There is a sector that does genuinely useful work. It takes the kind of problem — ecological, systemic, supply-chain, social — that an organisation cannot hold in one hand and makes it tractable. It translates genuine complexity into something a client can act on: a report, a framework, a set of recommendations. The client needs this. The complexity is real, and so is the gap between knowing something is complicated and knowing what to do about it. The business model works because that gap is real and someone has to bridge it.
The bridging has a name in cybernetics: attenuation. The consultancy reduces variety — takes a system with many states and many possible responses and compresses it into a model with fewer. This is not a failure of the work. It is the work. An organisation cannot act on all the complexity of an ecosystem or a supply chain; it can act on a simplified representation. The question is not whether to simplify but what the simplification costs, and who pays.
For most of the life of the transaction, the cost is invisible. The client receives a useful output. The consultancy is paid. If the consultancy is later acquired by a larger organisation, the acquirer is buying something that demonstrably creates value. Every step is rational. Nothing in the individual transaction reveals the problem, because the problem is not in any individual transaction.
The problem is structural, and it has two parts.
The first is that compression has a floor. There is a point beyond which further reduction no longer represents the system being compressed; it represents something else — something that has the form of knowledge but not its substance. A biodiversity metric can stand in for an ecosystem up to a point. A supply chain risk score can substitute for actual knowledge of what is happening in a supply chain only so far. Past that floor, the simplification no longer enables a genuine response to the system. It enables a response to the model. And those two things are increasingly different. The market does not know where the floor is. There is nothing in the price mechanism that signals when a compression has crossed from useful simplification into something that actively prevents engagement with the underlying complexity. Each acquisition, each new layer of integration, brings the floor closer without naming it.
The second part is where the cost lands. The parties to the transaction — the client, the consultancy, the acquirer — are made whole. The complexity was real, the service was useful, the transaction was fair. What none of them pays for is the range of responses that the system needed in order to remain capable of genuine self-regulation. As what-gets-removed-does-not-come-back argues, capability shed through simplification does not wait to be restored — it dissolves. The dissolution is not announced. It is simply absent when next required.
This is why the cost looks indirect when it arrives. It does not arrive as the failure of any transaction. It arrives as a landscape that is managed according to metrics but no longer attended to; a supply chain whose risk model is current but whose actual behaviour has become opaque; a community that has been served by proxies for long enough that nobody in the room any longer holds the direct knowledge the proxies were designed to represent. the-legibility-trade describes one version of this at industry scale: ecological complexity compressed into tradeable units, with the systemic side effects invisible inside the market frame that created them.
The underlying principle, stated plainly, is that a regulator must match the complexity of the system it is regulating. A simplified model can only govern a simplified system. As long as the real system retains its complexity — and it tends to — a governance model that has been attenuated past its floor of usefulness is not governing the system. It is performing governance of a representation of the system. The real system continues to accumulate states that the model cannot see and the governance cannot reach.
By the time this becomes visible, it presents as a sector in difficulty, an institution unable to respond, a policy instrument that does not connect to the thing it is supposed to regulate. It does not present as the accumulated terminal of a business model that sold complexity reduction one layer at a time until the floor gave way. This is not because anyone concealed it. It is because the causal chain is long and distributed across many transactions and many years, and every node in that chain looked like a success. Nobody was deceived. The business model worked exactly as designed. The cost was simply placed somewhere it did not appear on any ledger.
The business model works. It pays well and it attracts investment, precisely because it creates real value at each stage. The only thing it cannot do is account for the cost it defers — because the cost is never borne by any party to any transaction. It is borne by the systems that were compressed, and eventually by whoever depends on them.
Related
- the-legibility-trade — ecological complexity compressed into tradeable units; the industry-scale version of the same mechanism
- what-gets-removed-does-not-come-back — capability shed through overoptimisation does not wait to be restored; the dissolution is genuine, not temporary
- proxy-capture — when the simplified representation becomes the target, the underlying thing quietly degrades without anyone deciding that it should
- overoptimisation — systems narrow to the measurable and rewarded, shedding the unmeasured load-bearing capacity that provided resilience