When You Have to Decide Before You Can Discover

The three hidden costs of acting on a disordered process — and the paradox that turns them into a governance bottleneck

The first article in this series said that projects to simplify, automate, or improve an organization fail for lack of information about how the work actually runs. The second said that much of that information was never decided — it has to be declared, and declaring it changes the process in the very act of asking. This article is about what that act costs, and the strange order in which the bill comes due.

Three costs

Acting on a disordered process means paying three separate prices.

The cost of looking. Before anything, you investigate — if only to learn which facts are genuinely recoverable and which were never decided at all. This triage has a cost, and it usually ends in the discovery that the facts that matter are missing rather than merely hidden.

The cost of deciding. Where a fact was never decided, someone must now decide it — declare the canonical source, name the owner, assign the accountable party. This is neither free nor minimal, and the reason is precise: the binding question is who is able to make the declaration. The more the declaration entangles — the more systems, budgets, and departments its effects reach — the further upstream the decision must travel, toward the few people with authority to make a frame-defining call. Entanglement pushes the decision up; the higher it has to go, the scarcer the capacity to make it. This is the governance bottleneck, and its cost rises with the reach of the thing being decided.

The cost of absorbing. Once declared, the organization must absorb the consequences — the ripple from the second article working down the chain: budgets reallocated, inputs rewired, flows resynchronized, the parties who were quietly feeding the old arrangement adjusting to the new one. Where dependencies loop back on themselves, this can be heavy.

These prices are paid not in money but in organizational resource: people’s time, judgment, and the meetings it takes to align them.

The paradox

In the normal order of things, you analyze first and decide afterward. Here the order inverts.

To run the discovery — the process mining, the rationalization study — you must first settle the very things the discovery was meant to reveal. A provisional pick will get you moving: you can choose a tentative canonical source or case notion, cheaply and reversibly, and let the analysis run. But the instant the result has to drive action — retire a system, wire an automation — that tentative pick must harden into a binding declaration, one that carries ownership and consequence. That binding step is the costly, risky one, and it is a decision made in order to obtain the information that would tell you what to decide. A decision before the decisions — and, once it binds, a real architectural act carrying all three of the costs above.

The inversion that stalls the project

Take this decision to the governance layer and you meet a reasonable question: where is the target roadmap this decision serves? You cannot answer it. You are still in discovery; you need the decision precisely in order to produce the roadmap. So you find yourself asking a layer built to approve decisions after analysis to make a frame-defining architectural decision before any analysis exists to justify it.

That inversion is where projects stall — and they stall worst exactly when the decision matters most. The more entangled the declaration, the higher it must travel and the larger its ripple; but the larger its ripple, the less willing the upstream layer is to make it without the roadmap it cannot yet have. The decision is hardest to obtain in precisely the cases where it is most needed.

A contribution, put to the field

This is stated as a contribution, not a complaint. The existing literature tends to treat governance decisions as something that follows analysis, and discovery tools as if they ran on neutral data. The phenomenon named here is the inversion of both: an architectural decision that must precede the analysis, whose cost scales with its entanglement, and which changes the thing it is about. As far as we can find, this specific sequence is under-described — which is the reason to name it, and the reason it needs the field, not the archive, to confirm it.

So we put it to the practitioners who would know. If you have run a rationalization, a migration, or a process-mining effort: did it stall at the point where a canonical source or an owner had to be declared before you could even begin — and were you asked for a target you did not yet have the information to produce? Where this matches what you have lived, say so. Where it does not, say that too. This is the kind of claim only practice can confirm or refute, and it is meant to be tested there.

This describes a present condition — the order in which these costs fall — and makes no claim that it is worsening over time.