Americans love TV, that mistress who gets to camp in the living room. We spend a quarter of our waking days with the TV on, maybe the biggest lode of human attention in the country. Advertisers pay for most of it — they know where you’ll be sitting, so they have ads meet you there. The industry standard way of determining TV ratings (and thus advertising rates) is through companies like Nielsen, which has continually measured programs for more than 50 years. But that seniority could be harming TV in the digital age: These companies simply haven’t figured enough of a distinct method to keep up with the way that a modern audience — that’s you — watches television. Can they keep up?
To be fair, Nielsen’s innovation has endured because it did a tricky thing. How to count the number of Americans tucked in dens watching Gunsmoke, chopping vegetables in kitchens while half-watching Wheel of Fortune, or wrassling in basements while Spongebob Squarepants plays? Nielsen invented the idea and the technology behind ratings measurement, then sells those figures to agencies and companies that buy ads. As Nielsen says [on its website]((http://www.nielsen.com/us/en/solutions/measurement/television.html), “That great lipstick ad you saw during your favorite reality show — that was no accident — it was informed by big data.”
Other industries track product consumption using fairly straightforward methods. Movies are ranked by box office grosses, digital music and e-books are tracked by download numbers. You watch a thing, you buy a thing, it gets tallied. But TV can be a strange bird.
Specifically, Nielsen sends metering technology to a sample size of about 25,000 households to record their viewing habits each day.
Particular households are chosen as representative stands-in for a large demographic of the TV-watching population. Households must also keep their status as Nielsen guinea pigs a secret, so as not to skew the metrics — they don’t want your cousin coming over to stuff the ballot box by watching nothing but Scandal. Nielsen continually gathers data, and reports the ratings to the press each Tuesday. That way, L’Oréal knows roughly how many people, of what sex and age, are watching Survivor: Cambodia.
As in any survey, a given sample won’t necessarily represent the whole. Even as albums, DVDs and e-books get pirated, they can still count a hard number of sales. Nielsen ratings, meanwhile, can only extrapolate from a sample. The company, though, says that its methods make its data reliable; by insisting that viewers keep logs, Nielsen is confident people are actually watching shows, rather than leaving the TV on. And by getting granular on the demographics, the company knows who is watching, not just how much.
But how much are people watching? It ain’t the ‘60s; families don’t gather to watch one of four channels before tucking the kids in. Nielsen is trying to keep up, branching out from linear TV into VOD, DVR, and over-the-top content devices like Apple TVs or Roku. “Today, viewing video is a personal and online experience,” the company’s site explains. “This media fragmentation is both a challenge and an opportunity.” Technology has melted down the very concept of “viewing” and stirred it into a range of interacting, downloading, replaying, remixing, supercutting, and generally experiencing content in ways that defy easy count.
I asked Brian Yamada, Chief Innovation Officer of [VML]((http://www.vml.com/), one of the biggest full service digital ad agencies in the world, about the shift from more traditional ad-buys to the future of predicting how to get you to consume. “The majority of the media we buy is very much digital, not traditional,” he told me. “We still have a few places where we are buying traditional advertising, but the percentage of that becomes less and less because the marketplace continues to evolve for a need that’s less and less.”
Yamada explained that the need to do away with old ad-buying practices, like buying time during TV broadcasts based on where you think audiences would be and then delivering a company’s message in front of them, has caused senior analytics companies to scramble. “Nielsen is trying to become more diversified and more digitally focused like comScore,” he told me, referring to the internet analytics company. He continued: “A company like Nielsen has to get broader to make sure that they’re keeping up with that and not just reporting on who is watching what TV show at what time. It’s what they’re consuming and when they’re consuming it.”
Yamada sees older bastions of analytics being forced to shift their focus from an outdated model of what he called measuring “exposure” into a model that measures “impact.”
“A lot of the broadcast components were focused on understanding who was exposed to something,” he said. “So it was very much about who was aware, in theory, about who was paying attention and didn’t get up to go to the fridge when the commercials were on.” Now the digital expansion of ads and ad space forces brands to look beyond whether an impression took place, but rather to who is interacting with an ad and in what way. Nielsen’s outdated tracking model doesn’t account for this, but Yamada sees it as a necessary goal if they want to stay vital. “No question they’re trying to move there,” he said. “You have to be able to track downstream.”
This is where normal TV analytics can seem to come up short. For example, Nielsen can’t extrapolate specific DVR information, and those numbers are based on the industry standard that sample households with Nielsen boxes watch whatever they DVR within a week of recording the show. Any of us with DVRs bursting with old shows know that isn’t necessarily the case.
You know whose habits can be tallied pretty well? Older people. You know whose shows are maybe getting canceled because you’re saving them and watching them at odd hours? Yours, quite possibly.
Nielsen has only just begun to track streaming shows in what is currently still a semi-beta-testing period. Earlier this year the company announced it finally developed a method to parlay their set-top system for network TV into following streaming habits. It involves clients providing audio signatures for streaming programs from subscription services that Nielsen isn’t able to specifically track. Some would say it’s a convoluted but clever work around of the disconnect, but insiders would just say it’s progress. Either way it allows clients an in with streaming services via tried and true analytics partners like Nielsen.
Hulu was already working with Nielsen, and understandably: a large chunk of its content is network television shows. Amazon is onboard with Nielsen as well. But Netflix has balked. The notoriously cagey streaming company opposed explaining its viewership via anything but subscriber numbers and total streaming hours sitewide. If the numbers show that Netflix’s original programming isn’t drawing the audience of traditional network shows, investors might be turned off. This would undermine Netflix’s ideology, of never pandering to advertisers for original programming, and hamper its creative freedoms.
Case in point is comedy genius Tina Fey, whose show Unbreakable Kimmy Schmidt was dropped by her longtime network home NBC and resurrected on Netflix. When asked how Netflix compared to network TV, Tina Fey said, “…we don’t have any actual numbers … it’s very free to be free of that rating system, for sure.” Those “numbers” she’s referring to are a jab at stodgy data-tracking that could inspire Fey and other show creators to head to alternate channels of artistic semi-freedom like Netflix or HBO, which launched its own streaming service earlier this year. Netflix knows its numbers, but viewers and advertisers know shows only by their quality.
Online video viewership is still relatively small compared to live TV. But the push towards digital that Yamada sees taking over explains the need for companies like Nielsen and advertisers to move forward. Nielsen has dominated its niche since 1950, and understandably the future looks like a disruption. But streaming services are taking off while cable is shrinking. By tying itself so closely to cable, Nielsen eventually stands to go out with it when you inevitably cut the cord. Maybe that’s precisely what needs to happen, or maybe they’ll manage to keep their spot on top.
*Corrections 10/20/15: In the original version of this article, it was stated that Nielsen uses a sample size of 5,000 homes, when, in fact, the sample size is 25,000 homes. Also, no third party vets Nielsen’s data, and they publish ratings on Tuesdays to release to the press. An error in a previous version of this article incorrectly stated that Nielsen can only track DVR viewership for seven days, when, in fact, the TV industry chooses to cap these weekly ratings.*
You've read that, now watch this: "The Netflix Algorithm Is Killing Genre Bias, New Data Shows"