According to Benny Agbayani, the Mets were a lock to win the World Series. According to the eggheads of the 1950s, robo butlers and rocket cars should have been the norm by now. And if you’d listened to industry insiders last weekend, you would have expected ”Charlie’s Angels” to gross $28 million, instead of its actual, monster $40 million take. Whoops.
It’s one of Hollywood’s hottest topics this year: the high profile misfiring of ”scientific” box office predictions. It first arose on July 4, when research said Mel Gibson’s ”The Patriot” would seize the holiday weekend with $25 million, only to have George Clooney’s ”The Perfect Storm” rain on its Independence Day parade, grossing $42 million to Patriot’s $22 million.
Then ”X-Men” slashed its way to $55 million, well above its expected $29 million. And with ”Meet the Parents” outperforming projections — grossing $116 million so far — and ”Book of Shadows: Blair Witch 2” experiencing a frightening opening weekend of only $13 million, one can’t help but wonder if box office prognostications are being made with Magic 8 Balls.
The problem lies with ”tracking” — research designed to determine audience awareness of movies. ”Tracking describes a group of studies,” explains Joseph Farrell, cochairman of National Research Group (NRG), the leader in the field. ”We track children, parents, African Americans, Latinos — anyone they ask us to.” Studios typically have NRG rustle up info starting three weeks before the release of their movies, at a cost of roughly $250,000 per film.
The tracking firm then heads to the phones, polling at least 1,200 households a week to gather data on audience preferences. ”I pay attention to tracking,” says actor director Forest Whitaker. ”It gives a good indication whether you’re reaching the audience you’re hoping to reach.” What it doesn’t do, says Farrell, is predict how a particular film will do in the marketplace: ”We leave that to the studios.” But studios often find information is hard to control: Trade papers, websites, and loose lipped execs make sport out of interpreting the ”confidential” data, leaking box office estimates on an almost weekly basis.
And that’s where the fuzzy math comes in. Polling has serious blind spots. Teens, for example, are notoriously unpredictable because they tend to make last minute moviegoing decisions. That may explain why New Line isn’t panicking that Adam Sandler’s ”Little Nicky” is reportedly tracking lower than his last two films, which averaged $40 million openings. Furthermore, everyone has a different way of accounting for factors like star power, release date, and genre.
”Forecasting will always be like playing Jimmy the Greek,” says Tom Borys, box office analyst for ACNielsen EDI. ”Forty-seven percent of regular filmgoers select what film they’re going to see the same day; 25 percent pick [a movie] one to three hours before they go. So, even if you’re tracking two days in advance of the movie, you’re still too early.” Adds Josh Eliashberg, professor of marketing and operations management at the Wharton School: ”Sometimes they don’t ask the right questions [or] the right people. The motion picture industry’s [polling] is 25 years behind other industries’.”
That’s a lesson Artisan Entertainment learned the hard way. The studio commissioned surveys from NRG and a competitor, MarketCast, for ”Book of Shadows.” ”It’s like getting the CBS and Zogby [presidential] poll results,” says Artisan CEO Amir Malin. But doubling up didn’t deliver: Two weeks before the sequel hit theaters, the studio was expecting a $30 million opening. ”I never believed our tracking,” says ”Shadows” director Joe Berlinger of the lower than anticipated grosses for his film. ”Expectations were out of whack. But it doesn’t matter, really. At the end of the day, this movie should be close to $20 million in profit.”
Enough still believe in the data from the 22 year old company, however, that virtually all the studios employ it. ”NRG gets knocked unfairly,” says Marc Shmuger, president of marketing for Universal Pictures. ”The data’s not perfect, but it’s very useful.” Adds Revolution Studios’ Tom Sherak, ”Tracking is right much more often than it’s wrong. Trust me.” Should we? Reply hazy, try again.
(Additional reporting by Scott Brown and William Keck)