Part II — Seeing Organizational Flow · Chapter 27 · 4 min read · First Public Draft

Borrowed models

The Conditions Don't Come in the Box

The shape was part of the result. What produced it stayed behind.

Reduces frictionCreates valueOwnership clarityHandoff lossDecision latencyLearning cycle time

Somebody comes back from a conference, or finishes a good book, and what they bring back is a diagram.

It is usually a genuinely good diagram. That is what makes this one durable — nobody returns with a bad idea. They return with something that demonstrably worked somewhere, drawn clearly enough to put on a slide, and the meeting that follows is one of the more enjoyable meetings of the quarter. Everyone can see it. Everyone can see how it would fit here. Somebody volunteers to lead it.

Eighteen months later the boards are still on the wall and nothing moves any faster than it did.

I have been the person with the diagram. More than once, and with real enthusiasm, which is why I do not think this is about people being credulous. The diagram was accurate. It described something that was working. What it could not describe was everything holding it up.

The visible shape of an organization is a result, not a recipe.

The two most-copied models of the last few decades both make the point, and both make it in the same way.

Toyota’s production system does not arrive in the boards and the daily stand-ups. Those are what it looks like from outside — the parts that photograph well. What makes it work sits underneath: a capability — being able to stop the line at all — and the competence that capability is only ever as real as, built over decades. People who can see a problem, halt production over it, and be thanked rather than blamed for it. Lean programmes have been installing the artifacts without either for thirty years, and failing on exactly that distinction, usually while reporting successful adoption.

Spotify’s engineering model was published as a snapshot — how one company happened to be working at one moment, written up by people who have said since, repeatedly and in public, that it was never meant to be adopted as a framework. It was adopted as a framework anyway, in a great many organizations whose conditions had nothing in common with Spotify’s, several of which then concluded that squads and tribes do not work.

Squads and tribes worked fine. They were the shape of something, and the something did not come with them.

What actually gets copied is the part that is easy to see.

This is not a failure of intelligence, it is a property of distance. From outside an organization you can observe its structures, its ceremonies, its vocabulary and its org chart. You cannot observe how long a decision takes, who is allowed to say no, what happens to the person who raises a problem in week two, or which of the written rules everyone quietly ignores.

So the transferable-looking parts are precisely the parts that were consequences, and the parts that did the work are invisible from where the observer is standing. You end up importing the canopy and wondering why it will not stay green.

And the borrowed shape usually lands on soil that already had its own.

The conditions in the receiving organization are not neutral ground. There is already a way decisions get made, already a set of things it is unwise to say in a large meeting, already a funding rhythm that predates anybody in the room. The new model arrives on top of all that, and what emerges is a hybrid nobody designed: the ceremonies of one organization running on the incentives of another.

That hybrid is frequently worse than either. It has the overhead of the new thing and the constraints of the old one, and because everybody agreed to adopt it, questioning whether it fits reads as questioning the decision.

The honest version of this is uncomfortable for anyone who writes things down.

It applies to this paper too, which is worth saying plainly rather than leaving as an unstated risk. The examples here came from particular organizations with particular conditions, and they were never meant to be lifted. Anything in these pages that reads like a shape you could install is either badly written or being read as something it is not.

What travels between organizations is not structure. It is the question somebody asked that made the structure obvious — and questions are cheap to carry, survive translation, and cost nothing if they turn out not to apply.

Borrowing the answer is what disappoints. Borrowing the question rarely costs anything.

Cultivating this: Tending What Grew, in Part III.