A procurement team evaluates two products, chooses the cheaper one, and negotiates well. Eighteen months later an identity team is maintaining a separate set of credentials for it, because the cheaper product did not support single sign-on at the tier that was purchased.
Nobody made a mistake. The procurement decision was correct against the criteria procurement was given. The identity consequence was real, expensive, and invisible at the point of decision because identity was not in the evaluation.
This is the ordinary shape of technology failure in an organization of any size. Not a bad decision. A decision that was good in its own frame and costly in somebody else’s.
The chain is short and it is always the same
A cloud architecture decision determines the identity model available to you. The identity model determines what security controls are practical. The security posture determines how integrations have to be built. And the procurement decision that started it determines all three, along with what it costs to change any of them later.
Every link in that chain is well understood individually. Each has specialists, tooling and literature. What is poorly handled is the transition between them, because that is where the organization chart has a gap.
Why the gap exists
Organizations are structured around components because components are manageable. You can staff for infrastructure. You can hire a security lead. You can build a procurement function with clear authority and measurable performance.
You cannot easily staff for the relationship between them, and so it does not get staffed. It gets assumed, or it falls to whichever senior person happens to notice, or it surfaces in an architecture review that runs after the decision is already committed.
This is not incompetence. It is a structural property of dividing work into domains, and every organization that has domains has it.
The result is that each function optimizes correctly for what it is accountable for. Procurement reduces cost. Security reduces exposure. Infrastructure maximizes availability. Applications ship features.
Nothing in that arrangement asks anyone to optimize for the total, and the total is what the organization actually operates.
Why it survives
The reason the pattern is so durable is that the cost never arrives attached to its cause.
The identity team absorbing a credential sprawl problem in year two does not experience it as a procurement outcome. They experience it as their own backlog. The team that made the decision has moved on to other work, reported a saving, and been correct in doing so.
So there is no feedback. The decision is never revisited, the criteria are never revised, and the same trade is made again on the next purchase, by people acting reasonably.
