This is a post I’m plucking from my archive since it’s generated some interest recently and I thought I would see what other feedback I might get if I republished it. It was perhaps a bit premature when I originally posted it nearly 2 years ago, but maybe folks are now ready to leverage their collaboration and social network data for more varied types of analysis. I’d love your feedback :-)
Is active diversity what we should be measuring?
Since I started getting serious about social network analysis (SNA) and specifically since building our new engagement analytics solution (www.ibm.com/engage), which is underpinned by a flexible bigdata platform for implementing a whole slew of network analytics, I’ve been finding new use cases around every corner. The latest use case that has been rattling around my head relates to the measurement of diversity, but more specifically about understanding how diverse companies really are, not just in terms of hiring policy but in terms of how people work together. You might have diversity in your company (geo, skills, gender, demographic, …) but is doesn’t amount to much if it isn’t really “active” diversity; i.e. you are maximizing the value of that diversity through through collaboration between these disparate groups.
For example; imagine that geo diversity is important to your company and you build communities to facilitate global interactions. Even if the community does indeed have global membership that doesn’t mean global participation. It also doesn’t mean that you’ve real diversity in the conversations happening or ideas being shared. Are the interactions bidirectional or one directional? Are there clusters of co-located people engaging and you are not really getting the geo-engagement you were hoping for? Or perhaps there are pockets of true active diversity that you could learn from. Analysis is key since you can’t improve what you can’t measure…
“Innovation happens at the intersection of disciplines” which makes companies that have more activated diversity more likely to innovate; so how do we know if we have the right ingredients to be innovative. And different types of diversity will have a different implication for different companies at different times depending on their business objectives. Perhaps gender diversity is a priority this year, which means you really want to know if your female thought leaders have the optimal network topologies, are they connected with the right people, and getting the right opportunities to grow their influence? Or if you are growing into the emerging markets and knowledge and experience transfer is a key priority, do you have the right levels of active geo diversity across your collaboration networks? Or maybe organization diversity is a key objective as you are looking to generate greater cross-product value?
In our Engagement Analytics solution, which is the first solution built on our Enterprise Graph analytics platform, we applied a series of tunable social network analytics algorithms to the graph in order to generate a set of KPIs (scores) that gives a simple understandable and actionable measure of individual and organizational engagement (Activity, Reaction, Eminence, Network). These scores can be applied to every node in the network – a person (who is generating the greatest levels of reaction?), a community (which communities have the most active eminent users?), a blog (which blogs have the largest active quality networks?), or even a document (which documents generate the most activity, such as reads?).
Looking at the diversity challenge, I believe the identical approach applies, although in this case our algorithm factors in different dimensions and interactions in order to generate the diversity scores; so instead of our Activity, Reaction, Eminence, and Network subscores of engagement, with Diversity we might have Geo, Gender, Skills, and Organization, as an example. And these scores can similarly be attached not just to people but also to other nodes across the network, and of course aggregated up to organizational and corporate levels.
Is active diversity something that you would be interested in measuring for your organization?