Media Tech Summit: ‘Twitter Is A Pulse — But It’s A Biased One’

Facebook Users: Harvard Business School

One of the most well-received sessions at the Media Tech Summit went by so fast and was so stuffed with numbers that I wanted to have the slides in front of me before I did more than tweet about it. The Harvard Business School team of Mikolaj Jan Piskorski, associate professor of business strategy, and Bill Heil, a 2009 MBA, are in the midst of mining social network data and shared some early results:

Gender gap: In preliminary results still being vetted, based on all tweets of 300,000 random users up to May 2009 (gender tracking excludes bots and ambiguous names), women make up 54 percent of Twitter users but are more likely to follow men (44 percent compared to 56 percent). Men are more likely to follow men, too (35 percent compared to 65 percent). But genders tend to tweet at the same rate. Why is there a gender gap? Partly, they suggest, because women use Twitter differently — the majority use fewer hashtags, retweet fewer tweets, post fewer links and send fewer targeted (@) tweets. But even if women tweeted the same way as men, they would still have fewer followers. Piskorski and Heil say they are running sensitivity analyses “to ensure these results are solid.” Women get retweeted 50 percent less often, which may be one reason they get followed less often. One result, says Heil: “Twitter is a pulse but it’s a biased one.”

Using MySpace & Facebook: The numbers flip when you get into visual social networks like Facebook and MySpace. Women are still more than 50 percent of users but are viewed more — particularly by men who don’t know them. Piskorski: “Men looking at women they don’t know is on some level deeply, deeply reassuring (in terms of data), we behave online like any stereotype people have about men and women.” Women are more likely to keep their profiles private — 1/3 public versus 2/3 public for men. Women are most likely to view profiles of women they know and least likely to look at profiles of men they don’t know.

— Most activity is about looking: 70 percent of the time, social net users are reading or viewing, compared with 8 percent each for networking and writing. The remaining 5 percent is private. When it comes to profiles, the HBS team says people spend 12 percent of their profile time looking at their own, 44 percent looking at friends and family and a similar amount of time looking the profiles of strangers. Until this past August, Facebook users could browse networks they didn’t belong to — checking, for instance, single women in Boston — through http://www.facebook.com/b.php. That ability has been shut down.

— People are more likely to click on an image of people than they are the Avis ad next to it. Piskorski says that’s why click-thrus are so low on social media sites: “You need to do integrated marketing — always post in the right context.”

Piskorski explained as I was writing this that the data comes from a program that looked at MySpace profiles randomly through the month of June, mining demographic and location information from almost 20,000 U.S.-based private and public users (excluding bands and entities) who logged in at least once during the month. The resulting sample is representative of the MySpace population. The Facebook data comes from its ad platform.

Slides updated: I’ve posted updated versions of Piskorski and Heil’s slides showing the way people across the U.S. use Facebook and MySpace. It’s a little confusing so bear with me: Green is the base line — for instance, 12 percent of internet users live in a state; 12 percent use that service. On the Facebook slide, red state means traffic is lower than expected compared with internet users in the area; blue state means it’s higher than expected.

That’s reversed for the MySpace slide. MySpace hits its over-index peak in Texas and in parts of California (Sacramento, Modesto, Lodi, Los Angeles, but not Beverly Hills). Not surprisingly, MySpace also overindexes dramatically with Hispanic users – 194 percent. By comparison, Facebook hits baseline in more than 20 states, including Texas, and only overindexes in a handful. Facebook is used by more people but MySpace appears to have more breadth geographically.

loading

Comments have been disabled for this post