Authors: Dr. Morton Tavel, Mike Jensen, Gary Markovits, Devin Markovits & Blake Markovits
Facebook and Twitter are popular and successful examples of social media. Successes made possible by their underlying network structure. Networks enable connections and interactions that only a generation ago were impossible. Today, network insights are being used to understand complex systems such as disease epidemics, the electric grid and consumer adoption behaviors.
Why use networks to study innovation? No doubt innovation has a high degree of social content. Many innovations are created by the social interaction of groups and similarly, innovations are adopted by consumers in a socially based manner. So, it is reasonable that a network model should provide powerful insights.
At its core, a network is a simple concept; a pattern of connections between individuals, groups and organizations. Yet the insights gained from analyzing these patterns are remarkably powerful.
Network insights are now being used to study how innovations arise from organizational behavior. Innovation is the most basic systemic, value-adding activity of a modern corporation; it is an activity that every corporate CEO and CTO is eager to emphasize and improve.
Our network study of innovation has begun generating very interesting results. Using different network models, we have found important indicators of innovation. For instance, one model uses patents as a measure of innovation, to form what we call an innovation network. In this model of innovation, we construct a network of inventors, linked by the inventions they have created. The goal has been to untangle the social and technical aspects of inventing. This is done by forming networks of patented inventions across a wide range of industries and analyzing the roles of their inventors. The network structure reflects the social skills of particular co-inventors who are able to bring groups together for creative interactions. It also reflects the diverse technical skills required to bring an invention to fruition. We can then dig more deeply into the network structure produced through the invention process. This probe highlights methods for improving the innovation process.
Figure 1: A Basic Model of an Innovation Network
The set of differently colored geometric arrangements of dots and lines represent an innovation network of a company that is relatively segmented into different groups. Each dot is a “node” that represents an inventor. Each line between nodes is an invention on which both nodes collaborated or, possibly, some other form of interaction between them. Inventors A and B are closely connected and can easily transfer knowledge to each other. Inventor C is in a different group of the organization. Fortunately there is a connection between the green and yellow groups so A and B can share their knowledge with C, but they have to go through several other people to do this, decreasing the quality and quantity of knowledge flow. Inventor D is at the greatest disadvantage. D is in a group that is completely isolated from the rest of the organization. There is little chance for D to exchange ideas with the other inventors, limiting their value.
The individuals in an organization possess the knowledge that is key to innovation. The network structure enables the opportunities to share that knowledge. Arising challenges provide the motivation to share, and trust determines the willingness to share. In our interviews of many inventors it was repeatedly said that trust is a major factor in determining the links that will form between them and, therefore, trust determines the organization’s capacity for innovation.
The figure reproduced below is an actual and much more complex innovation network of a larger company in which all their 562 inventors are represented by blue dots and the gray lines linking them represent their collaborations through their inventions, jointly patented over a five year period. The resulting network structure is complicated, containing over five hundred inventors and more than one thousand patents, but even a cursory visual examination discloses certain facts about the innovation portfolio of this exemplary organization and its capacity for innovation.
For example, note the number of small disconnected groups of co-inventors that are not linked to the large, densely linked central core. This central core makes up the bulk of the set of inventions and is undoubtedly the storehouse of innovation within the organization.
Even within the core, certain structures can be discerned. Although it is connected by many links, it has a modular topology of smaller clusters. This kind of modularity can be related to the present and future strength of the patent portfolio and future prospects for innovation. Importantly, a few individual inventors appear responsible for holding different segments together.
Figure 2: Actual Innovation Network
When studying companies such as the one above, with broad-based patent portfolios that encompass multiple technologies, we use network-based metrics to help us identify individual inventors who have co-invented across several of these technologies. These inventors appear to provide coherence and structure to the overall network.
One of the network metrics that has proven to be particularly revealing is called “betweenness”. It is essentially a measure of how often an inventor falls on the shortest path between any other two inventors in the network. We interpret betweenness as a measure of how strongly that inventor is either involved in the generation or flow of new knowledge. When we locate those inventors with the highest values of betweenness , we find that they determine a special sub-structure, which we have called the “Innovation BackboneTM” of the company.
Below, we have plotted the betweenness for the inventors in the organizational network shown above on a scale between 0 and 100.
Betweenness drops rapidly in terms of the numbers of inventors that have a given value. We have seen this phenomenon in every corporate, academic and government innovation network we have examined. In fact it has led us to conclude an 80:20 rule for innovation, where 20 percent of the inventors possess 80 percent of the betweenness in a network. In the figure below, we have highlighted, in red, the thirty inventors with the highest betweenness along with their nearest neighbors. This structure is what we have called the Innovation BackboneTM.