Rating the Audience

Front matter

The Business of Media

Mark Balnaves and Tom O'Regan with Ben Goldsmith

This page intentionally left blank

Preface

Acknowledgements

Introduction The Single Number Summary

‘The Crossleys’ — Archibald Crossley

Arthur C. Nielsen (and the ‘Black Box’)

Bill McNair and George Anderson

New Forms of Knowledge about Audiences Theorizing the Convention

[he Ratings Intellectuals

Lazarsfeld

The Very Idea of Measurement

Single Source: ‘The Holy Grail’

Summary aming Error Invisible Audiences The BBC: Robert Silvey’s Thermometer and Barometer

The Diffusion of Ratings Technology

Proliferation of Channels and Measurement viii

100 107

[vi] Contents

Neuroscience, Neuromarketing and New Technologies of Measurement 11 Timeshifting and Technologies of Counting 15 The Increasing Technical Complexity of Audience Measurement 18 Calls for Harmonization 20 Summary 22

6 The Ratings Provider 23 The Official Truth 24 The Silent Revolution al ‘Superior Technology’: ATR-OzTAM and ACNielsen Controversy

in Australia 32 ‘Superior Technology’: Nielsen versus Hooper, Nielsen

versus Arbitron 37 Summary 42

7 The Networks (and Other Media Providers) 45 TV Economics 46 Standardization 49 Small Audiences and Set-top Boxes 55 United Kingdom 60 Summary 71

8 Advertisers and Media Planners T2 The Dual Persona of the Advertiser 72 The Media Planner 81 Cost Efficiency and the Curve of Experience 84 The Competent User 92 Summary 95

9 The Audience 97 The Modern Audience 98

The Average Household and the Representative Individual 201

Home Studies and the Public 205 Audience Consent 206 The Knowledge Aggregators 210 Summary 215

10 The Critics 217 The Broader Context 218 The Bogart Persona 221

Objections to Ratings 224

Contents [vii]

Setting Limits to Statistics 230 Problems with Increases in Scale 233 Impersonal Secondary Data 230 Deprofessionalization of Media Research 237 Summary 239 11 The Future of Ratings 241 Bibliography 256

Figures

1.1 W71NY FM station coverage 7

1.2 Creamer’s Letter to Clients 8

1.3 Radio homes in the United States, 1921-38 12

2.1 Extract showing description of methods 32

2.2 Early historical example of Anderson Analysis Radio Ratings Diary 33

2.3 Sample of in-house and fee-based analysis services 42

3.1 From construct to measure 56

3.2 Multi-stage cluster sampling — following the census tracts 61

3.3 Random selection from households for interview 61

5.1 Television meters for the metropolitan market in Australia 103

5.2 Television meters for National Subscription Panel 104

5.3 Canada’s declining network audience 106

5.4 Sands Research brain maps and eye gaze for 2011 Super Bowl advertisement, “VW Darth Vader’ 114

7.1 The money arrow 146

7.2 Projections of West European cable television versus advertising revenues, 2006-2012 148

7.3 Projections of United States cable television versus advertising revenues, 2006-2012 148

8.1 General model of the media planner 183

8.2 Campaign schedule for Renault 190

8.3 Nielsen Online Media Research Analytics 193

8.4 Nielsen BuzzMetrics 193

9.1 McNair Anderson Survey of Radio Listening household appeal, 1970s 207

[ viii ]

Tables

8.4 8.5 9.1 9.2

9.3 11.1 11.2

Breakdown of audiences by social classification (US$) 9 Advertising expenditure 1928-37 (millions US$) 11

Distribution of advertising expenditure, 2007 12

Nielsen panels 14

Structure of China’s TV panels 46

Structure of China’s radio panels 46

Table of random numbers 59

Panel result differences between providers 63

Trade-offs in audience survey research 66

Day-part comparisons 67

Channel comparisons 67

Equipment comparisons 68

Establishment of Peoplemeter panels by country 101

OzTAM audience breakdown 105

Australian radio diaries 106

Television Audience Measurement (TAM) panels 168

Homes viewing matrix 185

Audiences of four successive episodes (household ratings) 188 Nielsen ratings for 30-second advertisements in high-rating programmes 189

Cinema admissions 191

Optimedia’s Content Power Ratings 194

Comparison of Canada and US media households 199 Average time young people spend per day on computer and video games (hours and minutes) 200

Family use of media 200

Essential elements of the audience-ratings convention 248 Selective summary of early evolution audience ratings measurement in the United States, Australia and the United Kingdom 252

Preface few years ago, the Arbitron company introduced the Portable People

Meter (PPM) to measure radio listening in the United States. The

resulting controversy, which was resolved in April 2010 through

voluntary negotiations mediated by a senior member of the US

Congress, neatly encapsulates many of the issues raised by Mark Balnaves and Tom O’Regan in Rating the Audience.

When Arbitron substituted the PPM for diaries in several major US markets, measured listening to minority-oriented radio stations, in particular those catering to the African American and Hispanic communities, dropped significantly in many cases, compared to the earlier diary ratings. Minority station owners, trade associations representing them and a public interest group advocating for them joined together to form the Portable People Meter Coalition (PPMC) to protest. These stakeholders were concerned that undercounting the minority audience would reduce its value to advertisers, thus threatening the viability of minority-oriented radio programming.

Those claiming the PPM ratings were inaccurate pointed to methodological issues regarding the PPM samples. They suggested that the samples underrepresented the minority population, including the younger demographic. They pointed to the relatively low share of mobile-phone-only households in the samples, in contrast to the above-average incidence of such households in the minority community. This was attributed in part to Arbitron’s failure to employ street address-based procedures for choosing the sample. Moreover, in part because the PPM is more expensive to implement than the diary, PPM samples are smaller than diary samples. This could affect the precision of ratings estimates for subsamples of the radio audience. Arbitron argued that the PPM provides more accurate and detailed listening information.

The debate also involved the Media Rating Council (MRC), a private entity that evaluates audience research services and provides accreditation (although accreditation is not a legal prerequisite for operation of a ratings service). MRC is made up of users of audience research, including media outlets and advertising agencies. The Federal Communications Commission (FCC) also became involved, issuing a Notice of Inquiry (NOI) in May 2009. An NOI contains no regulatory proposals; rather, it seeks to gather information. The FCC related issuance of the NOI to its usage of Arbitron radio market definitions for enforcement and evaluation of its radio station ownership rules. The FCC sought comment on various aspects of

Preface [xi]

the PPM situation, including the agency’s legal authority to act. In the event, before any FCC action was proposed, Arbitron and PPMC came to an agreement. This included changes in the sample selection procedure designed to increase sample sizes and to increase representation of mobile-phone-only households in the PPM samples, plus other measures.

The US PPM episode reminds us that the nature of what the authors refer to as the ‘audience rating convention’ remains important and contested as audience delivery and measurement technology evolves. Rating the Audience provides a useful and entertaining look at the history of audience measurement, primarily in Australia, the United Kingdom and the United States. It includes discussions of the major stakeholders in determining the ‘audience rating convention’, the challenges posed to this convention by new technologies for delivering and measuring exposure to content, and changing attitudes of audience members with regard to privacy and willingness to participate in survey research. The analysis reminds us of the crucial role of audience measurement both for informing advertiser spending decisions and for guiding content production in the ‘two-sided market’ wherein broadcasters assemble audiences by providing attractive programming and then sell advertisers access to those audiences.

Rating the Audience includes illuminating interviews with many of the principal architects of audience research and reproductions of various historical artefacts. In the latter category is a 1970s letter that the McNair Anderson company sent to households it was trying to recruit for its Australian radio audience measurement panel. The letter described the purpose of the survey as follows:

The object of this survey is to obtain basic facts to enable those concerned to keep abreast of rapidly changing conditions and to guide them in their programme planning to give continually improved programmes and service to listeners.

So who are ‘those concerned’, and is it really the case that their only or primary purpose is to ‘give continually improved programmes and service to listeners’? The authors stress the ‘syndicated’ nature of much audience research, i.e. that it is supported by a variety of clients. If we follow the money, we find, of course, that prominent among ‘those concerned’ are the radio stations and the advertisers. The latter may well have some concern for the quality of programmes, but it is best to view that concern as instrumental. Ultimately, what they want to do is increase their profits by selling more product. Advertising is an input to that process. So is programming, to the extent that it ‘creates’ the audience for advertisements. Broadcasters themselves are also not insensitive to profits. They may be good citizens, eager to advance the public good, they may operate under licence

[xii] Preface conditions that require public service programming and desire in good faith to meet those conditions, but commercial broadcasters also want to make money. It may well be the case that the menu of broadcast programming that best meets audience preferences (hard, if not impossible, to determine precisely in practice) is different from the menu of broadcast programming that would maximize broadcaster profits (the difference between advertising revenues and broadcaster costs including, prominently, programming costs).

It is amusing and a bit ironic that McNair Anderson chose not to mention advertiser interests in obtaining ratings information, since these interests are so self-evident. And it is fair to say that the authors of Rating the Audience spend most of their time analysing the role of ratings in commercial transactions. As argued above, business firms advertise in order to increase their sales and profits, and they need some information about the impact of advertising on sales and profits in order to decide how much to spend on advertising and across which media and programmes to deploy it.

In drawing the distinction between collecting ratings information on the one hand and ‘audience research’ on the other, the authors remind us that a measure of audience size, even with some associated demographics, is just one input into figuring out what the advertiser really wants to know. Ratings only measure exposure to the advertising message. Ideally one would want to know something about the audience’s engagement with the advertisement. Does an advertisement or advertising campaign prompt audience members to DO something — gather more information or actually make a purchase? And how does the nature of the advertisement itself affect engagement and action? Survey research and research on consumer buying patterns can help to answer these questions, but the decisionmakers and their advisors clearly operate in an environment of some uncertainty. And speaking of uncertainty, how can we really be sure that the basic ratings data, based on sampling of the audience, are accurate? And if they are not completely accurate, are they still usable?

The authors address these questions by examining audience measurements from the viewpoint of various stakeholders — ratings providers, media providers (including stations and networks), advertisers and media planners, and those they refer to as ‘ratings intellectuals’ or ‘media critics’. They discuss the evolution of audience measurement techniques — from telephone recall to telephone coincidental to diary to meter to peoplemeter to Portable People Meter to the installation of software on audience members’ computers to measure online activity. Along the way they analyse a host of important issues. They examine sampling techniques, auditing of ratings results, how stakeholder consensus is reached regarding adoption of new measurement technologies, and the relationship between audience measurement

Preface [xiii ]

and market research. And they address changing public attitudes toward participation in audience measurement activities and market research. Balnaves and O'Regan pay particular attention to the tension between audience members’ desire for privacy and the desire of advertisers for the increasing amount of demographic and behavioural information about the audience that advances in technology are making it potentially possible to gather.

The authors also address the importance of the ratings to the audience. Free-to-air programming is what economists refer to as a public good, that is to say it is non-excludable (the television station cannot prevent individuals from consuming the programming) and non-rival in consumption (Balnaves’ viewing of American Idol does not reduce the amount of programming available to O’Regan). Since it is not possible to charge audience members a fee for free-to-air programming, providers do not benefit from the information about consumer valuation of a product or service that the price system usually provides. The primary piece of information about consumer valuation of free-to-air programming is the size and demographic composition of the audience, as measured by ratings.

So production of programming which appeals to a particular subset of the audience is dependent upon reasonably accurate measurement of the media usage behaviour of that audience segment. If it is accurately measured, then advertisers wishing to reach that market segment would have a basis for purchasing exposures in programming that appeals to that segment. On the other hand, if a demographic, for instance Hispanic Americans, is undercounted in the ratings, one consequence may be a menu of programming that does not cater for the tastes, needs and interests of that group. As one assesses the evolution and performance of the audience rating system, it is important to keep in mind this interest of the audience stakeholders in the process.

The timely look back provided by Rating the Audience offers a useful framework for understanding today’s audience ratings process. And it provides the tools for grappling with how technical change in audience delivery and measurement will Challenge the convention in the years to come.

Jonathan D. Levy*

“Jonathan D. Levy is the Deputy Chief Economist of the US Federal Communications Commission. The opinions expressed herein are those of the writer and do not necessarily represent the views of the FCC or any other member of its staff.

Acknowledgements he authors are very grateful for the cooperation and time of interviewees.

Interviewees for this book have been involved in the foundation and

operation of audience ratings in the United States, Australia and the

United Kingdom or in their critique. They were generous with time and

with help in documentation relevant to the development and operation of audience ratings.

United States

Helen Crossley is daughter of pioneer pollster Archibald M. Crossley and at the age of nine worked with her father counting radio listeners. She became an expert in public opinion polling in her own right and had a long association with the United States Information Agency (USIA), working with Leo P. Crespi to coordinate research surveys in Europe, Asia and Latin America. The surveys measured foreign public attitudes to US policy. Helen is a founding member of the American Association for Public Opinion Research (AAPOR) and the World Association for Public Opinion Research (WAPOR).

Gale Metzger is the retired co-founder and President of Statistical Research, Inc. (SRI). SRI created and provided a number of major media and consumer research services, including: the audience ratings for national network radio (RADAR: Radio’s All Dimension Audience Research 1972-2001); studies and audits of television measurement systems, including the development — in collaboration with major networks and advertisers — of a complete ratings service for the digital era (SMART: Systems for Measuring and Reporting Television); ongoing services to understand how consumers use media — including TV, radio and the internet — in everyday life (MultiMedia Mentor); many sponsored studies related to media usage and advertising, including reactions to programming and products and ad campaigns. The Market Research Council inducted Gale into its Hall of Fame and the National Association of Broadcasters gave him the Hugh Malcolm Beville award in recognition of his distinguished professional career in broadcast audience research.

Peter V. Miller, Associate Professor at Northwestern University, Illinois, is editor of the scholarly journal, Public Opinion Quarterly. He teaches and conducts research on public opinion and research methods. He focuses on issues of validity in social science research, including public opinion polls, government surveys and audience measurements, and on the role of survey research in public policy.

[xiv]

Acknowledgements [xv]

James G. Webster is Professor of Media Technology & Society at Northwestern University, Illinois. His research interests include audience measurement, the behaviour of media audiences and media industries. He has been a member of the Editorial Board of the Journal of Broadcasting & Electronic Media since 1985.

Australia

Don Neely started with Anderson in 1959 as a sales representative, later sales manager and then Managing Director; Ken Sievers was appointed Managing Director of Anderson in 1961. Following restructuring on 1 July 1966, the McNair research business was in the hands of the ‘three lans’, lan McNair as Managing Director, lan Muir as Media Research Director and lan Pilz as Consumer Research Director. In December 1973 McNair merged with Anderson to begin McNair Anderson with lan Muir as Managing Director, Don Neely as Media Research Director, lan Pilz as Consumer Research Director, Roger Layton as Consultant, Jim Grant as General Manager and Max Eve as Secretary.

The authors are also grateful to Gwen Nelson, to Ken Chesterfield for help on the origins of frequency and reach, and to Des Foster for the role of ratings in radio management.

lan Garland has helped the authors on numerous occasions to better understand developments in the subscriber/pay-TV market. He is the founding employee and managing director of Multiview Analytics. The company was established in 2009 to develop research and analytical services for the benefit of the subscription TV industry (STV) and to serve the broader media and marketing community. He was previously Managing Director of ACNielsen Media Australia from 1997 until 2001.

Professor Duane Varan, head of the Interactive Television Research Institute (ITRI) at Murdoch University in Perth, has been very generous with his time to provide explanations on his work and key issues emerging with changing models and metrics.

United Kingdom

Tony Twyman had a major role setting up Radio Joint Audience Research (RAJAR) as well as being Technical Director of Broadcasters’ Audience Research Board (BARB). He has been a Director of Television Audience Measurement (TAM), the holder of the first UK television audience research contract.

Nielsen Mark Neely and Nielsen assisted with checking statistics as well as general editing. The authors could not have provided the insights in this book without Mark’s help

[xvi] Acknowledgements and access to historical archives within the company. Historical and contemporary examples of analysis have been reproduced, with permission.

Permission to Reproduce

The authors drew on material from Mobilising the Audience, a University of Queensland Press 2002 text, in order to exemplify points about the rise of databases. The authors also thank all those who provided permission to reproduce pictures.

Finally, the authors wish to thank their respective families for their love and support. Warm thanks also to Emily Salz and Lee Ann Tutton, Bloomsbury Academic, for their support and counsel, and Howard Watson for his forensic copy-editing.

A huge apparatus is called into play.

Leo Bogart (Bogart 2000a: 301)

In this chapter the authors show why the measurement of audiences has become critical in modern society. Audience ratings are the currency by which decisions are made on whether an audience does, in fact, exist and the subsequent media buying and selling that happens as a result. Ratings, though, are also more than this. Like public opinion polls, they have become an established and socially accepted way of representing individuals and groups in markets or in political decisions. In order to gain this acceptance, modern ratings involves a large range of organizations, private and public, that have to agree on the methodologies and knowledge involved. This agreement is the basis of the ratings convention. As the number of media channels has grown, so has the capacity of audience ratings to represent everything, as the standard, come under pressure.