At Automattic, we believe that communication is oxygen, and we strive to make our communication as effective as possible. In my last post, “How Communication Density Fuels Automattic,” I introduced Automattic’s communication tools and explained the positive influence our annual company meeting — the Grand Meetup — has on our communication.
As a continuation, this post highlights key results of another study I did around our internal communications, which focused on the flow and consumption of information via our main and most transparent communication device, P2s. As described in my last post about our internal communications, P2s are the backbone of our interactions:
Our main communication platform is, of course, run on WordPress.com. We employ a network of blogs called P2s, that run the P2 theme. P2s allow us to post right from the blog’s homepage to have real-time, threaded discussions under the posts. We can cross-post to other teams’ P2s, mention other Automatticians by name to notify them of a discussion, like posts, and follow discussions. Everything is archived and searchable.
Flow of information at a remote company
We are all connected, and working in a distributed environment means that we have enough empirical data to trace these connections; we can answer questions such as, How do we interact with each other? How many cliques do we form? Who are our bridges and influencers? More on cliques, bridges, and influencers coming up.
To answer these questions, I created the interaction network of all Automatticians on all work P2s from the previous year. In the network, all nodes denote an Automattician, which is our internal term for an employee of Automattic. If an Automattician likes or comments on a post authored by one of their colleagues, then there is a directed edge going from them to the author of the post.
Clustering the network
To understand how many cliques we form, I employed a graph clustering algorithm called Louvain modularity. This approach helps us identify cliques (modularity groups) in the social network by looking at which groups of nodes are more connected to each other than they are to the rest of the network. Multiple Automatticians belong in the same clique if they interact way more with each other than they do with others. This is depicted in the visualization as the color of the Automattician’s node.
Influence and centrality
The size of an Automattician’s node in the network denotes its PageRank, which is a measure of influence in a network.
PageRank is a variant of Eigenvector centrality. Eigenvector centrality shows the expected amount of times the node will be visited over a random walk in the network. The PageRank algorithm introduces a randomness that makes sure that even less-connected nodes will be considered. In more subjective terms, PageRank says that nodes with higher quality connections are more important than nodes with lower quality connections. Developed by Larry Page and Sergey Brin to rank search results on Google, this is one of the most widely used methods for identifying influential nodes.
The network of information flow at Automattic
This network is a highly filtered version of the original that depicts only the strongest connections and most active people. (At Automattic, we are so connected that it would have been impossible to interpret the full network in any meaningful way without rigorous filtering. It would look like a massive hair ball. All the metrics are calculated for the non-filtered network, only the visualization is filtered.)
I have also removed the names of my colleagues from the network keeping only our CEO Matt Mullenweg᾿s name in the visualization.
We have seven larger cliques in our team and project-related discussions. Looking into who is connected to whom, we can see that we are more of a project-based company than a team-based one. While our team connections are strong, they are not necessarily stronger than our intra-team interactions. This was concluded via manual inspection where we note people from different teams being on the same cluster.
This network clearly shows us that we are a centralized company, centered around the CEO — besides PageRank, Matt ranked the highest in interactions (more incoming rather than outgoing), betweenness centrality, meaning that he is our top bridge/connector — most information passes through him — and in number of overall connections.
However, we also see a few influential nodes in each cluster, and strongly connected clusters with their own “leaders” show that while we are centered around Matt, each cluster also sets its own influencer for team and project-based discussions.
In an increasingly turbulent business environment where companies need to act fast, a purely decentralized innovation model won’t suffice. What is needed is an exact combination of knowledge and power among the rank and file, and the right level of power stems from corporate leaders. This doesn’t make the case for micro-management or corporate interference. But there are times when strategic intervention from the very top can prove critical. – “Google vs Amazon: How a Strong CEO Boosts Innovation” (International Institute for Management Development)
At Automattic, we work in a very autonomous manner. It is at first a little surprising to newcomers, however, as this and similar internal analyses have shown, we all strive to be connected and make our work discussions visible and transparent to the rest of the company via P2 blogs for teams and projects.
A network analysis like this also gives us opportunities to figure out how to best inform everyone of big changes in a fast way, whether our communication structure is robust enough to account for people taking longer leaves, or what the characteristics of above-average communicators at our company are. These are all things we have looked into, stay tuned for subsequent posts on our HR analyses!