WHEN THE SELLER no longer met the potential buyer face-to-face as craftsman or seller, but only indirectly through advertising, it was harder than ever for him to know what the buyer wanted, or even to know who the buyer was. The old direct democracy of demand—the customer telling the cobbler the style of shoe to make for him—had disappeared. In the everywhere market the merchant had to find indirect ways to answer his urgent questions. And manufacturers and merchandisers democratized the market by inventing ways for the consumer to vote his preferences. The vast market of unseen buyers gave rise to a new science for sampling the suffrage of consumers.
IN ITS BEGINNINGS, Market Research was directed simply to answering magazine advertisers’ questions about whom they were reaching. For in the early days of American advertising there was no generally reliable information about the circulation figures of newspapers and magazines. Since the advertising agent supported himself by buying space at wholesale rates and then retailing it to advertisers, the agent was always trying to sell space which he himself had already bought. Especially since there were no authentic figures, he was understandably tempted to exaggerate the reach of his advertisements. Unchecked by any impartial source, the publications themselves commonly issued circulation figures that were inflated.
From the point of view of the advertiser, then, buying advertising space was like buying a lottery ticket. Not only did he not know how readers were responding to his paid message; he had only the vaguest notion even of how many subscribers or purchasers or readers there were. Detailed questions about the character of the audience to whom the publication circulated were far outside anybody’s realm of knowledge.
An enterprising New Englander saw the advertisers’ need and aimed to make a business of providing them reliable information on what they were getting for their advertising dollars. In 1869 George P. Rowell, founder of the trade journal Printer’s Ink, who had got his start in Boston by a scheme of theater-program advertising, published the first extensive and impartial directory of American newspapers. He listed geographically and by class 5,411 United States publications and 367 periodicals in Canada.
After reliable circulation figures were available, it became more difficult to persuade advertisers to buy space which reached an undefined audience. Advertising in America was challenged to become a science. The decisive step was taken by Francis W. Ayer, a former schoolteacher who, with his father, had founded N. W. Ayer & Son in Philadelphia. Their main business was selling advertising space in religious papers. Soon after his father’s death a friend urged that since he was much too good a man to be “nothing but a drummer” for a product of dubious value sold to ill-informed customers, he should get into some honest business. “I will not be an order taker any longer,” Ayer determined. “I will not take orders for these lists of magazines and lists of newspapers and get my commission out of them and be satisfied just to make money. I will have a business, I will mean something to somebody every time I take any business, and I will have clients rather than people who just give me orders.”
In 1875 Ayer signed the first “open contract” (with a firm of rose growers in West Grove, Pennsylvania) and so gave a new character to American advertising. Under the open contract, the advertising agent became the agent solely of the advertiser. He offered expert advice on where, when, and how to place advertising. He shared with the advertiser all information about the price he was paying for the space, and he committed himself to secure space at the lowest possible price. In return, the advertising agency received a fixed commission (at first 12½ percent) and a contract to service the advertiser for a definite period. The advertiser then would not judge the agent in advance by the few cents’ difference between his bid and that of a rival agent for a particular advertising space, but would wait to see the effectiveness of the agent in increasing sales. In this way Ayer gave the advertising agency a new role in American life, and set a pattern which lasted into the late twentieth century.
Within twenty years, newspaper publishers, soon to be followed by magazine publishers, recognized the independent advertising agencies by announcing that they would give no special discount to direct advertisers. In return, the advertising agencies agreed not to try to beat down the publishers’ announced rates on behalf of their advertisers.
Advertising space in newspapers and magazines became a commodity in the open market, and publishers were finally under pressure to give full and accurate facts about the circulation and character of their publications. By 1914 the publishers themselves were subsidizing the new Audit Bureau of Circulation, fortified with authority to examine publishers’ records and to verify publishers’ statements. By the 1930’s about half the daily newspapers were members of the Bureau, and they accounted for some 90 percent of daily circulation.
NOW AT LAST the advertising agencies could be judged by the cost for reaching a customer. And the seller, for the first time, had expert help in the advertising he was paying for. In 1911, R. O. Eastman, advertising manager of the Kellogg Company, manufacturer of breakfast foods in Battle Creek, Michigan, persuaded fifty other national advertisers to support his postcard-questionnaire survey of magazine readership. He turned up some important new facts, for example on duplicate circulation, and so persuaded advertisers that they needed still more facts about what they were getting for their advertising dollar. When Eastman opened his own market research agency in 1917, his first client was the General Electric Company, for whom he tried to find out what the “Mazda” trademark meant to the consuming public.
Publishers now tried in earnest to discover what they were really selling. Circulation figures, they began to see, were not enough. They asked who bought their paper, where they lived, what they did for a living, and what they wanted to buy. This knowledge, destined to affect not only American advertising but even the products themselves and the development of new products, made the relationship between seller and buyer more self-conscious. The seller set up his own “intelligence agency,” and planned his strategy to conquer the market. The buyer meanwhile began to see himself as a significant unit in the mass of “consumers” who had a power to affect the flow of goods.
Cyrus H. K. Curtis, as publisher of the Ladies’ Home Journal and the Saturday Evening Post, had shown a remarkable talent for building magazine circulation. When he bought the “elderly and indisposed” Post in 1897 (price: $1,000), it had a circulation of less than 2,000, but within ten years he lifted its circulation to 1 million, and the circulation soon thereafter reached nearly 3 million. Like Bonner before him, Curtis enlisted the leading writers of the day by paying the highest prices for their work, and he succeeded in making the Post weekly fare for the great mass of Americans who considered themselves middle class. In the heyday of the Post, an age of optimism for “American business,” he gave to advertising about 60 percent of its 125 pages. Then in 1911, one of his advertising representatives persuaded Curtis that he lacked the information he needed about the wants and habits of consumers and about the dealers who sold the products that were advertised. A salesman for advertising space in the Post could talk glibly about the Curtis publications, but he knew very little about the customers whom their advertisers wanted to reach. Curtis agreed that more facts would make better salesmen. And he looked for a man “who could not only collect facts, but who knew one when he saw it.”
A young schoolteacher from Wisconsin was willing to take on this ill-defined task. When Charles Coolidge Parlin was hired in 1911, the name he invented for what he was trying to do was “commercial research.” (“Marketing research,” a more focused phrase, did not come into use until later.) Curtis had recently purchased Country Gentleman, but no one in the firm knew much about agriculture, so Parlin decided to begin by collecting facts about the agricultural-implement industry, which was buying much of the advertising in this magazine. By interviewing manufacturers, retailers, and farmers, within six months he had compiled a 460-page survey of the industry that revealed unsuspected facts about where agricultural tools were made, to whom they were sold, when, and where.
Parlin moved on to an ambitious study of the market for almost everything in the nation’s one hundred largest cities (at the time, cities of over 54,000 population). After 37,000 miles of travel and 1,121 interviews, he organized his information on the volume of business and on the kinds of merchandise in every department store, drygoods store, and principal men’s ready-to-wear store in all these cities. Department Store Lines (1912), his pioneer report, withstood the scrutiny of Marshall Field and other large-scale merchants of the day, and produced some useful distinctions:
Woman is a shopper. Out of that fact has come the modern department store. Partly by nature and partly by education, woman is a comparer of values….
A woman’s purchases may be divided into three groups: convenience goods, emergency goods, and shopping items.
Convenience goods are articles of daily purchase, such as groceries, aprons, children’s stockings and, in general those purchases which are insignificant in value or are needed for immediate use.
Shopping goods include all those purchases which require thought and will permit delay, such as suits, dress goods, high grade underwear, in fact high grade dry goods of all kinds. Values are compared and a serious effort is made to secure the best value for the money.
Parlin thus showed the manufacturer that if his product was a “shopping line” the number of stores which could carry his goods profitably might be so small that he could cover the market with his own salesmen; but if it was a “convenience line” the diffusion of the suburban and rural markets would require a jobber.
Parlin’s most remarkable product was his five-volume study Automobiles (1914). In the United States there were still fewer than two million automobiles, but one hundred widely sold makes. Parlin collected facts on manufacturing and distribution, on the influence of women on automobile purchase, and he asked questions which proved prophetic for the coming automobile age. He foresaw the tendency to reduce the number of grades and makes, he envisaged the dimensions and even the shape of the automobile market. As a result of his persuasive predictions, the automobile manufacturers increased their advertising in national magazines, and the Saturday Evening Post received a good share.
IN 1916 THE Chicago Tribune used house-to-house interviews to gather the marketing facts all over Chicago. By 1929 the United States Department of Commerce recognized the need for these facts with its first Census of Distribution, and then itself began regularly collecting facts on wholesale and retail establishments, on their size, and on the character and quantity of goods and services sold.
In the 1930’s a picture frequently reproduced in magazines showed a pile of hairpins which had been dumped on a vacant lot by a bankrupt manufacturer who had not kept up with the times. The promoters of the new science of market research argued that the manufacturer might have averted this fate if he had listened to the shrewd market researchers around 1920 who, when they saw the musical comedy idol Irene Castle featured on front pages with bobbed hair, forecast the change in hair styles and would have shifted their clients to making other products.
Some manufacturers used market research to go direct to the customers for their preferences. The design of the old-fashioned alarm clock had been so standardized as to become a comic-strip symbol. Then in the late 1920’s the Western Clock Company, drawing on recently discovered consumer preferences, offered a new model: smaller, thinner, of lustrous black and nickel, with svelte figures and graceful pierced hands. It was an immediate commercial success. But the makers of Listerine at first did not research the market and simply assumed that customers who liked Listerine mouthwash would like the same taste in a toothpaste; the toothpaste did not sell. Then, with market research, they discovered the consumer preferences in toothpaste flavor, which could have saved them large losses if known earlier.
Manufacturers had begun the practice of counting the market research ballots to discover the whims of their customers. “The consumer is king,” concluded C. C. Parlin. “His preference is law and his whim makes and unmakes merchants, jobbers, and manufacturers. Whoever wins his confidence controls the mercantile situation; whoever loses it, is lost.”
The efforts of craftsmen to please monarchs were crude beside those of modern American market researchers to anticipate the tastes, habits, and desires of their majesties The Public. In 1920 the Curtis Research Department sent investigators to the little town of Sabetha, Kansas, and interviewed all but twenty families in the 144 square miles surrounding. They secured information from these families about their buying habits and brand preferences. A few years later they took the first national pantry survey, based on inventories in 3,123 homes in eighty-five neighborhoods in sixteen states. Then, in Watertown, New York, they made the first “every-house-and-every-outlet” survey. Detailed reports on 28,203 of the town’s 28,930 homes provided a “sample” of 97 percent.
In a desperate effort to verify interviews and questionnaires on grocery consumption, the Curtis Publishing Company made a “Dry Waste Survey,” popularly known as the Ash and Trash Survey. Fifty-six Philadelphia families, fourteen in each income quartile, cooperated by agreeing to allow the Curtis researchers to collect and analyze their trash for a four-week period in July and August 1926. The families were instructed not to burn or destroy any containers and to keep a record of any wrappers or labels saved for premiums. Once a week Curtis researchers picked up all the trash from each of the fifty-six homes, then separately packaged and labeled the batch from each family. This packaged trash, finally totaling six thousand packages and containers, was taken to a nearby warehouse where it was studied and classified. The puzzled neighbors came to describe the work as that of the “Curtis Rubbishing Company.”
THE RISE OF radio made the relation between seller and buyer more indirect than ever, and deepened the mystery of who was receiving the advertiser’s message. In the 1920’s when David Sarnoff urged his associates to invest in the new wireless “music box,” they objected that it had no imaginable commercial future, because it depended on “broadcasting.” That meant, of course, sending out messages “broadcast” to persons who could not be identified, counted, or located. Since no visible connection was required between broadcaster and receiver, who could tell who was receiving the message? And who would pay for a message sent to nobody in particular? But there proved to be enough advertisers willing to send out their messages to everybody in general, taking their chances on who might be listening. Then, as the cost of radio advertising rose, advertisers, educated by market researchers, expected to know whom they were reaching.
Advertising on the airwaves offered the market researcher a puzzling new problem. While a publisher could identify his subscribers and could locate his newsstand sales, the radio broadcaster could only guess who might be listening.
The pioneer in this new field of audience research was Arthur C. Nielsen. Son of a Danish immigrant, and an engineering graduate of the University of Wisconsin, he devoted his life to unraveling the mysteries of the market. In 1933 he introduced a novel scheme of retail-sale indexes for druggists, who allowed him to collect the figures from their own records; then he went on to a “Food-Drug Index,” and to still others. Within four years he was receiving over $1 million annually from selling this information.
To test the audience for radio advertising Nielsen used the Audimeter, a device invented by two M.I.T. professors which could be attached to a radio receiver to record the times when the set was actually in operation, and the frequencies to which it was tuned. By 1950 he had applied this system to television. In a scientifically selected sample of homes, Audimeters were attached to television sets to record when the TV receiver was on or off, and to which channel it was switched; the records went to Nielsen in Chicago, where the information was correlated with the programs which had been broadcast, and was compiled into a report every two weeks. Each Audimeter family received a nominal sum for its trouble, and Nielsen agreed to pay half the cost of repairs if anything went wrong with the TV set. Since the machinery was expensive, only a small part of the audience could be monitored, and a great deal of effort went into making the sample representative. By the 1960’s the Nielsen Television Index Service annually offered 840 different reports with 14 million new figures, including estimates of the number of families viewing each broadcast, the number of different families viewing one or more of a series of broadcasts, each program’s share of total television viewing, and minute-by-minute fluctuations in the audience of each program. To establish this service on a break-even basis required seventeen years of work, and investments of $15 million. Just as Americans spoke of Dow-Jones averages on the stock market or Dun & Bradstreet ratings in the world of commerce, they now spoke of Nielsen ratings.
But there still remained mysteries about which of the listeners were really listening. Archibald Crossley’s studies in 1939 had shown that at any time as many as 25 percent of the radio sets left on were playing to an empty room. Later studies showed that during the day as many as 40 percent of housewives were not watching their turned-on television sets. More and more Americans were not listening to what they “heard,” nor looking at what they “saw.”
THE NEW TECHNIQUES for polling consumers were obviously useful, too, in polling voters to predict their opinions on issues and on their tastes in candidates. By the early twentieth century, academic students of politics had shifted their interest from “political philosophy” or “political economy” to “political science.” Now they aimed, in Harold Lasswell’s phrase, to predict “Who Gets What, When, How.” By the time Eastman and Parlin had begun to develop a science of “market research,” political scientists were ready and eager to develop their own predictive science of political opinions. As the century wore on, this new political science became more and more assimilated to merchandising, and shrewd politicians increasingly borrowed the techniques of market research so they could make themselves an attractive package for the democratic millions.
The first “social-scientific” survey of the opinions and attitudes of large numbers of Americans was the Pittsburgh Survey (1907–08) supported by the Russell Sage Foundation. By 1928 a bibliography of social surveys in the United States listed nearly three thousand titles. Alongside the social survey came the opinion survey. At first these had been amateurish, and the popular Literary Digest poll was taken from names listed in the telephone book. But after 1936, when the Digest poll predicted a landslide defeat of Franklin Delano Roosevelt by Alf Landon, the political-opinion surveys which commanded public confidence came to be based on scientifically selected samples.
The true ancestor of the modern scientific opinion poll was not the old straw poll but the market survey, which, in the early years of the century, had done much to improve the techniques of sampling. Elmo Roper’s first public-opinion survey, based on a carefully selected representative sample, appeared in Fortune magazine in July 1935, and from then on he regularly offered the results of his findings of public opinion about many issues. George Gallup founded the American Institute of Public Opinion, followed by the National Opinion Research Center, the enterprises of Louis Harris, and many others. The polls spread, their results were soon syndicated and became a popular feature of the daily papers.
John F. Kennedy’s possession of the resources, financial and intellectual, to employ opinion polls to the best advantage, was a major factor in his victory in 1960. With the constant advice of Louis Harris, a brilliant public-opinion analyst, in one episode after another Kennedy took shrewd advantage of the information he had obtained from the polls. After each of the TV debates, the polls enabled Kennedy to see which issues were likely to pay off in votes, and where opinion was most mobile. This knowledge had a good deal to do with his crucial decisions in the last ten days of the campaign.
But this role of opinion polls in American political life was something new. As late as 1935 a bill (H.R. 2729) was introduced in Congress which aimed to stop the “vicious practice” of opinion polling by prohibiting the use of the mails for taking polls. As the scientific polls became increasingly accurate they gained the confidence of voters and candidates. In the election of 1944, the predictions of none of the major polls differed by more than 2 percent from the popular vote actually cast. Elmo Roper’s Fortune poll predicted that President Roosevelt would win with a popular majority of 53.6 percent; the election figure was 53.4 percent. The errors of the pollsters in not predicting the election of President Truman in 1948 and the Gallup poll’s 5 percent margin of error in 1952 gave some support to critics. But in later elections, the polls showed such accuracy that in 1966 President Lyndon B. Johnson urged increasing the terms of congressmen from two to four years since public-opinion polls kept them so constantly and effectively informed of the wishes of their constituents.