Faking it on twitter, follower fraud

Hey there,

So, it's not Monday (which is my normal Social Media blogging day), but wanted to do a special post inspired by a tweet.

Twitter's great. Let me just say that. And while it's awesome to be able to connect with like-minded people, I love the opportunity to interact with people who might not be of like mind. One such person is @JamieGator, and he tweeted that President Obama's twitter followers were 70% fake, and a link to a USA Today piece:

"President Obama's Twitter account has 18.8 million followers -- but more than half of them really don't exist, according to reports." (read more)

The source of this was an article in the New York Times on celebrities buying their way to Twitter fame:

The practice [of buying followers] has become so widespread that StatusPeople, a social media management company in London, released a Web tool last month called the Fake Follower Check that it says can ascertain how many fake followers you and your friends have. (read more)

Now I might have read the latter piece wrong, but if not, then the "reports" mentioned in the initial quote is in actuality just a piece of data generated by this online tool Status People has made available, their Fake Follower Check.

Unfortunately neither article seems to shed any light on the metrics as to the methods or accuracy, even though Status People breaks it down on their Find Out More page:

• How does it work? We take a sample of your follower data. Up to 1,000 records depending on how 'popular' you are and assess them against a number of simple spam criteria.
  On a very basic level spam accounts tend to have few or no followers and few or no tweets. But in contrast they tend to follow a lot of other accounts.
• How accurate is it? For those of you with 100,000 followers or less we believe our tool will provide a very accurate insight into how many inactive and fake followers you have.
  If you're very, very 'popular' the tool will still provide good insight but may better reflect your current follower activity rather than your whole follower base.

To be fair, there is an update to the page that was just added this past week, on 8/22/12, with regard to their analyzing capacity, and how accurate it is:

We have managed to extend the accuracy of our App from 5,000 to 100,000 followers. So instead of grabbing the last 5,000 follower records and assessing up to 500 of them. We now, depending on the size of your follower base, grab up to 100,000 follower records and assess 1,000 of them across that base. This increases the accuracy of our App significantly. In fact it means that now 97% of the Twitter accounts checked will return an accurate set of scores. (read more)

So there it is. The New York Times articles starts of with a statements saying "If accurate", referring to this tool. Truth is, it's not a matter of "if accurate", it's "how accurate". And 97% is pretty good. Which brings me back to @jamiegator and his tweet.

70% of president's 19 million TWITTER followers fake... - 70% of president's 19 million TWITTER followers fake...(Th... ow.ly/1m8oHH
— James Cruce (@JamieGator) August 25, 2012

Now the the 70% fake stat is probably the one point of confusion with the Fake Follower tool. When you first go their page, they show Tweeps and the stat for their percentage of"Fake" followers. But when you actually give it authorization to access your account, they give you a number for your followers that are "Fake", "Good", and also "Inactive". Yes, I was curious and checked out my own stats.

And, yes, I'm quite happy with those numbers. But, just so you know, I've always been selective about who follows me, and have been even more so since I reached my twitter following limit for the first time, back in May. In fact, I actually block people if I don't follow them back. Doesn't happen that often, but the general gist is that if they're tweeps that I don't want to follow, I don't want them following me. And this is reserved for certain types of accounts, like those that are spammy, like real estate agents, porn bots, and anyone who's on #TeamFollowBack, among others.

But I digress. Back to the 70% stat, I had to respond to @JamieGators tweet by putting the number in context.

.@jamiegator 70% fake OR inactive. Still means @barackobama has > 6 million real/active followers, much > than @mittromney's 536,000 or so.
— JR Russ (@AWayofLife0) August 25, 2012

To be precise, 29% of @BarackObama followers were fake and 39% were inactive, as of the time of this post. 

So in the initial quote in the New York Times, it's not that "more than half of them really don't exist", it's that more than half of them are fake OR inactive. Or that almost a third of them don't really exist. And even with only 32% of his followers being "Good", that still left over 6 million "real" followers on Twitter, compared to 59% of @MittRomney's 908,000 or so total followers.

Did I mention I was a mathlete in high school? If we finally want to put some perspective and compare "Good" followers of one to the other, @MittRomney has only 8.87% of the real/active followers that @BarackObama has. I love numbers...especially when they're put into context.

Anyway, looking forward to seeing how Status People continues to improve this metric, as well as check out their other, paid services. In the meantime, this is a nice freebie. When coupled with other tools like FriendorFollow and Klout, it continues to become easier and easier to take charge of one's own social media presence, depending on how intentional you want to be with it.

Any questions? Also, do you find this helpful and useful with your own social media use? Or do you manage fake and/or inactive followers on Twitter another way?

Please share in the comments. And check out pieces on Status People's Fake Follower tool on Mashable, Facebook Has 44% Fake Followers, Twitter Has 33%, and CNET, How many of your Twitter followers are fake. Finally, keep an eye out for my weekly "Social Media Monday" post in a couple of days, discussing how SM has effected my upcoming Portland Trip in less than a week.

- JR