• United States
News Editor

Online amateur movie reviews predict box office success

Jun 07, 20044 mins
Enterprise Applications

Any business that depends on word of mouth – the film industry being perhaps the best example – should consider the powerful predictive potential of Internet forums

Have you ever read one of those reader-submitted movie reviews online and asked yourself: “Who cares what these chuckleheads think? Anyone with a browser and too much time can play film critic.”

Well, here’s who should care: Movie studio executives who want to know with a high degree of certainty how their just-released picture stands to fare at the box office – and take appropriate business action before it’s too late.

In fact, any business that depends on word of mouth – the film industry being perhaps the best example – should consider the powerful predictive potential of Internet forums, says MIT professor Chris Dellarocas, who has conducted an analysis of reader reviews submitted to YahooMovies and’s Internet Movie Database.

What Dellarocas and fellow researchers Neveen Awad and Michael Zhang demonstrated goes beyond the obvious supposition that movies generating positive ‘Net chatter likely will do better than those that get torched. (Access the study here.)

“Even though we cannot claim generality, this finding supports the viewpoint that online forums are emerging as a valid alternative source of information to mainstream media, replacing our societies’ traditional reliance on the ‘wisdom of the specialist’ by the ‘knowledge of the many,’ ” the report says.

The bottom line: A studio that sees a blockbuster brewing in its analysis of first-week online reviews could broaden the film’s distribution and crank up marketing, while those seeing a flop would have a chance to cut their losses.

The study also highlights the potential for corporate mischief, given that these review-and-ranking sites are forums for what is essentially anonymous opining. (For a PDF copy of the study, click here.)

“Manipulation of forums will become some sort of arms race between studios,” Dellarocas says. “They will all feel compelled to do some of it, in order not to be outdone by their competitors.”

However, he sees the risk to consumers as being minimal.

“Even though there will be some inflation of ratings across the board, trained consumers – i.e., consumers who have been reading/participating in those forums for a while and are familiar with their norms – will not be fooled,” he says.

While I can’t say I share that level of confidence in consumer competence, there should be no doubt that the studios will test the theory.

The study also acknowledges that analysis of first-week online reviews is not as accurate a predictor of ultimate performance for all movies. In particular, the study kicked out aberrant results for movies that opened to little critical fanfare – online or offline – yet went on to great success: “Chicago” and “My Big Fat Greek Wedding” being two examples.

“Sleeper movies tend to blossom more slowly, so first-week ratings cannot always predict their subsequent trajectory,” Dellarocas says. “However, we speculate that the poor performance of our model for sleeper movies is partly due to the fact that the mix of movies that we used to ‘train’ the model consisted mostly of blockbuster movies. We believe that conceptually the model is very much applicable to sleeper movies.”

One interesting finding in the analysis was that the correlation between online reviews and a film’s box-office performance was found to be stronger for reviews offered by women than their male counterparts.

What’s up with that?

“We’re not sure at this point,” he says. “YahooMovies users are 75% male, 25% female. One hypothesis is that women who participate in online forums are on the average better educated and more thoughtful than the average male user.”