 1 billion people around the world are regularly watching the entire ceremony.
The rise of streaming has made these sorts of numbers even harder to parse. It’s difficult to say, when HBO declares that 30 million people watch each new episode of Game of Thrones, where those numbers come from in total, because at least some portion of them comes from streaming services that HBO controls all of the data for. I don’t really doubt that 30 million people watch Game of Thrones each week — but I also know that beyond the numbers that are verifiably measured by Nielsen (an independent company), I can take anything additional with a grain of salt.
The numbers that Nielsen releases for TV shows are measured by “average audience” — which is to say, the average size of the audience spread out across the entire program. If 100 million people watch the start of a show but only 1 million are watching when it ends, its average audience will likely be somewhere around 50.5 million viewers. Nielsen also often releases breakdowns at the half-hour or even quarter-hour mark, the better to show how audiences taper or build throughout an event.
The Nielsens aren’t a perfect measure in this day and age. When they say that 12.07 million viewers watched the most recent Game of Thrones finale live, they’re really only counting viewers who watched the episode on TV that night, and missing many viewers who watched it live on streaming apps, or who watched it a couple of days later on DVR. (With that said, Nielsen does have a different measurement for people who catch up late, called L+3, L+7, and L+30 numbers, where the L means “live” and the “+X” means “plus viewership across this many days after the initial airing”).
But I can at least compare and contrast that live viewership number against other finales for other HBO shows and other TV shows more generally. It’s apples to apples. (If you’re interested in learning more about how TV ratings work, you can listen to a podcast about that very thing.)
What Netflix is trying to do is make what amounts to a number that counts everybody who’s checked something out for a bit seem like an average audience number. Indeed, Netflix’s shareholder letter admits this! It says that Netflix counts a “view” as an account that has watched 70 percent of one episode of a show. Would you equate “watching most of one episode” with being a “viewer” of that program? Netflix does, but would you?
I, for instance, am four episodes deep into the streamer’s Chilling Adventures of Sabrina and not particularly sure I want to continue watching (despite how much I love the show’s production design). Would you count me as a “viewer” of that program, when I’ve consumed less than 50 percent of the whole season that was available to me from day one? Netflix would, but I don’t think I would say I’ve “seen” Sabrina, except to say I’ve sampled enough to know it’s not really my jam.
Similarly, Netflix’s numbers are projections, projections that were released before either program had been streaming for a full month. (Indeed, as of this writing, Sex Education still hasn’t been out that long.) Even the most conservative projections can be horribly wrong — see also: Nate Silver on the 2016 election — and do you think Netflix is sharing its most conservative projections in a letter intended to impress its shareholders? I sure don’t. I think it’s putting up the biggest possible number it can coax out of its data without so blatantly manipulating that data as to open itself up to legal liability, because that’s what massive corporations do.
The thing is, Nielsen actually does measure Netflix viewership. It will admit that its methods aren’t exactly perfect just yet — Nielsen’s system, which uses audio signatures to determine which program you’re watching, has trouble differentiating between a movie being played on a streaming service and a movie being played on DVD, for example, and it still struggles to measure viewership on certain devices — but its numbers are at least consistent against each other and against its measurements for other TV providers in the US. And its average audience numbers for You (8 million across four weeks) and Sex Education (3 million across two) are ... quite a bit smaller than 40 million.
Neither of those numbers is bad. If a TV show were averaging between 3 million and 8 million viewers per episode on a broadcast network, it would probably be renewed without a second thought. But neither of those numbers is 40 million, either.
The story here is less about how many people are watching Netflix than about how Netflix is trying to position itself as a singular entertainment destination
:format(webp):no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13638091/BIRD_BOX_018.jpg)
The only reason I have the Nielsen numbers for You and Sex Education at the ready is that they were mentioned in a recent press conference, by FX Networks president John Landgraf — a persistent critic of Netflix who has been needling away at its refusal to make its data publicly known for years. So, yes, the Nielsen numbers for You and Sex Education were released by a frequent critic of Netflix who might not have done so had they been more reflective of Netflix’s initial projection of 40 million accounts for each series.
But his point that Nielsen and Netflix are painting very different pictures of the viewership for both of those series is still a smart one. While Netflix takes issue with the Nielsen numbers (despite the fact that they tend to be quite solid for Netflix), the network also has a tendency to renew shows that do well in the Nielsen ratings and to cancel those that don’t, outside of a few edge cases (which I talk more about here). Plus, the Nielsen numbers actually make it possible to usefully compare ratings for Netflix’s series to those for other networks’ TV shows. And even if the Nielsen numbers are off by as much as 25 percent — unlikely, as the Nielsen margin of error is probably more in the 10 percent range, but go with me here — that still means You is only being watched by 10 million people in the US. A good number but not a world-shattering one.
It’s also worth considering the context in which these numbers are appearing, which is first driven by practical concerns and then by more ethereal, hard-to-pin-down ones.
In terms of practicality, Netflix’s attempt to corner the market on its own data is an attempt to block out anybody else who claims to have viewership data for the service. Nielsen’s viewership measurements for Netflix aren’t yet perfect, but the company has made shocking strides in just the past two years, and it will very likely turn the corner toward being as accurate for Netflix series as it is for other shows within the next 18 to 24 months. If Netflix can start to introduce its own cherry-picked statistics that drastically differ from Nielsen’s, then it can create the impression that only it has access to the “true” data.
This has happened throughout TV history. In the pre-Sopranos days, HBO said it didn’t particularly care about Nielsen ratings, until it finally acquiesced to Nielsen’s measurements, and it’s not as though there aren’t plenty of other corporations in the world that insist independent statistics don’t reflect their internal data. Plus, since Netflix makes its money from subscribers, not ad revenue, it shouldn’t matter if some of its programs are less watched than others, at least in theory. It can simply cancel those that clearly don’t work, renew those that do, and make case-by-case determinations for the many in the vast middle.
But the more ethereal concern here explains Netflix’s need to boast about its massive viewership: It’s trying to assure a skittish entertainment industry that it’s the only game in town.
If you’re Bradley Cooper or Jennifer Lawrence or Jim Parsons and you’re trying to figure out what to follow up your most recent project with, you might not be inclined to do something with Netflix if its public image is that of a company that produces a lot of ultimately forgettable content that disappears into a massive void, as has long been the argument about Netflix’s movies in particular (at least before the debuts of the Oscar-nominated Roma and the apparently massive Bird Box). But if it’s such a massive service that even some of its more minor hits are being watched by 40 million accounts globally, well, that might make up for a lot.
The entertainment industry is, after all, one where cachet still counts. If Netflix can win a Best Picture Oscar for Roma, other people who want to win Oscars will be more likely to make a prestige picture for the service. And if it can convince enough of us that its shows attract so many viewers that they number among the biggest hits in the world, that perception will only help convince stars that appearing in the service’s movies and TV shows will bolster their profiles.
But all of Netflix’s image-boosting relies less on hard data than on selectively presented and interpreted numbers that are intended to argue a very specific point about Netflix’s place in the cultural firmament. And though that might seem like a Netflix-specific problem right now, we’re about to enter an era when more and more of our entertainment exists only on streaming services, when it might start to become very hard to get a sense of whether anything anybody says about their ratings is true. As with so many things in the entertainment industry in 2019, Netflix signifies the shape of things to come.
