• by Ross Fadner , March 29, 2004

A study conducted by the Online Publishers Association (OPA) and comScore Networks on the Internet media habits of 18-to-34-year-olds has reopened the debate over discrepancies in the data on the segment collected by comScore and Nielsen//NetRatings, its rival in the online audience measurement business.

The results of the OPA/comScore study, released Monday, reveal that 18-34s represent the most active online demographic segment. The study found that 72 percent of all 18-to-34-year-olds are actively using the Internet, which makes them the most active age group among those with Internet access. Nielsen reported that only 53 percent of 18-to-34-year-old Americans were active Internet users in February 2004.

The OPA report states that 18-34s "account for a significantly larger share of Internet usage relative to their proportion in the total U.S. population" than any other age group. The demographic represents only 24 percent of the population, yet it accounts for 38 percent of the total time spent online and 40 percent of total page views, according to the study. The finding suggests that 18-34s are extraordinarily active online.

Data from Nielsen//NetRatings points to the opposite conclusion. With the exception of the 12-17 and 18-24 demographic segments, Americans ages 18-34 were the least active online age group during February 2004, according to NetRatings data compiled for the MediaDailyNews. Adults ages 35-49 were online an average of 68 hours and 4 minutes per person, while 18-34s averaged 49 hours and 51 minutes. In fact, each successively older generation, including seniors 65 and older, had total online averages of more than 50 hours, according to the data.

Furthermore, NetRatings data states that for the month of February, 69 percent of Americans 35-54 accessed the Internet. At a glance, it would seem that one research firm says 18-34s are the most active demographic segment, while the other claims the exact opposite. Magid Abraham, president-CEO, comScore Networks, says: "I don't think there is any question that the data Nielsen is spewing out doesn't make sense. It doesn't pass the smell test," he says. "Do you really believe that 18-34s are the lowest group in terms of Internet usage?"

ComScore executives wonder whether there is a connection between Nielsen Media Research's under-representation of 18-34s (the TV ratings firm and a Nielsen//NetRatings sibling restated its data for 18-34s recently), and Nielsen//NetRatings' low figure for the demographic. "Are these two connected?" Abraham asks.

Manish Bhatia, senior vice president, products and services, Nielsen//NetRatings, admits that NetRatings estimates are often lower than comScore's across the board. "We stand behind our numbers," he says. "The way they do what they do is very different from the way we do what we do." Bhatia contends that "bigger is absolutely not necessarily better; bigger samples don't mean more accurate samples."

Media directors underscore the importance of a resolution for the interactive media community. David Cohen, senior vice president, Interactive Media, Universal McCann, says that "accurate and reliable research data is the foundation of any viable media business. It is incumbent upon us to understand the root of the differences between the comScore and Nielsen data and agree on a methodology moving forward."

Adds Cohen: "The question is not about why we should have two competing research companies delivering the same audience rating information, but why they are so divergent on basic usage information."

David Smith, president, Mediasmith and Christian Kugel, director, insights and analytics for Starcom IP, say that the real problem for media buyers and planners is that agencies aren't getting dependable reach and frequency tools from either research firm.

"We're not using the ratings themselves to make buys and we never have," Smith says. "You tell me one media buyer who uses NetRatings or comScore total audience data against a demographic to make a decision. The only way those numbers are used is to brag about who is number one, and for the lazy media buyer to say 'I bought the top sites'," he adds.

"Up to this point, it hasn't been that big of a deal," says Starcom IP's Kugel. "But as reach and frequency become more important in the online media planning process, accurate and stable population estimates are going to become increasingly more important, because that's how you project your reach and frequency."

Tobi Elkin contributed to this story