(Originally posted on October 20, 2011)

Below the results of a social experiment, where a new Twitter account was created, which made it explicitly clear that no content and only spam would be supplied, to obtain significantly more followers and social media influence than average, or proper. Here is the story…

Background

Recently I saw a posting that Britney Spears had managed to get more than 10 million followers on Twitter. In addition, whenever I speak with ‘iltwitterate’ acquaintances or read articles in newspapers, one’s influence in social media seems to be measured by the number of followers. Apparently, a measurement like Klout which takes into account the potential of getting people into action, requires too much social media-knowledge for most.

Regular social media

After having worked hard in creating content on customer service (in Dutch) and posting it every two weeks @peteralderliest, after 6 months I managed to interest 290 tweeps in following me. It is obvious that my cycle time of 2 weeks (‘Takt’in LEAN-terms) does not correspond with the cycle time of Twitter, where updates are required multiple times per day. Below is an overview of stats of several |Dutch social media parties which do much better. First some publishing organizations, then followed by some personal accounts. For fun, I included mine as well:

Note: any account starts with a Klout score of 10 out of 100, with a worldwide average Klout score of 20. Justin Bieber is the highest and scores 99 at this moment, the previously mentioned Britney Spears does 86. The highest-scoring Dutchman this week was Alexander Klöpping (@AlexanderNL), with 86.

Spam in Twitter

At the same time, ‘wondering through Twitterland’ I was confronted with the phenomenon of #Followback. Various claims were – and are – made that numbers of followers are up for sale, which would put the 10 million follower claim of Britney Spears – and others – into doubt. Therefore, I checked out the following accounts, each of which has made no significant contribution to content whatsoever. Virtually all tweets are limited to ‘Shout out’s of new followers, containing contents like “Thank you for following @abc12345” etc.. Most have no link with a Facebook-page, except @teamfollowwack, which accumulated 784 likes on their Facebook-page, which was also linked to the @Teamfollowwacky-account. @Loopinloops is linked to a Youtube channel with 581 subscriptions (many back and forth if I may believe the various messages).

It is obvious, though, that the relatively low numbers in followers provide no data to support any claim of illegitimacy of the 10 million follower-mark of Britney Spears.

Social experiment

However, would it be possible for a regular tweep to copy this ‘success’? Thus, on October 16, 2011 @lifeevents2 was started, it provided no picture and explicitly mentioned ‘no content promised’ in the headline. Then, I followed a couple of followback mailinglists, followed followers of these mailinglists, followbacked ‘incoming followers’, and Shouted Out anybody who followed me and unfollowed those who unfollowed me. In this experiment, I also followed accounts in other languages (e.g. Spanish) and using cyrillic and Chinese alfabets, since the fact that we could not read each others postings was not relevant. These efforts led to the following results:

Results

On October 20th, this account was determined to be a ‘Networker’: “You know how to connect to the right people and share what’s important to your audience. You generously share your network to help your followers. You have a high level of engagement and an influential audience,” according to Klout. I find that ‘interesting’.

Lessons learned along the way:

• ‘Shout outs’ do generate mentions from followers;
• Several ‘operational lessons’ which I will not mention here as not to facilitate further abuse of the system;
• The application truetwit.com is used to authenticate followers via a DM. It is fairly useless, since a bot may generate the followings, a human would deal with the ‘fall-out’ in the DM’s. At most it would hamper the speed of following.
• Followback is explained by some as a very simple game: you follow many accounts, wait for them to follow you back, and than quickly you unfollow them again so that you can follow new victims. This is ‘Twitterspam’. There are various tools to keep track of such behaviour. I used fllwrs.com to make the unfollowers visible and unfollow them as well.

Limiting abuse

Fortunately, Twitter has already introduced some limits to the number of accounts one can follow: initially, the total is limited to 2,000 until the number of Followers has reached a certain point. This is one measure to limit aggressive following and unfollowing. However, this has not stopped the following accounts which provided ‘nothing but Shout outs’ to gain significant Klout even with less than 2,000 followers:

Apparently these accounts were even more succesful than yours truly in obtaining influence in social media. Furthermore, Twitter has imposed (understandably kept secret) relative limits on the number of accounts that can be followed once Follower-numbers grow beyond the 2,000 level.

Next week, Klout is implementing further steps in improving their rating system, for which they claim transparency. (Maybe pressed for improvements with the advent of the new entrant Kred?) I have not been able to clearly find the Klout-algorithm, but undoubtedly that is the result of my ignorance. Maybe it would be an idea to keep that rating system as secret as the Google-algorithms. It also would prevent frequent explanations whenever an update is rolled-out, and in my opinion most legitimate users would consider that perfectly acceptable. But then quit the transparency claim.

Follow-up

Joe Hernandez of Klout has expressed the expectation that the reported scores will be negatively affected by the new algorithm. Next week I will come with an update to show what the Klout algorithm change had for effect on the Klout scores mentioned before, and to what extent they managed to filter out the abuse.