Secrets of the box-office smash
It's not the plot, it's not the star - it's technology that triumphed in 'Avatar,' observes Jehoshua Eliashberg, who has developed models for predicting popularity and success on the silver screen.
It's only been six weeks since the release of "Avatar," but it has already surpassed "Titanic" and become the biggest box-office hit ever. Both films were directed by James Cameron. "Avatar" has already grossed more than $1.8 billion worldwide, returning over half of the investment poured into it after its first week in the theaters.
How did "Avatar" make it so big? Many attribute the film's success to its spectacular technology, which allows the audience to see a 3-D movie made with a technique exclusively invented for it, and to experience the rich, imaginary world that Cameron created out of nothing - including a whole new language.
Aside from a few familiar faces like Sigourney Weaver ("Alien"), Michelle Rodriguez ("Lost") and Giovanni Ribisi ("Friends," "Saving Private Ryan"), the lead role is played by an unknown Australian named Sam Worthington, an actor who may be familiar to action-film fans from "Terminator Salvation."
"The 'Avatar' story proves that even if the lead actor is insignificant, the film can succeed," states Prof. Jehoshua Eliashberg, who is not an actor, director or a producer. But, as professor of marketing and information management at the University of Pennsylvania's Wharton School of Business, he is very familiar with the inner workings of the entertainment industry. Indeed, his best-known research relates to how films are made, financed and produced .
"An actor only has a nominal effect on the success of a film," he continues. "A doctoral student of mine conducted a study in which one of the variables was the lead actor, and the findings showed that his effect was marginal."
"Avatar" is not the only film that has achieved phenomenal success without a Hollywood luminary: Others in this category include "E.T." and the "Star Wars" series.
So which variables do have a substantial effect on a film's success?
Eliashberg: "I believe that the story is the critical thing. A good story with an average cast of actors and average publicity will have greater success than a bad story with a popular actor and the same marketing efforts. In the case of 'Avatar,' the fact that the film is 3-D adds another dimension of innovative technology, which also draws audiences. Here, the success belongs to technology and to an average story. It is difficult to assess what contributes most, but it isn't the stars."
How much of an effect does the film's director have?
"You could say there is a limited list of directors who have an effect: Steven Spielberg, Woody Allen, Martin Scorsese, Quentin Tarantino and, evidently, James Cameron, too."
Since the early 1990s, the Israeli-born Eliashberg, 64, has been constructing models that predict the success of films based on certain variables, and constitute the basis for recommendations as to how long a certain film should be screened, and at which times, how much money should a distributor spend on advertising, and even which screenplays might become box-office successes as films.
Indeed, in recent years with colleagues at Wharton, Eliashberg has been developing a technique for predicting the success of screenplays. This project began a few years ago, after he received a request for assistance from a major film studio, he says: "We developed a system that involves an analysis. The idea is simple: Take a screenplay, examine its 'DNA' and see how closely it resembles other screenplays that became successes."
Three readers, working over a period of a year and a half, went through 200 screenplays of films made in recent years, answering various questions related to the plot - such as whether there is a conflict in it, if the plot is logical or unpredictable, if the protagonist evokes sympathy, whether the general background is familiar to the viewer, or the ending logical or unexpected, and so on. At the same time, a computer program was developed that analyzed technical criteria, such as the number of words in a scene and the number of dialogues in the scripts.
What Eliashberg likes to call a DNA bank was created for each of the 200 screenplays. Fifty of them, whose levels of success at the box office were already known, were used as test cases to check the reliability of the model the researchers devised.
Eliashberg: "The results showed that we were reasonably on target. In 60-70 percent of the cases, we were able to predict success within a range of $10-15 million in revenues. There were also a few grating exceptions, and a few films about which we were way off."
What, for instance, were the mistakes?
"We did not correctly assess 'Chicago,' the musical made in 2002 with Renee Zellweger, Catherine Zeta-Jones and Richard Gere. We based our DNA test solely on one element, the screenplay, but there are other important elements besides, such as quality of acting and the public's previous exposure to the film, in this case, as a musical on stage.
"In 'Chicago,' the acting was good, the film received support from advertisers and it was widely screened. All of this contributed to its success, which we were unable to predict. Another miss, in the opposite direction, was 'Troy,' based on one of the bloodiest wars in Greek mythology, which starred Brad Pitt. We assessed that the film had potential because of the story, but it did not succeed as much as we thought it would. In retrospect, it is easy to explain why we were wrong. For instance, you could make the argument that Brad Pitt was not suited to the role."
How do you intend to improve the predictions?
"I am the first to admit that our predictions are not perfect and one of the reasons for this is our limited database. We only have scripts that you can download for free from the Internet. The hope is to create a database of 1,000 screenplays."
Eliashberg says he's received a request for help in choosing "winning" screenplays from a private investment fund. "As a result of the [economic] crisis, some funds left the film industry, and the [situation of the] ones that remained is now much more critical when it comes to selecting a project and investing money in it. We helped the fund's managers see what sort of chances these screenplays had."
Timing and screenings
Eliashberg did not set out to specialize in marketing. After completing his service in the Israel Defense Forces in 1966, he studied electrical engineering at the Technion - Israel Institute of Technology, before moving to business studies. He received his Ph.D. in business administration at Indiana University.
"I liked the mathematics in engineering and the marketing in business administration," he explains, "so I brought the mathematical tools into the marketing field."
Eliashberg has been on the Wharton faculty since 1982.
In the early 1990s, while supervising a doctoral student researching the world of entertainment, he began to take an interest in the film industry. At the same time, a Wharton alumna who managed a chain of movie theaters asked Eliashberg for assistance in deciding which films to screen. Along with colleagues Charles Weinberg of the University of British Columbia and Berend Wierenga of Erasmus University in The Netherlands, Eliashberg developed a system that helps movie-theater owners decide which films are worth screening, in which kind of halls they should be shown and for how long. Over the years, he has also worked with studios that distribute films.
"We created a model for movie-theater owners which, based on statistical data from already-released films, allows us to predict a new film's success. Our prediction is based on 20 or so variables, including the actors, the director and the expected date of release. An analysis is conducted and a comparison of the outcome is done vis-a-vis the information from our database."
Among Eliashberg's long-term clients are the Dutch movie-theater chain Pathe Cinema. "In the United States," he notes, "you can find a screen for every film, but in The Netherlands there is competition, including American and European films. A lot of films are competing for a few screens, so there the selection has particular importance."
At first, Eliashberg compared the profits made in a movie theater that used his model with those of two other theaters that did not use it. The intake at the one theater was 10-20 percent higher than at the others. Subsequently, his recommendations were implemented in six theaters in Amsterdam, with a total of 40 screens.
"In the initial stage, we told the theaters which film should be shown in which sort of hall. In the second stage, we determined the times the screenings should begin. We've been working with the company for 15 years. Now we have moved over to software that will help them make the decisions, and expect to begin a new project with Pathe France, which has theaters in Paris, Belgium and other countries in Europe. In parallel, we are talking with a movie theater company in India."
Based on your research, what advice do you have for owners of movie theaters?
"One of the most important suggestions relates to how long a film should be shown. According to most standard contracts between distributors and theaters, the bulk of the revenues at first goes to the distributor. The longer a film is screened, the higher the share of revenues enjoyed by the theater. The conclusion is that if you own a movie theater, it's worth your while to show fewer films, but for a longer time. This will give a larger audience an opportunity to see the film, and will increase sales. In the U.S., the revenue distribution process varies because there have been many mergers, and huge movie-theater groups have formed, with a great deal of clout. They demand half of the revenues, from day one.
"Another conclusion is that a film that is screened at two different movie theaters 10 kilometers apart will have a different level of success, for the simple reason that people in different environments don't have the same taste. This factor has a strong effect. I don't know if this is true for Israel, because it is a relatively homogenous country, but the conclusion is relevant to the U.S. and Europe. If we take, for example, two different regions of Amsterdam, in one they will want an action film, and in the other, a comedy.
"The third conclusion is that theater owners make a lot of mistakes and leave a lot of money on the table. They err when they choose films that are unsuited to local tastes, or show them at inappropriate times, without taking into account what the audience is doing then, or they don't show the film for a long enough period of time.
"Another issue is the characterization of a film by gender. There is a market in Hollywood for 'chick flicks,' for example, 'Pretty Woman.' Because of the genre, an advertising strategy directed toward women is usually used, but in one of our studies, we noted that it doesn't make a difference."
Straight to DVD?
Based on your studies, what is considered a winning genre?
"It's impossible to put your finger on any one thing. One of the conclusions we reached from the DNA analysis of films is that a winning screenplay is one that has action, with characters to whom you can relate, is based on a 'clear premise' (for example, saving the world from evil), and also has a multidimensional protagonist."
Eliashberg and his colleagues have also conducted studies with ramifications for film distribution: "Early on, distributors have to decide whether to take the film to the theater or go directly to DVD. If they decide on the former, they have to determine how much to allocate to marketing, how many screens the film will be shown on, and how much TV and other advertising is needed. Then another variable enters the equation: how many viewers will recommend the film to one another. If a lot of people talk about and recommend a movie, less money needs to be spent on advertising, as the viewers are already doing the work.
"We managed to create a mechanism that enables us to predict the nature and the extent of word-of-mouth communication, and to make recommendations about the amount of resources to invest in advertising and about whether a film should go to the theaters or be distributed on DVD directly."
To determine how much the viewers will talk about a film, the researchers constructed a set of questions, such as "To what degree is the film a subject of conversation?" or "Is this a film that you would want to watch alone?" Two hundred people agreed to answer the questionnaire in return for seeing the movie for free.
Eliashberg: "Based on these studies, we found that if a film is rich enough, with a good plot line, it does not require much advertising. Most of the distributors in the U.S. and Europe now spend more than necessary on publicizing movies. We also learned that the relative effect of critics on film-watching is practically nil. We examined a lot of critics and films, did statistical analyses to assess their influence, and deduced that they carry almost no weight at all.."
As part of his work with distributors, Eliashberg says he succeeded in almost perfectly predicting the success of the 1993 comedy "Groundhog Day," starring Bill Murray, about a weatherman from Philadelphia who relives the same day over and over: "We predicted that revenues in the American market would be $68 million, and in actual terms, the sales were $70 million in the U.S."
The bottom line, concerning movie-goers in the U.S. and Europe, Eliashberg emphasizes, is that recommendations by someone who sees a film are the most important factor when it comes to drawing an audience. After that, in order of influence, come the advertisements and the trailer, and the critics' recommendations.
How can you generate buzz for a film?
"I don't know if it's done with films, but there are companies that specialize in 'buzz marketing.' They give incentives to people whom they see as opinion leaders, with the aim of getting them to talk with their friends. For instance, Nike has a sophisticated system that identifies young leaders, and they receive the company's newest sneakers for free."
Like us on Facebook and get articles directly in your news feed