Matthew Cornell

March 25, 2011

Between 2007 and 2009 I spent a ton of time in Twitter before it finally hit me that 1) the net improvement to my life was zilch, and 2) had I thought of it going in as an experiment, I would have quit a long time ago and freed up energy for more effective efforts. Of course social media tools can provide plenty of value, but, as Alex said, Social media is an addictive time suck.

How do we go about measuring the value of Twitter? Business calls it ROI, but I think of it as simply what you hope to get out of it. The key is deciding why you’re using it. In my case I was dabbling, which is a fine motivation, as long as it’s done experimentally. After all, how many discoveries came from just getting curious and trying out something new? But here I should have set a time limit, and I’d still want to have something quantified, even it it’s as soft as “perceived value.”

But for more specific uses, coming up with measures is important. Are you trying to get more customers? Do you want to hear from people who can give you ideas for your product or book? Or maybe it’s more of a social pulse use – keeping in touch. Some metrics are straightforward, such as # inquiries about your business, or number of tweets from others that made you smile. However, I think a major challenge is latency – the time delay between action on your part and resulting effects seen in your life. For example, it might be months before you hear from someone who’s been silently reading your tweets. Maybe in those cases we could make the measure more direct by asking them explicitly what the impact is. I’m not sure.

While I didn’t treat using Twitter as an experiment per se, I managed a few times to use Twitter itself as a platform for experimentation. For example, I had a business trip to North Carolina coming up, so I tested using the tool to make business contacts for the trip. I subscribed to a bunch of local folks, followed their tweets, and replied when appropriate. The result (measured in # business meetings set up) was zero, but the “at least now I know” feeling was satisfying. One thing that’s vexing is irreversibility. Any experiments we might do with our followers has permanent effects because individuals have memory (“Oh yea, isn’t he the guy who tried giving away the iPads?”), and the account has memory (I might lose or gain followers as a result). One workaround is to make separate accounts for testing, but unfortunately you can’t clone followers. Then again, making us start from scratch is probably a good way to keep the experiment clean.

Others have used Twitter this way, such as a much-publicized psychic experiment. (I found the comments on Richard Wiseman’s Blog announcing the experiment to be fascinating from the experimental design perspective.) A little searching turned up other posts like Twitter for Research, The start-up chronicles: Experiments with Twitter, Using Twitter for Market Research, and a Twitter Tip Sheet for Experiments. I don’t think I’d count The Anybody/Everybody Twitter Experiment as one, though; what was measured?

What do you think? Have you used Twitter in one of these ways, i.e., as itself an experiment, or as a platform for running them? And if you regularly tweet, I’m curious to hear what you get out of it, and what you’ve measured to determine that.

[Image from horslips5]

(Matt is a terminally-curious ex-NASA engineer and avid self-experimenter. His projects include developing the Think, Try, Learn philosophy, creating the Edison experimenter’s journal, and writing at his blog, The Experiment-Driven Life. Give him a holler at matt@matthewcornell.org)