Unlike non-electronic media, television has no simple way of determining the number of viewers who watch a particular show. There are no tickets to count, coupons to clip or unsold magazines to return. And yet the industry's raison d'etre is not to deliver a show to an audience but a market to an advertiser. A network or local station sells (or, more accurately, rents) its audience to clients like a herd of cattle, currently at a cost per thousand (cpm) of about $4 for prime-time fare. How to count the heads became one of the industry's first problems.

Into the breach stepped the A. C. Nielsen Co. of Chicago, Ill., a marketing research organization that makes 90 percent of its earnings surveying supermarket shelves but has gained its public notoriety from its TV audience survey, the high-sounding National Television Index. From a Census Bureau master list the company picks a secret sample of 1170 TV homes in randomly selected locations. Nowadays each home is outfitted with a tiny electrical device called an audiometer that records, every 30 seconds around the clock, the channels that all TV sets are tuned to. The audiometers are hooked up to special telephone lines that permit Nielsen computers to simply dial in for the viewing information. A ratings number is the percentage of all "Nielsen families" tuned to a particular show-at least all those that are counted during each biweekly rating period, generally far less than 1170. A "share" is the percentage of Nielsen families actually watching TV that is tuned into the show. Generally speaking, no regular prime-time show stays on the air very long with much less than a 30 share in its prime time.

For its cooperation each family receives about $25 a year plus compensation for half of its television repair bill. But the real rewards must be psychological – the genuine, if anonymous, power that the family has over the country's viewing choices. The Nielsen sample represents a mere seventeen-one-thousandths of one percent of the more than 70 million homes with TV sets. Thus each Nielsen family stands in for some 60,000 households. A mere 12 Nielsen families make up a single rating point. And yet a lead of one rating point in a single show can mean hundreds of thousands of dollars in advertising revenue to a network over the course of a season.

Most statisticians agree that the Nielsen system provides a reasonably accurate representation of aggregate viewing patterns. The key words are "reasonably" and "aggregate." Somewhere in every ratings report Nielsen says that they are wrong one-third of the time. But beyond such normal "sampling errors," there are more serious flaws in the system. For one thing, the audiometer only measures the number of sets tuned to a particular program, not how many people in a household, if anyone, is watching it. A few years ago Dick Adler, then the L. A. Times' TV critic, penetrated the Nielsen Company's veil of sworn secrecy by locating a man whose house was equipped with an audiometer. Under the cover of a pseudonym, "Deep Eyes" admitted that he consciously manipulated the system by leaving his set on some 50 hours a week tuned to his favorite shows-whether or not he was watching them or even home at the time.

Then, too, the company's estimate of the total viewing audience is grossly inflated. Nielsen counts every set connected to an audiometer as a single household even if several of them belong to one Nielsen family. To compound this inflation, for a TV set to be credited to a given program it need only be tuned to it for six minutes. Thus a person who switches channels in the middle of a show is counted twice. Moreover, the entire sample is heavily weighted toward white – middle-class families. Nielsen itself admits to a "light sampling" of blacks and Hispanics – one that is not representative of their proportion of the total population. And by placing audiometers only in homes and excluding bars and college dorms or other places where single people might congregate, the sampling vastly underrates the percentage of Americans living alone. (Indeed, when an audiometer is occasionally placed in a single person's home that person immediately becomes known as a "Nielsen family.") Since more than half of the people who live alone are over 65, the Nielsen sample seriously undervalues the country's older population. It is no coincidence that minorities and older people are the two demographic categories least desired by major advertisers.

This brings us to the concept of demographics – or the qualitative analysis of raw audience data. Throughout most of the history of television, advertisers bought a "bulk" rather than a "select" viewership, the larger the better. Then in the '60s sponsors came increasingly to realize that not every member of the audience was equally valuable to them. No longer were they merely interested in reaching the largest number of people – they had to be the right people – young families with plenty of "disposable income" at their disposal. "Young marrieds," "target audiences" and "purchasing power" became the latest Madison Avenue buzz words, the ideal being that if you could concentrate your advertising on people between the ages of 1849, you were likely to be reaching two parents with, say, three kids – or, in more basic terms, rive sets of teeth that needed brushing.

Nielsen itself set the rules of the game. In 1962 the company began to provide some demographic categories along with its aggregate ratings. At first the categories were men and women ages 18-34, 35-49 and 50 plus, but by the late '60s they were much more finely divided. The way Nielsen collected such demographic information was by choosing an additional 2000 homes and paying them to keep diaries of which programs members of the family watched and who watched them. But if the aggregate ratings are based on a small enough sampling, imagine how few Nielsen subjects represent, say, all 1824 year olds (much aside from the difficulty of getting people to accurately record their viewing habits on a daily basis.)

The networks themselves began to encourage advertisers to seek out younger and younger viewers. The way this came about was that ABC, finding itself perennially unable to compete with the other networks in terms of the "bulk tonnage" of the product (you and me) it could deliver, began to promote itself as "the network of the young." With a string of youth-oriented shows like American Bandstand, Mod Squad and The Young Lawyers on its schedule, the network would put out a press release after every season saying that a higher percentage of young people watched ABC than the competition (which claim CBS would legitimately rebut in its press release by saying that percentages be damned, more young people watched its shows in absolute terms).

In 1971 ABC released a marketing study it had sponsored which claimed that younger people were much better advertising prospects because they absorbed commercial messages better than older people, were more apt to experiment with different brands and learned about new products more quickly. The CBS research department rebutted the report with a study showing that the best target audience for most products was 25-64 year olds rather than 18-49 year olds and that the 50-64 year group was "far more valuable to advertisers" than the 18-24 year group. But by then the network knew that it was fighting a losing battle. NBC, also behind CBS in terms of gross numbers, began to pick up the demographic beat, as did all major advertisers. It got so that viewers over 49 years of age – in other words, families consisting largely of two sets of teeth – hardly counted anymore, unless an ad agency happened to be pushing Polident. If we are television's product, they were clearly distressed goods, "seconds." Things came to such a pass by the early '70s that CBS began canceling perfectly successful shows that their research department told them "skewed old" – Iong-time favorites such as Greenacres, Petticoat Junction and Gunsmoke (the latter with a healthy 30 plus when it was canceled in 1975).

Bulk circulation became almost irrelevant as television began to assume what the medium's preeminent historian, Erik Barnouw, has called "the demographics of the supermarket." Sponsorship became a matching game, a matter of correlating demographic information on retail customers from the Brand Rating Index with similar demographic reports from the Nielsen service. Networks began to give sponsors "assurances" guarantees that if, at the end of a three-month period they have spent more than an agreed – upon cost per thousand for a certain demographic group, they would be given bonus spots. From being physically present in the "sponsors booths" in the early days of television, advertisers had reached the point where they didn't even have to watch the shows that promoted their products. It was a lucky break for them, too, since their influence on the quality of these shows became far more deleterious at a distance than it had been up close.

Increasingly, TV programmers are catering exclusively to the young and the rich. The emphasis on younger and younger target audiences has made some of the most popular prime-time fare virtual cartoon shows for post-pubescents. And since the same qualities, which make younger people more receptive to new retail products, make them more receptive to new shows of whatever quality, the cartoons seem to come and go with increasing frequency. In recent years some new shows have been canceled before they even got on the air because they did not have desirable demographic breakdowns in preview tests, while many new series were canceled after only a few showings. Almost every TV writer and producer I talk to complains that this new network policy forces them to reach for immediately recognizable clichés since they do not have the time to develop a character over a period of time.

In catering to younger audiences, the networks have almost completely disenfranchised the old – that is, unless the old happen to be rich as well. (Golf, which attracts an older, wealthy audience, is televised far out of proportion to viewer interest because "big ticket" advertisers like car manufacturers are willing to pay handsomely for its select audience.) The two most popular shows in syndication, Hee Haw and The Lawrence Welk Show, were canceled on network television because they skewed old (and rural – double sin). Moreover, the fact that documentaries and news programs also attract an older audience (the average age of the audience for network news shows, in fact, is over 50) has been used as an argument to keep the former off the air in prime time and the latter from expanding beyond its present headline skimming half-hour format.

The irony of all this is that the emphasis on pinpointing younger and younger target audiences may have already backfired even on the industry's own profit-making terms. All the time that the networks have been single-mindedly devoting themselves to serving the young, nobody in television seems to have noticed that the population of the United States has been growing steadily older. According to a recent report of the Senate Committee on Aging, the number of Americans 65 or older increased by 18 percent between 1970 and 1977, as contrasted to a total population growth of only five percent. If this trend continues, by the year 2030 the median age of Americans will rise more than eight years from the current figure of 28.9. Some major advertisers are beginning to catch on to this population shift and to view their prime target audience as 25-54 year olds rather than 18-49 year olds. They are even designing products specifically for these long-neglected "mature adults." A recent issue of Advertising Age reports that Pfizer has just brought out a new shampoo "for the special needs of older people" called New Season, and that Alberto-Culver is testing a similar product. Now that there is virtually nothing left on the networks' schedules that any self-respecting older person would care to watch, the new Alberto-Culver shampoo will be called Prime Time.