One of my colleagues is doing research on social network use, and she asked me to help get people to take her survey. It takes about 8 minutes to fill it in. I was glad to help, and to do it, I set up a test.

First I posted the link on my Facebook page and asked my friends to take the test. About 20 out of 188 did so.

A few days later, I posted a link on Twitter, and asked everyone following me to take the test. Another 20 did, out of around 3000.

And now, here’s an interesting question: how do you blog readers stack up against my Facebook Friends and Twitter Followers? My bet is that you’ll win – to prove me right:

Click here to take the survey yourself!

The contrast in results between Facebook and Twitter illustrates some important lessons about influence. A lot of people have been talking recently about how to best measure online influence. Like innovation, influence is another thing that is awfully hard to measure.

One big problem is that influence is pretty hard to define in the first place. What does it mean? To me, influence is about getting people to take action. If that’s the case, you might think that I am lot more influential on Facebook (where about 11% of the people on my list of friends took the survey) than I am on Twitter (where about 0.7% of the people on my list of followers took the survey).

But I’m the same person – so am I influential or not?

One of the best thinkers around right now on the topic of influence is Valeria Maltoni – here is what she says about Klout’s attempt to measure influence:

I can tell you that Klout knows squat about me and my behavior. Zero, nothing, niente, nada. Got it? The fine folks behind the algorithm have no idea of who gets my emails and calls, which are the tools I use most to conduct my real business.

They know nothing about what I read and why I read it, because they are not reading these articles or talking with me. They are just tabulating the keywords and volume of my Twitter activity. Twitter. Shrink me into 140 characters. Or maybe they are 134 more than those in Pirandello’s play (more context was the lesson there, it is here, too).

Are the people in my life even on Twitter? You don’t know that.

Am I the person you read here every day? (And I thank you humbly and sincerely for reading and thinking about this content.) You are not just the person who is reading. You are much more than one thing you do, so why would I be just the person who is sharing here?

Martijn Linssen has done a lot of good work assessing the success of Klout in measuring influence.

This experiment illustrates some important points about influence:

• You can’t reduce a complex phenomenon to a single number: influence happens in person, online, with people we know well, and with people we’ve never met. This makes it very tricky to measure. This leads to:
• Don’t mistake the metric for what you’re trying to measure: the real problem with things like Klout is that once we have a metric, people will start trying to game the metric. You can do this, but it doesn’t increase your actual influence. The only way to do that is to do things that have a strong, positive impact on people, and to do it consistently. That’s a system that you can’t game – and if you focus too much on managing the metric, you’ll actually get worse at the thing that really matters.
• Influence really happens in networks: Duncan Watts has done a lot of excellent research that shows that the main thing that causes ideas to spread within networks is the extent to which the people in the network are likely to spread the idea. Here is how he put it in a recent post:
  

  When we hear that a raging forest fire has consumed millions of acres of California forest, we don’t assume that there was anything special about the initial spark. Quite to the contrary, we understand that in context of the large-scale environmental conditions — prolonged drought, a buildup of flammable undergrowth, strong winds, rugged terrain, and on so — that truly drive fires, the nature of the spark itself is close to irrelevant.

  Yet when it comes to the social equivalent of the forest fire, we do in effect insist that there must have been something special about the spark that started it. Because our experience tells us that leadership matters in small groups such as Army platoons or start-up companies, we assume that it matters in the same way for the very largest groups as well. Thus when we witness some successful movement or organization, it seems obvious to us that whoever the leader is, his or her particular combination of personality, vision, and leadership style must have supplied the critical X factor, where the larger and more successful the movement, the more important the leader will appear.

• Consequently, understanding how ideas spread through networks is essential to understanding influence: this is an idea that Greg Satell has incisively written about. Here’s what he says:
  

  In effect, starting an epidemic is similar to a broadcast search. You are better off casting your net as widely as possible and reaching influential people as well as less influential ones. (See this article for more about broadcast and directed network searches)

  Some paths will fail, but the more paths you initiate, the more likely that your idea will infect those who are susceptible to it. Just like delays at any airport can affect large hubs, influence can originate anywhere in social networks.

So the real answer to the question of whether or not I’m influential is: yes. Or no. Or maybe. The one thing that we can say is that my Facebook friends seem to be a lot more willing to act on a request for help than my twitter followers are. But this again is a network effect, and doesn’t actually have that much to do with me personally. The connections on Facebook are different, and people use that network to meet objectives that differ from those that Twitter users are trying to achieve.

Influence is very important, but measuring it is hard. The best way to increase your influence is to keep producing ideas that help people.

Also, if you could retweet this, that would be cool – it would really help my Klout score…

(just kidding – I’d much rather have you fill in Sabine’s survey – and the number of people that have gone to the survey from here is now higher than we got from either Facebook or Twitter!)

(that’s The Minutemen playing their great song Take Our Test)