Why Nielsen Ratings Are Inaccurate, and Why They’ll Stay That Way
It’s easy to understand why people are ambivalent, or even hostile, toward ratings. The thinking goes something like this:
1. People love the shows that they love, a lot
2. Some of these shows get canceled “too soon”
3. They get canceled, almost invariably, because of what some mean old bully named Nelson said. Nielsen. I mean Nielsen.
To intensify your heartburn, the validity of these ratings is seemingly taken for granted. When you find out that your favorite show is getting canceled for having a 1.1 rating, there’s never any “but” — just the number. A network might keep a show around for prestige, or on the hope that it’ll grow, but that’s the executives’ call. There is no real defense against The Number.
And it feels terrible! Comedy fans know this better than anyone else, or perhaps they just suffer with the least grace. Talk to a diehard Arrested Development or Party Down fan, and you’ll hear anger; at the greater public for not watching the show, then at the network for canceling it, then, perhaps, at the people who provided the ratings in the first place.
It’s natural for a TV buff or comedy junkie to feel a bit of malice toward these ratings, and more specifically, toward the company that produces them. Nearly everyone has a passing familiarity with this old company — you know, Nielsen boxes, the Nielsen metrics, the Nielsen Families. And seeing as it’s an image that was first rendered many, many decades ago, it’s easy to resent, or even write off. It sounds foreign, especially to web-savvy viewer who watches as much comedy on his laptop as on his TV. I mean, do you know any “Nielsen Families?” Me neither.
A broader look at the company does little to stem a forsaken fanboy’s skepticism. “There are billion-dollar companies invested in preserving the Nielsen model,” New York Times media reporter Brian Stelter told me. “Nielsen’s clients are incumbents, protecting existing business models.” NBC President of Research and Media Development Alan Wurtzel was a bit more blunt: “Listen, Nielsen is a monopoly. They’re the only game in town. [Their ratings] are the only currency.”
Even more discomfiting, in the context of just definitely knowing that this fantastic show has like a hidden audience, because it must, is the idea that Nielsen isn’t solely concerned with measuring a show’s fan base. It’s not, and neither network nor advertiser seems to mind much.
“People don’t want to think about whether to data is true or not,” continued Stelter. “But as long as we all agree to go along with them, it’s relatively comfortable for the industry.” He recalled balking at the numbers for Obama’s inauguration, which claimed just 37.8 million people watched — fewer than watched Reagan’s three decades ago. This didn’t jibe with anything, really; personal anecdotes about how everyone seemed to have seen the speech were just anecdotes, but they were uniform and universal.
Nobody I talked to in the industry — be it on the ratings side or the advertising side — would dispute this story, or others like it. Some even granted that the numbers were probably off. Unanimously, though, they agreed on one thing: to focus on events like that is to miss the point.
How Ratings Work
Nielsen’s ratings, in the most basic sense, are like a poll: a sample of people, which the company claims fully represents the TV-watching public, is asked to report their viewing habits back to Nielsen. These people, of which there are about 50,000 (in 20,000 households), are approached by Nielsen to participate, and paid a token amount for their time and effort. (This broad sample differentiates Nielsen from smaller competitors like TiVo, whose subscribers tend to be wealthier than average.)
Under the title, “On Being a Nielsen Family,” blogger Crone and Bear It writes:
Being a Nielsen household brought with it some nice perks. They actually paid us. We got a check several times a year for $50. Now, that’s not a lot of money, but every little bit helps.
Plus, the folks at Nielsen were just super nice. Someone would visit occasionally to check the equipment and sometimes they would bring goodies along — a cheesecake or a nice apple pie.
So there may be pie involved. Probably not. But still: maybe.
Some households are monitored passively, but the real valuable Nielsen data — the stuff the networks pay for — comes from its People Meters. These things:
Each People Meter sits atop, under, beside or behind the family set, and records whatever happens: what is watched and for how long, what is recorded, and even what’s fast-forwarded. The box even knows which family member is watching, by means of a small remote control. Each family member, even the children, has his or her own little “I’m watching” button.
This data is sent back to Nielsen each night and parsed into the various ratings, including the live stream count, and the live + same day DVR numbers — the number most often quoted to the press. (Ratings are available for up to seven days after the initial air date, a number Nielsen settled on a compromise between the networks’ desire for more more data, and advertisers’ need for viewers to be current.) The numbers represent share of total viewers: a raw 1.0 rating, for example, indicates that one percent of the 115.9 million estimated TV-watching households, or 1,159,000 TVs, were watching a program. The data is also broken off into different demographic ratings, the most important being people ages 18-34.
Here’s where things get a bit weird. The press and public is interested in knowing how many people watch a show, because it’s the most obvious indicator of its success. This belief drives the way we think and talk about ratings. It also happens to be wrong.
To be sure, networks are interested in knowing how many people watch their programming, and freely tout or play down Nielsen’s wider audience measurements. But the numbers that networks and advertisers actually use — to sell ads, to set prices, and to decide on the fate of a show — are commercial ratings. In other words, advertisers don’t care how many people are watching a show nearly as much as they care how many people are watching their ads. Nielsen provides this number, which takes into account everything from next-day DVR viewing to fast-forwarding through commercials. If every Nielsen Family watched a show the day after it aired but skipped through all its ads, that show would probably be canceled.
In reality, a large majority of viewing still takes place live, when commercial skipping isn’t possible. (Though channel flipping during commercials can have a noticeable effect on C3 ratings.) Within the small-but-not-insignificant set of people who watch recorded shows, around half of them skip through commercials.
Absent from these ratings, and most others, are types of viewing that serious TV fans — and especially comedy lovers — are well acquainted with: Netflix, Hulu, iTunes and even on-demand cable viewing don’t count toward the ratings totals, either because they don’t have ads at all, or because the ads shown differ from the ones shown during initial broadcast. A show with 10 million weekly Hulu viewers and one million live viewers is only slightly better than a plain one-million-viewer show, in terms of revenue. Nielsen does track online activity, and plans to increase its coverage dramatically over the next few years. But unless the ads shown during an online broadcast are the same as the ones shown during the TV broadcast, online viewers will continue to be broken off into a separate and much less important figure, and not a show’s main rating.
This, more than anything else, explains why viewers feel uneasy about ratings. Networks, advertisers and Nielsen itself all know and acknowledge that the ratings reported and transacted with don’t reflect a show’s full audience. Back to the example of the inauguration: it strikes people as problem that the event’s viewership didn’t seem to be measured properly, but for the people who buy Nielsen’s data, it’s really not. It’s tough to make much money on an one-off event with no ad breaks.
NBC’s Alan Wurtzel underscored this discrepancy. “Ratings are a currency, so they’re just as important now as they were ten years ago. It’s how we get paid. But in this new media environment, do these numbers reflect accurately how many people are viewing this content? The answer is no.” As viewing habits become more fractured, and viewers become less bound to TV schedules, he says, the current ratings system’s focus on live and recently recorded TV will become a problem. “Are we happy with the way we’re following technology and being able to measure it? No. We’re way behind. On the other hand, are Nielsen ratings important and critical to the industry and as important to the industry as they even were? Absolutely, when you consider that if we didn’t have them, we wouldn’t get paid.”
What Comes Next?
Nielsen, networks and advertisers are all wrestling with the problem of media fragmentation, and nobody has charted a clear path forward yet. A number of competing companies like TiVo, TRA and Rentrak, are selling supplementary data to some of Nielsen’s clients, but none has a product that’s competitive with Nielsen’s core TV ratings. The Coalition for Innovative Media Measurement, or CIMM, hopes to glean better research from set top box data, and has the backing of major TV networks. Nielsen, which is in the process of going public, has extensive plans for increasing its measurement on online activity, and is running a pilot program that uses existing set-top boxes for audience measurement.
But all of these companies are charged not just with figuring out how to measure audience, but also with predicting the very ways people will consume TV in the future. Some point to digital set-top boxes as the future, with on-demand content and DVR usage eclipsing live broadcasts. Others, said Nielsen’s senior vice president for insight and analysis Patricia McDonough, would say that the set-top box is obsolete, and that we’ll all be getting our TV off the Internet in five years. “I’m not saying that I believe either one or the other, but I think our strategy is to be able to measure the stuff that’s coming into the set no matter how it gets there.”
It’s clear that the way networks and advertisers track viewership is going to change dramatically over the next ten years, as viewing habits are transformed by technology. For now, though, little is likely to change. Americans spend more time than ever before watching basic cable and broadcast programming, at an average of 34 hours per week — an hour less than that maximum legal work week in France. Networks are happy with this. Advertisers are happy with this. Viewers, apparently, are still happy with this. When one of those three facts changes, so will the way ratings work.
Until then, take what little comfort you can from this: It probably wasn’t your favorite show’s terrible ratings that got it canceled.
It was your favorite show’s ads’ terrible ratings.
John Herrman is a lovable ginger who writes about technology for Gizmodo, amongst other places.