Reuters reported recently on a study published by an Italian communications professor that suggests that up to 46% of Twitter followers for brands may be bots. It’s not entirely surprising as companies have been using autoresponders to “tailor” advertising messages to followers – augmenting follower counts is a logical extension of that.
The problem that I find with the study is the methodology. Here are some key aspects:
- 39 brands were analyzed
- Since bots can’t be identified, account behaviors were analyzed to assess whether the account was bot or not. Some of the criteria: no profile picture, no headline, only interfacing from one platform – attributes that could likewise identify an inactive user.
- A sample of 10,000 followers was analysed for each account.
While 10,000 random followers may be a statistically significant sample of each account, 39 accounts is most certainly not a statistically significant sampling of brands on Twitter. In order to make any conclusions, the companies studied would have to be a significant (random) sample of both. The ambiguous difference between the bot definition and inactive users also adds some uncertainty to the findings. Given the odd methodology, the 46% number is meaningless. The true statement is that up to 100% of brand followers could be bots. It’s a possibility, but has very little relevance to anyone following a brand on Twitter (though if you follow a brand on Twitter, the likelihood that 100% of its followers are bots decreases).
All things considered though, it appears that Reuters and the study authors tempered their conclusions (more or less) to the constraints of their data set. The same cannot be said for subsequent reports about the study:
- Nearly Half of all Twitter Followers are Bots
- Companies’ Twitter Followers Mostly Robots: Study
- Almost half of Twitter brand profiles’ followers are ‘robots’: Study
- Didn’t you know it? Half of the followers of businesses on Twitter are actually robots
- Study Finds 46% Of Company Twitter Account Followers Are Bots
The data of course doesn’t substantiate any of these headlines, but you can see how a slightly flawed anecdotal study can be blown out of proportion.
The reality of a study on bot followers is that it’s of zero relevance to anyone using Twitter. It does however underpin social advertising tactics, showing that there is real opportunity for businesses to genuinely engage people in a space where many businesses will not.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Did you enjoy this post? Consider sharing future posts with a Facebook task or Twitter task through ifttt
Not into long-term commitments? Please consider sharing via the share buttons above.
Thanks for your support – please shoot me a note and tell me what I can do for you (can I like you on Facebook, follow you on Twitter, share your blog, do your dishes?)! ***disclaimer – I hate doing dishes***
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
