The popularity of a particular film largely depends on word-of-mouth recommendations according to a new study by statistical physicists in the US and Chile. César Hidalgo of the University of Notre Dame and colleagues have also developed a quantitative indicator of a film's quality, which they say could be used by film producers and studios to estimate the commercial value of a movie (New J. Phys. 8 52).
Hidalgo and co-workers have developed a model that can explain the different types of behaviour observed in filmgoers. The model takes into account how many people go to the opening of a movie and whether they go to see a film alone or in a group. It also describes the rate at which individuals lose interest in a film once it has opened, which means the probability of them going to see it decreases.
Using these parameters, the researchers came up with a quantitative indicator that provides a good estimate of a film’s commercial value. This indicator does not consider polls and surveys about the film, such as top-ten lists, and is based on how many people go to see a particular film during the time it is showing in cinemas.
The Chile-US scientists say their quantitative indicator could be used by film producers and studios when deciding to produce a sequel, for example, or when investing in a film of a certain genre. “At the moment, the film industry bases such decisions solely on revenues but I believe an indicator such as ours would allow the industry to discover hidden markets that could be exploited,” says Hidalgo.
The team constructed two main equations in its model by considering two points. The first is that individuals are not likely to go and see the same film twice. The second is that the probability of someone going to see a particular film depends on his or her interactions with a person who has already seen the film. In other words, if this person liked the film, others in his entourage are more likely to go and see it. The reverse applies if the person didn’t like it.
The model, which the researchers validated by comparing it with box-office data for the 44 movies with the biggest budgets in 2003 (see figure), could also be extended to other form of entertainment. These include going to concerts, plays, museum exhibitions and buying a particular CD or best-selling book. “Our paper belongs to a larger class of papers that are starting to show that human behaviour is not as random and unpredictable as some people think,” says Hidalgo.
