Part II — Seeing Organizational Flow · Chapter 18 · 7 min read · First Public Draft
Human conditions
Nobody Writes Down the Cost of Not Trusting
Trust, empathy and the willingness to say the awkward thing are not decoration. Their absence turns into structure.
Jump to a chapter
Narrated with Microsoft's neural voice, not me — a recording.
Two teams, the same building, the same quarter. One of them hands work over with a message that says here it is, shout if anything looks wrong. The other hands over with a document, a walkthrough booked a week out, and a checklist somebody built after the last time.
Same work. Same competence. The second one takes about three weeks longer a year, and if you ask anybody why the process exists you get a completely reasonable answer about quality.
The part worth sitting with is what happened when I asked about the checklist. Something had gone badly wrong between those two teams — a release, a missed dependency, a bad week that turned into a bad quarter. It was six or seven years ago. Nobody involved was still there. The person who built the checklist had moved on, the people it was built in response to had moved on, and not one person in either team could tell me what had actually happened, only that there was an incident once.
The relationship had been replaced twice over. The caution outlived it. Everyone was now carefully working around a wound that no living person in the room had received.
I spent a long time treating that sort of thing as a cultural matter and therefore as somebody else’s department — the soft end, adjacent to the real work of boundaries and ownership and decision rights. That was a mistake, and a slightly embarrassing one for a paper about structural friction, because the soft end is where a good deal of the structure comes from.
What is actually going on is a substitution.
Where trust is present, a lot of coordination simply does not need to happen. You send the thing. The other side deals with it. If something is wrong they tell you, and neither of you experiences that as a failure worth building a control around.
Where it is thin, all of that has to be reconstructed out of process. Not because anyone decided distrust was the policy, but because a check is the cheapest available substitute for confidence, and it is always available. Nobody has to argue for it. Nobody looks unreasonable asking for one.
So the meeting exists because somebody was not sure. The document exists because somebody was burned. The approval exists because a relationship went wrong in 2019 and the person who fixed it left in 2021, and the approval stayed.
It Grew Like That describes structures nobody decided on. This is one of the mechanisms that grows them.
It shows up as three different absences, and they cost differently
Trust is the expensive one, and the slowest to rebuild. It sets how much coordination a piece of work needs before anyone is comfortable. It is also the one that compounds in both directions: every uneventful handover makes the next one lighter, and one bad experience buys a control that outlives the memory of what it was for.
Empathy is the quiet one, and it decides whether a boundary is workable. Not warmth — the plainer thing: some picture of what the other side’s week actually looks like. Its absence is why so many interfaces are designed by people who have never had to consume one, why a “simple request” arrives at a team already three weeks deep, and why the answer we are at capacity gets heard as obstruction rather than as information. Two teams that each understand the other’s constraints negotiate in an afternoon. Two teams that do not escalate, and escalation is a handover in a nicer outfit.
Courage is the rarest, and the only one whose absence is completely invisible. Everyone Left Agreeing describes the room where nobody says the thing. The cost of that is not the meeting; it is the six months afterwards, spent on a plan several people privately doubted. There is no artefact anywhere recording that the doubt existed. That is what makes it the hardest of the three to argue about with numbers.
An organization cannot be made only of comfortable people
There is a version of all this that reads as an argument for niceness, and it is worth blocking off, because the opposite failure is just as expensive and much better disguised.
An organization where everybody is comfortable does not look broken. It looks calm. Meetings finish early, nothing is contentious, and the plans get approved. What is missing is anyone willing to be the first through the snow.
That is the picture I keep coming back to for this. Somebody has to break trail — go first into unmapped ground, at walking pace, spending far more effort per metre than anyone following will ever spend, and be wrong about the route often enough to look foolish in front of people who are warm and behind them. Everybody else gets a path. Whoever broke it gets tired legs and, in most organizations, a quiet reputation for being difficult.
So a place made entirely of comfortable people has no trail-breakers, and a place made entirely of trail-breakers has nobody consolidating the ground behind them — which is its own kind of exhausting, and produces a great deal of half-finished new direction.
And breaking trail is only half of it, which is the part that gets missed.
A track through deep snow is not a path until somebody can follow it. Go first without any feeling for the people behind you and you produce a route at your own stride length, over the ground you personally found interesting, at a pace nobody else can hold — and then find it irritating that nobody is using it. I have watched that happen more often than I have watched a genuine shortage of courage. The exploring was never the constraint. Bringing people with you was.
So the empathy is not a softening of the courage. It is what converts a personal route into a shared one: going at a pace the rest can actually walk, saying out loud what you found and what you got wrong, and treating the people who did not want to go first as sensible rather than as ballast. They usually are sensible — somebody has to hold the ground while you are away from it.
The mix is the point, and the two halves need each other more than either usually admits. The ones following need enough courage of their own to say when the route is wrong rather than trudging along behind it in silence.
Where I have seen this go wrong, it is rarely that the trail-breaker was missing. It is that the organization had one, found them uncomfortable, and gradually taught them to stop.
Why none of it appears anywhere
The friction is real and the accounting has no line for it.
An approval has a name and an owner and shows up in a process map. The reason it was introduced does not. Six months of quiet doubt after a decision leaves no trace at all. And the person who would be best placed to say this control exists because two people fell out is usually one of the two.
So the cost gets attributed to something else. It becomes governance, or bureaucracy, or — my least favourite — culture, said as though culture were weather rather than the accumulated residue of a few hundred specific moments.
There is a version of this chapter that would be sentimental about it, and I would not defend that version. This is not an argument that people should be nicer to each other. It is an argument that these three conditions are load-bearing in the same way a boundary is, that their absence converts into structure, and that the structure is considerably more expensive and much harder to remove than the conversation would have been.
What makes it worth naming separately
Because it is the one kind of friction that cannot be fixed at the level it appears.
Move a boundary and the handover goes. Move a decision and the wait goes. But you cannot remove an approval that exists because somebody is not confident, and expect the confidence to arrive afterwards. Take the check away and the person who wanted it will find another one, or will simply do the checking privately, which is worse because now it is invisible.
That is the observation, and it is not a comfortable one: some of the friction in an organization is not structural at all. It is structural-looking. It is what a missing relationship turns into when it has been left alone long enough to harden.
Cultivating this: Growing What Cannot Be Installed, in Part III.
Editor's Notes
Psychological safety belongs to Amy Edmondson, and *The Fearless Organization* is where the evidence lives — this chapter is downstream of her work rather than adding to it. What I would add is only the accounting: that the absence of these conditions does not stay a feeling, it converts into approvals, forums and handovers that then look like process rather than like what they are.