Your developer community is probably pretty diverse – comprising of unique members working on little side projects, contributing to open source, poking at APIs, and integrating new tools. They're interacting with various topics in many channels – and somewhere along the way they might engage with you. Whether or not that relationship sticks and what it means depend on a lot of things.

Regardless of what they're thinking (if they think about you at all) - understanding the members of your community in terms of how they are engaging in your ecosystem is vital to working with them towards meaningful outcomes. But – communities are big, and it’s tough to do this at-scale.

If you meet a developer one-on-one, you can quickly gauge their relative status, activities and adapt your communication accordingly. Maybe you can pitch in on the conversation they're having, or maybe you recognise them from your forum or another event. But to network and build relationships in the digital realm alone is a whole different ball game. Conducting individual outreach via Twitter, Stackoverflow, your forum, or other channel is the default approach.

Now, I don't blame you if you don't get excited about manually scanning different platforms to track down people engaging with, or about you or your products — our job is to connect with people, empower them in our community and with our products, and build relationships. Not to be a human version Google Alerts. Besides, DevRel is a strategic job and we should avoid only engaging with our audience when they initiate.

That raises a problem, though. It's likely that your community consists of hundreds, if not thousands of members. Sure, some of them might be lurkers, or people who poked at your product and decided to move on. But if some of them had stuck around long enough to see the feature they were missing being deployed, or gotten the right engagement at the right time, they might've been active members by now. So how do we know what to do and who to focus on?

Segmentation

By segmenting your community into smaller groups, such as Promising or Adrift, you can create different strategies for each segment as a starting point for your next actions. Whether you want to craft more relevant content, build coding programs, or a communication strategy that's geared to engage a particular audience.

For this we're using the Community Grid. It's a frequency-recency model that separates your community in 8 main segments, supplemented with the Gone segment and a Not Seen bucket. The idea is simple; the most engaged people in your community are on the top, the least engaged people are on the bottom. If you come across people that would be great additions to your community, but haven't engaged with you yet, they will be in the Not Seen segment.

So what about recency? With recency we mean the number of days since last engagement. This is reflected on the Grid on the x-axis — the most recent members are on the left. It means that if your members don't engage with you, they will slowly drift away to the right. We all understand the importance of recency as being active a year ago isn't the same as being active today. But instead of just separating people into active and non-active buckets, it's more interesting to understand who is slipping away. Recency splits the segments into different shapes, cutting them off where characteristics start to change.

It is simple yet powerful — it shows you where groups of your members fall in terms of their relationship to your organization, so you can direct your efforts based on insights. It's also just the first step towards a deeper community strategy, but gives you and your team both a framework to hold on to, and a goal to work on (increase retention, grow the rights part of your community, and engage in the right way).

Getting Started

So that's all nice and theoretical - but how do you use it? The challenge is getting the right data to calculate engagement. Ideally you're capturing all touch points with your community, even when you're not in the room (e.g. a conversation on a subreddit, an answer on StackOverflow mentioning you).

Starting with your organization's Github repos and your Stackoverflow tags, we created a tool that does this for you. You can head over to comgrid.com, enter your details and after some number crunching we'll send you your personalized Community Grid. It'll show you how many unique members you have, your retention rate, the top members in each segment, and some more in-depth information about each segment.

The Community Grid is a great first step to get some insights in your community. If you're interested, you can transition your Grid into a full-version of Port (free for the first users) to get insights in the members behind the interactions. The model gets even better with more data sources and the interaction weighting model we use at Port.

We're math nerds, so we're working on more data models to give you insights. The Community Health Index to indicate how things change; our Port Score to qualify community members based on their engagement, relevant activity, and reach; and more.

Ultimately, it's best for you to see it with your own eyes. Try the Community Grid and let us know if the model works for you, or what other platforms you'd like to see.