Are social networks just pretty snowflakes?
Have just been to TedxTuttle. Interesting. Eclectic certainly. The clash of styles between a TED sit back and listen approach and a Tuttle sit around and talk approach didn’t quite meld together. I felt that a TuttlexTED event might have been better. Now there’s a thought. To which I can hear @lloyddavis saying “well, just give it a go then”.
The presentation that got me thinking the most was by Mat Morrison (@mediaczar). Mat had been doing some interesting work looking at social media networks and influence. Mat made the very pertinent point at the start of his presentation that in PR (indeed in all forms of communication and marketing) the objective was to find the smallest number of people in order to influence the largest number of people. His approach therefore to analysing social networks was based on this premise – i.e. find the most influential people and points of influence and understand the channels and connections that link networks together. He showed that even large networks often depend on only a few people to actually hold them together and make them work. Logically this would suggest that the focus in looking at influence in networks should be identifying and targeting these people.
Finding interesting patterns of influence within networks looks to be a very, err, interesting. The size of the network and the volume of data within it presents huge opportunities for number crunching and pretty graphics. But I couldn’t escape a nagging suspicion that while this might be interesting, it may not be useful. My starting point with social media is always the idea that it is different. Answers and approaches that work are very rarely conventional. Social media is a black swan – a never seen before type of thing. Up until now we have had individuals and we have had crowds, but now we have a hybrid – a connected crowd. The connected crowd is a very big thing, but you can only communicate with it as though it were a very small thing. Mass, one-to-many messages just don’t go anywhere. It is back to this old idea of mine that influence lies in spaces, not places.
Now at one level it is very easy to regard the connected crowd as a network, or series of networks. It certainly looks like that and many people have done some very good work looking at it as such (Antony Mayfield’s study springs to mind). But as with the ‘all elephants are large and grey but not all large grey animals are elephants’ sketch I get a sense that using conventional network analysis in social media may help us create interesting patterns, but like snowflakes, each of these patterns may be unique and therefore not something we can replicate or rely upon.
In the old days of traditional media, patterns of influence were fixed. A route that worked today will work the same tomorrow. The Financial Times as a source of influence was essentially the same yesterday as it was today and will be tomorrow. If we get our message into the channels that flow from the FT (paid for or mediated) we will always have the same level of influence. Social media isn’t like that. As I am fond of saying “in Twitter, all tweets are equal”. It is the content and the context that defines meaning and influence. The man who tweeted the picture of the US Airways plane on the Hudson River was not an influential twitterer before or after that event. He, or rather than particular bit of content, became hugely influential for a very short period of time. His star blazed across the network firmament for a day or so and then faded away. In a world where content has become separated from a means of distribution (i.e. the world of social media, especially the world of Twitter) content is free to carve its own distribution path or create its own influence. We can’t necessarily predict it or control it. In the ‘old world’, where we fixed the message / content in advance, the game was all about how we created and shaped its distribution. Now that is increasingly out of our hands.
As part of his study to find patterns and influence, Mat had taken a particular tweet that had achieved a significant level of distribution. I can’t recall it exactly but it was something like “London Twitter festival ends in chaos as Twitter crowd attacked by Facebook mob”. The tweet itself was seen as the dependent variable, essentially no more than a marker to highlight a particular distribution path. But fundamentally, the reason this tweet spread a long way was not really down to the dynamics of the network, it was because it was funny, it told a story, it created a mental image.
Might it be possible, that in the social media network, all parts of the network have a potential path to all other parts of the network. Note the word potential. The path is activated by relevant information. Now while there may be clusters within this network, these clusters are formed by repetitive flows of similar information. Change the information and the clusters will change. The shape of the network is determined by the information that flows within it, not the other way around. Traditionally we have become accustomed to the idea that distribution shapes information (the medium is the message and all that). But that is what social media is changing – here it is the message that defines the medium.
So the key to influence does not lie in studying the medium – the channels and patterns of distribution within networks – because these are the effect, not the cause. Influence lies in the information itself. We should be using the pretty graphics as the marker to understanding the information, not using the information as the barium biscuit to light-up a pretty network.
Anyway – food for thought.