by Dave Zornow
Published in Cable Avails magazine, September 1993
As the cable industry rushes toward a multimedia/multiplexed future of 500 channels, the ratings services are under pressure to keep up with an increasingly complicated viewing environment. Most of the spot cable business is measured by a household television diary, a more than 40-year-old paper-and-pencil measurement tool which consistently underestimates cable TV viewing. The cable network business uses Nielsen's people meters, a system many researchers question because of low cooperation rates and inconsistent button pushing. Top local markets use a combination of household meters and paper diaries to create demographic viewing estimates.
Arbitron, Nielsen and several other firms are researching new technologies to keep pace with the viewing environment. Although all of the solutions offer potential improvements over current measurement tools, they also raise new concerns about how we measure television viewing.
Nielsen is developing a "passive meter" to replace the 4000 people meters currently used to measure broadcast networks, syndication and network cable. People meters rely on household members to push a button when they're watching TV, but researchers question how accurately the system captures viewing to demos like children and teens. Fatigue is also a factor, since the amount of data reported by long time people meter households is different from panelists who have recently joined the sample. Nielsen's passive system is designed to overcome these concerns by reducing the burden on the respondent.
The passive system uses a computerized image recognition system which stores a coded image of the faces of family members. The meter continually searches the room, comparing each person's face to images stored in the passive meter. When a match is found, the demographics of that viewer are assigned to the program being watched.
Nielsen is the first to admit that moving passive technology from the lab to the living room is a big step. "We've made great progress overcoming problems of different lighting conditions, changes in the appearance of the respondent and viewing from varying distances, says Ceril Shagrin, Senior Vice President of Market Development at Nielsen Media Research. The passive meter program was originally announced in 1990 with roll out scheduled in 1993. The revised schedule emphasizes increased testing and agency and media acceptance of the technology. "We're not going to roll out until it meets the degree of accuracy we've set for ourselves," says Shagrin. "We need to show the users of national TV data that the system works." Nielsen plans field tests in 1995 with 1996 the earliest date for full implementation.
Although the passive system requires virtually no effort by the respondent, its "big brother" implications are potentially more intrusive than the current button-pushing system. But Nielsen Media Research President John Dimling doesn't think it will be a problem. "The passive meter can only determine if a panel member's face is turned toward the set - it cannot determine what they are doing," says Dimling. "Once people understand what the passive meter is -- and what it isn't -- we find that the cooperation rates are about the same as the existing people meter."
Arbitron is testing new technology which gives survey participants portable meters that can detect inaudible tones added to the audio channel of all program sources. TV, radio and cable programmers embed a signal identifying the program source with its broadcast/cablecast time. Respondents carry a cigarette-sized pocket people meter which picks up all of the codes it "hears." At the end of each day, survey participants plug their pocket meters into a receptacle to discharge all information and recharge the batteries. Data is sent back to Arbitron via modems and standard telephone lines.
The pocket meter approach offers several advantages over current measurement techniques. The same meter can measure radio as well as television providing a single source for electronic media. Because the pocket people meter is carried by the respondent, it is also suitable for out-of-home measurement not possible with a household meter.
Audio encoding has its drawbacks because it requires the cooperation of the media being measured. Arbitron can't identify program sources electronically unless programmers add the hidden audio code. Another concern involves whether respondents will carry the pocket people meter with them throughout the day. Arbitron's Director of Research Jim Peacock says he's encouraged by the 'carry around' rates of prototype meters in field tests. "The pocket people meter has the potential to be accepted like a diary but have the data return rates of a meter." Peacock says field tests show the meter detecting motion among 80% of the respondents during a typical three hour daypart. Arbitron plans to begin live tests in first quarter 1994 with the goal of installing two live markets by the end of the year.
Eight years ago, The PreTesting Company, Tenafly, N.J., introduced a portable metering approach which included all electronic media and magazines. CEO Lee Weinblatt and media consultant Steve Douglas have since developed a prototype which adds a quarter-second one khz tone to each program at inaudible levels. This signal is then picked up by a portable meter which doubles as a watch worn by the respondent. By adding a special microchip to a magazine's binding, the same meter/watch can also tell when the respondent picks up a magazine.
Despite the promise of creating a true "single source" system for measuring TV, cable, radio and magazines, Weinblatt and Douglas have not received wide support from agencies and media. "Agencies and networks say it's wonderful, but when we leave the presentation I think they hope our elevator cable snaps on our way out of the building," quips Weinblatt. Concerns about potential lower viewing, listening and readership levels strike fear into the hearts of media sellers. "Advertisers are the only parties which are anxious to implement the technology," Weinblatt adds.
The most advanced technical solution is offered by MicroMeter, a San Francisco-based startup venture. President Dennis Spragg says MicroMeter is developing a system which uses proprietary acoustic matching technology to identify viewing and listening sources without encoding. "The Micron is a passive audience measurement device no larger than a pocket pager," says Spragg. "By sampling radio, television and cable audio a minimum of four times a minute, the Micron can determine when respondents are exposed to any electronic media by comparing its data back to a recorded program source." Spragg says this approach is superior to the other portable technologies because no media cooperation is needed.
Although MicroMeter has no industry track record, it has garnered support for its approach in the agency, media and computer industries with current and former executives from IBM, US West, MTV Networks, Chiat/Day, National Semiconductor and Intel on its Board of Directors. Roll out plans include live tests in 2nd quarter 1994 and local market roll out by the end of the year.
All of these systems pack more technology into a wristwatch-sized package than Dick Tracy could have ever imagined. But new measurement technologies bring new concerns about the ratings. Nielsen's passive meter can tell whether respondents are in the room and if their faces are turned toward the set. But what about respondents who don't face the set, or listen to television from another room? Are they legitimate "viewers" who should be counted for advertising exposure? Arbitron says their pocket people meter can detect embedded audio from as far as three rooms away. But are these people "viewers?" Humans can pick one audio source -- such as a familiar voice -- from a cacophony of music, noise and extraneous sounds. Will a portable meter be able to determine which source a respondent is paying attention to?
Today's audience measurement techniques allow the user to define viewing by pushing a people meter button or writing an entry in a diary. The new technology will change the rules of the ratings game and the results for all of the players. ##
Dave Zornow is president of TNG Research, a media research and software company in River Edge NJ. You can send him comments on Compuserve at 75060,1335 or care of Cable Avails.
Return to: Home • Published Articles