Tuesday, August 28, 2012

Faking Twitter Numbers - and Exposing Them

It's been a dirty little secret for a while - a lot of Twitter followers are "fake."  Offers of generating a thousand "followers" for as little as $5 have been floating around the Net for a while and a number of websites advertise generating Twitter followers for pay.  And some news reports have been quick to question large jumps in followers - often asserting (without evidence) that they must have been bought and paid for.
  I should note that jumps in the numbers of followers can be fairly common for politicians, celebrities, and others in the public's eye.  A big story or new release can focus attention on public figures, putting their names (and twitter accounts) to the attention of thousands or millions of other Twitter users, sparking an interest in starting to "follow" them.  I'd be much more suspicious if the jump in followers occurred before a big event (especially if during that event, the number of Twitter followers are touted) than if it follows something that triggers public attention.
  There are two basic methods for generating "followers" for the "target" Twitter account.  In the first, software looks for existing accounts of people with similar interests.  The software then uses the "target" account to "follow" the generated accounts, hoping that many will reciprocate.  The other approach is to use bots to generate new Twitter accounts just to "follow" sites that pay for them.
  Now, a UK firm, StatusPeople, has developed a means of identifying fake followers.  The measure tracks the accounts of followers back and looks at a number of measures of account activity.  A follower is considered to be fake if there is little activity beyond following. As a company spokesman described it:
 "A fake account is set up to follow people or send out spam. They normally have no followers, but follow large numbers of people. An inactive account is one in which there has been no activity for a while. They could be real people, but we would describe them as consumers of information rather than sharers of information. A good account is everything that remains."
The company tested its process on a number of Twitter accounts for politicians and celebrities - among the highest reported levels of "fakes" were the Twitter accounts of soccer star Cristiano Renaldo (79% of his 12.6 million followers were fakes) Lady Gaga (71% of her 29 million followers were fakes), President Obama (70% of his 18.8 million followers were "fake") and major professional sports leagues (MLB was 75% fake followers, NBA has a 73% fake rate, and NFL had 68% fakes),
  Unfortunately, the measure doesn't indicate what level of followers were bought and paid for, and by whom.  Also, let me note two methodological caveats with the measure - first, the metric differentiates between "fake", "inactive" and "good" accounts before combining "fake" and "inactive" and reporting the total "faked"; second, for the accounts with large numbers of followers, they don't actually test and validate each follower - they examine a sizable number of the most recent followers and extrapolate that to the whole population of that site's followers..

  You can check the "fake level" of Twitter accounts through Fake Follower Check.

Sources - How To Expose Fake Twitter Accounts Of Celebrities, International Business Times
Buying Their Way To Twitter Fame, New York Times
10 Sports Twitter Accounts With a Shocking Number of Fake Followers, Mashable Social Media

edited twice to correct typos

No comments:

Post a Comment