Leading Web Measurement Firms Merge to End Industry Infighting.
But is the Devil in the Details?

by Dave Zornow

Here are excerpts from my November 1998 column for Cable Avails magazine. For the complete story and lots more on the business of buying and selling Cable Television locally and nationally, call Cable Avails at 303 837-0900. 

Ad-supported web sites and ad agencies are breathing a sigh of relief after the two leading providers of World Wide Web usage have agreed to merge their operations. RelevantKnowledge, based in Atlanta, and Media Metrix, with offices in Manhattan and Long Island will join forces to create a single company with 90 employees and about 250 clients. The new company will be called "Media Metrix: The Power of Relevant Knowledge" and will offer a combined sample of 40,000 business and home users covering 15,000 web sites.

The merger is in response to frustrations expressed by online advertisers and agencies for uniform measurement standards and more consistent usage reporting. Clients to the new company are just happy that the two firms have agreed to terms and will cease publicly sniping at each other. "This medium is getting slaughtered because it can't be reliably measured," says Richy Glassberg, senior vice president at Turner Interactive Sales. "The very least this will do is keep them from fighting with one another. We are getting killed because the agencies and advertisers can't quantify what we have to offer." Glassberg says the confusion created by methodological and reporting differences discourage agencies from spending money on web advertising.

Clients hope the new Media Metrix will overcome some of the limitations of its two predecessors. For example, Media Metrix had a relatively small representation of business users as compared to RK. RelevantKnowledge always struggled with measuring America OnLine, a service which some estimates say accounts for 50 percent of U.S. usage of the World Wide Web. Their methodology for recruiting respondents was also different: RK used a researcher-preferred random probability sample designed to properly represent the web population; MM preferred a "random mail" approach using targeted mailing lists and members of parent company NPD's diary sample. Each company differed in the way they collected data from respondents with RK using the web to collect data in real time and Media Metrix requiring panelists to mail back a diskette containing information collected with its PC Meter software. As a result of these differences, the monthly report of top 25 web sites produced by each company showed contradictory lists of the web's most popular pages.

Client reaction to the merger has been overwhelmingly favorable. "This creates the motherhouse of internet research companies," says Lynn Bolger, senior vice president and media director for APL Digital, the online division of ad agency Ammirati and Puris. "Both companies are combining their strengths into a single strong source instead of using their energies to rip each other apart." Cable networks like Discovery agree that the merger can only work to strengthen the fledgling industry. "We're very happy about it," says Doug Pulick, vice president for research at Discovery Communication. "The two systems were so disparate that it was a case where competition wasn't helping the industry. There were huge differences in how they performed their work."

Is the devil in the details?

Can two companies with different methodologies, differ data collection schemes and different cultures merge their systems into one unified product? "The differences in our top 25 lists are not as great as some people believe," says Mary Ann Packo, president of the new combined Media Metrix. "Once we normalized the data and started comparing the results on an apples to apples basis, they were really close." Some of the discrepancies Packo cites include different measurement weeks, differences in what population was measured (Media Metrix measured users two years and up compared to RK which used a 12+ sample) and differing results from the business and AOL portions of the panel.

The new Media Metrix plans to merge the results of both services beginning with its November 1998 Web Report in mid-December. Media Metrix will then develop and deliver a hybrid meter which combines Relevant Knowledge technology with Media Metrix's PC Meter. The new software will be rolled out to the 40,000 members of the combined panel in the second quarter of next year. Packo dismisses concerns about merging RK's random-digit-dial (RDD) telephone sample with MM's sample built from mailing lists. "When we compared the usage from both panels, the results were very similar," she says.

Clients aren't so sure the transition will be that easy. "Its more complicated than that it seems," says APL Digital's Bolger. "You need to carefully plan how you are going to merge the panels to make sure you don't screw up the whole system."

Other clients remain skeptical about the merged companies intentions, fearing that they will drop RK's more scientific approach for MM's less expensive recruiting method. "I find it hard to believe that Media Metrix's sample -- as a sub sample of NPD's national PC sample -- is a better representation of the web universe than RelevantKnowledge's," says one client. "It would be a step backward if they were to give up on probability sampling."

The assertion that both panels have similar profiles is somewhat suspect, " says Craig Gugel President of media research consulting firm Manhattan-Pacific Multimedia and former executive vice president of Bates USA. "They have produced different numbers in the past year and I would be suspicious if the differences would suddenly disappear." Gugel says that RK's better methodology was their recognized advantage over Media Metrix and that many clients would not look favorably at joint venture if they rejected the random probability approach. "One of the reasons we were intrigued by RelevantKnowledge was that their methodology was so polished. We will have to see what the new company decides to do."

Back to Articles page

© 2003 TNG Research. All Rights Reserved.