Build from Zero Followers then Alienate Half – A Twitter experiment
Twitter Experiment Proposal:
Build up a following from zero then alienate them, dropping to 50% of peak number.
Start with a new Twitter account, then do whatever is necessary to build from zero followers to some number x followers. Then without using @mentions or DirectMessages, send out tweets that are designed to make people unfollow/block you. The goal of the experiment is to tweet in such a way that at least 50% of your followers decide to unfollow you.
My gut feeling is that this goal would be unattainable once you build a base of more than 80 followers (x > 80). Sheer offensiveness becomes its own self-perpetuating draw, so it would be tricky to tweet messages that are individually objectionable but not sensational or titillating to others. The other main, perhaps primary, factor making this difficult is that people tend to not opt-out of things unless they are repeatedly annoyed (see methodology comment #1 below).
Some comments on experimental methodology. Each could end up being an additional variable in a larger experimental model — each would lead to a different build-drop trending profile:
- Settle on a reasonable tweet-frequency, e.g., 4 tweets per day — it should go without saying that the spirit of the experiment is not honored if, for example, you tweet every 10 seconds after peaking with the goal of driving people off.
- Decide whether you are going to be a generalist tweeter or if you will focus on some hashtag-based group, community, or topic.
- Decide on if and how you will block spammers and promiscuous/indiscriminate followers. These people (and robots) typically do not cull friends, so once they are following you, they won’t stop no matter what. It might be beneficial to decide on a “recipe” for automatically blocking new followers; e.g., auto-block if a follower (1) has over 50 followers but (2) the number of people following them is less than 20% of their number of people they are following; or auto-block if a follower seldomly tweets but is following a large number of people (although this could just as easily be an honest lurker not a robot). This could be resolved with a generalized method for identifying twitter spammers, robots, and junk tweeters — what approaches already exist?
Fancier automated approaches:
- To drive away followers: It would be interesting to create a robot that surveyed the tweets of each of your followers and generated tweets designed to be turn-offs to each one. Extra points if the robot takes your other followers’ discussions and moods into account and tempers each response to either maximize multiple unfollows or minimize the chance that (for example) Follower #1 is put-off while inadvertently appealing to Follower #2.
- For the initial build-up to peak number: Without doing a thorough search, I’m sure there already exist robots designed to create tweets that appeal to lots of people with the hopes that they start following. If you know of any, please share.