Skip to main content

Testing, testing

There's a whole lot of testing going on at Nielsen Media Research as the company, which owns the TV ratings business in the U.S., prepares to measure viewing in the digital age.

The tests, which began last year, are designed to address the myriad of unanswered questions concerning how to measure viewing in an environment where viewers will have access to hundreds of TV signals, including HDTV, multicast channels, interactive components, personal video recorders, over-the-air data streams and video-on-demand.

The two biggest tests are being conducted in the Northeast, where Nielsen is testing a new meter that executives say will probably be required to measure multicasting of digital TV signals. The company is also in the early stages of a test in Boston for local peoplemeters, which would, like the national peoplemeter, measure both household and demographic data electronically, thereby eliminating the need to collect demographic data from handwritten diaries kept by viewers.

But local peoplemeters are controversial. TV researchers say stations fear that their ratings will go down-just as ratings for the networks dropped when Nielsen first switched to peoplemeters on a national level in 1987.

The company is also testing lower-cost local household meters, which, if implemented, could bring metered measurement to smaller markets. Nielsen now has metered 48 markets covering 62% of all viewership, with five more markets scheduled to be metered in the next two years.

In local metered markets, household ratings are measured electronically by the meters, while the demographic data is still collected by diaries. Not the ideal, admits Nielsen Research Media President John Dimling. The ideal, he says, would be to have peoplemeters in local markets to complement the national peoplemeters. And ultimately that is the goal, says Dimling.

For almost five decades, Nielsen has had two completely different ratings systems measuring network television and TV stations in local markets. The national service is called Nielsen Television Index (NTI), which currently uses 5,000 peoplemeters nationwide to measure all national over-the-air and cable TV signals.

At the local level, the Nielsen Station Index (NSI) measures TV station viewing. In local markets without meters, roughly the bottom 150 TV markets, all ratings-household and demographic-are based on diaries that are kept in the four so-called "sweeps" months of February, May, July and November. The local stations use the data from sweeps to price their advertising inventory.

The long-term goal would be to merge NTI and NSI, says Dimling. As to the timetable for merging the two services, he says, "That will depend on how quickly we can roll out local peoplemeters. And that will be a function of customer response. But that's definitely the direction we're going. I think it makes sense for everybody."

But local Boston TV stations aren't happy with Nielsen's decision to use their market as the guinea pig for local peoplemeters. "It will be disruptive and unreliable, at least at first, and I wish it were happening elsewhere," said Paul LaCamera, vice president and general manager, WCVB-TV Boston.

LaCamera says the biggest concern about the test is that it is being financed by "an interested party," namely MediaOne, the cable company that operates a major system in the market. And local TV stations haven't been asked to participate, he says. Together, that makes the entire project "very suspicious." When the peoplemeters were first introduced nationally, cable got higher ratings, while broadcasters took a ratings dip, which probably explains MediaOne's interest, says LaCamera.

Other broadcasters in Boston aren't as worried about MediaOne's participation. WBZ-TV Vice President and General Manager Ed Goldman chalks it up to salesmanship on Nielsen's part. Nothing wrong with that, he says. But he also says the local peoplemeter represents a "Band Aid" approach to the problem, although he's not sure what the best remedy is.

It would be years, of course, before the two services become fully integrated. And the local peoplemeter is just one outstanding issue. The national ratings are also more precise; they measure viewing literally minute-to-minute, while the local diaries and meters measure viewing by the average quarter hour. Again, local stations resist changing that for fear of a ratings decline.

On the digital front, Nielsen has many issues to address. In addition to multicasting, Nielsen has to come to grips with the new breed of personal video recorders, whose brand names include TiVo and Replay. NMR recently came to an agreement with Replay to develop software that would be installed in Replay PVRs and allow Nielsen to measure the viewing in households with such devices. Right now, it can't.

Dimling says that they are in talks with TiVo to work out a similar agreement.

But despite the many outstanding issues, time is on Nielsen's side, given the much slower than anticipated rollout of digital television. When asked how far away Nielsen is from being ready to measure viewing in the digital-TV age, Dimling replies, "I think we are on schedule to measure it as it is likely to develop."

The digital rollout, he notes, "seems to be moving a lot more slowly, let's say, than [former FCC Chairman] Reed Hundt had anticipated three or four years ago. The equipment change-out and additional programming are expensive."

The industry's slow transition to digital has worked in Nielsen's favor, Dimling admits. "It has given us more time with the a/p meter," the next-generation device known as the active/passive meter.

The test of the meter in the Northeast in 150 homes will probably last another six months to a year, he says. In that test sample, Nielsen is using both the existing peoplemeter and the a/p meter in order to get direct ratings comparisons.

It's still unclear what meters or combination of meters Nielsen will use to measure digital TV. "We will be changing the technology we use for collecting data if stations go to multicasting," Dimling says. But changing to what will depend on the outcome of a number of ongoing tests.

The a/p meter is just one option. Another possible option is the Portable Peoplemeter (PPM), which has been developed by Arbitron.

The current peoplemeter measures channels that TV sets are tuned to. In the digital age, where broadcasters can send multiple signals in one channel, meters have to monitor individual programs by reading codes embedded in them by the producers. That's what the a/p meter does, using what Dimling calls an updated version of the AMOL (Automated Measurement of Lineups) technology, currently used by Nielsen to verify network TV lineups on affiliates.

Arbitron's Portable Peoplemeter is also based on code-reading technology-specifically an audio code that can monitor both TV viewing and radio listening. The beauty of the PPM, researchers say, is twofold. First, it's on all the time and doesn't require the constant logging in and out that the current peoplemeter does. And the PPM is mobile-it goes where the user goes and picks up any broadcast signal within earshot of that person all day long. The current peoplemeter is restricted to in-home viewing and excludes all the viewing that occurs in hotels, vacation homes, college-dorm rooms and bars.

But there is a potential drawback: The person using the device has to download data each night by plugging into a docking device. The unknown is whether users would have the discipline to do so.

That's one of the questions Arbitron hopes to answer in a test of the device in Philadelphia to start later this year.

Earlier this year, Nielsen and Arbitron signed an agreement that could lead to the two companies forming a joint venture to use the PPM to measure local TV and radio. But that will depend on the results of the test, which Nielsen has agreed to help shape on the TV-ratings side.

While a joint venture is possible, Dimling stresses that it's "all hypothetical" at this point. "We believe the peoplemeter is the standard for measuring television," he says. So, basically, the burden is on other devices, through testing, to demonstrate they do it better.