What to do instead
The whole argument in one page
For somebody who arrived here first: the claim, the measure, the three things it says about one team, and the shortest useful thing to do about it.
The people this page is about, and which of them have ever worked together
- 01The senior engineerThe pair that came out of a shared on-call week
- 02The second engineer
- 03The designer
- 04The data analyst
- 05The product manager
five people · 10 possible pairs · 1 of them has worked on something together · 3 have only been in the same meeting · 6 have never done either
Five of the eight, with the manager left out so that what remains is what the team has without them.
A fair pilot starts with a written purpose and shared baseline; managers can learn more on this page before comparing what changes after a new working practice.
For a separate perspective on teams and working practice, consult IMD.
The grid this entire site is built on, drawn once more at the end
- 01The team's managerFive of the six
- 02The senior engineerThe sixth
- 03The second engineer
- 04The engineer who joined in MarchStill no pair with anybody
- 05The designer
- 06The data analyst
- 07The product manager
- 08The QA engineer
eight people · 28 possible pairs · 6 of them have worked on something together · 4 have only been in the same meeting · 18 have never done either
Eight people, one quarter, counted from the work itself. Every figure quoted anywhere on this site comes from this grid or from the two panels beside it on the home page.
If you have landed on this page before any of the others, this is the whole of it, and the rest is elaboration.
The claim
A team is the set of people between whom work actually runs. A box on an organisation chart is a reporting arrangement, and whether it is also a team is a separate question with a separate answer.
The answer can be checked. Take the people in the box, consider every possible pair, and ask which pairs have done a piece of work together that needed both of them. On the team drawn above, six of twenty-eight pairs met that bar in a quarter, five of the six involved the manager, and one person had no pair with anybody.
Why this measure rather than another
Because it records events rather than opinions, so two people who disagree about everything else about the team still agree about the number.
Because nothing else about a distributed team can be counted without asking people how they feel, improved by saying it has improved, or settled without an argument.
And because every reading of it names two specific people and suggests a specific thing to do, which a score out of ten does not.
It is a narrow instrument. It says nothing about whether the work was good, whether anybody enjoyed it, or whether the pair should have needed two people at all. It goes stale in a quarter. It depends entirely on where you set the bar, so two teams' numbers cannot be compared. None of that stops it being the only honest picture of the shape.
What it usually shows
Three things, on most teams, most of the time.
- The network is sparser than everybody thought. The first count is always a surprise and the surprise is always in the same direction.
- Most of it runs through one person. Not because that person is hoarding, but because work arrives centrally and is distributed centrally, and every individual act of distribution creates a pair with the distributor.
- At least one person has nothing. They are usually competent, self- sufficient and causing no concern, which is why nobody has noticed.
What does not change it
The quiz night. The offsite, as usually arranged. The random coffee pairing, which decays by about week six. The icebreaker. The weekly all-hands. Values, charters and handbooks that nobody edits. Anything where everybody attends and nothing passes between any two of them.
None of those are worthless and none of them are what they are sold as. An event is not a pair, and a team that has had twelve of them in a year has had twelve pleasant occasions and the same grid it started with.
What does
Work between two people, with a date, that somebody is waiting for. Everything that works is a version of that.
The cheapest versions are already on your calendar doing something else: a review rota where who reviews whom is decided rather than chosen, a shared duty with a real handover, a weekly note whose author rotates so that each person in turn has to go and ask the other seven what they are doing.
The sharpest version is two named people given one deliverable and one deadline, four times a quarter, aimed at pairs the grid says do not exist.
What this site is not saying
It is not saying that social events are a waste. They are pleasant, they signal that somebody spent attention on the team, and they make the first real request between two people slightly cheaper. The claim is only that they do not do the thing they are usually bought to do.
It is not saying that a sparse team is a failing team. Some groups are correctly a list of specialists with a common manager, and the only thing wrong is the word.
And it is not saying that counting pairs tells you how a team is doing. It tells you what shape it is. Whether people are happy, whether the work is any good and whether anybody wants to stay are different questions, and the honest way to find out is to ask.
The shortest useful thing to do
Write down everybody on your team. Write down every pair. Mark the ones that have done something together in the last three months. Write the date next to the answer.
Then pick one pair that does not exist and give those two people one small piece of work that needs both of them, this month.
Count again in a quarter. If the number moved, you will know what moved it, which is the whole reason for writing down the first one.