Services Try To Track “Buzz”

Brandimensions, TV*BuzzMetrics and others monitor viewer chatter on the Interne

Nielsen may say that CSI was the most-viewed show during sweeps, but My Name Is Earl can claim the most Internet “buzz.”

According to a midseason report by Canada-based Brandimensions, which bills itself as “a market-intelligence and brand-protection-solutions” company, Earl commanded an 11.7% share of Web audience discussion and a 4.41 score on a five-point sentiment scale this fall.

A new breed of ratings services is pitching the networks to subscribe to reports using their buzz-tracking technology, offering data they say measures viewer engagement and chatter just as it bubbles to the surface. Like Brandimensions, TV*BuzzMetrics from BuzzMetrics, which is owned by Nielsen parent VNU, tracks Internet talk about TV.

But the emergence of these services raises a central question: Can anyone really track something as elusive as buzz, which can just as easily occur at the watercooler as on the Internet?

Brandimensions, which based its findings on Internet comments recorded between September and November, says its robots, crawlers and spiders traverse 20 billion Web links to fan sites, blogs and chat rooms in a 30-day cycle to determine buzz rankings for TV programs. The service “gives us a unique measure of viral energy,” says CEO Bradley Silver.

Some in the industry question that a program's buzz can be determined from findings on the Internet, whose message boards can be manipulated by influential chatter by shows' fans or even by stealth messages from the networks themselves.

NBC, for one, used a Yahoo! buzz-tracking service before the 2002 Olympics to determine the best time to begin promoting the games but has seen distortions in the trial research from the newer companies. Fox has declined their pitches, says a network spokesperson.

“As this has become more public,” says Alan Wurtzel, NBC Universal president, research and media development, “we're becoming increasingly skeptical about the accuracy because people have started to game it.”

TV networks have long tried to quantify viewers' early engagement with shows through research other than the size estimates offered by Nielsen. Finding out which characters or storylines charm viewers can help a network determine a show's best time slot or direct its plot or marketing materials.

Brandimensions' research helped ABC decide to tone down the political slant of its marketing for Commander in Chief when it indicated that viewers perceived the show to be a political drama like The West Wing, says ABC Senior VP of Marketing Mike Benson.

“One of the things that makes programs successful is when they become part of the social and cultural fabric—when people not only watch them but talk about them and the characters and become involved,” says David Poltrack, executive VP of research and planning for CBS and UPN, who also advises BuzzMetrics. CBS is weighing syndicated services from BuzzMetrics and Brandimensions after using their buzz-monitoring to target marketing.

“What has traditionally been word of mouth has migrated to the Internet, where you can now capture it,” says Poltrack. “These firms are able to capture what people are talking about.”

The services, whose research has been used in the consumer-products and pharmaceuticals industries, aggregate the fan feedback, using computerized language filters to separate the positive from the negative and even to ascertain the posters' gender, age and region.

Media-services company Initiative Media, which has an enterprise deal with TV*BuzzMetrics, uses the service to help direct advertisers' product integration—for example, to tell a lingerie maker which character on Desperate Housewives gets the most Internet chatter from males, says Stacey Lynn Koerner, Initiative executive VP/director of global research integration.

The services say they can root out anyone trying to game the system.

Brandimensions says it stands out from its competition because it filters Web postings through a team of 400 individuals—college students, retirees and stay-at-home parents—who pick up colloquialisms that computers miss.

Still, says CBS' Poltrack, “this is never going to be scientific. It's always going to be anecdotal. This is not a pristine product.”