In a world where a single rating point can be worth $150 million a year, it's hard to find an objective observer. Cable companies, whose audiences keep growing, downplay their dissatisfaction. "Problems with Nielsen need to be addressed," notes USA Network research VP Tim Brooks, "but the [broadcast]networks are screaming fire in a theater where somebody is smoking a cigarette." Given the many tales of wonky ratings, however, confidence in Nielsen is down. "Nickelodeon is having a banner year," says Susan Nathan, senior VP of research at ad agency McCann-Erickson. "Are the huge increases real? I don't know."
"There isn't a sector of this business that isn't pulling its hair out," moans one high-ranking nonnetwork exec. "No one knows what the hell is going on, whether ratings shifts are real. I mean, we're going nuts."
Behind the palpitations are numbers that, to the networks, look as sturdy as a Larry King marriage. According to CBS executive VP of research David Poltrack, if you compare ratings for Late Show With David Letterman in the 36 markets where Nielsen conducts both local and national surveys by meter, the local measurements are 15 percent higher. "These are two measures of exactly the same thing," he says. "They should yield identical results, and they don't."
Fox has seen similar discrepancies in ratings for Martin, which draws higher local numbers than national ones in the Southeast. Sports broadcasts are feeling the fluctuations too. NBA games on NBC are down 10 percent so far this season--a strange drop in usually consistent numbers; NBC, ABC, and Fox all say their NFL fall ratings dipped inexplicably. And a July episode of CBS' Saturday-morning cartoon Santo Bugito drew 539,000 viewers--but no 2- to 5-year-olds.
Another suspicious example: Soap opera weddings traditionally light up the ratings. Last July 6, the day before Julia and Noah got hitched on ABC's All My Children, the show scored 2.73 million women 18 to 49 years old. The day after the wedding, the figure was 2.69 million. The day of the wedding: only 2.49 million.
Nielsen insists that these instances, like CNBC's goose egg for the Bulls' postgame show, are normal statistical anomalies. "The overall picture is so consistent it's almost scary," says Nielsen senior VP of research Barry Cook. Those raising doubts, he adds, "didn't see their research analysts throw out 100,000 comparisons that looked okay, just to find the one that didn't. It's a no-win situation."
Ironically, the current wave of criticism stems in part from Nielsen's "improving" its sample audience from 4,000 to 5,000 homes using a new recruiting method. "Nielsen families" are chosen at random from the 97 million U.S. homes with TVs; they're paid $50 to $100 to have their homes wired. But the networks say recent samples do not represent a true cross section of U.S. viewers. According to ABC, Nielsen's September-January data deviate from the ratings company's own estimates of all U.S. households by underrepresenting single- person households (by 15 percent) and low-income households (by 26 percent), and overrepresenting homes with college graduates (by 31 percent). "It's hard to blame this on the weather," says ABC marketing exec Alan Cohen.
