Before TV shows air, they have to survive ... The Lab

Let's be clear: I like volleyball. But when a TV pilot throws three 20-something, scantily clad, superbuff women into a show about an undercover female volleyball team tracking a nefarious kidnapping mob, well, it's the red button for me. That's the spot on my response pad where I tell ASI Entertainment, Tinseltown's oldest and most widely used audience-testing facility, that I would change the channel if I were watching at home.

I'm here taking part in a demonstration of the science behind the art of today's entertainment world. Virtually everything that Hollywood produces – TV shows, movies, commercials, even infomercials – is tested here. Though the industry's creative teams cringe at the idea of subjecting their creation to market research, most admit they have no choice since TV shows can cost $4 million per episode and movie budgets often hit nine figures.

"Given the kind of money involved," says Jonathan Shapiro, executive producer of "Justice," an upcoming show on Fox, "it would be irresponsible not to do it." Industry veteran Tom Werner ("The Cosby Show," "Roseanne") adds, "I'm not happy about it, but I know the networks rely on it."

Often, they must also bow to what Shapiro grudgingly calls the wisdom of the group, acknowledging the usefulness of collective wisdom in everything from TV shows to jury testing and political polling. Testing on Mr. Werner's new show, "Happy Hour," revealed results that he says his team already intuited. "The test audiences really liked the character Amanda," he says, referring to a woman who uses comedy to compensate for her insecurities. "So the networks asked us to give her more to do. So, you'll see her character play a bigger role this fall."

The independent ASI has been dialing up audience responses ever since 1966, when there were only three TV networks. The team recruits viewers through random phone calls to households within a 50-mile radius of its site, nestled next door to the Television Academy of Arts and Sciences in North Hollywood. Audiences, who are paid between $50 to $75 per person, can be tailored to client demands. Testing a new drama? Bring in the drama fans. Same for sitcoms. If a client wants a Midwest audience, ASI can take the dog-and-pony show on the road, putting together a perfect test group in Chicago or Detroit.

The testing process begins in one of the two 48-seat theaters at ASI. Members of the Television Critics Association, invited to ASI for a demonstration, gather in the black screening room, no food or drinks in hand. "The equipment is very sensitive," says Neal LaVine, the theater director. Two large tinted glass panes stare back at us from the front of the room. In a normal test, clients who have paid some $20,000 for two-hours worth of testing watch the proceedings from plush, black-leather chairs behind those windows. Two average-sized TV screens blink down at us from high on the wall. No high-def or giant screens here, because that's not what the average viewer owns, says CEO David Castler. "We're not going to go with plasma to enhance the program," he adds.

The seats have fold-up tabletops. Velcroed to the side is a pencil and the wired dial pad with five degrees of "like" to "dislike" on its face. We also have a phone pad of buttons on which I find the red button telling ASI that if I were home, I'd hit the remote.

But that's only two minutes or so into the show, which turns out to be an unaired (gee, really?) UPN pilot called "I Spike," from the 2000 season. The show speeds on, full of silly car chases, scantily clad young women, and lots of serious posing and pouting reminiscent of early Aaron Spelling fare. By the time the show finishes, the room has dissolved into banks of hoots and snickers. A graph with falling stock-market-type spider lines shows up on the screen. Mr. Castler explains that the red line represents the women's responses while the blue line shows the men's. When this graph is superimposed in real time over the show itself, the clients can actually see, second by second, exactly which bits the audience liked/disliked.

"This is probably one of the worst scores I've seen," Castler says, laughing. "The good news is, you're not going to see this on television."

The graph tells the story the clients want to know – who likes what (actors, relationships, dialogue, setting, etc). Within two minutes, more than two-thirds of the women in the group had changed the channel. Most of the men, however, hung in to the bitter end. Not surprisingly, women in bathing suits score higher with men, notes Castler. Women's scores, by contrast, tend to rise when characters develop relationships (because women like to watch that sort of interaction).

In an actual test with real clients, we would break to fill out questionnaires and then head into a focus-group room to answer specific questions. A TV team will use the data specifics to retool the show, as they did with "Happy Hour." Often, the networks will bring episodes back for another test. They might bring in subsequent episodes to ensure that the series is delivering on the promise of the pilot.

"We're here to tell you where the weakness is – what we see as the problems, and what we see as the area for development," says Castler.

Testing is not infallible. The creative types who shiver at the mere mention of audience testing tell stories about world-altering hits that bombed in the testing barracks. "'Seinfeld' and 'Roseanne' are two good examples," says TV vet Werner, adding that it was only his clout as the producer of "The Cosby Show" that put "Roseanne" on the air relatively unchanged.

ASI freely acknowledges this industry truth. "Norman Lear's 'All in the Family' bombed in testing," Castler reveals. But that's part of the point. The TV producers took some of the tips, warmed up the Archie-Edith Bunker relationship a bit, and went on to become perhaps the biggest sitcom in history.