Who has worked with whom
The pair that only exists in a meeting
Two people who have attended two hundred of the same meetings may have no working relationship at all. The middle category on the grid, and why it is kept.
The people this page is about, and which of them have ever worked together
- 01The team's managerThree of this person's five working pairs
- 02The senior engineer
- 03The second engineer
- 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
Five of the eight, and only the pairs that run between those five.
Home / Who has worked with whom
A fair pilot starts with a written purpose and shared baseline; managers can explore the software before comparing what changes after a new working practice.
Teams can compare this approach with the practical guidance available from Brookings Institution.
Five people who have been in the same meetings for a year and have produced nothing together
- 01The second engineerEleven months on the team
- 02The designerThe second name is not on this team
- 03The data analyst
- 04The product manager
- 05The QA engineer
five people · 10 possible pairs · not one of them has worked on anything together · 4 have only been in the same meeting · 6 have never done either · one name in the lists is not on this team and was left out
One name in the lists was left out. It belongs to somebody outside this team, so there is no cell in this grid for the pair it would make. The denominator here is calculated from the five people on the roster, and a connection quietly dropped would shrink the fraction without saying why.
The grid has three states. Worked together. Been in the same meeting and nothing more. Neither. The panel above draws five people among whom the first state does not occur at all: four of the ten possible pairs are in the middle category, and six are empty.
Those five have been in the same room, repeatedly, for the better part of a year.
Why the middle category is kept
It does not count toward anything, and an earlier version of this grid did not have it. It was added because the distinction turned out to matter for what you do next.
A pair that has never met is a cold start. Arranging work between them means arranging an introduction, a reason, and a first conversation in which both people are working out what the other is like.
A pair that has met is warm. They know the other exists, roughly what they do, and that they are not alarming. The only thing missing is a reason. Converting a near-miss into a pair costs one task; converting a cold pair costs a task plus everything that precedes a task.
So the middle category is a list of the cheapest available improvements to the grid, which is a more useful thing to have than a tidier picture.
What a meeting actually transfers
Very little, and it is worth being specific about why, because the intuition that two hundred hours in a room must produce something is strong.
- A meeting transfers announcements. One person says a thing, several people hear it. That is a one-to-many transfer, and it works.
- It transfers the existence of people. After a few months everybody knows who is on the call, what they are called and approximately what they do.
- It does not transfer how anybody works, because nobody works in a meeting. It does not transfer the cost of asking somebody a question, because nobody is asked anything unexpected. It does not transfer judgement about anybody's output, because no output is produced. And it transfers almost nothing about the people who do not speak, who on a call of eight are usually five of them.
The two engineers in the panel have watched each other speak for a year. Neither knows whether the other sends drafts early or late, whether silence from them means agreement, or what they are secretly good at.
The recurring meeting as a substitute
This is where the harm is, and it is harm by substitution rather than by the meeting itself.
A standing meeting is the easiest thing to add to a team that feels disconnected. It is visible, it is cheap to arrange, everybody attends, and afterwards the calendar shows that something is being done about the problem. It occupies the slot in everybody's mind marked "connecting the team", and because that slot is occupied, nothing else goes in it.
Six months later the meeting is still running, the grid has not moved, and the conclusion drawn is usually that the team needs a better meeting.
Converting a near-miss
The conversion is always a task and it is always small. The things that work are the ones where the pairing is incidental to something that needed doing anyway.
Send one person's work to the other for review, specifically, rather than to the channel. Ask the two of them, jointly, for a decision that genuinely has two sides. Put them both on one small thing with a date close enough that they have to talk this week rather than next month.
What does not work is naming the relationship. Telling two people they should work together more, or pairing them as buddies, or scheduling them a recurring half hour, produces a half hour in which two people who have nothing to do together try to find something. Most of them quietly stop after a month, and the pair is now slightly worse off, because both of them have evidence that it does not work.
The exception worth knowing
One kind of meeting does produce pairs, and it is worth naming so the argument above is not read as blanket.
A meeting where two specific people have to settle something between them, in front of others, is a real exchange. A design review where the author defends a decision and one colleague presses on it. An incident review where two people reconstruct what each of them did. A planning session where two people disagree about sequencing and have to resolve it.
The common property is that two named people have to deal with each other about something determinate. The hour is not addressed to the room; the room is watching two people work, which is a different thing and an unusual one.
The one in the panel that was thrown away
One name in the lists above belongs to somebody outside this team. It was in the record because the pair is real — the designer does work with that person, frequently and well.
The grid refused it, and the refusal is right. This picture is about the twenty-eight pairs inside one box, and its denominator comes from the roster. A connection to somebody outside it does not have a cell, and quietly folding it in would make the team look better connected than it is by counting a relationship that belongs to a different network. The note on what the chart cannot tell you has more on why those outward connections matter anyway.