Audience measurement calculates how many people are in an audience, usually in relation to radio listenership and television viewership, but also in relation to newspaper and magazine readership and, increasingly, web traffic. The term is sometimes used with regard to practices that help broadcasters and advertisers determine who is listening, rather than how many people are listening. In some parts of the world, the resulting numbers are referred to as audience share; in other places, the broader term market share is used. This broader meaning is also known as audience research. Measurements are broken down by media market, which corresponds to large and small metropolitan areas.
The diary was one of the first methods of recording information. However, this is prone to mistakes, forgetfulness, and subjectivity. Data is collected down to the level of listener opinion of individual songs cross-referenced against age, race, and economic status in listening sessions sponsored by oldies- and mix-formatted stations. IBOPE was the first real-time service for audience measurement in the world, beginning in São Paulo in 1942.
See also: Nielsen Media Research. The audience measurement of U.S. television has relied on sampling to obtain estimated audience sizes in which advertisers determine the value of such acquisitions. According to The Television Will Be Revolutionized, Amanda D. Lotz writes that during the 1960s and 1970s, Nielsen introduced the Storage Instantaneous Audimeter, a device that sent daily viewing information to the company's computers using phone lines and made national daily ratings available by 1973. Although the audimeters did not supply sufficient information about audience demographics, it allowed Nielsen to establish diary reports that presented insight into the audience. According to Lotz, the Nielsen sample included approximately 1,700 audimeter homes and a rotating panel of approximately 850 diary respondents. Nielsen was the controlling factor of audience measurement for national network television.[1]
Networks blamed Nielsen for inaccurate rating measurements in the mid-2000s, and the company implemented its automated Local People Meter (LPM) technology.[2] The LPM marked the shift from active, diary-based local measurement to passive, meter-monitored measurement of local markets. Technologically, the LPM is similar to the original Nielsen People Meter; the key advance was that the LPM provided accurate measurements of local markets.[3] [4] The LPM system has allowed the industry to measure year-round, rather than the quarterly "sweeps" periods. Researchers believed that the LPM more accurately reported the full range of programming watched, including channel-surfing. Arbitron's Portable People Meter uses a microphone to pick up and record subaudible tones embedded in broadcasts by an encoder on each station or network, and has been used to track in-store radio.[5]
The introduction of digital terrestrial television (DTT) complicates audience measurement. In a multisignal context, with new content in front of technological convergence, the correct representation of viewing behaviors faces methodological challenges. New methodologies (audio or video matching, water marketing) are needed to measure digital television audiences.[6] Measurement with audimeters faces the dual challenges of analog and digital measurement in a mixed television broadcast.[7]
Because of the Internet, many businesses can sell outside their local markets. This helps them offer niche items that would face challenges in finding customers in their specific market area. In the Journal of Advertising Research, Chris Anderson writes: "For some internet-based businesses, locality no longer regulates the market." Offered wider choices, consumers award fewer of their "votes" to the big hits and more to specialized niche choices. According to Anderson, people always wanted more choices but their desires were obscured by distribution bottlenecks imposed by cost or locality.
New digital technology initially complicated in-home measurement systems. The DVR seemed incompatible with a Nielsen box, which was designed to measure the frequency of a television signal to ascertain the channel being viewed.[8] Since a DVR always produces the same frequency, an active-passive (A/P) meter could be developed to read audio tracks of a particular program instead of the frequency of the television signal.[9] Other challenges to the industry were digital cable, the Internet, and viewing devices other than televisions. As new ways of measurement became available and users could be monitored for content and use, concern arose that sampling techniques might become obsolete. The increasing fragmentation of viewing with different technologies posed difficulties in reporting viewer numbers for content. Nielsen began rolling out its "anytime anywhere media measurement" initiative in 2010, which includes DVR views in its television figures.[10] GTAM (Global Television Audience Metering) is based on the development of audience-metering technology to deal with the challenges in measuring the viewing behavior of consumer households across several platforms (TV, Internet, mobile devices). A/P meters would be replaced by GTAM meters, which were expected to use a combination of active and passive measurement technologies. Unlike A/P meters, however, they would not require a physical connection to a media device.[11]
MediaWiki software can be equipped with the HitCounters extension as a form of audience measurement and determining wikiFactor, a rough measurement of a wiki website's popularity. With the increased popularity of webinars and video conferencing due to the remote work requirements of the COVID-19 pandemic, a set of direct and anonymous audience engagement tools such as Mentimeter and Actymeter became popular.[12]
Nielsen//NetRatings measures Internet and digital media audiences with telephone and Internet surveys. Nielsen BuzzMetrics measures consumer-generated media. Other companies collecting information on Internet use include comScore, Wakoopa, and Hitwise, which measure hits on Internet pages. Visible Measures focuses on measuring online video consumption and distribution across all video advertising and content. GfK's Cross Media Measurement Solutions measures offline sales after TV, Internet, and mobile exposure.[13]
Sightcorp, TruMedia, Quividi, relEYEble, stickyPiXEL, Cognitec, goCount and CognoVision provide real-time audience data, including size, attention span and demographics, using video analytics to detect, track and classify viewers of digital displays. Networked Insights measures online audiences, and released a report[14] ranking television shows by social-media interaction. According to the study, half of the shows on Networked Insights' top-10 list did not appear on the Nielsen Media Research (NMR) list.
The demographic of a show's audience is also measured, and is often notated in abbreviated form:[15]
A ratings point is a measure of viewership of a television show. One television ratings point (Rtg or TVR) represents one percent of television households in the surveyed area in a given minute. In 2004, there were an estimated 109.6 million television households in the United States; one national ratings point represented 1,096,000 households for the 2004–05 season. When used for a broadcast program, the average rating for the duration of the show is typically given. Ratings points are often used for specific demographics, rather than households. For example, a ratings point for the key 18- to 49-year-old demographic is equivalent to one percent of all 18- to 49-year-olds in the country.
A Rtg or TVR differs from a share point in being the percentage of all households; a share point is one percent of all households watching television at the time. The share of a broadcast is often significantly higher than the rating, especially when overall TV viewing is low. A low share may lead to the cancellation of a TV program.
Gross rating points (GRPs) or target rating points (TRPs) are chiefly used to measure the performance of TV-based advertising campaigns, and are the sum of the TVRs of each commercial spot of the campaign. An ad campaign might require a certain number of GRPs in a particular demographic for the duration of the campaign. The GRP of a campaign is the percentage of viewers multiplied by the average number of ads viewed.
Net reach is the total of all audiences exposed to a vehicle, excluding duplicate viewership.[16]
| Weeks | Gross reach | Total reach | Duplication | Total duplication | Net reach | |
|---|---|---|---|---|---|---|
| Week 1 | 1,000 | 1,000 | -- | -- | 1,000 | |
| Week 2 | 2,000 | 3,000 | (300) | (300) | 2,700 | |
| Week 3 | 1,500 | 4,500 | (900) | (1,200) | 3,300 | |
| Week 4 | 1,200 | 5,700 | (1,000) | (2,200) | 3,500 |
Standard audience surveys collect data on audience size and basic demographics. Audience quality is inferred from an audience's demographic profile, but do not measure audience engagement.[17]
Audience-research methods were questioned during the 1990s with the arrival of new media, particularly digital media, and changes in public media habits. People meters, which recorded household viewing by noting the station to which a TV was tuned, failed to capture new viewing habits such as recording programs for playback at a later time or watching a podcast or download on a device such as a tablet or computer. Video on Demand (VoD) enables consumers to decide when to watch programs, and smartphones enable consumers to choose where to access content.[18] Media research companies have been forced to devise new methodologies capable of tracking new and unprecedented viewing habits across a diverse range of different platforms.[19]
Audiences often use several media at the same time; a teenager might be plugged into a radio with earphones while working on the Internet, where they are scanning online newspapers and magazines. Media research companies were poorly equipped for the challenge of simultaneous consumption across several platforms, but is slowly adapting.[20]
Data collected by people meters is available the morning after programs have aired, but data collected in diaries requires more time to analyse. Radio surveys, still used for radio ratings in a number of countries, are normally available quarterly. The time lag in reporting fails to give advertisers sufficient lead time to take corrective action during a campaign.
Other questions have been raised about diary-based data-collection methods. The expectation that diarists participating in a radio sample will accurately record their listening at 15-minute intervals has been challenged.
Surveying listeners for preferences has been criticised as inflexible. Listeners complain that radio lacks variety and depth, but measurement methods facilitate further refinement of already-minutely-programmed formats rather than an overhaul. In the US, listeners hear old favorites rather than new music. Data obtained by some audience-measurement methods is detailed for individual songs and how they are reacted to by each age, racial, and economic group the station seeks to attract. This approach leads to recognizable songs (such as those by the Beatles) which score well with a cross-section of listeners.[21]
In most countries, the advertising industry endorses a media-research company as its provider of audience measurement. The methodology used by the provider becomes known as the "industry currency" for audience measurement. Industry members fund audience research and share the findings.[22] In a few countries where the industry is fragmented or where there is no industry association, two or more competing organisations may provide audience measurement services; such countries are said to have no industry currency.
Four methods of data collection are used to survey broadcast audiences: interviews, diaries, meters, and scanning and modelling. Research companies use different methodologies, depending on where (and when) media is used and the cost of data collection. All methods involve sampling: taking a representative sample of the population, recording media use, and extrapolating it to the general population.[23]