Sep 9, 2016

Watson Looks at Organizational Networks

We have been mapping and measuring people networks for more than 20 years, mostly using our own software.  Last year we were chosen by IBM as a "social influencer" in the cognitive computing space and we started to explore uses of Watson Analytics. We were curious -- how will Watson reveal the hidden patterns in organizational data?

Once we got a feel for Watson, we loaded up some of our client network data sets to see what Watson would do with them.  Fortunately, Watson liked our plain text CSV(Comma Separated Values) files showing links/flows between various nodes.  We were pleased to see the results and to see how easy Watson showed us the network structures in our data. Yet, Watson is currently not a substitute for professional social network analysis software.  Not today, but some day!

Various other IBM social influencers were also busy "kicking the tires" of Watson.  Soon there was enough group experience to compile a short e-book:  23 Reasons to Get Excited About Data.

Figure 1

Below are two excerpts of Orgnet's portion of the e-book.  We visualize a client's organization within Watson.  Figure 2 shows the hierarchy -- who reports to whom.  The larger nodes have more direct reports.  Watson's default network visualization changes node size by the common network metric called Degree Centrality.

We see an interesting pattern! First we see that a hierarchy is a network!  Mathematicians call that type of network a "tree."  We see what is commonly called a hub-and-spoke network is the same as a hierarchy(tree).  A hierarchy is a hub-and-spoke network and pure hub-and-spoke network is nothing more than a hierarchy.  Many organizational consultants today make the argument of Hierarchy versus Network -- you have to pick one.  But, hierarchies are networks with specific properties!  Hierarchies and networks are on a continuum -- they are not separate species! They don't collide, they meld.

Figure 2

Next we look at the emergent organization, or what is sometimes called a Wirearchy.  The emergent organization in Figure 3 shows us how work actually gets done, how information and knowledge is shared, and how learning and adaptation happen in organizations.  Figures 2 and 3 show the same group of employees but with a different set of relationships and flows -- prescribed and emergent.

Want to know how well your organization is performing?  Don't look at the Hierarchy -- any competitor can duplicate your formal structure.  Look at the patterns of emergent connections -- that is your competitive advantage!  Smarter organizations have better emergent connections.

Figure 3

In Figure 3 Watson again changes node size in the visualization based on the number of connections each node has.  That is interesting, but not the most useful metric in organizational network analysis -- quantity ≠ quality, and more ≠ better. Here is where professional organizational network analysis software shines -- it goes beyond simple metrics to map and measure more complex organizational patterns that we often see in high performing organizations.  It is the non-obvious patterns in an organization that often separate the good and the great!

The  23 Reasons to Get Excited About Data e-book has other great examples by some very smart people including: Jessie Liu, Deborah Berebichez, Juntae Delane, John Cook, Anil Batra, Christopher Penn, Bob Hayes, Randy Krum, William McKnight, Jeremy Pincus, and Emilio Ferrara.  Download the book and get familiar with Watson and what you can do with it!

No comments:

Post a Comment