We’ve spent the last 6 years building Kumu into an incredibly flexible and powerful network and system analysis platform. Over the years, many people have pushed for integrations with other platforms. We’ve held off on these integrations and have instead focused our efforts on improving Kumu’s core platform.
That’s now changing with the introduction of Compass.
Compass combines Kumu’s technology and deep experience in social network analysis with data on your team’s communication patterns from Slack. Slack has taken the world by storm, quickly overtaking previous favorites like Atlassian’s HipChat and moving millions of users from an email-centric world to a chat-centric world.
With many teams now using Slack for the majority of their communications, Slack becomes a goldmine for harvesting insights on how your team collaborates. But this information is almost impossible to glean from just watching activity across different channels (and with the sheer volume of activity taking place in Slack, most people are struggling just to keep up).
Visualizing communication patterns
Compass allows you to step back and see the pattern of who’s communicating with each other and how frequently. Rather than being limited to a linear conversation history, we’re able to map the web of conversations taking place across your Slack team.
We can then use social network analysis metrics to deliver insights about who plays a key role based on the structure of that network. These metrics include powerful calculations like betweenness centrality, which highlights key influencers and potential information bottlenecks in a network. Betweenness centrality is calculated by identifying all the shortest paths between each pair of people in the network and then counting the number of shortest paths each person lies on.
By running betweenness and a handful of other metrics, Compass can help identify where changes may improve collaboration on your team or highlight people who are playing a key role that was previously hidden. If you’re new to social network analysis and would like to learn more about each of the metrics, make sure to check out Social Network Analysis Made Easy on Speaker Deck.
To give you an idea of the type of insights we’re planning to provide, here’s an early look at what we’re considering for our “leaderboards":
- Influencers. People who are in an influential position in the network. These are the people to invest in while keeping a watchful eye to make sure they don’t become bottlenecks. Indicated by high betweenness centrality.
- Extroverts. People who are the most connected or who are centrally located to other connectors. These may be potential influencers, but they are just as likely to be the social butterflies of your team spending a bit too much time chatting on Slack. Either way, they have the ability to spread information quickly and are an important part of any change effort. Indicated by high degree and high closeness centrality.
- Bridges. Bridges are a particular breed of influencers. They may not be in a central position within the network but they do play an essential role connecting two groups who would otherwise not be connected. Pay attention to these folks because if you lose them, collaboration may cease to exist across certain teams. Indicated by high betweenness and low closeness centrality.
- Hermits. Those connected to the fewest number of people. Sometimes hermits will be new hires. Other times they’ll be people in positions you don’t expect to be on Slack very often. Other times it will be people intentionally on the periphery. What is important is that you know who these hermits are and make sure you don’t have people who are stuck on the periphery trying unsuccessfully to get more connected. Indicated by low degree and low closeness.
- Chatterboxes. Sometimes it’s helpful to know who sends the most messages on Slack. Chatterboxes is that list. If you’re incredibly productive and you top this list, good on you. Not so productive and still top this list? Might be time to reduce the number of gifs you post. Indicated by number of messages sent.
- Spammers. There are great reasons to use Slack’s @channel/here/team functionality to notify people in bulk. And there are not so great reasons. We’ll let your team be the judge of whether the people topping this list are actually spammers or stand falsely accused. Indicated by number of times @channel/here/team is used.
You’ll be able to track the top users for each of these categories across your Slack team or for any individual channel. Even better, you’ll be able to see how these change over time (last week, last month, last year).
Have feedback about the leaderboards? Add a comment!
Direct messages and private channels
A non-trivial amount of communication on Slack happens in direct messages between users and private channels. These messages are intentionally private. If we leave out these messages, you won’t get an accurate picture of your team’s communication patterns. So how do you include these messages while respecting team privacy?
Our approach is two-fold:
- Opt-in. It’s your choice whether to activate direct messages and private channels for your account. If you do choose to activate, each member of your team has the ability to grant or deny access (via the Slackbot) to their own private messages.
- Who, not what. If activated, we only use these messages to identify who’s involved in the conversation. We never read or store the underlying message content for direct messages and private channels.
We feel this approach minimizes privacy concerns while maximizing potential insights.
The communication data Slack provides is powerful on its own, but it becomes even more powerful when you layer in additional data about each person. Consider a few of the possibilities:
- Team/Department. Use the Slackbot to ask “Which team(s) do you work on?" and then identify communication patterns across teams. Are Marketing and Sales communicating with each other? If so, who’s the primary go-between? Who’s out of the loop?
- Projects. Use the Slackbot to ask “Which projects are you working on?" to build a visual map of the project ecosystem and identify key coordinating points across projects.
- Org chart. By asking a simple “Who do you report to?" we can build the org chart on-the-fly and then compare that to the actual communication patterns for your organization (hint: org charts lie).
- Locations. Create a geographic map of member locations to identify whether your team communication is regionally bound.
The Slackbot becomes an easy way to not only collect this information but also keep it updated over time. Create your questions, set the desired update interval, and Compass will take care of the rest.
Sentiment analysis and language processing
Until this point we’ve focused on the who. But the what of message contents (for public channels) can provide valuable insights as well. We can track patterns in keywords over time, helping to flag significant changes that hint at potential problems (before they blow up in your face).
We can also use sentiment analysis to identify the balance of positive vs. negative sentiment used in specific channels or by individual people. This can be another helpful warning sign (or an indication that it’s been too long since your last team retreat).
We want your feedback!
That’s the essence of what we’re building with Compass. We’d love your feedback on which insights would most compelling for your team.
Please add a comment or send an email to firstname.lastname@example.org answering the following questions:
- How many members are on your Slack team?
- Are you interested in visualizing DMs and private channels too?
- What are the biggest challenges your team is facing with Slack?
- What’s the most valuable insight you hope to get from using Compass?
We’re currently finalizing the architecture for Compass (see Single vs. multi-tenant SAAS if you’d like to learn more) and will be ready with a beta soon. If you’re interested in early access, make sure to sign up for the waiting list.