Untangling fake followers, Newt Gingrich and Twitter’s Suggested User’s List

If most of former United States Speaker of the House @NewtGingrich‘s followers on Twitter are “fake” or inactive, they’re more likely to have primarily come through a gift from Twitter in 2009 than any clandestine campaign purchases. Remember when Anil Dash wrote that “Ryan Osborn does too, (@Rozzy) over at NBC News, who reminded me of Dash’s post today.

The issue of “fake followers” arose today after an incendiary post by John Cook over at Gawker sourced the comments of an anonymous former campaign staffer alleging that Gingrich paid “follow agencies” to create some 80% of his followers. Mashable has picked up the story, asking if the former United States Speaker of the House bought most of his Twitter followers.” The allegations surfaced today a few weeks after Politico published a glowing story about “Gingrich being miles ahead in the Twitter primary,” with “an impressive 1.3 million followers.”

Unfortunately, covering this election season as a “social media horse race,” with social media followers numbers as a 21st century success metric, holds some risks. (OhMyGov, to be fair, generally takes a much nuanced approach than most mainstream outlets.)

All three writers (and their editors) have apparently forgotten that Newt Gingrich was added to Twitter’s Suggested User List back in the fall of 2009, about three weeks after an AP story reported that Government” list. He has retained over a million followers since then in the way that Politico depicted — but he didn’t get there through deepy policy wonkery and quick wit alone.

On the one hand, if it’s confirmed that the campaign has been pumping Gingrich’s follower numbers through third parties, it will be yet another case study to add to the rapidly growing pile of political social media snafus and I’ll eat some crow about the origin of those hundreds of thousands of accounts. As Vanessa Fox tweeted, “both methods likely contributed. people/orgs fall back on fakes to drum up #s way too often, but suggested list def much of it.”

On the other hand, if it turns out that media outlets simply developed amnesia about Twitter’s early history, there might be enough crow around to bake a Blackbird Pie or two.

UPDATE: An analysis by social media analytics firm PeekYou supports the anonymous staffer’s accusation of “follower fraud:”

The Consumer Ratio measures the percentage of a Twitter audience that is identified as a “consumer” or “voter” in Newt’s case, vs business, private/anonymous and spam accounts. The average range sits anywhere between 30-60% human depending on this type of account. Newt’s was 8% — the lowest the team had ever seen by 5%.

“We have seen some pretty low ‘Consumer Ratios’ in our testing, but Newt Gingrich’s was the lowest we had ever seen. At first, we actually thought it might have been a bug on our side, but a quick manual look at the data showed our analysis was true,” said Michael Hussey, CEO and founder of PeekYou.

“Once the news broke yesterday the team went back to look at the report. The data supported that out of Newt’s 1.3 million followers only 8 percent (2 percent less than claimed in recent media reports), are identified by our algorithm as humans, meaning Newt’s follower count is really closer to 106,055,” continued Hussey.

When assessing those approximately 100,000 followers via PeekAnalytics, the company established that 67% are male, 41% are over 35 years old, and 61% have less than 100 online connections. Using normalized population data, the top 5 states for legitimate followers are Georgia, North Carolina, Pennsylvania, Texas, and Wisconsin. In addition, more than the majority write a blog (typically supporting press/blogger interest) and favor Facebook as their social network of choice.

UPDATE II: Indiana University has posted an analysis of the 2012 presidential candidates’ Twitter followers that offers some more insight. In the chart below, pulled from Indiana’s post and data, the data from a random sampling of 5000 accounts following the candidates is displayed. Smaller values correspond to bluer colors, larger values corresponding to redder colors. Draw your own conclusions.

*I was added to Twitter’s “Technology” recommendations in early 2010, after the SUL was scrapped for an algorithm that I suspect is at least partially based on Lists. My Twitter account @digiphile) now has nearly 90,000 followers, up from 7,500 or so. While I think most of you are real, I suspect some of those accounts are similarly inactive or (gasp) robots. I haven’t had the time to weed through tens of thousands of notifications to groom them out. If you’ve made it to the end of this post, know that I’m much more interested (and grateful) to see how many of you reply to me, read me or share my work than my follower count. It’s ok. Nobody really has 1,000,000 followers, after all, with the obvious exception of @JustinBieber, whose intensely loyal fans have been known to put an unexpected stress test on young photosharing startups.

[Photo Credit: Gage Skidmore]

4 thoughts on “Untangling fake followers, Newt Gingrich and Twitter’s Suggested User’s List

    • PeekYou is adding the last tweet date metric to the data set too … not live yet.

      We ran Marshall’s follower report last week — his consumer ratio is 47% (which is very healthy — above average) and his audience reach is actually the same size as Newt’s — his 32K followers are much more robust web users than Newt’s 100K actual followers.

      46% of Marshall’s followers have over 1000 social connections (across various social networks).
      61% of Newt’s followers have less than 100 social connections

  1. I remember being at TwitterCon back in 2009 and @amberlynne:twitter  mentioning that  @peacecorps:twitter  had jumped from thousands to hundreds of thousands of followers seemingly overnight (bumping them to top 10 of GovTwit IDs tagged “agencies” http://govtwit.com/list/all/tags/agencies. Pinged someone at Twitter and they confirmed ID was recently added to SUG: http://twitter.com/#!/GovTwit/statuses/5073118811

    NYTimes did a couple of stories on effect of SUG back in 09 as well: http://www.nytimes.com/2009/06/14/weekinreview/14cohen.html and

    • We have run many other accounts that were on the suggested user lists — and they have normal “consumer ratios.”  Newt’s 8% ratio was a significant anomaly…

      Take a quick manual look at his followers — and you’ll see for yourself.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.