I haven’t thought this all the way through, and there are a lot of people (Jon Husband, Harold Jarche, and a guy I had beer with on Monday who is the founder of Simplexity Systems and the inspiration for this post, for example) who are way more knowledgeable about Social Network Analysis (SNA) than I am. I’m really just riffing a bit here.
Monday evening I had beer with a couple of people, one of whom I’d just met (thanks for buying, by the way). His official title is something like Manager of Enterprise Architecture, but his real mandate is to shake things up and make some changes for the good of the organization (a client of mine, BTW). Anyways, after a bit of chit chat, and my two companions finishing up with what they were talking about before I got there, the conversation turned to the topic of Social Network Analysis. What the heck is SNA? Well, my very simple understanding of it is something like …
Analytics and algorithms are used to mathematically prove the strength of relationships between nodes (people) in a network. For example; by examining aspects of an email chain between multiple people it is possible to map the relationships between the various participants and to see how strong those relationships are. One thing that’s really cool about the whole SNA thing is that it not only measures the numbers of emails flying about and their sources and destinations, it also measures and evaluates elapsed time. What’s missing (or we just didn’t talk about it) is the sentiment of the relationship, since the analysis is focused on emails going back and forth and not the content and tone of the emails.
Before I forget … you ought to check out Wirearchy for some more in-depth stuff about SNA and how it can be applied …
Anyway, after chatting for a couple of hours, and the conversation being cut short (babysitters, feeding kids, family nonsense) I went into head scratching mode for a bit. I started thinking about other types of connections that could be mapped, using SNA principles. Could we map relationships between people and content, and then make inferences about those relationships? Could we make suggestions about potential relationships? For example, could we make inferences and suggestions about a relationship between two people (content author and content consumer) based on the consumer’s relationship (activity) with the author’s content, even though the people may not know each other? To what end would we apply these insights?
I also started thinking about what would happen if we added content and semantic analysis to the mix. Could we draw conclusions about the tone of the relationships? Could we figure out if a relationship between individuals was positive or negative? What else could we infer about the relationship?
What if, instead of looking at relationships between individuals, we aggregated the findings to look at relationships between departments in an organization? Could we identify relationships and dependencies where we previously assumed none existed? If we could, could we also then use this information to restructure certain elements and systems in the organization? In effect, could we use the combined results of Social Network Analysis, Content Analytics, and Semantic Analysis to tear down silos and improve information flows, thereby positively impacting the organization? My gut says we can.
As I said, I know very little about SNA though I am convinced that if it were applied in concert with other analytic approaches there’s a lot of good stuff we could do. For the moment I’d really like to spend more time with my new drinking buddy, some wine or beer, and a whiteboard to learn more about this whole Social Network Analysis thing.
The stuff of HR nightmares! The privacy paranoia police will be on your case 😉
I think you are onto something though… Qualitative, rather than quantitative, modelling of relationships is something I’ve been thinking for a while (there’s a couple of flipcharts full of squiggles somewhere in my study… ), but I got stuck trying to identify a way to articulate the business value and monetising the idea.
More beer definitely required! 🙂
Ya know, I was thinking about HR’s response to the potential privacy issues. Specifically, I thought “screw ’em”.
LikeLiked by 1 person
Yep. Screw ’em. There should be little to no expectation of privacy when using corporate email, other than between employees and HR.
Regarding tone, I recently heard (on CBC Radio I think) about a project to specifically identify sarcasm, so I don’t see why other tones (respect, disrespect, impatience, etc) could not also be identified.
Regarding silos, analytics could be very beneficial in determining if two different silos are talking about the same thing, but not with each other. Not that that ever happens in any business with more than five people.
Email analysis could also be used to improve the speed of communications. For instance some companies are implementing rules that limit the number of email messages on any given subject. Tracking would make sure the rules are followed. Word counts could also be used to analyze communications issues – i.e. chronic verbal diarrhea.
LikeLiked by 1 person