Part II — Seeing Organizational Flow · Chapter 19 · 4 min read · First Public Draft
Alignment
Everyone Left Agreeing
Nobody lied. Nobody misheard. Everyone simply left with a slightly different picture in their head.
Jump to a chapter
Narrated with Microsoft's neural voice, not me — a recording.
The meeting ended a few minutes early, which everyone took as a good sign.
Eight people, an hour, a genuinely productive discussion. Nods at the right moments. No unresolved argument. Somebody said “great, I think we’re aligned,” and everyone agreed that yes, we’re aligned, and went back to their desks.
Three weeks later it turned out that two of those eight had been building toward noticeably different things. Not opposite things — that would have been caught immediately. Adjacent things. Different enough to matter, similar enough that every status update along the way sounded fine.
Nobody was careless. That’s what makes this one worth a chapter.
Agreement in a room is mostly a feeling, and the feeling is unreliable.
What actually happened in that hour was that eight people heard the same words and each constructed their own picture from them — filled in with their own context, their own assumptions about scope, their own sense of what “we’ll handle that later” covers. Everyone left holding a picture. Everyone assumed it was the same picture, because the words had been the same.
Words are extremely low-bandwidth compared to the mental models they’re standing in for. Most of the model stays in the head it came from, unspoken, because it’s obvious — to the person holding it.
The gap is invisible at exactly the moment it’s cheapest to close.
This is the cruel timing of it. In the room, with everyone present, the correction costs one question and ninety seconds. Three weeks later it costs whatever was built in the meantime, plus the conversation about how we got here, plus the small dent in trust between two people who now each privately suspect the other wasn’t listening.
The cost of the assumption is paid entirely in the future, which is precisely why nobody spends the ninety seconds in the present.
Nobody asks the checking question, for a very human reason.
Asking “can I say back what I think we just decided?” feels, in the moment, like admitting you weren’t following. It slows a meeting that is going well. It risks looking slow in front of people whose opinion you care about. And everyone else is nodding, so presumably it’s just you.
Except everyone in the room is running that same calculation. Which is how eight people who each had a small uncertainty all decide, independently and simultaneously, to keep it to themselves.
Handovers are the same failure with a longer fuse.
A handover is a meeting where half the participants aren’t in the room. One side has a rich, detailed model built over months. The other side receives a document and a conversation. Both parties agree the handover happened — there was a session, there were slides, questions were invited and none seemed pressing.
What transferred was the vocabulary, not the model. The receiving side now has enough language to describe the thing and not enough understanding to make judgement calls about it, and that gap only becomes visible when the first non-obvious decision arrives.
This is friction that generates no friction at all, right up until it generates all of it.
Nothing feels slow. Nobody is waiting. Everyone is productively building — which is what makes false consensus the most efficient way an organization has of moving quickly in slightly different directions.
And it is worth noticing how much of this paper it quietly touches. The handover that loses context, the decision that meant something different by the time it arrived, the team that built the adjacent thing. Several of the shapes elsewhere in these pages are this one, wearing different clothes.
The most expensive gaps are the ones where everyone was sure they agreed.
What moves this is cheap and slightly awkward: someone saying back, out loud, what they are going to go and do, while the people who could correct them are still in the room. Half the time it costs ninety seconds and nothing happens. The other half recovers three weeks. What makes it hard is not the technique but the social cost of going second — which means it works when whoever has the most standing in the room does it first, and unreliably otherwise.
Cultivating this: Reducing Friction, in Part III.