Automated Tweets: Results of a Two-Week Test

Two weeks ago today, I began an experiment using Hootsuite to automate tweets daily. I wanted to know if it would increase conversions, result in fewer followers, or earn me a stern talking-to from Twitter purists. These are my findings.

To begin with, let’s take a snapshot of what my Twitter presence looked like two weeks ago.

  • 1,494 followers
  • 10,054 tweets
  • I had scheduled maybe half a dozen tweets in the past couple of years. To say that I never automate would be fairly accurate, if a bit hyperbolic.

Here’s where I stand today, two weeks later.

  • 1,522 followers
  • 10,254 tweets

For this experiment I wanted to promote a few things. First and foremost, that I’ll be providing podcast production consulting services at Affiliate Summit East, coming up in about 3 weeks’ time. This was my main focus, and the subject of 66% of the automated tweets. Second, that I’ll be leading a panel at Affiliate Summit called Affiliate Improv. Third, the Be a Better Podcaster show. Lastly, my other shows.

Now, there’s something that I freely admit to: this is a very small audience I’ve got here. I’m no Brogan. I’m a Clark, but I’m not Brian Clark (and we’re not related). With just about 1500 followers, my results here are probably wildly different than someone with a zillion followers would experience. Granted.

I used Hootsuite to schedule three tweets daily for two weeks. Two of the three would be advertising my consulting service at Summit, and knowing that some people only check Twitter in the morning or only in the evening, I decided to alternate messages. The first message was:

2-for-1 podcast consulting at Affiliate Summit. bring a friend or colleague, split the cost :) http://qaqn.com/csg #ase11

The next morning, this ran:

going to Affiliate Summit? interested in #podcasting? i’m open for podcast consulting while we’re there! http://qaqn.com/csg #ase11

Those two messages alternated morning and mid-day. For example, the first ran every other morning beginning on Sunday. The second ran every other morning beginning Monday. The first ran every other afternoon starting Monday. The second ran every other afternoon beginning on Sunday. Get it?

The third messages ran late night. These were far less structured and only served to break up the other two so that Sunday afternoon’s message wouldn’t be repeated Monday morning with nothing between. I’m a pretty prolific Twitter user, so this wasn’t likely to be a problem, but just in case. I didn’t want to be extremely obvious about automating these things.

A promotional tweet for Affiliate Improv ran 5 of the 14 nights. A tweet for each of the (then) six episodes of Be a Better Podcaster ran 6 of the 14 nights. The remaining three late night tweets were for my other shows – one each for Geek Dads Weekly, Inside Internet Marketing and Yet Another Weight Loss Show.

The Results

So, what did I find? Well, first, as you can see above, I gained followers. This is not a result of automated tweets, of course, but a natural increase. I scheduled 42 tweets, but tweeted myself another 158 times on top of that. The important distinction here is that I didn’t lose followers – not even some people that I know that are always saying they’ll unfollow anyone that automates (don’t go away now that you know I automated, guys – this isn’t a regular thing for me, and you’ll see why shortly).

My messages were retweeted only 17 times in total – less than half the 42 messages got a single retweet.

I did gain a potential client, but it wasn’t due to these tweets – it was someone who made contact with me based on a conversation held elsewhere.

What about my traffic? I’m afraid this is the point where my experiment completely breaks down. I have no traffic data. My analytics package was not set up correctly and hasn’t been collecting data. I never noticed because I’ve always been far more interested in the number of podcast downloads than the number of site visitors. Google Analytics was set correctly on July 13th and I can tell you that since then, Twitter traffic has accounted for 13.06% of my site’s traffic. Unfortunately, I don’t know if that’s up or down from the period before I started automating.

My Conclusion

It took me about 90 minutes to figure out how to set up Hootsuite for bulk automating, create a CSV file with all the appropriate fields and customize every single tweet. You see, Hootsuite actually requires you to upload unique tweets. You’re not allowed to upload the same tweet 50 times. I did not want to come up with dozens of variations of the same tweet, so after some trial and error, I figured out that I could upload garbage, then edit each scheduled tweet individually – pasting in the correct tweet text. This all took the bulk of a morning, and my return? Nothing… well, nothing noteworthy. The automated tweets seemed to have no effect on anything.

I guess the most important takeaway from this for me is that automation isn’t bad. It was not, in this case, particularly good, but it wasn’t bad. I don’t think I’ll spend any time doing it again anytime soon… maybe if I ever hit 10,000 followers and I have a larger base of people to work with, but until then, it’s pretty much a waste of time.

Your milage may vary.

About Daniel M. Clark

Daniel M. Clark is a podcaster and proprietor of QAQN, a writer at danielmclark.com, and an all-around cool dude everywhere else.

God, I hate talking about myself in the third-person.

Vinny O'Hare August 1, 2011 at 12:37 pm

I wish you would have told me what you were doing I could have saved you half the morning. The cvs upload tool sucks and I could have given you my template that works.

The first fail here was scheduling them for the morning. Most of the people on the West coast probably didn’t see it depending on what time you sent it. Same for the East coast if you did it too late.

Comments on this entry are closed.

Previous post:

Next post: