What a team is
The team that is really three teams
A box on a chart often contains two or three real groups with little between them. Counting the pairs makes the fault lines visible without anybody having to say so.
The people this page is about, and which of them have ever worked together
- 01The team's managerFive of this person's six pairs are with people not in this grid
- 02The senior engineerThe one working pair on the whole team that does not involve the manager
- 03The second engineer
- 04The engineer who joined in MarchNo pair with anybody
four people · 6 possible pairs · 2 of them have worked on something together · 4 have never done either
Four of the eight, and only the pairs that run between those four.
A team testing collaboration workload reporting should agree on purpose, access, and interpretation first, then use the resulting data to improve work rather than police individual moments.
An independent source worth comparing with this example is Interaction Design Foundation.
Five of the eight, and the pairs that run between those five
- 01The team's managerThree of this person's five working pairs
- 02The senior engineerThe one pair here that does not involve the manager
- 03The second engineerHas worked with one person and met two
- 04The designerShared with two other teams
- 05The QA engineer
five people · 10 possible pairs · 4 of them have worked on something together · 2 have only been in the same meeting · 4 have never done either
A sub-group of the team counted on this site, drawn on its own so the clusters inside it are visible.
Take the eight people in the box and draw only the pairs that have done work together. Quite often what appears is not one connected shape but two or three small ones with nothing between them: three people who build a thing, two who run an adjacent thing, and a few singletons attached to the manager and to no one else.
Those sub-groups are teams in the sense this site uses. The box is not. It is a container holding several teams and some individuals, and a great deal of confusion follows from treating it as one unit.
How the fault lines appear
They are almost always inherited from the work. Two people share a codebase, so they review each other constantly and a strong pair forms in a month without anybody arranging it. Two others share a customer and talk daily. The designer is shared across three teams and belongs properly to none of them. The data analyst reports here but spends most of their time answering questions from a different part of the organisation.
Nobody decided this. It is the shadow cast by whatever the team was asked to do, and it changes when the work changes, usually about a quarter later than the work does.
Why it is worth naming
The confusion costs real things.
Goals get set for the box. A single quarterly objective is given to eight people, three of whom can affect it, three of whom cannot, and two of whom are not sure which group they are in. The three who can affect it do the work. The others attend the reviews.
Rituals get sized for the box. A standup for eight people where only three pairs have anything to say to each other is eight people's time spent so that three people can coordinate. It runs for months because cancelling it would look like giving up on the team.
Activities get bought for the box. The offsite brings all eight together to build the kind of bond that already exists inside the clusters and does not survive outside them.
And difficulties get read as personality. When two sub-groups have no pair between them, the absence gets explained in terms of character — they do not get on, they are protective of their area, there is history. Usually there is no history. There is simply no task that has ever needed both of them, and nothing else creates a pair.
What to do once you can see it
Three responses, in increasing order of effort.
- Accept it and adjust the apparatus. If the box genuinely contains three groups, stop setting one goal, stop running one standup and stop buying one activity. Let each cluster have its own rhythm. The manager remains the connection between them, which is an honest description of the situation rather than a failure.
- Bridge it deliberately. Pick the two clusters whose separation costs the most — usually the two that keep discovering each other's decisions late — and create one small piece of work that needs one person from each. Not a liaison role, not a standing sync: one task, finite, with a date. One bridge pair is worth more than any amount of shared ceremony, and it is the only kind of thing that shows up on the grid.
- Redraw the box. Sometimes the right answer is that the chart is wrong and should change to match the work. This is the most effective and the least available: it needs somebody with the authority to redraw, and boxes are expensive to redraw because they carry budgets and titles.
The sub-group that is not on the chart at all
One more thing the overlay turns up, and it is worth looking for: a pair or trio that spans the boundary of the box.
Almost every team has at least one person whose strongest working relationship is with somebody in another reporting line entirely — the designer and the engineer on the product they are embedded in, the analyst and whoever consumes their numbers. That pair is invisible to every instrument the organisation has, because the chart draws it nowhere and the team's own grid stops at the roster.
It is usually the most productive relationship either person has, and it is the first thing damaged by a reorganisation that nobody thought to check against it.
The honest version of the question
Before trying to connect the clusters, it is worth asking whether they need connecting. Two groups that share a manager, a budget line and nothing else may be perfectly well arranged. The cost of separation is only paid where the work actually crosses: where one group's decisions land on the other, where the same problem is solved twice, where a handover happens.
Where it does not cross, leave it alone. A team is not better for being more connected in the abstract. It is better for having a pair at each point where the work needs one, and worse for having ceremonies at the points where it does not.