My Twitter experiment has been going through the weekend. It’s time to do a check-in on its status. As mentioned in the initial post, I’m running an experiment with a fake Twitter account to see just how many followers I can get by playing along with the #TeamFollowBack type hash tags. As of my last post, I had 282 followers on the fake account. It’s been less than 24 hours and I now have 430 followers. The strategy has been essentially the same for the whole day – I have a number of 20,000+ accounts followed that seem to just spew the #TeamFollowBack (or #TeamFairyRow or #FollowBack or #500aday) hash tag periodically  through the day. I for the first two days of this experiment, I kept TweetDeck running and would just hop in every now and then and retweet a handful of tweets.

At some point today I found myself with a bit of free time, and I wanted to see if I could take this to the programmer’s logical conclusion – automation. Twitter has a great API and there’s a ton of clients that make it super easy to work with it. I’m using ntwitter for node.js as my interface to the API. The strategy I used for my little script is to use the streaming API to read my fake account’s “home page.” Any time a tweet comes in where the author has more than 20,000 followers, I use the REST API to retweet it blindly. I don’t even check for the #TeamFollowBack hash tag. This took less than an hour to implement.

The results at first were pretty encouraging. My script will retweet up to 60 times and then bail out, reporting the net gain or loss of followers that it detected during its run time. At first, I was easily getting 20 or so net new followers per run (and each run would take maybe 10-15 minutes). That continued for around 100 followers.

At some point, the rate of new followers dropped dramatically. I wasn’t watching it super close the whole time (I was playing more Dungeon Defenders with the boy), but I’m now down to 2-5 new followers per run. I have three theories about what happened. First, it’s entirely possible that I’ve been followed by everyone that’s playing along right now. The second possibility is that I’ve been followed by everyone that plays along *at all*. Finally, it’s possible that because I’m not playing the game 100% (remember that I don’t actually follow anyone back, even though that’s how it’s supposed to work), people aren’t following me to begin with.

I’m guessing that my first theory is what’s happening. I’m going to turn the script off while I put the kids to bed and then fire it back up in a couple of hours. We’ll see if my follower count goes back on the rise then. For the record, in the time it took me to write this article I gained 8 more followers, so the experiment is far from dead regardless. 438 and counting!