A consultant proposes a diagnostic built on a well-known management theory. The deck cites the founding book. The method has five levels, a maturity grid and a scoring sheet.
Open the founding book and search it for the five levels. A framework’s reputation and its source are two different objects, and only one of them has evidence attached.
Why these three models
The decision is whether to adopt a framework, buy the engagement built on it, or restructure a team around its recommendations. The features that fire are a claim whose cost of assertion is near zero, a name that has become valuable independently of its content, and a frequency question about whether adoption changes outcomes at all.
Three lenses. Certification produces an equilibrium answer about which claims carry information. Goodhart’s structure produces a complex answer about what happens to a framework once its name becomes the asset. Base rates produce a cycle answer about the only question that decides whether to spend. The first supplies a test you can run in ten minutes on any framework, including the ones you already believe.
1. A claim is worth what it cost to make
The general principle is old and it is the whole of the method. A statement carries information in proportion to what it would have cost someone to make it falsely. Statements that are free to assert separate nobody from anybody, which is why verification systems exist and why they are always expensive.1
Applied to frameworks: attributing a method to a famous book costs nothing, because almost nobody checks. The check itself is cheap and that asymmetry is the opportunity.
Here is the procedure, and it takes about ten minutes. Get the primary text. Search it for the signature terms of the framework being sold to you. Count the occurrences. Then search for the terms that actually dominate the text and see whether they appear anywhere in the deck.
A worked example, and I picked one where the attribution error is extremely common and entirely checkable. Stafford Beer’s Cybernetics and Management was published in 1959, and it is routinely cited as the source of the Viable System Model, of the Systems One through Five architecture, and of algedonic signalling. It is not. Those belong to Brain of the Firm in 1972 and the work that followed it.
Here is the count from the 1959 text itself. Algedonic: zero. System Five: zero. Brain of the Firm: zero. The word viable appears seven times and every one of them is ordinary adjectival use, a viable model, viable neurons, adaptiveness in viable systems. Now count what the book is actually built from. Cybernetic: 502. Variety: 121. Black box: 66. Homeostat: 47. Requisite variety: 13. Ashby: 22.2
The 1959 book is a genuinely useful book. It is simply a different book from its reputation, and the difference matters commercially: the evidence and the practical record for the Viable System Model belong to the later work and to the decades of application after it, none of which is inherited by citing the earlier title. A deck that gets this wrong has told you that nobody on the other side opened the source.
Ten minutes with a search function settles a question that a whole engagement is priced on.
The result of the check has three readings. The terms are present and used as described, in which case you have verified the lineage. The terms are absent, in which case the framework is a later accretion and you should ask what it is actually based on. Or the source is unobtainable, which is itself informative about a method being sold as established.
The check has a failure mode worth building in from the start. A term can be genuinely absent from one book and present in the author’s work, because ideas develop across papers, later editions and translations, and a framework does not become fabricated by being younger than the title on the slide. So the finding is never that the framework is false. It is that the framework does not live where the deck says it lives, which is a question about which text carries the evidence, and somebody should be able to answer it.
That sets up what to do when the seller reframes, which most will. Told the terms are absent, they will say the method is based on the source rather than taken from it, and that is frequently true. Ask the follow-up: which claim on which slide came from which chapter. A supplier who has read the source answers in a sentence. A supplier who has not offers a description of the spirit of the work, and the gap between those two answers is the whole of what you were trying to establish.
2. What happens once the name becomes the asset
The second lens explains why the drift happens at all, and it is structural rather than a matter of individual dishonesty.
Once a framework’s name carries commercial value, the name is what gets optimised. Certification programmes appear. Maturity grids get added because clients want a score. The version that spreads is the one that is easiest to teach and to sell, not the one that is most faithful, and each retelling is a small simplification made by someone with a reason to simplify.3
Notice that this is the same structure as a metric becoming a target. The framework began as a description of something real, the description acquired a name, the name acquired value, and the thing now being propagated is the name with whatever content travels most easily attached to it.
What spreads is not the theory, it is the part of the theory that fits on a slide.
There is a specific version of this worth naming because it is invisible from inside. When a framework is imported from another field, the metaphor carries assumptions with it that nobody restates. Borrowing a term from biology brings selection pressure and inheritance along with it. Borrowing from engineering brings a designer and a specification. Those assumptions may not hold in your company, and because they arrived attached to a word rather than as claims, they never get examined.
So the second question to ask of any borrowed framework is what came across with it that could be wrong. If the answer is nothing, either the framework is genuinely domain-neutral, which is rare, or nobody has looked.
The second-order effect lands inside your own company and it is the expensive one. Once you buy the engagement, people get certified in it, and a certificate is a personal asset that outlives its usefulness to you. Two years later the framework is being defended by staff whose standing rests on it, and dropping it costs them something real. The cheap protection is to certify one or two people rather than a cohort, and to say at the outset that the trial is a trial. A department trained all at once cannot report that the method did not work.
There is a version of this you may not be able to refuse. Where a ministry tender or a donor’s procurement rules name a certified methodology, the framework is a requirement rather than a choice, and arguing about its provenance only spends the meeting. Buy the certification as a licence to bid, keep it separate from how the work is actually run, and do not let a compliance purchase become the company’s operating theory by default.
3. How often does adoption change anything
The first two lenses test the provenance. Neither tells you whether the framework works, and a faithfully transmitted framework can still be useless.
That is a base rate question. Of the organisations that adopted this method, in conditions resembling yours, what proportion saw the outcome change. Not whether it can work, which is a statement about possibility and carries no information, but how often it did.4
Most framework purchases are argued entirely from vivid specific cases, which is precisely the evidence that the base rate should be discounting rather than replacing. One transformed company in the deck is a selected observation, and the selection was made by the person showing it to you.
You will usually find that the base rate does not exist. That is a legitimate finding rather than a dead end, and it changes what you should buy: not a transformation, but a bounded trial with a stated outcome and a date, sized so that the failure is affordable. A framework with no base rate is not disqualified. It is simply not something to restructure around before you have generated your own evidence.
A bounded trial has a shape, and stating the shape before you sign is what keeps it bounded. One team rather than the company. Eight weeks rather than a year. One outcome measure chosen in advance, with a number attached that would count as the method having worked. A named person who decides at the end, and a written note of what happens if the answer is no. That last item is the one usually left out, and without it the trial cannot fail. It can only be extended.
Keep one distinction sharp while you read the result. A trial run by a capable facilitator tests two things at once, and the framework is the less likely of the two to have produced the improvement. To separate them, have your own people run the second cycle with the consultant out of the room. If it still works you bought a method. If it does not you bought a person, which is worth knowing, because a person can be hired and a deck cannot.
The falsifiability test belongs here too, and it is the sharpest of the three. Ask what observation would show this framework to be wrong in your context. If no observation could, you are not looking at a theory. You are looking at a vocabulary, and vocabularies are sometimes worth having for the coordination they provide, but they should be priced as vocabulary rather than as method.
What the three say together
- Search the primary source for the framework’s signature terms. Ten minutes, and it settles the lineage question.
- Ask what came across with the metaphor that might not hold in your business.
- Ask for the base rate, and treat its absence as a reason to buy a bounded trial rather than a transformation.
- Name the observation that would show it is wrong here. If none exists, price it as vocabulary.
Where they disagree
The provenance test and the base-rate test can point in opposite directions, and the case where they do is the interesting one.
A framework can be a badly attributed simplification of its source and still work, because the simplified version is easier to run and consistency of application often matters more than fidelity. Equally, a framework can be perfectly faithful to a rigorous source and change nothing, because the source was descriptive rather than prescriptive and no amount of accuracy makes a description actionable.
When they conflict, the base rate should win, because it measures the thing you are buying while the provenance check measures the thing you are being told. The provenance check keeps its value as a signal about the seller: a supplier who has not opened the source has told you something about the diligence you can expect in the engagement, whatever the method’s own merits.
What none of them contain
None of the three accounts for the coordination value of a shared vocabulary. A framework that is theoretically thin can still be worth adopting because it gives forty people the same words for the same problem, and that benefit is real and is invisible to every test in this article.
None of them handles the case where you are the one selling. The same asymmetry that lets a supplier cite a book nobody opens applies to your own deck, and the number of founders who have checked their own citations to the primary source is small.
And one property this ensemble will not produce: frameworks change the people who use them even when the framework is wrong. A team trained to ask about feedback loops asks better questions afterwards regardless of whether the specific model was faithful, and none of these three lenses can see a benefit that arrives through the users rather than through the method.
The one action that survives the ignorance: take the framework your company currently runs on, find its primary source, and search it for the terms your version uses. Do this on your own framework before you do it on a supplier’s, because the finding is more useful when it is about you and there is nobody to blame for it.
Who has to move
Whoever signs the engagement, and the check has to happen before the proposal is socialised rather than after, because once four people have repeated the framework’s name it becomes awkward to ask whether it is in the book. The instinct when a method sounds rigorous is to evaluate the logic of the deck, which tests the consultant’s coherence rather than the method’s foundation. The cheapest first test is the ten-minute search of the primary text, and it can be delegated to anyone who can use a search function.
Sources and notes
- Jeffrey Carpenter and Andrea Robbett, Game Theory and Behavior, MIT Press. Costly signalling and certification, and the general result that a claim carries information only in proportion to the cost of making it falsely, are developed in the chapters on asymmetric information and signalling. Used in section 1 as the reason a cheap attribution carries no information and a cheap check can therefore recover a great deal.
- Stafford Beer, Cybernetics and Management, English Universities Press, 1959, compared with Stafford Beer, Brain of the Firm, Allen Lane, 1972. The counts in section 1 were taken from the 1959 text directly. Method, stated so you can repeat or contradict it: the copy is a 116-page scan with no text layer, so it was rasterised and put through optical character recognition, and the terms counted case-insensitively across the whole book. OCR introduces errors, so treat the small counts as approximate and the zeros as the load-bearing result, since a term appearing dozens of times would not vanish entirely under transcription noise. Every occurrence of the word viable was read in context and all seven are ordinary adjectival use. The Viable System Model, the Systems One through Five architecture and algedonic signalling are developed in the 1972 book and in the work that followed. The procedure recommended in this section is the procedure by which this note was written.
- Donella H. Meadows, Thinking in Systems: A Primer, Chelsea Green Publishing. The trap in which a measure becomes a target and the system then optimises the measure rather than the purpose is set out in the chapter on system traps, under seeking the wrong goal. Applied in section 2 by analogy to a framework whose name rather than whose content becomes the object of optimisation. The analogy is stated as an analogy, since Meadows is describing metrics inside a system and not the transmission of ideas between organisations.
- Sanjit S. Dhami, The Foundations of Behavioral Economic Analysis, Oxford University Press. Base rate neglect and the substitution of a vivid specific case for the underlying frequency sit in the part on bounded rationality. Used in section 3 for the direction of the error, with no magnitude claimed.
A note on a number this article does not give. There is no share of management frameworks that turn out to be misattributed. Nobody has counted, and a figure invented for rhetorical effect here would be the exact failure the article is about. What transfers is the procedure, and its cost is ten minutes per framework.
Joshua Agonya Pi’Rwot, Founder.