TV execs to Nielsen: get SMART; can the Nielsen ratings survive 500 channels, the Internet, and a new competitor? - Nielsen Media Research Inc. copes with Systems for Measuring and Reporting Television rating laboratory

American Demographics, Oct, 1997 by Brad Edmondson

Can the Nielsen ratings survive 500 channels, the internet, and a new competitor?

In the world of market research, it can cost millions of dollars to get incomplete answers to simple questions. Here are two: how many people watch television programs, and what are their demographics? Answering these questions earns Nielsen Media Research more than $300 million a year. Billions in advertising revenue depend on the answers, because the cost of a television ad depends on how many people watch it.

Measuring television audiences was relatively simple in the days when "television" meant ABC, NBC, and CBS. But now there is a fourth broadcast network (Fox), and two-thirds of U.S. households subscribe to cable services that can offer more than 100 viewing choices. Also, many television signals will soon switch from analog to digital transmission, which takes up less of a cable's capacity. When this happens, the number of channels available in a typical cable home could rise to 500 or more.

The explosion of program choices is only one of Nielsen's problems. Human behavior is another. More than eight in ten U.S. households now own VCRs that can tape television programs for later viewing. Nielsen recently estimated that on a typical weekday, about 3.7 million women watch TV when they are away from home. Even when they are at home, about half of women and four in ten men who are "watching TV" are also reading, talking, folding laundry, or doing something else. Now add in the Internet, Nintendo, and other videoscreen diversions. Measuring television use in the 1990s is like trying to shovel smoke.

Nielsen tries to capture TV viewing in its many forms, and it has a plan to cope with the 500-channel future. But as the questions grow more complex, Nielsen's customers grow less satisfied. "I don't trust their numbers at all," said Don Ohlmayer, president of NBC's West Coast division, in a recent interview. "They're trying to measure 21st-century technology with an abacus."

The feud between Nielsen and the networks has been building for eight years, and next year it could become a war. Three years ago, a group of TV networks and advertising agencies hired Statistical Research, Inc. to set up an independent TV ratings laboratory that would come up with improvements in audience measurement techniques. They hoped Nielsen would embrace the improvements. So far, Nielsen hasn't. Now SRI's ratings lab, also known as Systems for Measuring and Reporting Television (SMART), is in the final stages of developing its own ratings system. By next spring, Nielsen could have something it hasn't seen in a decade--competition.

PEOPLE METER PUNCH-OUT

Nielsen says that its numbers are reasonably accurate, and that the television networks are unhappy because their ratings have been going down for a long time. Of the 100 network TV programs with the all-time highest Nielsen ratings, 18 are episodes of "The Beverly Hillbillies" that ran in 1963 or 1964. Only nine of the highest-rated programs aired in the 1990s, and six of those were Super Bowls.

"The real issue is finding new and innovative ways to measure what's on the screen," says Nielsen vice president Jack Loftus. "And you have to do that at a low cost, so you can increase the sample size of your survey." The networks say Nielsen is missing the point. "Measurement is the technical issue," says NBC senior vice president Nicholas Schiavone. "But the real issue is how people use television." Nielsen, he says, "pays more attention to technology than psychology."

Nielsen measures the TV use of a national sample of 5,000 households in more than 200 markets. Many of those households still record their viewing choices on paper diaries, as they have for almost 50 years. But for the last decade, Nielsen has asked respondents in the largest markets to punch personal identification codes into remote control-like devices called "people meters" when they turn on the tube. The ID codes, channel choices, and time of day are stored in a box that is wired into each television and VCR in the household. The box is also connected to a telephone line. Every day, the box automatically dials a data-processing center and downloads its record.

Many Nielsen customers have been suspicious of people meters from the get-go. In 1989, a review by the Committee on Nationwide Television Audience Measurement (CONTAM), an industry group formed in 1963 to monitor Nielsen practices, found that audience estimates drawn from people meters were 6 percent lower than those drawn from national telephone surveys. For some kinds of viewers, the differences were much greater. People meters missed half of visitors to households, one in four men aged 18 to 34, and one in eight women aged 18 to 34, according to the phone surveys. People meters also missed a high proportion of children, especially teenagers.

In spring 1990, the first Nielsen ratings that relied on people meters were released--and the networks screamed. The numbers showed a sharp, across-the-board drop in television viewing. The networks lost $300 million in advertising revenue in the first quarter alone. "It was an accident waiting to happen," says Schiavone, who has been president of CONTAM since 1990. "We paid dearly because of a methodological artifact, not because of any real change in behavior."