Nielsen sweeps

Nielsen sweeps are four annual television audience-measurement periods that use paper diaries to produce local-market ratings. They give local stations and advertisers data for negotiating advertising rates, yet their low response rate, reliance on viewers’ memories, and dependence on atypical programming make them a poor guide to ordinary viewing. Their long persistence, despite the availability of people meters, is a case study in how a flawed system can survive when the participants who could replace it cannot coordinate—and when one faction benefits from keeping it in place.

The Measurement System

Four times a year—in February, May, July, and November—Nielsen sends 2.5 million paper diaries to randomly selected people in almost every television market in the country. Recipients record the programs they watch over the course of a week. Nielsen also gathers information about the people who complete the diaries, then uses the results to construct demographic portraits of local markets. Local network stations, or affiliates, and local advertisers rely on those portraits when negotiating advertising rates for the months ahead.

The system addresses a genuine gap in audience measurement. Nielsen’s people meters can provide a relatively accurate national picture of television viewing, but there are only 5,000 of them, scattered across the country. That sample cannot show what viewers in a particular city are watching. A local business, such as a health club in Fort Wayne, Indiana, might want to know which Tuesday prime-time program attracts local viewers aged eighteen to thirty-four. National figures cannot answer that question, and sweeps diaries were designed to supply local demographic detail.

The Limits of the Diaries

The diaries provide no guarantee of accuracy. Only about 30 percent of the diaries Nielsen distributes are returned. Because those who respond may watch different programs from those who do not, the low response rate creates what is known as cooperator bias. The people whose habits are recorded may therefore be a poor stand-in for the larger audience the ratings purport to describe.

The method also depends on delayed recall rather than contemporaneous recording. Viewers commonly fill in their diaries at the end of the week, remembering what they watched rather than documenting it as they watched. High-profile programs are easier to recall, so the records tend to inflate network ratings and deflate those of smaller cable channels. The diaries also struggle to capture the restless habits of viewers who switch channels.

Even a perfectly completed diary would describe an unusual viewing environment. During sweeps, the networks load their schedules with one-time specials, expensive movies, prominent guest appearances, and new episodes of their strongest shows; reruns are not allowed. In February 2003, ABC, NBC, and Fox each spent millions of dollars on programs about Michael Jackson, while that month also featured the climaxes of The Bachelorette and Joe Millionaire. Such schedules make sweeps ratings a measure of viewing under exceptional conditions, not a representative picture of the other eight months.

Costs Across the Television Industry

The ratings help determine advertising prices, so advertisers may pay on the basis of data shaped by both nonresponse and stunt programming. Allen Banks, an executive media director at Saatchi and Saatchi, North America, called sweeps a sham and a subterfuge, arguing that the ratings offered little resemblance to what happened during the rest of the year. Some advertisers try to account for the distortion, but in many local markets sweeps remain the only hard data available, and consequently retain disproportionate influence.

The system also distorts the networks’ programming choices. Because a sweeps month is so valuable to local stations, networks compete by airing their strongest new programs at the same time. Viewers can watch only one program at a time; if the networks staggered their most attractive shows, more people might watch them across the schedule. Instead, simultaneous competition can reduce their combined audience. The concentration of new and special programming in sweeps months also leaves other periods crowded with reruns: a viewer drawn to November’s spectacles may find January far less varied.

The better alternative is not costless. People meters could supply a more reliable way to measure local audiences, but installing and maintaining them across thousands of homes in each of the country’s 210 television markets would likely cost at least nine figures. Nielsen did not disclose the expense. Yet the scale of that cost must be set against the money already at stake: about $25 billion in annual advertising spending is based on sweeps data, and the networks spend hundreds of millions of dollars on sweeps programming. The book’s argument is that an investment in people meters, or a comparable system, could serve the industry better than continuing to pay the costs of unreliable measurements and schedules built around them.

Why the System Persists

The central obstacle is not simply a lack of a technical alternative. Replacing sweeps would require networks, stations, and advertisers to cooperate and share the costs, while each could hope that someone else would take the lead. In that situation, a participant who waits can still benefit if others act; if everyone waits, however, the old arrangement continues. The common gain from better ratings does not automatically produce a joint investment to obtain them.

The way the diary system is funded makes the coordination problem more severe. Affiliates pay 90 percent of the cost of collecting and analyzing sweeps diaries, and the party paying the bill has the power to dictate what happens. The diary system favors familiar networks and recognizable programs, thereby increasing affiliates’ ratings at the expense of smaller stations. Affiliates also avoid paying the hundreds of millions of dollars that networks spend on sweeps programming, and the ratings from the other eight months matter less to them because those months are not being measured in the same way. Their incentives diverge from those of networks and advertisers, even though all participate in the same industry.

That divergence was visible when Nielsen introduced people meters in Boston in 2002: not a single affiliate signed up. The stations preferred to have no ratings rather than accept figures from the alternative system. The example shows how a group’s decision can be shaped not only by the overall merits of a measurement method but by who controls it and who gains from its existing biases. Had networks and advertisers historically shared more of the cost of local ratings, they might have had more leverage to change the system; instead, they were effectively left to follow the affiliates’ preferences.

A Slow Change and a Broader Lesson

By 2003, Nielsen announced that it would roll out people meters in the country’s ten largest television markets. The move suggested that the diary system might change as cable grew in importance and paper diaries looked increasingly outdated. It did not erase the long-running mismatch between the industry’s shared interest in reliable local ratings and its failure to organize a transition.

Nielsen sweeps thus serve as a counterexample to the idea that many individually capable participants will necessarily make a group decision intelligently. The affiliates, networks, and advertisers each had reasons for their choices, but those reasons did not combine into a solution that benefited the industry as a whole. The result was a multibillion-dollar business continuing to rely on a backward, inaccurate technology because its major players could not cooperate. The contrast is pointed: actors may be sensible within their own incentives and still sustain a collectively wasteful system.