The Value of Working Like a Network
About a decade ago (or more – its all so foggy now), Google came out with an ad campaign that stated “It’s all about search.” I remember mocking the slogan at the time as over-simplifying things just a tad. As a collaboration guy, I argued that navigation and taxonomy and governance and a dozen other factors were equally as important. The more I worked with SharePoint customers and understood the impacts of poor planning and the sometimes shortcomings of the technology, the more I realized how important search was to successful collaboration. It really is all about search. Just how important is search to SharePoint? According to my good friend and CTO at SharePoint search ISV BA Insight, Jeff Fried:
Search is now a fundamental capability used by everyone, and a core layer that data-centric applications leverage heavily. Microsoft has leading edge technology AND the largest brain trust in enterprise search on the planet and continues to invest strongly. And search is still an unsolved problem, with lots of innovation ahead.
But this post is not about search. It’s about working “like a network.” But don’t worry – I’ll connect the two in a minute.
I’m trying to remember when I first heard this phrase. I think it was a keynote given by then Director of all-things-SharePoint, Jared Spataro, during which he shared the collection of slides you see in the image above, pushing out what sounded like platitudes. However, the phrase has come to grow on me – and with what we now know about Microsoft’s product roadmap, the availability of the Office Graph and Delve and the underlying machine learning that helps power some of the coolest features within Office 365, AND the growth and adoption of social technologies across just about every business system, it makes total sense. But maybe not in the way you are thinking.
The problem with the way that many people think about network science is that they are overly reliant on search, and technology, in general, to pull everything together and make things work without an understanding of their own role – and what they can do to optimize their results. I think it is universally recognized that small teams are better at collaboration. When you are working with 5 to 7 people, communication and collaboration can be fast and effective. Those same practices and tools may not scale to a larger team, making communication and collaboration sluggish, and often creating silos of data. The idea of a single network, with all nodes connected to all other nodes, is a small-team concept – and simply does not translate to large organizations. And yet that is how we handicap ourselves in enterprise collaboration, assuming that as the network grows, with every node (person, document, artifact) connected to every other node, search will “just work” and social collaboration across this flattened, two-dimensional organizational concept will somehow make people more….well, collaborative.
According to Ron Burt, the Hobart W. Williams Professor of Sociology and Strategy at the University of Chicago Booth School of Business, your network is actually a set of clusters – not one giant network. Burt talks about clustering being one of the basic patterns within network science, and how we all naturally participate in cluster. Some clusters come from our roles and professional circles – communities of practice, like being a business analyst or a project manager, for example. Other clusters form around age, musical tastes, educational backgrounds, sports, and so forth. Information is created and travels around within the cluster, but much of that data never leaves the cluster.
But there are some individuals within each cluster who act as brokers between clusters. These are people who see value in sharing information outside of a cluster, and who bring new ideas into the cluster, or group, from other groups. There’s a great article by Forbes contributor Michael Simmons (Why Being the Most Connected is a Vanity Metric) in which he interviews Ron Burt, and provides some additional insights into how networks work. For example, about brokering, he wrote:
A key insight from network science is the power of brokering, the act of moving information from one group to another. Burt explains, “What a broker does is make a sticky information market more fluid. Great ideas will never move if we wait for them to be spoken in the same language.”
Network brokers (i.e. – connectors) have three advantages:
Breadth. They pull their information from diverse clusters.
Timing. While they may not be the first to hear information, they are first to introduce information to another cluster.
Translation. They develop skills in translating one group’s knowledge into another’s insight.
Combined these three advantages give an individual an overall vision advantage to see, create, and take advantage of opportunities.
Well worth a read if you’re interested in learning more about network science and some of the advances made in recent years, but my point here – and back to the connection between working like a network and search – is that simply connecting yourself and your data into a machine learning-based technology platform is not enough to ensure collaboration and search are efficient and effective. Much can be automated, for sure, but the mistake is thinking that some of the old ideas about taxonomy development, site structure and navigation, and governance (which includes community management) are still as critical as they have always been. To work like a network means that each of us acts like a broker, adding value to the clusters in which we participate – and then connecting data and people and ideas across clusters, translating each body of knowledge for those other networks. In the end, search will be improved because our social interactions will make connections and attach context beyond what machine-learning can derive from our content and connections.
I am fascinate3d by the topic. I don’t feel like I’m doing it justice here in this stream-of-consciousness post. But it’s what is on my mind tonight. Hope someone finds some value here.