The readout lands and it is genuinely good. Twelve weeks of interviews, a clear-eyed map of where the operation is losing time, and a recommendation: you need a transformation lead, two change managers, and a data engineer. Everyone nods. Somebody says finally, someone who gets it.
Then the same firm staffs all four roles, and nobody in the room says the obvious thing out loud.
The pattern underneath
This is not a scandal and the people involved are not cynics. It is arithmetic. A firm that earns on placement has one honest way to grow the account, and the assessment is the only instrument that can widen it. Nobody has to bend a finding. The pressure sits earlier than that — in which questions get asked, which gaps get called structural rather than fixable, and how many names it takes to close one.
I have watched this from both sides for twenty-four years, and the tell is never a lie. It is an absence. The recommendation arrives with a headcount and no alternative that costs the firm money.
An assessment you can't take anywhere isn't a diagnosis. It's a proposal wearing a diagnosis costume.
The cost of it does not show up as a bad hire. It shows up as a program that runs eighteen months because it was scoped to run eighteen months, staffed against a problem that a smaller, sharper answer would have closed in eight weeks. You will never see the counterfactual. That is precisely why it persists.
Three tells, before you sign
The diagnosis is free. A firm that gives away the assessment is telling you where it earns. Free is not generous — it is a customer-acquisition cost, and it gets recovered in the shape of what the assessment recommends.
You cannot take it anywhere. Ask whether the readout is yours to hand to your own team or to another firm. If the answer is procedural — a license discussion, an addendum — you have your answer. A finding that only works when its author executes it was written to be executed by its author.
Nobody will say who owns the number. Ask what happens to the fee if the number does not move. If accountability stays entirely with you while the invoice runs on time, then you did not buy an outcome. You bought people, priced by the month.
The fix is one clause
Pay for the diagnosis on its own, and require that it stands without the firm that wrote it. Not as a negotiating trick — as a design constraint. A diagnosis written to be executed by anyone has to name the leak in your language, size it, and sequence it, because that is the only version another team could pick up. The moment it has to survive being handed over, the padding falls out of it.
Then ask the second question, which is the one that sorts everybody: who owns the number? There are only three honest answers. You own it and I advise. You own it and I supply the people. I own it, and the fee is at risk if it does not move. All three are legitimate. Only the vagueness is not.
I am on both sides of this, which is why it is worth me saying it. I write roadmaps and I put people on them. So the discipline is the same one I would want as a buyer: the roadmap is paid, it is yours, and it is written to be executed by anyone — your own team, another firm, or mine. Plenty of them end there, and that is a complete engagement.
The fastest way in is to point at the leak you feel. You'll get the honest read: what it is, whether software actually fixes it, and how far it moves.