Reuters

If you live somewhere with easy access to Variety or an I-95 exit, it might be impossible to imagine finding somebody who hasn't heard of (or hasn't sat, bleary-eyed, ingesting the entirety of) House of Cards, the sleek and entertaining political drama on Netflix. According to YouGov's Brand Index, a rough measurement of pop culture attention, the quantifiable buzz around Netflix reached an all-time high last weekend, when the second season premiered.

But how many people actually watched the show?

Netflix doesn't share (and doesn't care about) live audiences, and neither do its advertisers, because there aren't any. So rather than rough Nielsen figures, we have to go by even rougher broadband analytics. But here's our best guess: "Anywhere from 6-10% of subscribers watched at least one episode of House of Cards," Procera Networks found, and in the U.S., "the average number of episodes watched during the weekend was three." Fascinatingly: There was no appreciable increase in Netflix’s overall traffic. 

Given that Netflix has just under 30 million domestic subscribers, that means that two to three million people watched House of Cards in its opening weekend. (A previous Procera estimate went as high as 16 percent of Netflix subs, or nearly 5 million.)

Two million, three million, five million people. Whatever the real number is, that's an impressive audience for a streaming network supposedly cultivating a long tail of entertainment. But is it an enormous audience for a supposed "hit" show? 

Approximately 115 million households own a TV and about 100 million pay for a cable subscription. So we should expect typical TV audiences to be higher than Netflix. NCIS, the Navy-police juggernaut that attracts its gray viewership like boomer catnip, draws an absurd 17 million viewers a week. But compare House of Cards to an average show on a broadcast network. According to Nielsen, CBS ended last year with an average prime-time delivery of 12 million nightly viewers. 

It's awkward to compare streaming estimates to Nielsen estimates, but it seems safe to say the average CBS program has at least twice times as many viewers as House of Cards.

There are a couple interesting points to make off this observation:

1. Popularity is weird. 

Popularity has become both easier to measure and harder to measure at the same time precisely because there are so many metrics. The most essayed-about show might be Girls. The most tweeted-about show is, statistically, Pretty Little Liars. The most talked-about, right now, is House of Cards. But the most popular show (which is barely essayed-about, rarely tweeted-about, and hardly talked-about) is NCIS, whose audience is literally as big as those three other shows—combined ... times two.

In a wonderful essay on this, our age of pluralist popularity, Adam Sternbergh wrote, "we’ve turned off Top 40 and loaded up Spotify; we’ve clicked away from NBC and fired up Netflix." Yes and no. "We"—Sternbergh, me, and everyone we know—might have clicked away from broadcast. But even stuck in what appears to be structural audience decline, CBS still pulls down ratings that make Netflix hits seem like quaint Acela-corridor niche series. The sliver of pop culture we've slid under the media microscope bears little relation to what's sampled by the rest of the country. 

2. Netflix and HBO are lucky...

Netflix and HBO need original programs, because their catalogue of movies and TV shows doesn't differentiate them sufficiently from Showtime or Amazon. But on a show-by-show basis, it actually does not matter how many people watch House of Cards (or True Detective). And, indeed, far fewer people do watch them than we'd imagine.

Broadcast networks, however, need massive audiences to watch their shows because they live and die off adjacent advertising, which is priced in viewers. Netflix and HBO aren't selling eyeballs. They're selling subscriptions. They don't need to obsessively target the throng assembling daily to gobble up crime procedurals and zombies, and it's no coincidence that the programs selected to please a small, educated audiences are celebrated by the small, educated TV writers who ignore what everybody else is watching.

3. ... broadcast is not lucky. 

The mega-cable bundle—that is, the regular cable package plus premium channels like HBO plus over-the-top services like Netflix—supports a pluralistic audience by allowing different channels to specialize in different audiences. According to conversations I've had with network executives, this has a trickle-down effect, as networks seeking specific audiences build specific reputations that become crystallized within the artistic community. Showrunners selling boomer-candy crime procedurals go straight to CBS; writers of juicy romantic dramas go directly to ABC; complex story-lines attract bids from AMC, HBO, and Netflix.

The outcome is sort of weird. Pop culture critics, who tend to be attracted to the thing that's most popular, mostly ignore the most popular shows on TV, which are lower-brow fare crafted to get high ratings. Meanwhile, a handful of networks whose business models rely on subscriptions rather than advertising amass all the most-talked-about shows on television. And that's how the people reading about TV and the people watching TV live in two separate worlds. 

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.