In January, leading Internet rating service Nielsen/NetRatings reported the number of men between 18 and 34 who are considered "active" Internet users was 16.2 million, an Internet penetration of about 47%, based on U.S. Census Bureau figures. NetRatings' own user panel, whose Internet activity is automatically tracked each month, registered that 56% men aged 25-34 were actively using the Internet in January 2004.
In contrast, rival ComScore Media Metrix reported in January that active male Web users between 18 and 34 years old numbered 27.1 million, an Internet-penetration factor of 78%. ComScore's panel data showed 77% of men 25-34 were online in January.
The gap has ratcheted up frustration over inexact measurement by Internet companies which rely on the research to sell advertising.
"The online advertising industry could reach $15 billion someday, but discrepancies and inconsistency in data could hinder our ability to get there," said Greg Stuart, president-CEO of the Interactive Advertising Bureau. "It would seem, based on these numbers, that Nielsen isn't any better at measuring guys online than they are at measuring guys who watch TV." (Nielsen/NetRatings is the Internet audience measurement service run by NetRatings Inc., which is 64% owned by VNU, the parent of TV ratings leader Nielsen Media Research.)
Discrepancies in online ratings have been an issue on and off since 2000. But as the Internet matures, reliable standards are becoming increasingly crucial for wooing online advertisers.
"NetRatings numbers are not useful and highly irregular. Their penetration rate for young men is absolutely crazy," said a major online advertiser who requested anonymity.
"These discrepancies between the two major Internet ratings services give me pause, because I want to be able to tell advertisers that the Internet medium is measured as accurately as any other medium," said Eric Duong, founder of Optio, San Francisco, which buys Internet time on behalf of KSL Media for clients including Twentieth Television. "If we don't have standardized Internet audience measurement we feel confident about, it hurts our business."
In addition to the striking disparity in penetration rates for young male Internet users between the two measurement services, NetRatings has raised eyebrows with adjustments to its estimate of the total active Internet audience between 2003 and 2004.
Between those years, NetRatings' estimate of that audience rose 12.8%, from 130 million to about 150 million users (see chart, above). NetRatings attributes that increase primarily to internal revisions of its estimate of the total number of people with access to the Internet. Meanwhile, ComScore's estimate of the total Internet population increased only 6% during that time period, from 145 million to 152 million.
"Our clients have been bugging us for an explanation about NetRatings' data. We can't figure out how they came up with such as steep rise in Internet users in over one year," said Magid Abraham, president-CEO and co-founder of ComScore Networks, Reston, Va. "The big jump in their numbers should throw off trend data," he added.
"I can't account for ComScore's numbers, but I feel our methodology is very solid," said Michael Saxon, NetRatings' VP-media products. "I would agree that higher numbers are not necessarily better, and we strive to do what's best."
NetRatings says ComScore's own methods are suspect. "ComScore is our major rival, and they have an agenda in stirring things up," said a NetRatings spokesman. "ComScore also has a history of revising its own data." In fact, ComScore said it announced to clients in early 2003 that it was restating its Internet measurement numbers to be more accurate, following its acquisition in June 2002 by ComScore Networks.
`a lot of flak'
"We took a lot of flak for that, but we were very forthright about it, and since then we have had very solid data to stand on," said Mr. Abraham.
The methods NetRatings and ComScore use to poll consumers about Internet access, and to track their actual usage, play a key role in the issue.
NetRatings' NetView panel consists of 41,547 individuals (106,000 globally) who are considered "active" Internet users, who were recruited by random-digit telephone dialing, whose computers automatically measure their real-time Internet usage. In addition, every quarter NetRatings performs a random telephone survey to estimate the total number of U.S. individuals with access to the Internet to arrive at its estimate of Internet penetration. NetRatings' home-computer panel participants receive a $50 savings bonds every six months in exchange for participation; its at-work users receive a $100 savings bond.
ComScore's Internet user panel consists of 1.5 million (1.1 million in the U.S.), who are recruited through random-digit dialing and random online recruitment. Participants are rewarded using a combination of cash and non-cash incentives.