You've already had the conversation in the hallway. You've already noticed that nobody can tell you what the number proves — only what it is.
The renewal is coming. The board is asking. The acquirer is closing. The claim you're being asked to rely on was produced by someone else, and you will be the one who signed.
Before you put your name behind it, find out whether it can carry the weight.
01 — A contract is up for renewal. You have to approve the next term, but no one has checked whether the claimed savings are real.
02 — You have inherited green dashboards. Now you have to explain what actually improved—and how you know.
03 — Your board, audit committee, or investors ask what the results really prove. You need an answer you can stand behind.
04 — You are buying a business, deploying AI, changing the workforce, or releasing funding—and you will own a decision built on results you did not produce.
05 — The program is called a success, but nobody has checked whether its business case still holds.
06 — Everyone close to the numbers already knows which relied upon claim won't hold up. So do you.
07 — The objective is written as a project. The key result is written as a deliverable. The work that was supposed to produce the outcome is funded somewhere else, and nobody in the review is looking at it.
08 — Three people define the same objective three different ways. Each definition is reasonable. None of them is the same.
09 — You ask what you would have expected to see if the program had no effect. Nobody in the room can answer.
10 — Your AI program is about to accelerate decisions against metrics nobody has independently verified — and the model that produced the claim has already been replaced twice.
The question that matters remains unanswered: which of these relied upon claims are true?
Green dashboards. No believers.
The review cycle closes. The question stays open.
Risk travels down. Authority stays up.
Success enters the record. The business case never reopens.
Time passes. The relied upon claim survives. The capacity to verify the claim doesn't.
Then someone asks the question the reporting or deliverable was supposed to answer.
The problem is not necessarily a bad metric, result or claim. It is that someone is relying on it—or is about to—to make a consequential decision before its evidentiary basis is clear.
The pattern you've seen runs the same way in organization after organization:
01 — Introduce a new methodology.
02 — Create new terminology.
03 — Create new rituals.
04 — Create new metrics.
05 — Report progress.
06 — Discover contradictions.
07 — Add more process. More dashboards. More governance. More terminology.
08 — Continue — because stopping would expose that the prior cycle didn't establish anything.
↺ Repeat. Each cycle reports green. None establishes whether anything is true.
What rises with each cycle is investment — process, dashboards, governance, terminology. What doesn't rise is the validity of the claims. The divergence is the spiral.
Each cycle makes the organization more sophisticated at representing progress — not necessarily more capable of producing it.
If that pattern is familiar, you don't need another framework. You need to know which of your numbers are still true.
The spiral is one instance of a broader problem: organizational claims systematically outrun their evidence. That problem has a name in the research — the assurance gap.
This is the category of Decision Evidence.
A claim is fit to support a decision only if the evidence can bear the weight of that decision. Decision Evidence is the independent examination of whether it can — before the decision-maker puts their name behind it.
The spiral has a precondition.
It runs where the people setting the goal are far enough from the work that they cannot see whether the intervention caused the result—and where the organization has attached a rating to the goal, so the rational move is to report progress rather than to discover the truth.
Where the decision-maker is close to the work, where the feedback loop is short, and where the upside is aligned, the artifact often does what it says. That is the narrow band. The spiral is what happens outside it.
The AI connection.
The same pattern is running now in artificial intelligence (AI). The construct is looser, the capital is faster, and the people who can see the truth have even less authority than they did in the OKR program. The model changes every quarter, so the claim you are relying on was made about a version that no longer exists.
An objective can be a coordination device — a way for different communities to work together without agreeing on a definition. That is legitimate, and it is why the artifact survives.
It can also be a measurement instrument — a construct that is operationalized, tested, and held to a standard. That is also legitimate.
It can also be an accountability device — a way of attaching a consequence to a named person, so that someone answers for the result. That is also legitimate, and it is what the board actually needs.
What no artifact can do is all three at once.
Coordination requires interpretive flexibility. Measurement requires a single enforced definition. Accountability requires a named owner who bears the consequence.
The artifact that coordinates cannot measure — the definition has to stay loose enough that everyone can read it. The artifact that measures cannot coordinate — the definition has to be tight enough that only one reading survives. And the artifact that attaches accountability can do neither, because the person held accountable will manage the number rather than establish the truth, and no one shares information freely with the person who will be blamed for it.
That is why the dashboard shows progress and the outcome doesn't arrive. The artifact is doing coordination work, being relied on for accountability, and the measurement it claims to do is the job that got dropped.
The finding is about the instrument, not the people who adopted it.
The person closest to the work knows whether the pilot actually worked. But they are not accountable for the decision, they are not rewarded for surfacing the contradiction, and they may be punished for it. So the contradiction stays in the corridor.
The person accountable for the decision sees the dashboard. It shows green. They have no way to know what the person closest to the work knows.
The truth of the claim belongs to no one — and that is why the green dashboard persists. Not because anyone is lying, but because no one in the chain has the knowledge, the incentive, and the authority to establish the truth at the same time.
The review is the only role that has all three. That is what independent determination means.
What would you have expected to see if this claim or program had no effect?
"The artifact, framework, or process was fine. The implementation was poor."
You've heard it. You may have used it yourself.
It is the defense that protects the framework by attributing every failure to the person holding it. It works because the construct was never specified — so every failure can be attributed to another layer: implementation, execution, culture, leadership.
The review names the defense, because a claim that cannot be falsified at any layer cannot be shown to work at any layer. Naming the defense is what makes the finding stand.
1. Can three people define the objective in one sentence, independently, and agree? If not, the artifact is coordinating, not measuring — and the review will say so.
2. What would you have expected to see if the program had no effect? If nobody in the room can answer, the program has no claim, hypothesis, or other belief that is contrary to the facts — and no one can tell you whether the program caused anything.
These are the first questions the review asks. If both tests come back clean, you may not need the review.
Decision Evidence is the practice of independently examining consequential claims before a decision-maker relies on them.
The examination informs the consequential decision. It does not replace the decision-maker's judgment, implement the program, or guarantee the business outcome.
:Copy-right/copy-claim (c) 2026 OKR Strategist™. | All Rights Reserved. | This site issues determinations of temporal permissibility under constitutional constraint. Nothing herein constitutes advisory, fiduciary, managerial, or execution services. | No authorization, delegation, or ratification of action is provided or implied. | Unauthorized use, reproduction, or misrepresentation of this material is strictly prohibited. | New York, USA. | Do NOT scrape, sell or share this material or my personal information.
