|
Com 316: Fundamentals of Broadcasting: Audience Research |
Ratings are a way of determining audience effects --- Do people like the show? Are they watching it? That's the simplest type. But there are many other types of audience research, and field has become very complex, if not more competitive over the years.
Television Audience Research:
In the beginning there were two key television ratings companies. Nielsen and Arbitron. In 1993, Arbitron decided to get totally out of the television business and concentrate on radio. That has left NMR (Nielsen Media Research) with a monopoly on broadcast television audience research. While that may have been good for them, the broadcasters, advertisers and media buyers are not particularly pleased....
For the first several decades of audience research in television there were primarily two methods used, the diary and the autometer. Arbitron relied on the diary method in both its national and local market measurements. Nielsen used diaries in local markets too, but gradually moved to the autometer in the major markets.
The audimeter ( actually, storage instantaneous audimeter) was a boxed device that was wired into the television set. It monitored to what channel the television was tuned and sent that information back periodically by telephone line to a mainframe computer in Dunedin, FL, Nielsen's headquarters. The device was passive, in that the viewer didn't have to do anything, and this was seen to be an improvement over diaries. There was concern that people didn't always record their viewing accurately if at all.
But there was concern about the audimeter too. The information provided gave no clue as to whether or not anyone was WATCHING...only if the TV was ON and to WHAT CHANNEL it was tuned. Advertisers weren't really tickled about buying TV HOUSE HOLDS when no one was home....
So in the late 1980's Nielsen introduced the PEOPLE METER. This device was supposed to solve the problem of "TV-ON-But-No-One-Watching." Viewers who participate punch in a code number on a hand held device to indicate who is watching. The set-top unit records what channel is being viewed. Immediately upon implementation, there was a hue and cry from broadcasters! Ratings among specific audiences dropped radically and ratings in general fell by several points!
Nielsen argued that problems from the previous methodology were being corrected; networks argued that their audiences were being under-counted! There were some specific problems:
Nielsen continues to defend its methodology, increased the NSI sample size to 5000 in 1996, and is developing a passive people meter. This device would "recognize" various family members by key facial features and record who is in the room. However, how many of you would want such a device hooked up to a bedroom television? There are serious concerns about the invasion of privacy. Another passive device under consideration has been an infrared identification system, but it had a hard time telling the difference between a human being, the "warm spot" on the couch, and the family dog.......Not many dogs buy groceries.....
Sample selection has been a controversy for Nielsen as well. While they go to great lengths to ensure as completely a random sample as possible, there have been problems.
Nielsen uses a multistage random selection method. Sections of the country are selected at random, then sections within that state or community, then blocks or their equivalent in the city or country and then random selection of the housing unit within the block. They try to be as faithful to that system as possible. But in the past there have been accusations that blacks and other minorities were under-reported because Nielsen reps may have been uncomfortable going into certain neighborhoods to place the meters or diaries. Whether that was the case or not, the company has made efforts to correct those problems.
Be sure that you understand the significance of random sampling (that everyone has an equal chance of being selected) and probability.
It is also important to know that as sample size increases, the probability of error decreases.
Such error would be the result of a sampling problem. Usually + or - .05 is acceptable, but two or three points can make tremendous difference to a small market station - whether or not national advertisers look at them for their advertising or not. Rating point difference can be caused by as little as one or two viewers in small market studies.....That's why sampling is so important and why accurate information is so crucial. If you are a independent station manager, struggling to get ratings, you don't want somebody's buddy sitting down and filling out their ratings diary with bogus information. That diary could cost you thousands or hundreds of thousands of dollars if you lose business because of it.
Nielsen Market Research was SOLD in 1999 to Dutch publisher VNU NV for $2.7 Billion dollars. The deal will be examined by the U.S. Justice Department for possible anti-trust violations some time in October of 1999. So far, we haven't heard anything about final approval.
Why would Nielsen sell to a Dutch publishing company? Because that publishing company is already heavily into audience research and media publications. VNU's Marketing Information Services, which had revenues of more than $337 million in 1998, already owns Scarbrough Research, a radio audience measurement service. It also owns Media Reporting, an ad tracking service, as well as Sound Scan. VNU also publishes SRDS a publication which provides information about local station commercial rates and other data (SRDA = Standard Rate and Data Service). Other media publications include Ad Week, The Hollywood Reporter, and Mediaweek. All are widely read trade publications. (McClellan, 8/20/99)
Students are strongly encouraged to visit the Nielsen website at http://www.nelsenmedia.com . Be sure to check the page, WHAT TV RATINGS REALLY MEAN.
NABSS: Nielsen Agency, Broadcaster and Syndication Service: This service combines two former services, the Nielsen Television Index (NTI) and the Nielsen Syndication Service (NSS) and provides continuous information about broadcast network and national syndicated programs. These data are gathered with People Meters and provided to networks, advertising agencies and programming executives. You may heart them referred to as the "over-nights" and these are the ratings that are reported on the web and in the newspaper the day or so after a program airs. Remember, though, these numbers do not represent the entire country, only key major markets. Nielsen now has a national sample of 5000 homes with People Meters.
Nielsen Homevideo Index: This service provides information about use of cable, pay cable, VCRs, DVD players, Satellite Dishes and other developing technologies in television.
NSI: Nielsen Station Index: This service has been provided since 1954 and will expand to 53 major markets with daily audimeter Service in 2001. It also provides diary data during "sweeps" periods for 210 DMAs in the U.S. More than 100,000 diaries collected during each sweeps period.
NMS: New Media Services: NMS offers research for syndicated programs, especially at the start-up point. It also offers SIGMA which electronically verifies air play of PSA, VNRs and program promotions.
NAS: Nielsen Advertiser Services: This service provides data to advertisers to help them design more effective campaigns.
Nielsen Interactive Services: Includes two specific Internet-related services:
NSMS: Nielsen Sports Marketing Service: This service which tries to meet the needs of sports teams and companies as well as networks, stations, and others involved in sports marketing on the national and local level.
Ethnic Reports: Nielsen also offers specialized reports on the viewing habits of ethnic audiences. All ethnic groups are sampled in proportion to their percentage in the national population in the regular Nielsen samples. But Nielsen will draw from those samples to provide specific reports about any specific ethnic group. African-Americans are the largest minority segment of the viewing audience, and actually watch more television than any other segment of the population. Nielsen established the National Hispanic Television Index in 1992 to measure the viewing habits of Spanish-speaking Americans.
Nielsen Media Research Canada: Offers television ratings services to Canada.
For additional information about the Nielsen procedures for market research, see "What TV Ratings Really Mean" at http://www.nielsenmedia.com/whatratingsmean/
Day-part is very important in local ratings and Nielsen defines them as a follows:
|
Day Parts
|
These are EST, so adjust for CST
LOCAL PEOPLEMETERS: Nielsen announced late in October, 2000, that it planned to introduce local peoplemeters into the top 25 markets by 2010. The local people meter will be tested soon in Boston with 420 meters installed by April and 600 by July. The "demonstration service" will be run from October 2001 for at least 12 months. At the end of the "demonstration" period the current meter/diary service will cease in that market.
The implications of local peoplemeters are several:
For more information, see the article by Steve McClellan, "Get Ready for Local People Meters," in Broadcasting and Cable, 10/30/00 p. 38
Local peoplemeters were rejected by television stations in Boston earlier in 2002. Even though Nielsen was pulling all other meters and diaries out of the market, as late as the end of April (before May sweeps) none of the local stations had signed up. The reasons? All of the things listed above plus some additional ones:
Nielsen is committed to the local people meter system, even though it has not be received with enthusiasm....
Sweeps:
Ratings are not taken 52 weeks per year. Rather, selected periods are scheduled when both local ratings occurs. These are part of the NSI service. In some larger markets, there are extended sweeps, but in most of the 210 DMAs sweeps occur in November, February, May and July (only 207 DMAs are rated in July). We are entering the November sweep period right now.
For 2001-2002 season, sweeps occurred as follows:
The advantage of the extended sweeps periods in larger markets is that it reduces the temptation to "stunt" and also the effect of stunting is reduced. It also gives a more accurate picture of normal viewing habits, and allows programmers to make decisions based on trends longer than four weeks at a time. Further, if something happens to change viewing patterns, it may not always happen in a rating period and thus could go undetected for three or four months. For more specific information about sweeps periods, see the Nielsen site.
In fall of 2002, Nielsen had to get "serious" with some stations that were stunting, and one stations was 'delisted' for targeting Nielsen families during the sweeps period. KSWB-TV in San Diego had sent videotapes out to 75,000 homes before sweeps began. It told Nielsen families to watch for their station and "write it down" in the diary. (Staff, Broadcasting and Cable, 12/11/01) Another station sent a message to viewers saying, "Write us down, or we may go dark," but it responded to warning from Nielsen and was not delisted. (McClellan, 11/20/02)
Television Audience Measurement Terms You MUST know:
Other methods of audience research can be used by Nielsen and other research firms:
Diary Method: limitations are:
Sampling terms:
Arbitron:
For many years, Arbitron also did television audience research on both the national and then only on the local level. In 1993, they decided to get out of television and concentrate on radio.
Arbitron refers to individual markets ADIs, "Areas of Dominant Influence."
Arbitron now has three core businesses:
Services Offered
Radio Reports in local ADIs. These data are collected at regular intervals and provide the basis for advertising rates for local radio stations. Arbitron is the largest company gathering radio research.
Diary method, provides local market reports.
National Network Audience Estimate Report: gathers information concerning radio network listenership. 75 radio markets. 22+ radio networks covered in this report.
80% of all radio buys are based on Arbitron ratings.
In radio research, some different terms are used:
SPA: Syndicated Program Analysis : demographic information on syndicated programs.
Pathfinder: An ongoing comprehensive research project that evaluates consumer media behavior and preferences. Data are collected through the use of telephone interviews and mailed surveys to a sample of 4500 consumers.
Arbitron NewMedia Internet Services: Will now provide services to track internet users listening to streamed radio programs. A new agreement with RadioWave.com will facilitate this service for advertisers and others who want to know how many people listen to radio webcasts.
Broadcast Advertiser Reports: BAR: Manual notation of commercials run in each market [for verification purposes]
Portable People Meter (PPM)
Arbitron and Nielsen are testing a new technology which addresses many of the problems with standard audience research technologies. The portable people meter is a pager-like device which consumers wear all day long. It is designed to recognized inaudible audio signals imbedded in the audio portion of programming. Those signals are imbedded in the programs using a decoder provided by Arbitron. At the end of the day, the consumer puts the portable people meter in a special base unit which downloads the information it has collected all day.
Arbitron has been conducting tests of the device in Philadelphia during 2000 and has been working in cooperation with Nielsen on the test. Thirty-one radio stations, two TV stations and 6 cable services and networks are participating. If the test works out, the two companies may work together to deploy the system throughout the U.S.
The portable people meter has four components:
The Portable People Meter needs to meet several specific needs if it will be successful:
In addition, Arbitron wanted something efficient in cost and management. They spent nearly ten years developing the PPM, and it was tested in Great Britain before domestic tests were begun.
The advantages of the Portable People Meter:
The initial tests in 2000 and 2001 went very well, showing larger electronic media use than standard Arbitron and Nielsen methodologies. The next phase will involve installing PPMs in 1500 homes in Philadelphia's market, paralleling the full-market research conducted by Nielsen and Arbtron. This will allow direct comparison between the services for ratings of cable, TV and radio and other media.
In June, 2002, PPM data was released for the Philadelphia market and showed significant increases in viewership. Some of the results....
(Trigoboff, 6/14/02)
Sometime near the end of 2002, Arbitron and Nielsen Media research will get together to see determine whether they will launch a joint venture rating service using the PPM. Nielsen just introduced new state of the art people meters, but they have not been enthusiastically received. Some stations appear to be waiting for the PPM. (Trigoboff, 6/10/02) The two companies will expanded the number of PPMs in the Philly market with a new 1000 panel of consumers and there may be another trial with Hispanic audiences. (Trigoboff, 7/25/02)
Be sure to visit the PPM site at the Arbitron web site at http://www.arbitron.com
For more information see the PPM site at the Arbitron web site. http://ppm.arbitron.com/
Visit Arbitron's web site at http://www.arbitron.com
SRI
RADAR: Radio's All-Dimension Audio Research
Conducted by Statistical Research, Inc. (SRI), RADAR has been around since 1972. They provide reports twice a year, and gather data by using telephone recall interviews. A sample week covers 12,000 respondents during which researchers call the SAME people every day, several days in a row. The primary focus is on radio network ratings.
See the RADAR site at http://www.sriresearch.com/RADAR.htm
SRI provides a number of valuable services to marketers and advertisers. They not only look at programs but how new technologies are being adopted and utilized by consumers.
See their site for information about THE HOME TECHNOLOGY MONITOR. These reports address how people use new technologies, what factors lead people to buy a new or existing technology, and how fast new technologies are being adopted. For example, they report that DVD players are being adopted much faster than VCRs.
They also conduct general marketing and opinion research. For example, after the WTC attack, they conducted research that showed Americans liked patriotic ads and that consumers were ready to return to TV programming and advertising. They also study market segmentation and characteristics of those segments.
Griffin Report:
ACCURATINGS:
New service in 1992 that challenged Arbitron in the radio ratings field. They offered a new concept: radio listeners tend to be loyal to one station at a time, and created a new category of data: "partisans" or devoted listeners. Station GM's really like it.
Different methods than Arbitron:
Result: younger skewing stations tend to perform better
Weekly faxed reports.
Still unknown whether it will be accepted in the advertising community, but it's still operating.
OTHER TYPES OF RESEARCH:
TVQ Rating: conducted by Marketing Evaluations of Port Washington, NY and their goal is to measure likeability and recognition of personalities and programs by audiences. They offer
ASI: AUDITORIUM TESTING:
Most famous location was PREVIEW HOUSE in L.A. which was operated by ASI. The testing is discussed in Head and Sterling. ASI pioneered auditorium testing for television commercials and programs. Viewers are seated in an auditorium and given a handheld device with knob which they turn to indicate whether they like or dislike what they are seeing on a preview screen. A Mr. Magoo cartoon is shown before the content being tested. The cartoon serves as baseline for the researchers for comparison with other tests/audiences. Information has been used to determine whether a commercial, a television series or a film would work as presented. Sometimes, the results will bring about a major change in the content of a work. The ending to "Fatal Attraction" was changed because audiences at Preview House didn't like the original ending.
ASI has recently opened a new facility in Los Angeles called the New Media Center with all the newest technological innovations. They have also constructed six regional testing centers.
Visit ASI's web site for a tour of the facility and excellent information at http://www.asientertainment.com/tour/
Station Consultants:
In TV: concerned with the "look" of personalities; format of programming, "look" of set, promos, IDs, etc. FRANK MAGID is the biggie here. For an excellent article about the company see: http://www.gazetteonline.com/money/mon615.htm "
Others include
In some cases these organizations collect their own data, analyze it and recommend appropriate changes. (Could mean your hair, your wardrobe, or your JOB!)
You can visit the Magid site at http://www.magid.com
Radio station stations also depend on consultants for format information, how to improve personality performance, etc. Visit a radio consulting firm created by a former professor of mine at the U.T., Rob Balon. The company is Benchmark, and you can find it on the web at http://www.austin360.com
A link on the site goes to an interesting article about Auditorium Music Testing and can be found at http://www.austin360.com/ads/benchmark/article1.htm
Station Rep Firms:
These companies are used to sell time to local, regional and network advertising clients. At the present time there are more than a dozen different companies which compile this type of information (sometimes buy it from those mentioned above) and put together profiles of businesses, viewers, prospective clients so that specific programs, spots can draw their attention.
Music Testing:
Auditorium Testing: designed to evaluate recurrents and oldies and how they work in a play list.
72 to 200 people are invited (promised T-Shirt, etc. or small pay) to an auditorium where the researchers play "hooks" of songs (5 to 15 sec). People categorize the song as "unfamiliar," "tired of it," "hate it", etc. They may play between 200 and 400 hooks (more than that not reliable) Stations usually do this once or twice per year. It's great research, but it's expensive.
Call-out research:
This method is designed to test currents and telephone interviews are used. The song hook is played and the listener is asked for a response. Only 20 cuts per call maximum can be played before the respondent gets tired.
This is usually done at the local station on a monthly basis.
Some other methodologies you need to know:
Telephone coincidental: Call and ask what people are watching RIGHT now. Sometimes used by station management to get immediate feedback. Ideally, a random sample should be use
VALS: leader in psychographic research
This method was established by Marshall Marketing, Leigh Stowell & Co., and SRI (those are the RADAR people). It stereotypes consumers by value systems and lifestyles, and also examines how people maintain those lifestyles. It breaks consumers into categories like: actualizers, fulfilled, believers, strivers, experiencers, makers, strugglers. (you don't need to know these) motivations behind consumer decisions.
Scarbrough: Provided a service which gathers information about media usage and product usage. The method used is telephone interview, self-administered questionnaire, and TV diary.
Current issues in audience research:
There were also concerns about local ratings (Arbitron in local radio; Nielsen in local TV)
Concerns over validity of ratings and legal liability if they are indeed, inaccurate, led to FTC ruling and courts have ruled that ratings are estimates and opinion, not fact.
In 1992, NAB committees examined and made recommendations (too complex for discussion here)
CONTAM: NAB's Committee on Radio Audience Measurement
COLTAM: NAB's Committee on Local Television Audience Measurement.
Concerns about Arbitron:
NAB and RAB established task force to investigate complaints from radio industry about Arbitron. They were concerned about 8% annual increase in fees yearly between 92 and 97 and that Arbitron is a monopoly.
Concerns about ratings continued to grow.....
1993: Arbitron dropped its local TV and Cable ratings. One of the reasons was the expensive failure of ScanAmerica. (A service which combined television ratings with information about what consumers purchased. Consumers were supposed to scan the UPC code of everything they bought. They didn't do it. So Nielsen has a virtual monopoly on television ratings.
Where were the out of home viewers?: 1993 study found 28 million viewers watch TV away from home. For those viewers, about 23% of their viewing occurs away from home: workplace, college, hotels/motels, restaurants, second homes, airports, etc . In 1995, that number was down to about 23 million who watched 21% of their television viewing outside of home. If out of home viewing is calculated, MONDAY NIGHT FOOTBALL gets a 7% ratings boost! This is significant because viewing in certain demographics increases when these numbers are added. Prime time viewing among 18-34 adults increases 4% when these numbers added. (That could mean more dollars for networks, stations, etc.)
1994: Networks, due to concern of cost, effectiveness and reliability, committed 30 Million $$$$ to create a new measurement system.
Complaint against Nielsen: Began in 1987 with new people meter system overstates some households, understates others and is skewed toward cable households
Don't want to get rid of Nielsen, just offer some alternatives, other or additional options for information.
The new system would be run by CONTAM and managed by SRI (yes, the RADAR people) A goal was to develop new ratings methodologies which will be effective in new multi-channel /multimedia environment.
In 1996, Fox seriously considered filing a breach of contract suit against Nielsen because of inaccurate numbers. One reason: huge HUT numbers for children's programming but data shows no kids actually watching! (Bstg 4/22/96p15)
Hypoing is another concern: This occurs when programmers use stunting techniques to draw abnormally large audiences during sweeps periods so that they can charge more for their advertising time. Advertisers want to get around the practice by buying time based on 52 wk/yr demos rather than sweeps demos. (Bstg 3/31/97p.32)
As of May, 1997, The American Association of Advertising Agencies submitted a proposal to Nielsen for a 52-week measuring system. The system would be continuous and would be used in all 211 local TV markets. If adopted, the system would eliminate sweeps all together and with it the need for hypoing. (Mediaweek, 5/19/97p65) Nielsen did not embrace the idea.
Meanwhile still more errors keep popping up. In Columbus, GA, the station manager, Lee Brantly (formerly the manager of WAFF-48) found the ratings showed his station 10 points off the competition. It turned out the time periods in the "book" were out of sync and the numbers therefore faulty, but the competition has been using the numbers to distinct advantage and to Brantly's DISadvantage. BSTG 5/12/97p112.
In 1996 S.M.A.R.T. (Systems for Measuring and Reporting Television) was introduced. This is the alternate system commissioned by CONTAM. SMART used 500-home ratings "lab" in the Philadelphia area. It used an easy to use system which includes:
In the summer of 1998 a number of advertisers, ad agencies and the four major networks all subscribed to the service. Broadcast and advertising industry professionals have been concerned about what they call the "arrogance" of Nielsen and its unwillingness to solve the problems with its measurement service.
In 1999, SRI announced it was concluding the S*M*A*R*T project, while some of the research would continue in existing and future media projects. ( http://www.sriresearch.cm/press/pr19990527.htm )
SRI, "SRI Concludes TV Metering Project; S*M*A*R*T Continues." Press Release 5/27/99 http://www.sriresearch.cm/press/pr19990527.htm
The complaints about the current system aren't going away.
April, 2001, brought a problem with Nielsen's mainframe computer and when it was turned back on, it was not programmed for the "time change." So when the ratings were reported, some of them were an hour out of whack....Fox noticed the problem with the X-Files new episode scored much lower than normal. It was pretty embarrassing..... (Schlosser, 4/16/02)
In November, 2002, the chairman of Turner Broadcasting, Jamie Kelner told investors that Nielsen Media Research "didn't do its job very well" because it missed younger viewers. The census reported that the 18-34 audience had grown substantially, but Nielsen did incorporate that into its recording of ratings. That cost Turner-owned WB $30 million in lost revenues because they appeal to that age group. What that means is that probably more people 18-34 were watching WB but because their numbers for the total population were reduced, the percentage they reported that watched WB was also reduced. So advertisers got more than they paid for. (Broadcasting and Cable, 11/4/02)
Lynne Schafter Gross, in her book, Telecommunications: an Introduction to Electronic Media (1997) lists a series of important issues facing audience research and electronic media industries. I'll provide some explanation and illustration to those points:
Accuracy: Many channels have ratings of less that 2%; with such small samples in most local markets, there's no way to know what the real rating is. Any small error will pollute the result.
Total Sample: Random sampling doesn't provide an accurate picture of who's really watching. Some sort of info needs to be gathered on every TV in every home.
Sample Size and Composition: what about rural populations, minorities, at work and out of home viewers? They are either under-represented or not counted.
Non-respondents: Even if you get an accurately drawn sample, some people won't participate, so the randomness is thrown off. If they do agree, they will probably forget to record their viewing at times or not remember accurately if they use a diary. NAB has been tracking the response rates of both Nielsen and Arbitron for several years. You can see those results at the NAB site at http://www.nab.org/resarch/Reports/Nielsen_ArbitronRates/sld001.htm
Technical problems: technology is not foolproof.
Diary: not filled in; not completed or fabricated
Interviewer bias: "I like Star Trek but I'm leaving the house, so I'll tune the TV to that station so they'll get the rating even though I'm not really watching....:"
Hypoing: When you manipulate the programming to get a good rating you don't end up with a representative rating.
Dependence: Billions of dollar depend on the way one company operates. That's too much power in too few hands.
Secretiveness: Research companies don't make their methods public because they're afraid other companies will copy them. As a result, we can't really check the reliability or validity of their methods.
I'll add another:
New Media Technologies: with new technologies offering access to more and different content both in and out of the home, audience measurement research must address not just television, radio and cable new technologies which continue to develop and draw audiences.
Audience measurement is a complex but extremely important part of this industry.
Copyright, 2002
Dr. Janet McMullen
Return to Com 316