Eyes Focus on ‘People Meter’ as It Gauges TV Viewing

Brainstorming researchers at A. C. Nielsen Co. decided back in 1957 that what they needed to plumb the mysteries of American TV viewing habits was a sort of electronic pillow.

Their idea was to put a specially wired cushion on father’s easy chair, a second on the sofa where mother reposed, and others where the children usually sat.

A pressure-sensitive sensor in the pillow would register the viewers’ presence, giving the researchers data not only on whether the set was on, but on exactly who was watching “Playhouse 90,” Phil Silvers in “You’ll Never Get Rich” or other programs of the day.

Before long, someone realized that family members sometimes sat in each other’s chairs, or on the floor, and the device--by now tagged the “Nielsen whoopee cushion”--was abandoned. But researchers and the companies that advertise on television have never lost their desire to find out exactly whose eyes are turned to the set at any moment.

This year, Nielsen has decided to start relying on a new device to gather such demographic data and, in so doing, has brought the biggest change in audience nose-counting since the advent of television.

Innocuous-Looking Box

The introduction of an innocuous-looking black box called the “people meter” has filled the air with a fog of statistics, and set off sharp exchanges between the advertisers and ad executives who advocate its use and the network officials who complain that it has arrived too soon.

The advertisers and ad agencies hope that the new system will yield more accurate audience numbers--and possibly, by showing lower viewing levels, allow them to win some price reductions. In a year when the ABC and CBS television networks are already suffering losses, that’s what the networks fear most.

CBS may have the most to worry about, for data from the new device suggests that CBS audiences in some key time periods may have been smaller than previously believed. Nielsen results for the past season show that for the prime-time hours, CBS household viewing levels were 7% lower with the people meters than they had been with the old, so-called diary method, while NBC’s was 1% lower and ABC’s was 3% higher, according to CBS officials.

Even such small differences translate to huge sums in the network advertising business, where a single prime-time rating point is worth more than $90 million in revenue over the course of a season.

The demographic data collected with people meters supplements other data that provide estimates of the number of American households tuned in at any particular time. The demographic data is particularly valuable to advertisers because it helps them to target their commercials to the specific audiences they wish to reach.

“We’ve had an orderly market here for 30 years, and suddenly it’s in chaos,” said William Rubens, vice president of research for NBC. “We’re going to have millions of bits of additional information floating around, and nobody to say what they mean. . . . The industry has been seduced by the glamour of the button.”

The moment of truth is rapidly approaching. The networks and the advertisers next month will begin negotiating advertising sales for the “up-front” buying season, in which the broadcasters will sell about $3 billion of their $8 billion in advertising time for next season.

“It should be one of the tensest, toughest negotiations we’ve seen,” said Alec Gerster, media director at Grey Advertising in New York.

As this drama has developed, the industry has also watched the unfolding of a separate plot centering on the competition between the companies that provide the all-important ratings research. Nielsen, a Dun & Bradstreet subsidiary that has dominated the ratings business since the 1940s, was forced to begin use of the people meter because a competitor, AGB Television Research of Great Britain, began lining up clients of its own with people meter data.

How People Meter Works

This year, both companies will offer people meter services, and while the much-larger Nielsen seems to have the edge, the advertising and broadcast industries are waiting to see which system proves fastest, cheapest and most accurate. “Nielsen has already been one of the big losers in all of this,” said Hugh M. Beville Jr., a longtime broadcast ratings researcher and author of a book on the subject. “Their domination of this business is facing its first serious challenge.”

There is little mystery in the people meter’s operation.

About the size of a cable-TV channel selector, the meter has buttons assigned to each family member, and others for guest TV viewers who may wander into the room. Viewers are supposed to push a button when they begin watching and each time the channel is changed.

If they fail to do so, a red light flashes as a reminder.

Nielsen began publishing people meter data from a sample of 1,000 households in January, 1986, after collecting preliminary numbers for two years. The company plans to increase its sample size to 4,000 by September, 1988.

For the just-ended 1986-87 television season, Nielsen also continued to collect data under the diary method, in which viewers wrote longhand accounts of the programming they had watched during the previous week. Often, one family member handled the task for all people in the household. (In addition, Nielsen also keeps tabs on viewer habits by means of electronic devices attached to TV sets of a specially selected group of households. The devices tell Nielsen when the set is on or off and to which channel it is tuned.)

That method was used to gather demographic data for 30 years and seemed adequate in the days when network affiliates and a few non-network stations dominated viewing. Viewers’ memories were not taxed when they tried to remember which shows they had watched.

More Independent Channels

But the diary method has seemed less and less reliable as program offerings have increased to include more independent and cable channels, and as well as videocassettes. Some analysts say as many as half of the nation’s TV sets are now equipped with remote control devices that allow viewers to quickly move from channel to channel.

“You just can’t expect somebody to remember what was watched two days ago when there are 50 channels and people are moving constantly from one to another,” said Robert Eigel, head of the network TV advertising department of the N. W. Ayer agency.

And the diary method had other weaknesses. Viewers sometimes admitted writing that they had watched shows they felt they should have watched, rather than the shows they actually did.

In a broadcast last year of the “McNeil-Lehrer News Hour,” for example, a young member of a Nielsen survey family told an interviewer that he often reported watching the science show “Nova” even if he had not, because he thought the educational program was worthy of a higher rating.

“Around here, we really cringed when we heard that,” said Jo LaVerde, a Nielsen spokesman in New York.

But the people meter has its own problems, mostly due to the effort required for viewers to push and repush its buttons. Research already shows that children, teens and less educated adults are less inclined to follow the instruction conscientiously, network researchers say.

The people meter results, for example, show the Saturday morning audience of children to be 12% lower than the audience measured by the diary method. Researchers assume that children--generally unsupervised by their parents at that hour--simply don’t bother to push the button.

“They’re just lazy, or they’re bored,” said Marvin Mord, ABC’s research chief.

And some say the 1,000-household Nielsen research sample has been skewed toward better-educated people because that group is more likely to agree to participate in such experiments. Twenty-seven percent of the surveyed households are headed by college graduates, while only 21% of American households are headed by people with college educations, said David Poltrack, CBS vice president of research.

“The less educated are just less likely to agree to cooperate, and they’re more likely to get fed up and drop out once they’ve joined,” said Poltrack. “Some people in society are just less willing to interact with technology, and that’s going to skew the results.”

The people meter’s heavy demands on the survey group is also reflected in the high percentage of data that has to be thrown out because the viewer has slipped up somewhere along the line, according to Poltrack. Nearly a quarter of the data is rejected because viewers have failed to properly record their viewing habits, he said.

The volatility of the ratings is another sign of the system’s weaknesses, in his view.

Evening news programs have generally stable ratings because their audience appeal varies little from one day to the next. Yet one month, the people meters showed CBS viewing levels 4% higher than with the diary method, while in another month, audience levels were 11% lower, Poltrack said.

Many of the biggest prime-time hits have registered lower audience levels with the people meter than with the diary method. Researchers believe that this confirms their suspicion that viewers filling out diaries tend to remember and write in big hits, even though they actually watched less well-known shows.

The audience level found by the people meter for “The A-Team,” for example, was 12% lower than that recorded by viewers who filled out diaries. For “The Cosby Show” and “60 Minutes,” it was 2% lower.

Viewing Information Defended

Nielsen officials defend their people meter results, saying they have been generally borne out in corroborative telephone surveys. While the education levels of the sample tend to be somewhat higher than in the general population, they say that the public tends to fib a little anyway about their level of academic achievement.

They do acknowledge, however, that results from children have not been satisfactory, and have recently begun a program to try to improve those results. Nielsen plans to offer children in surveyed households gifts worth a few dollars to recognize their efforts.

The research company will also spend more time briefing children before the survey work begins, and following up more quickly with “coaching” calls when results show that the children have not been conscientious.

“The fact is, it’s difficult doing any research with kids,” said John A. Dimling, a Nielsen senior vice president.

Nielsen’s research has seemed to satisfy most of the major ad agencies and big advertisers who urged adoption of the people meter.

“They’re not without their problems, but they’re better than the paper diaries for the fragmented audiences of today,” said David Braun, director of media for General Foods Corp., in White Plains, N.Y. “I don’t think, frankly, we had too much of a choice.”

But the debate and data-analysis of the past year has evoked expressions of concern from other groups, including associations representing cable television and the non-network, or independent, television stations.

While the independents would not be affected by the national audience surveys that gauge network ratings, there is a general assumption that the people meters will be used before long to measure audience levels in some large markets during the local rating surveys, or “sweeps.” How the people meter will affect this year’s ad sales season is still far from clear. CBS officials have warned that their financial results for the second half of the year may be hurt by the switch to the new ratings system.

But network officials believe that sales will have more to do with the general strength of the network advertising market. “If demand is there, these numbers may not mean that much,” ABC’s Mord said.

At next month’s negotiations, the networks are expected to ask for an easing of the traditional guarantees they offer advertisers.

Under such guarantees, the networks must give advertisers compensatory, or “make good,” ad time if their programs don’t reach audience levels that are specified at the time of the advertising sale. This year, with the ratings system under a cloud, the two sides may need to set some compromise, say ad agency and network officials.

Many believe that the people meter system actually represents an interim technology that will soon be supplanted by a ratings system requiring no action by the viewer to record viewing habits. Nielsen, the ScanAmerica research unit of Arbitron, and R. D. Percy & Co., a Seattle market research firm, are each developing data with such systems, which use infrared or sonar sensing devices to find out who is near the TV. But they have their own problems.

The infrared scanners that count bodies by sensing warmth can be thrown off by the presence of a household dog or heat pipes, noted Dale Metzger, president of Statistical Research Inc., a New Jersey research firm. And naturally, they can’t tell if viewers are reading the newspaper or dozing, he said.

“I wouldn’t want to predict how long it will take till we find a totally passive system,” Nielsen’s Dimling said. “It’s safe to say, though, that we’re going to be in the middle of change for a long time.”