When the CBS television series Limitless premiered on a Tuesday night last September, 9.9 million people tuned in to check it out. In today’s TV world, which is nothing if not limitless—given the deluge of programming on everything from Netflix to cable to YouTube and myriad ways to consume it—10 million pairs of eyeballs for a new show is impressive.
And that was only the beginning. After factoring in about a month’s worth of DVR playback and video-on-demand viewing following the show’s premiere, plus streaming data from CBS’s website and app, the network’s total Limitless audience estimate jumped to 16 million. "That’s a big difference," says David Poltrack, chief research officer for CBS. "That’s the new way that television programs become successful. They don’t become successful just on network viewing, on live TV. They don’t become successful just on the Internet. It’s an accumulation of this different distribution. You need to measure everything."
It sounds so simple. Just add up all the numbers, right?
Hardly. The new math of TV ratings and audience measurement has never been more complicated. As online video platforms proliferate—and audiences scatter across smartphones, tablets, laptops, and connected-TV devices, such as Apple TV and Roku—getting an accurate read on viewership has become a Sisyphean task. There is no single, authoritative provider of data; networks have to cobble together sometimes contradictory information from an array of sources. What’s at stake: Without a full picture of a show’s audience, neither producers nor advertisers know the true value of what they’re selling and buying.
Nielsen, the ratings stalwart whose weekly reports have been deciding the life or death of TV shows for six decades, wants to ride to the digital rescue with a new "total audience measurement" that promises to account for viewing across all platforms. This effort, which has been rolling out for the past few months, may seem improbable considering the company has been claiming it could tame the digital chaos since at least 2007. That’s when Nielsen, at the behest of its industry partners, introduced the so-called C3 metric to factor in the ascendance of DVRs; it calculates the average viewership of ads during a show within three days of airing. Five years later, advertisers and agencies started embracing Nielsen’s C7 rating, targeting a week’s worth of views. But both figures are quickly becoming less relevant, as people watch shows weeks, even months, after they air. What’s more, Nielsen’s metrics have traditionally not looked at which programs are being watched over subscription streaming services, such as Netflix and Amazon. Since these platforms are commercial-free, they have no incentive to provide ratings—leaving a black hole of information on who is binge-watching Transparent or old episodes of Parenthood (data that’s useful to content producers, if not advertisers).
Nielsen’s success with total audience measurement would be nothing short of a reinvention for the company, which has become a punching bag for media executives who carp that its ratings systems are antiquated. "The idea that every Tuesday you get a rating for what happened on Monday night—one that only represents 60% of what the total audience is going to end up being—is a frustration for us," says CBS’s Poltrack. Linda Yaccarino, head of ad sales for NBCUniversal, went on a diatribe against Nielsen at the Consumer Electronics Show in January, dismissing the C3 rating as "practically useless."
These sorts of complaints—which Nielsen doesn’t necessarily disagree with—are what spurred the company to action six years ago, when total audience first went into development. Nielsen, after all, is a nearly $17 billion company, with ratings accounting for a significant part of its revenue. Steve Hasker, Nielsen’s COO, calls total audience "the biggest product development project that Nielsen has ever undertaken." Scores of engineers have been brought on to build plug-ins for the multitude of devices that Nielsen is now tracking and to create a software developer kit for media companies to install in their apps and online video players.
So what’s included? Nielsen says it can now measure all TV and digital platforms—Netflix, Yahoo, and YouTube among them. (Its work-around for the tight-lipped Netflix: track the audio fingerprints of some 6,000 individual episodes.) It can parse numbers for a show into metrics that are comparable across platforms, from views to total time spent on a program to average audience size. It can also break down the numbers by device. And thanks to a partnership with Facebook, it can deliver age and gender demographics for online video viewers, by correlating them with Facebook’s anonymized user-identification database of 157 million people in the U.S. alone.
Many in the television industry are rooting for Nielsen to succeed, particularly if its results counter the bleak refrain that TV ratings are on a downward spiral. It would, says Hasker, "explain to advertisers, to agencies, and to Wall Street that the viewership for my programming has gone up, not down." It also creates an opportunity for networks to monetize all those "extra" digital views that haven’t been counted.
The $70 billion advertising question is how marketers will respond once Nielsen’s numbers finally place television and digital audiences side by side and in a comparable format. Nielsen’s traditional metrics track the average audience size per minute of a show, while online video metrics have revolved around views. For a sense of the distortion this creates, recall the chest-thumping by Yahoo back in October, when it reported there were 33.6 million streams of an NFL game between the Buffalo Bills and the Jacksonville Jaguars. But only half that number were unique viewers, and once you spread them across an entire game, well, in TV terms, you end up with an audience of just 2.4 million.
Whether or not a perfect solution emerges, Nielsen’s new total audience measure seems most likely to succeed . . . in fanning the debate. "When marketers think about marketing, they’re not thinking about TV versus online versus whatever," says Brad Smallwood, VP of measurement and insights at Facebook, which last fall introduced an ad-buying product that provides Nielsen ratings for digital ads. "They just want to know how many people saw [their ad]."
Not all marketers agree. "To some degree, the normalizing of ratings across platforms makes sense, because it helps advertisers compare apples to apples," says Bre Rossetti, VP, director of strategy and innovation at Havas Media. "But it can also be dangerous." Rossetti argues that marketers instead need to start "trading in attention," or how intently people engage with their content.
One benefit of Nielsen’s effort is that it could give television studios added muscle in negotiations with streaming services. "My theory," says one veteran TV executive, "is that Netflix is driven by broadcast and cable shows" more than original content. While Netflix itself has data that answers that question, it hasn’t opened itself to outside measurement. (Though Netflix would not comment for this story, CEO Reed Hastings recently told investors: "Our titles are watched on the go and at home on a wide range of devices, making measurement of the viewing of any given title difficult for third parties.") But if Nielsen’s data does align with this exec’s theory, he contends, "Studios that sell to Netflix will be able to say, ‘Hey, wait a second here. What’s driving your subscribers are my shows, not yours. Your shows are the perception of what you are, but my shows are the reality.’ " That means the balance of power in this new world order could be disrupted—by one of the ultimate symbols of the old one.
A closer look at how we watch television today
Using its new total audience measurement tool, Nielsen broke down six ways people consumed a single episode of a network drama after it aired last September. It tracked viewers who watched live, as well
as those who used DVRs, video-on-demand services, connected TVs, and computers and mobile devices in the 35 days that followed. Across all demographics, 45% of viewers caught the show live (see chart, below). But when you look at different age groups, that figure varies dramatically, with 15% of 25- to 34-year-olds watching the show as it aired versus 64% of the 50-plus audience set. Surprisingly, the youngest viewers (18 to 24) were more likely to watch the episode live than digitally or via connected TV or VOD. Digital does have a hold on one age group: 18% of people ages 25 to 34 watched on a computer or mobile device. They were the only age group with more digital than live viewers—for now.
Total viewership of a single TV episode from September 2015:
A version of this article appeared in the April 2016 issue of Fast Company magazine.