Nye argues that it is a combination of lots of things, like economic success, technological prowess, good governance, lack of corruption, etc.
In a similar fashion,”influence” has become a central topic of discussion in marketing circles. The opportunity to get messages across with less expenditure of resource and effort is enticing indeed. However, defining people’s influence (excluding, of course, celebrities) is problematic.
Nevertheless, measuring influence is certainly not impossible and there are several valid approaches. Here’s an overview:
A Network Theory approach
Certainly the most comprehensive method of evaluating influence is the network theory approach. It is used extensively in counter-terrorism, consumer analysis in closed networks (i.e. telecom) and organizational consulting.
A particularly impressive application of network theory to decipher influence was used in the aftermath of the 9/11 attacks to determine the leadership structure of the terrorists. Of course, the actual methods are classified, but people familiar with the technique point to this fascinating paper by Valdis Krebs as a reasonable approximation.
On his Orgnet site, he gives an overview of how it all works, which I will summarize below:
The method incorporates three objective measures of influence:
Degree Centrality: This is simply a measure of direct connections. In the network above, Diane would have the highest degree score because she is linked to the most people. However, this measure can be misleading because, as I’ve explained before, just adding followers won’t increase your influence and can, in fact decrease it.
Betweeness Centrality: Another way to gain influence is to simply be well positioned. For example, the personal assistant of the CEO can function as a gatekeeper and therefore exert influence even without a lot of connections just like, in communist times, the woman who ran the meat store was a mini Duchess considered worthy of patronage.
Despite her popularity, if Diane wants to get to Ike and Jane, she will have to go through Heather, who has the highest betweeness score.
Closeness Centrality: This measure is somewhat of a synthesis of the previous two, but might be more important, as this paper suggests. It measures the amount of social distance a node would have to travel to get to anyone else in the network using both direct and indirect links. Fernando and Garth, in our sample network, have the highest closeness scores.
Implementing and augmenting the Network Theory method
The problem, of course, with the network theory method is that it is very difficult to implement, especially with large networks. Even in the small network of ten people shown above, the number of calculations is significant. For thousands, or even millions of nodes, the costs will often be prohibitive.
However, Christakis and Fowler, in their book Connected, have found a reasonable work around. Through a quirk in network math, our friends tend to be more central than we are. So simply asking people who their friends are and then targeting them has the potential to carry buzz more efficiently.
This Economist article uncovers even more clever ways to utilize insights from social networks. For instance, one telecom company discovered that bosses tend to make long calls, but receive short ones. Therefore, by examining bi-directional communication, influence can be uncovered.
Social networks are an increasingly active area of serious academic research, so we can expect greater insights to be uncovered in the coming years.
Do you have Klout?
The rise of social media has led to a cottage industry of influence evaluation. There are web sites popping up everywhere purporting to measure your influence. While their methodology isn’t as sophisticated as the network theory approach, they do measure how people react to social media activity, which is a plus.
Valery Maltoni, over at the Conversation Agent blog, gives a nice overview of the major evaluation services. The most popular of these, Klout, uses the following three metrics:
True Reach: This is similar to degree centrality in that it is a measure of first degree links. However, because Klout focuses on social media, they factor out things like bots and people who aren’t actively engaged.
Amplification: Klout also looks at what people do with your tweets and status updates. You score improves when people retweet, mention your Twitter name, hit the “like” button or comment to your status updates. Because it focuses on actual results, this is one area which probably improves on the basic network theory approach.
Network Score: Klout also analyzes how influential your connections are. This would be a partial surrogate of betweenness and closeness in the network model.
The online influence measurement services have become somewhat controversial and many people think they are a step backward. I, however, believe that they are a welcome addition to web metrics. They still have a long way to go, but do seem to be serious about making a significant contribution. Joe Fernandez of Klout seems especially constructive.
Big seeds
Despite the increasing sophistication of influence measurement, virality is still a crap shoot. As Duncan Watts showed in this paper, while some of us are more influential than others, the likelihood of an “influential” to set off a viral chain isn’t much greater than anybody else.
His solution? Mass media.
While many social media advocates have done their best to show their contempt for conventional media tactics, Watts feels differently. Moreover, as he is no media mogul, but one of the true network theory pioneers, those that would ignore him do so at their peril.
As I explained in an earlier post, he calls his approach Big Seed Marketing. His reasoning is that since influence is so hard to track, it is much better simply to start with a lot of reach (i.e. a big seed) and use social media to amplify it. It seems to me to be an incredibly reasonable and sound approach.
Influence targeting
So where is this all going? It’s hard to tell, but if I had to guess I would say that we’ll end up with an approach that looks a lot like the marketing tactics of the past.
Generally speaking, we target not because we are sure that we have the right people, but we want to exclude those that are least likely to be valuable. In other words, we just want to gain some efficiency by improving our averages. Services like Klout can help us do that.
On a larger scale, social media such as Facebook, Twitter and LinkedIn are beginning to change the data infrastructure of the Web by exposing connections that heretofore were hidden from marketers. Undoubtedly, this new data will be incorporated into demand side platforms that are increasingly driving online media buying.
One thing is for sure. We are embarking on a new media paradigm where old truths are being infused with fresh possibilities.
Greg Satell is a blogger and a consultant at the Americal online media Digital Tonto. You can read his blog entries at http://www.digitaltonto.com
