Before the Hard Conversation With Your Co-founder

The document is usually called something like operating principles. Two founders wrote it after a bad quarter, at an offsite, and they meant every line of it at the time. It sits in a shared drive. Neither of them has opened it in a year.

By the time one of them books a session with me, they have tried considerably more than the document. They have read the books on difficult conversations. They know to lead with curiosity, to speak from their own experience, to assume good intent, to schedule the discussion instead of letting it ambush a Tuesday. Some of them have run the exercise properly, with an agenda and a timer and a genuine intention to listen.

It works for about eleven minutes. Then one of them says a sentence in a voice that is a half step flatter than it needs to be, and the other hears the whole of the last two years inside it. After that the agenda is a formality. Both of them are still using the words from the book. Neither of them is receiving anything.

The technique was not the problem in that meeting, and neither was the sincerity. Two co-founders in a room like that do not have a communication problem. They have a regulation problem. Every framework either of them has been handed for a hard conversation assumes two settled people, and there are not two settled people in that room. On a bad day there is not even one.

Every framework assumes two settled people

John Gottman and Robert Levenson spent years wiring couples up and measuring what happened in their bodies while they argued, and the physiology turned out to predict the outcome better than the content did [1]. Gottman later put a rough number on the part that matters here. Somewhere past about 100 beats per minute, you stop being able to take in what the other person is saying, however much you want to and however well they say it [2]. The machinery for listening is not available. It has been switched off in favor of machinery for surviving.

Dan Siegel has a name for the band you have to stay inside: the window of tolerance [3]. Above it, you get heat: the sharp reply, the sentence you assemble while the other person is still talking. Below it, you get the flatness, the shrug, the founder who goes agreeable and absent at the same time. Neither state can do anything with an I-statement. Both states can produce one.

So the order of operations runs the other way from how it is usually taught. The first step is not to be reasonable. The first step is to tolerate, and to stay regulated in each other's company while something difficult is in the room. Awareness comes after that. Communication skills come after that. No script jumps the queue, which is precisely why the ones they have already tried did not hold.

Capacity for this can be built. Both people's can. It takes longer than anyone wants, and it does not look like progress at the start, and that is not evidence that it is not working. Founders have unusual trouble with this last part, because in every other domain of their working life the absence of a visible result within a quarter is exactly what a failed intervention looks like. They are correct about that everywhere else.

You cannot invite someone into a room where they are the problem

One co-founder books the session. Which means there is a problem, and the problem is the other one. Everybody in the arrangement understands this, including the other one.

He walks in braced for an hour about everything he has got wrong. Often that is exactly what the hour becomes, and then he does not come back, and the founder who booked it concludes that he was never willing.

Whatever the initiating founder says in the first five minutes, however carefully she has worded it, if the premise underneath the words is that he was wrong, the conversation converts into a referendum on his adequacy and ends there. He will defend himself. Anyone would. That conversion is the thing to name, and the founder who called the meeting is usually the only person in a position to see it happen, because she is the one who chose the frame.

The fix for this is not a softer delivery, and it is not a better opening line. It is upstream of the wording, in the premise itself.

What rehearsal is actually for

I have written before that rehearsing conversations you never have is a sign that a co-founder relationship has gone quiet. That is still true, and it is a different thing from what I mean here. Private rehearsal at one in the morning is a way of not having the conversation. It runs the scene until you win it. You wake up with the argument sharper and the relationship no closer.

I am a registered drama therapist, which means rehearsal is a working method for me and not a metaphor. In a session it runs the other direction. You practice the conversation before you have it, and you do not practice it in order to script it. You practice it so that the body has met the size of the feeling once already, somewhere other than the room where it counts, and has somewhere to put it when the moment is real. You practice it until you lose, and then you find out that losing it was survivable.

This makes things worse before it makes them better, and founders should be told so before they start. The feeling has to move through somewhere, and it will not stay politely inside the fifty minutes. Small pieces, slowly, and everyone's pace is their own. If a founder tries this once, feels terrible for two days, and concludes that it does not work, what she has actually learned is that it started.

Some of this happens with one founder alone, which is often where it has to begin. Some of it happens with both of them in the room, which is work I do with couples and which the format was built for. If you have never done any of it, here is what a first session looks like.

Two kinds of no

In any single bad meeting, the two look identical. Over months they separate cleanly, and which one you are dealing with decides whether the partnership is worth the work.

One co-founder wants this and cannot do it. He has no idea how, and very little tolerance for the attempt, so he runs out of capacity fast and the meeting ends badly. Then it shows up somewhere else. A gesture. A thing he does a week later that he does not explain and does not take credit for. That is not a no. That is not yet, and not this much.

The other one does not want to change. You go one way, he goes the other. Whatever you bring reads as an attack, gentle or firm, and so does saying nothing. Nothing lands, because landing is not the game.

When both people want to change, this work is much easier than either of them expects. When one wants it and the other does not, it rarely works, and no quantity of skill on the willing person's side closes that gap. One-sided effort turns into pursuit, and after long enough the pursuing is no longer something she is doing about the partnership. It has become the shape of the partnership. That happens at home too, on the same mechanics.

Which is the answer to the question the founder reading this is actually holding. She has stopped asking whether to fix it. She is asking whether it can be fixed, and whether it is on her to keep trying. That question is answerable, and it is not answerable from inside one bad meeting, and the thing she should be watching for over the next several months is not whether he does any of it well. He will not do it well. She is watching for whether he wants to. Done badly and wanted is something two people can work with. Done well and not wanted is not a partnership, it is a performance of one, and she will be the only one rehearsing.