And the winner isn't....
As the movie industry nurses its collective head in the aftermath of last night’s Oscars, those in the business of making predictions based on social media chatter will be scratching theirs and wondering why their forecasts were so wide of the mark.
Despite critics hailing The Artist as almost a dead cert to take home the Best Picture award, social media analysis in the run-up to the event suggested it might not turn out to be the one-horse race many were expecting.
At the University of Southern California’s Annenberg Innovation Lab, Professor Jonathan Taplin studied the sentiment in tweets to predict that Woody Allen’s Midnight in Paris would take the award. Meanwhile social media agency Banyan Branch’s analysis of tweets and Facebook ‘likes’ tipped The Help as the front-runner. But in the end, it was the critics who called it correct.
Banyon also used tweets and likes to pick the likely winners for the Best Actor and Best Actress categories. Again they proved incorrect, with The Help’s Viola Davis and Moneyball’s Brad Pitt failing to turn their social media popularity into an Oscar. Eventual Best Actress winner Meryl Streep came a close second in Banyon’s analysis, but it was way off the mark for Best Actor. It had The Artist’s Jean Dujardin in third place.
General Sentiment was another company playing the predictions game – and it did so better than others, thanks to its approach of analysing social media and Twitter sentiment alongside the bookies’ odds for each nominee. First let’s start with the missteps. It made The Artist’s Bérénice Bejo the favourite to take home Best Supporting Actress ahead of The Help’s hotly tipped – and ultimately successful – Octavia Spencer. It also tipped The Help’s Davis to beat Streep.
Yet General Sentiment was right on the money in the Best Picture, Best Actor and Best Supporting Actor categories – the last of which went to Christopher Plummer, as predicted, for his performance in Beginners.
All in all, it wasn’t a great night for demonstrating the predictive abilities of social media. However, it should be noted that the public has no input into which films and performers actually win the Oscars – that job is handled by the Academy Awards panel. With that in mind, this year’s presidential campaign and the upcoming mayoral election in London are the two bigger tests of social media analysis as a means of gauging the public mood.
Until then, forget your sorrows with this amusing little skit from last night’s ceremony: supposedly a long-lost recording of the first screentest and focus group for the Wizard of Oz.

We hope you enjoyed this article.
Research Live is published by MRS.
The Market Research Society (MRS) exists to promote and protect the research sector, showcasing how research delivers impact for businesses and government.
Members of MRS enjoy many benefits including tailoured policy guidance, discounts on training and conferences, and access to member-only content.
For example, there's an archive of winning case studies from over a decade of MRS Awards.
Find out more about the benefits of joining MRS here.
12 Comments
Francesco D'Orazio
13 years ago
Hi, we actually got it right using social media data. Here's how http://www.facegroup.com/is-a-silent-black-white-movie-the-strongest-oscars-favourite-ever.html As per the studies you mentioned, also referenced in our study wrap up, they both looked at sheer general mentions of a movie, which to be honest is not a very useful way of looking at social data in this case because getting an Award does not equal being socially relevant (as you get voted on by a panel of practitioners). Which is one of the reasons why for example we looked at mentions of a movie in relation to a specific award and also looked at external data in order to 'augment' the online audience feedback (social media) with offline behaviours (box office) and the expert views (the critics, the bookies, the practitioners) More coming soon on which data sources got closer to the right results... cheers fran
Like Reply Report
Brian Tarran
13 years ago
Thanks Fran. That's a great post – and congratulations on getting the prediction right.
Like Reply Report
Brian Tarran
13 years ago
Just to add: I would like to see the breakdown of which data source was closest when you have it. Looking at your blog, it made me wonder whether the social media data really added that much. It seems you would have arrived at The Artist as your predicted winner based solely on bookies' odds and critic reviews. Which, I guess, begs the question: what's the point of social media predictions?
Like Reply Report
Francesco D'Orazio
13 years ago
We started seeing The Artist run away when we isolated social media mentions of a movie in relation to each award, before looking at any other data stream. Sentiment was only helpful when layered on top of Social mentions in relation to an award, not when layered on top of the movie's general mentions. And with bookies, while the odds was fairly consistent in the UK, in the US there were quite a few bookies pointing at The Help or Moneyball or The Descendants, so not a clear picture. It's only when we married social mentions in relation to the awards with critics and bookies that things started shaping clearly in favour of The Artist. So I guess social in this case helped providing bottom up validation or grounding to a top-down betting exercise (either by critics or the bookies)
Like Reply Report
Brian Tarran
13 years ago
I stand corrected. Thanks for explaining that, Fran.
Like Reply Report
Francesco D'Orazio
13 years ago
pleasure. to be honest when I saw those social media predictions based on sheer volumes of tweets and likes to facebook pages I couldn't really understand WHY they were doing it that way :)
Like Reply Report
Brian Tarran
13 years ago
I saw that article last night too. They also got Best Actress right, but Best Actor and Best Director wrong. A 50% hit rate, then, versus General Sentiment's 3/5.
Like Reply Report
Francesco D'Orazio
13 years ago
that's because they didn't mash it up :) to be honest we didn't have the time to look at the other categories in our study, but we collected the data so hopefully we'll be able to review it all at some point. I guess with this one we wanted to see if there was a way to use social media for something that isn't inherently social, and then pilot data mashups to see what sticks... nothing more ambitious than that really
Like Reply Report
Brian Tarran
13 years ago
Also, Radian6 scored 2/3: http://www.radian6.com/blog/2012/02/the-oscars-nearly-3-million-social-media-mentions/
Like Reply Report
Francesco D'Orazio
13 years ago
Yea saw that, good job, again, by looking at specific social data, not just general mentions.
Like Reply Report
Bruce Tedesco
13 years ago
Why would anyone think that any social media method would be aligned with votes form the members of the Academy? It is cute to see how close results are, but there in no scientific reason that the general population has a view identical to the industry voters.
Like Reply Report
Raz Khan
13 years ago
The Radio 5 Live show, Kermode and Mayo's Film Review, did something similar by polling their listener base. They typically get 1 million podcast downloads each week so it's a sizeable audience, and with a clear interest in movies. Their listeners got most of their predictions wrong, but the reviewer Mark Kermode was substantially correct on his personal predictions. The reviewer is more of an industry insider and likely to be more aligned with the actual voters in the Academy. As you say Brian, it will be more interesting to see how good SM is at predicting more public decisions such as elections.
Like Reply Report