Nielsen, ComScore Taking A Measure Of Their Web Metrics

August 28, 2007: 08:05 PM EST

Aug. 28, 2007 (Investor's Business Daily delivered by Newstex) --

Companies in the business of tracking online usage are searching for measurements that reflect new technologies and ways people interact with Web sites.

Tracker comScore SCOR has introduced a slew of metrics this year that measure everything from site visits to time spent and what kinds of hardware and software users have. ComScore this month introduced qSearch 2.0, a service that expands the company's measure of online searches beyond leaders such as Google GOOG and Yahoo (NASDAQ:YHOO) YHOO to include searches on commerce sites, via downloaded toolbars and through local desktop search.

The idea is to give clients -- companies that operate Web sites and/or advertise on sites -- a clearer picture of user behavior and into how to best to spend their search marketing budgets.

"When we see there's a better way to be tracking something, we're going to go out and do it," said comScore analyst Andrew Lipsman.

ComScore rival Nielsen/NetRatings, a unit of privately held Nielsen Co., last month added a total-minutes metric to its measurement service. It also added a service called GamePlay that measures use of console and PC video games.

ComScore and Nielsen/Net-Ratings provide research that helps Web site publishers set ad rates and helps advertisers get more for their money. The trackers hope their metrics will boost revenue. That will likely happen only if the metrics help sites gauge their worth to advertisers, help provide better online ad pricing and more effectively get ads in front of the right viewers.

What's driving the changes?

One factor is ways Web sites are created. A rising number of sites use Asynchronous Javascript And XML, or Ajax, a way of using Java, XML and other technologies. Ajax makes it possible for pieces of Web pages to be updated on the fly without refreshing the whole page. One result is visitors call up fewer Web pages -- and the site falls in rankings based on standard page-view measurements.

The rise of video is stripping additional relevance from the page-view metric, as people spend long periods viewing a single site page.

Another change in Web viewing, says comScore, is that up to one-third of Internet users regularly delete their cookies, the bits of software that many sites inconspicuously store on users' computers so users can be identified on future visits. That's how sites greet you by name, for example.

Without the cookie identifier, users will then be counted as new, or unique, visitors the next time they come to a site. That artificially inflates another standard measurement of Web site popularity, the number of unique visitors a site gets per month. (A person who enters the site once and another who enters the same site 1,000 times should both count as one unique visitor.) Cookie deletion is causing sites to overstate their audience by as much as 150%, estimates comScore. That can lead sites to overcharge for advertising.

Alas, many people in this field concede there is no ideal way to measure online audiences. For instance, it's pointless to compare the audience that flocks to YouTube, which strives to hold onto visitors as long as possible, with that of a site like Google that's trying to get people in and out of its search engine as quickly as possible -- preferably by having users click its search ads to link to an advertiser's site.

The latest changes help, many say. Page views aren't "a very good gauge anymore of user behavior," said Scott Ross, product marketing director for Nielsen/NetRatings' NetView service. "Time spent (per site visit) is a much better common denominator."

Peter Daboll, chief of insights for Yahoo and former CEO of comScore, agrees.

The page view "does seem a bit archaic and doesn't seem to measure anything well," he said. "We're making progress in trying to understand what matters."

What matters is a combination of getting users to a site in the first place along with the extent to which those users are engaged once they get there.

That is best measured by time spent on a site, says Jack Wakshlag, chief research officer for Turner Broadcasting. "If people spend lots of time on a site, then it's an engaging Web site," Wakshlag said.

But relying on time spent on a site isn't perfect. Time Warner's (NYSE:TWX) TWX AOL family of sites moves up in the rankings because of its instant messaging service. It inflates measurements of time spent because it registers as site engagement while users have it running, and many users keep it running in the background.

Conversely, Google ranks low in time spent on a site, per session, because the site is designed to move people to other sites fast.

Such differences in what sites try to accomplish is why comScore plans further changes to its metrics.

It's preparing to launch a service that will measure ad impressions, or how many times an ad is "served" to a Web surfer.

It's also looking to refine its measures. If a user has an IM window open, along with streamed video of a baseball game and a text-based news site, comScore wants to measure how much time each application is in the foreground.

The new measurements from comScore and Nielsen/NetRatings could help online advertisers who face a confusing market of wildly fluctuating prices.

"The advertisers are going to see less variability in the prices they see from sources A, B and C," said Bill Cook, senior vice president of research and standards for the Advertising Research Foundation.

Advertisers strive to track the effectiveness of online ads.

The goal, says Yahoo's Daboll, is to best gauge the impact of advertising. "There's nothing worse," he said, "than throwing money at an ad campaign and not knowing how well it works."