142 West 57th Street
Arbitron has three core businesses: measuring radio audiences in local markets across the United States; surveying the retail, media and product patterns of local market consumers; and providing software applications that give its customers access to Arbitron's media and consumer information via proprietary databases, and enable them to more effectively analyze and understand that information.
The Arbitron Company is a market research firm that gathers information about broadcast media and consumer spending for use by client broadcast media, advertisers, and advertising agencies. The company's principal method of gathering information is through diaries that are filled out by television viewers or radio listeners. Initially the company surveyed the television viewing habits of American consumers in the 1950s. In the 1960s it began surveying radio audiences as well. In the 1990s the firm abandoned TV ratings to its rival, Nielsen Media Research Inc. (eventually renamed ACNielsen Corporation), and began monitoring Internet radio usage.
Early History: 1940s-70s
In the late 1940s, a George Washington University student named Jim Seiler wrote a thesis that described a method using personal diaries to survey television audiences to find out what they were watching. Only one household in ten had a television set then. Seiler, who lived in Laurel, Maryland, was subsequently hired by the National Broadcasting Company (NBC) to survey its audiences. NBC soon developed a business that marketed such information to other broadcasters and advertisers.
Seiler set out on his own in 1949 and founded the American Research Bureau, which later became Arbitron Ratings Co. and then The Arbitron Company. Initially focused on the Baltimore-Washington, D.C. area, the company eventually spread nationally. In 1960 it became a subsidiary of Control Data Corporation, and in 1964 Arbitron entered into radio audience research.
Measuring Radio and TV Audiences: 1980s
In the early 1980s Arbitron had about $80 million in revenue. It was not as well known as its rival, ACNielsen. The two companies were involved in a controversy over measuring cable homes. Nielsen surveyed cable homes four times a year, while Arbitron only did it annually. Arbitron also measured radio ratings and local TV ratings. Only Nielsen measured national network TV ratings, although Arbitron was planning to mount a challenge in that area. Over the past several years Arbitron had spent more than $23 million to meter eight cities electronically in addition to New York, Los Angeles, and Chicago. With 11 metered markets and 5,700 machines, it had more than Nielsen, making it attractive to the networks. In addition, the company planned to add six more cities in 1984 for a total of 7,200 meters. For Nielsen, broadcast ratings accounted for about 11 percent of its $644 million in 1982 revenue.
Arbitron used its Area of Dominant Influence (ADI) to classify its ratings, while Nielsen used its Designated Market Area (DMA). Arbitron provided annual rankings of ADIs in terms of number of TV households; it would also add or combine ADIs each year. In smaller markets, where it was important for a television station to be identified with an ADI, changes made in ADIs could result in complaints and even litigation. Usually Arbitron's and Nielsen's respective rankings of television markets would not be in complete agreement.
For radio ratings, Arbitron issued quarterly market rating reports with such information as audience trends, discrete demographics, audience composition, demographic buyers, hour-by-hour average quarter-hour listening estimates, listening locations, loyal and exclusive audience, overnight listening, and ethnic composition. The company used an Arbitron Radio Advisory Council, which channeled feedback from the radio community back to the company.
Radio stations typically entered into one-, three-, or five-year contracts with Arbitron to provide monitoring and rating services. In 1986 Arbitron had about 1,750 radio station customers. It was facing new competition from Birch Radio, established in 1978, which had grown to cover 230 radio markets. Birch's clients were primarily advertising agencies, advertisers, and smaller market radio stations, but it was planning to go after contracts with major market radio stations as well. While Arbitron used its diary system to measure listening habits, Birch employed a telephone recall system in which respondents were interviewed about what they had listened to the previous day. Birch was undercutting Arbitron's pricing by about half and was also battling for acceptance of its methodology. In 1987 Birch merged with Scarborough Research Inc.
Arbitron introduced its people meter, ScanAmerica, in late 1986, by testing it in Denver. It was planning to introduce people meters for the 1988--89 TV season. Meters were expected to provide more accurate information, as television viewership was becoming fragmented between broadcast network, independent, syndication, and cable television. Arbitron's ScanAmerica was designed to package national TV audience ratings with household purchase data. The service would require participants to scan the bar codes on purchased products with a wand. Primary customers were expected to be packaged goods advertisers. Critics argued that ScanAmerica was too ambitious, and that it would be too difficult to collect both household product purchase information and national TV ratings. Arbitron committed $125 million to develop ScanAmerica.
Introducing New Services, Discontinuing Others: 1990s
By 1990 Arbitron had nearly 2,000 full-time employees and more than 1,000 part-time employees nationwide. It introduced a new electronic monitoring system called MediaWatch in the spring of 1990. Targeted to advertisers who wanted to monitor their commercials and those of their competition, MediaWatch would enable more accurate identification and classification of TV ads on a full-time basis in most of Arbitron's 75 markets.
Arbitron was one of three firms that provided single source data that tracked consumer behavior to determine the influence of advertising exposure among individual consumers or groups of consumers. ACNielsen and Information Resources Inc. were the other two. All of the systems were scanner-based, generating a database at the store level to determine the effectiveness of marketing promotions. Arbitron planned to expand its ScanAmerica service from its initial test in Denver to provide local coverage for the top 25 markets by 1995.
During the 1990s sample size would be a controversial issue for Arbitron's radio ratings. The company announced in the fall of 1990 that it would increase the sample size of its radio audience in all markets. A ten percent increase would be included at no charge, with customers having the opportunity to purchase extra samples in addition to the ten percent increase in standard markets. The company had been advised by the Arbitron Radio Advisory Council to increase its sample size by 33 percent over a three-year period.
Arbitron also faced strong competition for its TV rating services from ACNielsen. A 1990 survey by Petry Television indicated that stations were discontinuing the common practice of using both Nielsen and Arbitron ratings services. In the top 50 markets, more than 40 percent of TV stations subscribed to just one service for the November 1989 ratings book, double what it was three years ago. Cancellations were estimated to cost Arbitron and Nielsen between $12 and $15 million annually in lost revenue.
In October 1990 Arbitron dissolved its supermarket tracking unit, Sales Area Marketing Information (SAMI), and entered into a cross-licensing arrangement with Information Resources Inc. of Chicago. SAMI had tracked supermarket sales data through warehouse withdrawals. Now, Arbitron would use IRI's data, which was gathered through the use of scanners at retail checkout counters. IRI would pay Arbitron $7 million over four years and receive any revenue from Arbitron's clients for the service. IRI also got access to Arbitron's broadcast tracking data.
Arbitron's 1990 revenue was estimated at $187 million by Marketing News. The company's principal businesses were TV audience measurement services in 209 local markets and radio audience measurement services in 262 local markets. In addition, two ScanAmerica markets were in operation, namely, Denver and Phoenix. The firm's MediaWatch system recorded TV commercial exposure primarily for advertisers and their agencies. Additional revenue came from the publication of specialized reports on TV and radio usage.
After testing and fine-tuning its ScanAmerica service in Denver for three years, Arbitron was prepared to install ScanAmerica in 1,000 households in Atlanta, Chicago, Dallas, Los Angeles, and New York in the spring of 1991. At the time Arbitron, like many other broadcast industry-related services, was undergoing financial difficulties. Arbitron announced company-wide layoffs in early 1991, eliminating 52 positions.
In 1991 the company's advisory board continued to press Arbitron to increase its sample size, this time by 20 percent over the next two years. It also wanted the company to reduce the reporting service from four books to three books a year. Arbitron had about 1,800 radio station clients at this time. Radio subscribers were given the opportunity to vote on reducing the number of ratings periods from four to three. Ultimately, the radio and advertising subscribers voted to continue the four 12-week audience surveys. However, the company vowed to continue to try to find ways to increase the sample size and, thus, the reliability of its surveys.
TV stations continued the trend of opting for only one ratings service. CBS chose Nielsen over Arbitron for its five owned and operated stations. Out of a total of 811 TV stations, 54 percent subscribed to a single ratings service, with only 43 percent subscribing to both Nielsen and Arbitron. Of the 811 stations, 592 (73 percent) subscribed to Nielsen and 543 (67 percent) to Arbitron.
Later in 1991 CBS contracted with Arbitron for its ScanAmerica service for network ratings as well as for local market ratings for its owned and operated TV stations and its seven AM and 11 FM radio stations. The ScanAmerica service would consist of 1,000 households in five markets: New York, Los Angeles, Chicago, Dallas, and Atlanta. CBS was the first network to sign up for ScanAmerica, which Arbitron rolled out on November 4, 1991, as a result of the contract with CBS. CBS had just recently gained the number one spot among broadcast television networks and would also continue to use the Nielsen ratings service. Within a week Arbitron signed a multi-year contract to provide all seven Fox-owned TV stations with local ratings services.
At the end of 1991 radio stations were applying pressure to Arbitron to improve its radio ratings system. They wanted Arbitron to broaden the sample size, improve diary reporting, and make the system more economical. Arbitron identified its top priority to improve the ratings system as improving response rates from diary-keepers. Weak reporting was especially prevalent among Hispanics and young males.
Starting in 1989, Arbitron began holding a consultants fly-in every December. At the 1991 meeting major issues involved ways to improve reporting among 18- to 24-year-old males. Arbitron's policy at the time was to exclude group quarters, such as college dormitories and the military, from its samples. The company planned to introduce two new services in January 1992. One was a radio version of MediaWatch, that would monitor radio advertising. The other was a disc-supplied version of Radio Fingerprint, which broke down raw diary data into age, sex, zip code, and format preferences.
Arbitron's main competitor in radio ratings, Birch Radio Inc., ceased operations at the end of 1991. The ratings service was shut down by its parent company, VNU Business Information Services Inc. Arbitron and VNU signed a five-year contract that made Arbitron the exclusive marketer of VNU's Scarborough report to radio and television stations. The demise of Birch left Arbitron as the sole company providing data on local radio audiences.
In March 1992 the Federal Communications Commission (FCC) announced it would use data supplied by Arbitron to determine radio station ownership limits. The new limits would prevent broadcasters from purchasing a new station if it would give a broadcaster more than 25 percent of the total market, based on audience share, in markets with 15 or more stations.
The National Association of Broadcasters (NAB) announced it would hold a summit in November 1992 to discuss problems with Arbitron's rating system for radio. The chief complaint seemed to be that audience samples were too small, due largely to low response rates from some audience segments. Arbitron had consistently rejected increasing the sample size as being too expensive. Instead, it wanted to lengthen the current three-month reporting periods. Some hoped that new technology would provide a solution.
In television, Nielsen continued to gain market share at Arbitron's expense. Between May 1991 and May 1992, Nielsen increased the number of exclusive clients from 286 to 323, an increase of 13 percent. Arbitron lost eight percent of its exclusive clients, dropping from 199 to 184 during the same period. Nielsen served a total of 641 commercial stations, up from 626, while Arbitron served 502 stations, down from 539 stations. With the addition of Lifetime and TNT cable networks to its MediaWatch service, Arbitron was monitoring about 80 percent of all cable TV advertising, including CNN, ESPN, Family Channel, MTV, USA Networks, and WTBS.
Focusing on Radio Audiences: 1992-2000
After four years of research and ten months of operation, Arbitron dropped its ScanAmerica service for lack of a viable subscriber base. The company said it planned to continue to offer local ratings service in six markets using ScanAmerica technology. The move left Nielsen as the only company measuring national network TV audiences.
Control Data split into two companies in 1992, Control Data Systems Inc. and Ceridian Corporation. Ceridian would become Arbitron's parent company. In anticipation of becoming a public company, Arbitron released revenue figures indicating that annual revenue from radio was $94 million out of total revenue of $194 million. The ratings company had 1,900 radio station customers.
In September 1992 another rating service entered the radio market. Strategic Radio Research's AccuRatings service used weekly telephone calls to survey radio listeners. At first it targeted radio stations rather than advertisers for clients. In 1993 AccuRatings was introduced to New York City. The company offered the service free of charge for one year to advertising agencies. The AccuRatings service was currently available in 14 markets representing about 20 percent of the national radio audience. By March 1994 AccuRatings was available in 25 markets, including eight of the top ten metropolitan areas.
In December 1992 President and CEO Anthony 'Rick' Aurichio was replaced by Stephen Morris. Morris was formerly president and CEO of VidCode, a media company that tracked broadcast commercials, and head of the Maxwell House division of General Foods. As president and CEO of Arbitron, Morris would report to Lawrence Perlman, chairman, president, and CEO of Ceridian Corporation, the successor to Control Data. Aurichio's sudden departure was clouded by concurrent setbacks at Arbitron, including the end of ScanAmerica. In addition, several radio station groups, including Hearst, Westinghouse, and Post-Newsweek, had cut back on Arbitron service.
In December 1992 Arbitron announced it would test a new passive measurement system using a device called a portable people meter. The device would pick up inaudible codes etched into radio or TV soundtracks to record viewing and listening habits. Pretesting Co. already had a patent on a wristwatch device that used an audible code and was watching Arbitron closely. Arbitron planned on introducing its people meter in two markets in late 1994. The device utilized technology developed by Martin Marietta Corporation, as well as the BBM Bureau of Management of Canada, AGB McNair of Australia, and Intellisys Automation, in addition to Arbitron.
In April 1993 Arbitron presented another plan to increase the sample size of its radio rating service. The company proposed replacing the current system of 12-week surveys with 18-week rolling average reports that would be distributed every six weeks. Thus, the new system would have eight ratings periods instead of only four per year. Although the new system would involve additional costs, Arbitron did not plan to charge more under the new system.
Arbitron revised its plan to increase sample sizes after receiving negative feedback from broadcasters. The company scrapped the concept of 18-week rolling averages, which was not well-received by radio stations, and announced it would increase sample sizes by distributing more diaries. By the end of 1994 the number of diaries in continuously metered metro areas would increase by 30 percent. Then, the sample size would be increased another 40 percent in 1995 to achieve a 70 percent increase over 1993 levels. In noncontinuously metered markets, the sample size would be increased by a total of 40 percent through the introduction of new diaries. As part of the plan, radio stations would be charged an additional four percent and Arbitron would begin using a computer imaging system for diary storage and retrieval in the fall of 1994. Developed jointly by Arbitron and IBM, the system captured exact images of every diary page and sorted it on CD-ROMs for retrieval during processing and review.
Arbitron announced it would terminate its broadcast and cable TV ratings services effective December 31, 1993. A company spokesperson noted that competition from Nielsen resulted in fewer contracts and declining price per contract. Arbitron had lost money for several years in the television ratings part of its business. It would cut more than 700 positions from its workforce, leaving the company with about 550 employees. The company planned to continue developing information tools for TV and cable stations to work with advertisers, and it planned to continue development of its people meter for radio and TV. In May 1994 Arbitron joined with GE Capital and investment banker Veronis, Suhler to make a multimillion-dollar equity investment in ADcom Information Services. ADcom used a technology that combined pay-per-view with passive measurement to install household meters for cable TV ratings.
In November 1994 Arbitron acquired a 50 percent interest in Scarborough Research Corporation from VNU Business Information Services. In exchange, VNU received a 50 percent interest in Competitive Media Reporting, a subsidiary of Ceridian Corporation. Scarborough conducted qualitative research on the habits and media usage of consumers in 58 markets. It was expected that a joint venture between Arbitron and Scarborough would bring new research techniques to Arbitron's radio ratings service.
In March 1995 Arbitron released its first study of new media preferences. The company collected information about consumer electronic and entertainment preferences from 4,000 U.S. consumers. Among the technologies studied were home entertainment media, online services, and CD-ROMs. The company also entered into a joint venture with ASI Market Research and Next Century Research to create the Interactive Information Index, which would rate interactive media in terms of consumer preferences.
Arbitron was plagued with data collection errors in 1994, when books from nine markets had to be reissued. To address this problem the company created a new position, director of data collection, and rehired Pierre Bouvard as general manager. Beginning with its spring 1995 radio survey, Arbitron began calling college and military residences that were equipped with private phones in an effort to better measure the traditionally hard-to-measure 18-to-34-year-old male group. More than 400,000 such numbers were added to the sample base in 51 markets.
In 1996 Arbitron won a patent infringement lawsuit brought against it by Pretesting Co. in 1993. The suit centered on Arbitron's portable people meter, which Pretesting claimed violated its portable sound-technology patents. Because of the lawsuit, Arbitron had only developed a working prototype of the people meter, which was a passive measuring device that recorded inaudible sound signals encoded in radio and television programs. Arbitron's revenue for 1996 was $153.1 million.
At the end of 1997 Arbitron acquired Continental Research, located in London, England. It was the first step in Arbitron's plan to expand outside of North America. The company planned to bid on the official U.K. radio ratings contract when it came up in 1998. However, it was unsuccessful in its bid, but still planned to develop a presence in the United Kingdom. In a related acquisition, Arbitron purchased the overseas TV and cable software products of Tapscan Inc., a software developer based in Birmingham, Alabama. Later in 1998 Arbitron began testing its personal portable meters in Manchester, England.
In 1999 Arbitron released a study on Internet radio habits. By the end of 1999 the company had released its first monthly report on Internet radio usage. By 2000 the monthly reports covered nearly 400 channels on the Internet. Arbitron would form an alliance with Lariat Software in mid-2000 to use its MediaReports software to collect data directly from servers to report the number of unique listeners to an audio channel on the Internet. A competitor, MeasureCast Inc. of Portland, Oregon, began its streaming audience measurement service in August 2000.
In 1999 low response rates became a crisis of sorts for the diary method of tracking radio and television audiences. As a result, both Arbitron and Nielsen were under pressure to develop alternatives to the diary method. Arbitron's average response rate on its radio listener surveys was 35.9 percent, compared to 39.1 percent in 1998. The lowest response rate in any single market dropped to 26.3 percent. Among the steps Arbitron took to boost response rates was to offer higher cash payments, make more telephone calls to survey participants, and repackage the diaries. Advertising buyers were disgruntled at the start of 2000 when Arbitron had to delay the release of its fall 1999 ratings report by 21 days. The company explained that the delay was due to its new diary-processing system and issues related to Y2K compliance.
In mid-2000 Nielsen and Arbitron announced they would cooperate in testing Arbitron's new portable people meter. The test would take place in the Philadelphia-Wilmington, Delaware, market with 300 people during the fourth quarter of 2000. For the test, Nielsen would contribute its expertise in TV surveys.
In July 2000 it was announced that parent company Ceridian Corporation would spin off Arbitron in 2001. Ceridian and Arbitron would split off into two publicly traded companies. In 1999 Arbitron's $215 million in revenue was about 15 percent of Ceridian's $1.3 billion in revenue. Arbitron consistently had annual profit margins of 30 percent or more, making it a cash cow for its parent company. It was speculated that once it became a public company, Arbitron would be acquired by another media firm, possibly the Dutch-based VNU, which acquired Nielsen one year after Nielsen went public. Taylor Nelson Sofres of the United Kingdom was mentioned as another possible buyer.
As Arbitron prepared to become a publicly traded company, it held a near monopolistic leadership position in radio ratings. The company had also established itself in rating Internet radio usage. During the 1990s it nudged along new technologies to improve audience measurement, culminating in its portable people meter being tested in 2000. In the new millennium, Arbitron would likely continue as the most reliable radio ratings service in the industry.
Principal Subsidiaries: Tapscan WorldWide; Continental Research (U.K.).
Principal Competitors: ACNielsen Corporation; Strategic Radio Research Inc.; MeasureCast Inc.; Information Resources Inc.