Increasingly, companies and organisations are using social media as a crystal ball to predict the future. Negative spikes in sentiment to predict a drop in stock prices, explosive volumes of mentions to predict the election of a candidate (or a hung parliament, as Tweetminister predicted at the last elections. Check out a couple more examples here and here)
So far the trick worked: high levels of social media mentions and engagement = social relevance. But this case is different. Nine films are nominated for Best Picture at Sunday’s Oscars. According to many sources, “The Artist” is the favorite to claim the big prize. But the Academy choses the winners, not the general public. Or does it?
Yes and no. The members of the Academy are members of the audience too, and as such they are influenced by the people who surround them, especially the ones that are most similar to them, and share similar tastes. However, there are many other factors that come into play in this case, and a simple prediction model based on social relevance (= high levels of mentions, sentiment, engagement) will probably not do the trick.
First of all, sheer volumes of mentions in this case are less relevant than they are in a political election or in any other public event shaped by the audience.
A few other studies on the Oscars have used volumes of tweets or likes on Facebook as indicators. One study is predicting The Help to be the absolute favourite. Another one predicts the Midnight in Paris to be the favourite. There seems to be a little confusion around.
Our data points elsewhere. First of all, we didn’t just measure volumes of mentions of the movie, we looked at volumes of mentions in relation to the award nomination. And not only at that: we looked at the sentiment of those mentions, their visibility and the engagement they generated.
Second, this can’t be just about social media as the final judgement will be expressed by a panel of experts/practitioners. We think social media data is most useful when mapped against other data streams, because social media doesn’t happen in a void.
This approach is part of what we call Augmented Research. In this study AR meant combining the following streams of data harvested for two weeks (Feb 7th – Feb 21st):
1) Volume of tweets, status updates, blog posts, forum posts, news articles, images and videos.
2) Odds for each movie nominated against each Award.
3) Box Office Data for each movie.
4) Experts ranking via Polls and online ratings.
So we have been looking at something like this for each movie:
We are not going to delve into the details of the graph above, but what is interesting is that there seems to be a correlation between the box office data and the social media data. Peaks at the box office anticipate peaks in social media in smaller and smaller increments. We haven’t seen any of the opposite: peaks in social media anticipating peaks in the box office data. Which could potentially indicate something interesting in terms of influence dynamics and the relationship between traditional media and social media, at least for now.
But let’s not digress. We wanted to see if any of the above could be of any use to predict which film is going to win the Oscar for Best Picture on Sunday.
We started looking at volumes of buzz around each of the nine nominated movies (Feb 7th – Feb 21st). The doughnut below shows days as circles and within each circle the proportion of buzz associated to each movie.
According to this model, The Help should be the Oscars favourite, but the ranking is rather balanced:
1) THE HELP
3) THE ARTIST
4) THE DESCENDANTS
We then introduced the Sentiment of those conversations in order to weight volumes. But the landscape got even more balanced. Unfortunately.
We decided to try something else. When it comes to the Oscars, social relevance doesn’t necessarily mean being Award-worthy. So we then looked at just the conversations that were related to the Oscar nomination for Best Picture (“movie title” + “Oscars” | “movie title” + “Best Picture” and so on for 15 stings per movie, Feb07th-Feb21st). We started seeing some clear movements in the chart.
1) THE ARTIST
2) THE HELP
4) THE DESCENDANTS
Although The Artist looks solidly ahead (more than double the volumes than any of the contenders), there is still a good chance of a catch up, especially since all the top contenders are extremely close to each other.
We needed another opinion, and we asked it off the people who are actually closer to it all: the critics. We pulled some good data off Metacritic and layered the critics score on top of the social media scores. We used the Metascore, based on 40+ critics globally for each movie. And this is the result.
The Artist is now clearly running away and the competition lags behind in a rather compact front of four movies including The Descendants, Hugo, The Help and Moneyball.
In search for an even safer bet we then looked at the betting experts. We layered the daily data coming from the bookies for each movie on top of the social media data. And this is what happened…
Well, this kind of helps. I guess we will be placing our bets on The Artist as Best Picture at the 2012 Oscars.
A few people have been campaigning in support of The Artist. We mapped them out and found out that one of them is Bret Easton Ellis.
We will be watching the Awards Ceremony tomorrow night and check whether our prediction was any good. Not that we are going to make any money though, looks like this is the safest bet ever.