Set-Top Data: Taming the Wild West

Author:
Publish date:
Updated on

On our Webinar yesterday, “The Changing Face of Viewer Analytics,” Alan Wurtzel, NBC Universal’s president of research and media development, questioned the value of set-top box data today to measure TV viewing today: “Set-top box data is not, by any means, a panacea.”

One example he cited: NBCU compared reports from three different research providers based on the same set-top data from Charter’s Los Angeles system — and found 6%-7% difference in the ratings.

“Nobody was ‘wrong,’” Wurztel said. They were just using different measurement rules to account for things like when to assume a TV isn’t be watched even if the set-top is still on. Research vendors that collect and resell Charter Los Angeles data include Nielsen, TNS Media Research and Rentrak.

That variation in set-top box measurement practices, Wurztel said, was a big reason why NBCU and 13 other media companies, ad agencies and advertisers formed Coalition for Innovative Media Measurement (CIMM). He said CIMM can provide focus to measurement providers, rather than everyone providing a “laundry list” of priorities.

Wurztel, though, has other issues with set-top box data, including that it doesn’t provide demographic information about who’s watching a given TV program. “Why is this a step forward, when in fact it’s giving us less information than conventional Nielsen data is right now?” Wurtzel asked rhetorically.

Tracey Scheppach, SMGX’s SVP and video innovation director, responded that set-top box data can be correlated with household-level demographics from companies like Experian: “What we’re really after is, we’re trying to buy true audience at scale.”

Scheppach in general was more bullish on the promise of set-top data — and its utility today. Her perspective is that it can provide better insight than the traditional panel-based ratings system, which she said was “essentially archaic in terms of its potential to measure our true audience.”

“We have finally gotten to the place with audience fragmentation that the opportunity [for set-top box measurement] is now meeting the need,” Scheppach said, noting that the agency has been working with set-top box data for the last three years.

George Shababb, president of TNS Media Research, acknowledged that there are limitations with set-top data today. One big issue: data is available for only 35% of the U.S. “One broad shortcoming is coverage. We do think that is going to improve over time,” he said. “But sitting here today, we are measuring pockets of television viewing” with data from DirecTV, Charter and others.

But Shababb said there’s also tremendous potential. Besides representing a very large data set — in the millions of viewers, versus the 18,000 in Nielsen’s national panel — set-top box data is free of non-response bias and “respondent fatigue” over time. “So there’s inherent quality to set-top data that we believe will result in a better measurement tool long-term,” he said.

Later in the discussion, Wurztel clarified that he wasn’t opposed to using set-top data for media research, just that there’s work to be done: “I’m not arguing we don’t want it. I’m simply saying that, we’re not really close to that yet.”

You can access an archive of the Webinar here (free registration required): www.multichannel.com/analytics.

Related