Popularity on social networks, measured by fan count, turns out to be a good predictor of stock prices, according to a study by a Pace University researcher who worked with the social media tracking service Famecount to gather statistics on fans and followers. The results: As mentions in social media increased so did the stock price.
This is actually a fairly preliminary "pilot" study that looked at three big consumer brands -- Starbucks, Coke, and Nike -- but the results so far were enough to prompt press releases from both Famecount and the university.
Of course, there remains plenty of room for skepticism. Comments posted at the bottom of Famecount's press release on the study begin with one critic who wrote, "I call BS. Fan counts are correlated with other things that are also causing the stock price to move." Several other comments were also dismissive, although one stock trader argued that, regardless of cause and effect, he would be interested in any correlation that might make him money.
The author of the study is Arthur O'Connor, a doctoral student in management who also works as an IT consultant (he declined to say where). O'Connor said he had read other studies that applied more exotic sentiment analysis tools to analyzing what people are saying about a company on blogs and social media sites. Often, the results seemed to boil down to simple popularity, so he decided to look at that alone.
"To my surprise, actually fan counts were statistically significant," O'Connor said. To control for cause and effect, he decided to look at fan counts on a 10-day and 30-day lagged basis, and the statistical correlation between fan counts and stock prices remained. "I am not ready to say that fan counts drive stock prices, but it's not hard to believe that the people who are doing the analysis of stock market fundamentals on Wall Street may factor this into their models someday soon."
O'Connor said the reason he has looked at only three companies so far is that he has had to do a fair amount of manual grunt work to normalize the data, factoring out things like stock market holidays. He hopes to apply some automation to the next phase and have it cover at least 30 firms.
As for the guy who called his study "BS," O'Connor said, "To be honest with him, I don't blame him. A big piece of me has felt that way and continues to feel that way. But when I lagged the data, and it still statistically predicts the prices . . . well, the data is the data. I'll learn more as I expand the study."
Note that the study was limited to Facebook fans, Twitter followers, and YouTube company page views. It doesn't say anything about whether people are saying anything good or bad about your brand, only whether they're paying attention.
By that measure, O'Connor's stock is rising.