Two researchers at HP Labs have established that they can use Tweets to predict how well a movie will do - research that can be applied to all manner of events including how well a product or ad campaign may perform.
Although the researchers, Sitaram Asur and Bernardo Huberman, have only applied their methodologies to Hollywood movies, the accuracy rate of their tests is startling - essentially it was more accurate than the current gold standard, the Hollywood Stock Exchange, which the industry uses, writes Fast Company.
Asur and Huberman started by monitoring movie mentions in 2.9 million tweets from 1.2 million users over a three month period. They then looked at two different performance metrics: the first weekend performance - when attendance is driven by buzz - and the second weekend performance - which is usually based whether people actually like the movie.
For the former, they built a computer model that factored in the rate of tweets around the release date and the number of theaters, Fast Company said. That model was 97.3% accurate, compared to the Hollywood Stock Exchange, which had a 96.5% accuracy.
For the second-weekend performance, the authors built a ratio of positive tweets to negative ones, which they blended with the Tweet rate metric in another prediction algorithm. "This time, the method was 94% accurate."
The research could be used to predict the outcome of all kinds of events, according Huberman - including how well major new products would be received and the winners of major political races.
A similar model might work especially well for products or trends that lack prediction markets such as the Hollywood Stock exchange, Popular Science noted.
The findings also pose an interesting question for advertisers and executives, Technology Review said: "Can they change the demand for their film, product or service buy directly influencing the rate at which people tweet about it? In other words, can they change the future that tweeters predict?"Return to marketing news headlines
View Marketing News Archive