Team Connectivity draws the network of who interacts with whom, and labels each connection low, medium or high. It's there to surface isolated people and overloaded hubs.
Source: The New Science of Building Great Teams, Harvard Business Review.
What it measures
Interactions between pairs of people, not totals per person. Every line in the network is one direction of one pair, and the label belongs to that line.
Leanmote counts the number of interactions in the period and divides by the number of weeks, so the underlying value is interactions per week for that pair.
One source at a time
The chart shows a single source, which you choose. It does not add them together:
Source | What it reads |
Communication tool | messages between people |
Git | pull request comments and review activity |
Meetings | calendar events people shared |
Tasks | comments on tasks |
Two gaps worth knowing about. Linear and Notion do not write task interactions, so an organization working only in those tools sees an empty Tasks network — that's missing input, not a quiet team. And Leanmote only reads metadata: who interacted with whom and when, never the content. See the privacy policy.
How the labels are assigned
This is the part most people get wrong, and it changes how the chart should be read.
Leanmote sorts every pair by its weekly average and cuts the list into thirds:
Position in the sorted list | Label |
Bottom third (up to the 33rd percentile) | Low |
Middle third | Medium |
Top third (above the 67th percentile) | High |
The labels are relative, not absolute. They describe a pair's rank inside the current selection, not a level of engagement. Three consequences follow, and they are not intuitive:
There will always be "low" connections. Roughly a third of the pairs get that label no matter how evenly the team collaborates. A perfectly uniform team still shows a bottom third.
"High" does not mean a lot. In a quiet period, the top third is still labelled high.
The labels move when the filter moves. Change the team, the date range or the source and the same pair can change label without anybody's behaviour changing, because it is being ranked against a different population.
Read the network for its shape — who is connected to whom, who sits at the edge — and read the interaction counts for magnitude. Don't read the colour as a measurement.
Why this matters
Isolated people. Someone at the edge of the network may be blocked, onboarding, or working on something nobody else touches. The chart tells you to go ask; it doesn't tell you which.
Hubs. A person connected to everyone is a single point of failure and a burnout risk, whatever the label says.
Missing edges. Two sub-teams that never interact are usually a structural fact about ownership, not a people problem.
What to do about it
For people at the edge: pair them up, bring them into review rotations, or check whether their work is simply invisible to the source you selected.
For hubs: redistribute the dependencies rather than asking them to do less.
Before acting, switch sources. Somebody invisible in Git may be central in Tasks.
Related metrics
Review participation
Activity distribution
Tasks Collaboration