Understanding
Whether a person can comprehend what is on offer — what it is, how it works, why it matters, and whether it is for them. Not awareness, which measures whether they have heard of it.
Economists design markets that already have participants. We work on the question upstream of theirs: what allows a market to emerge at all.
When participation fails, the cause is always one of five constraints. At any moment, growth is limited by whichever of them is currently most binding.
Two sentences doing two jobs. The first makes a diagnostic possible, and it is falsifiable in a single move: find a participation failure that is none of the five. The second is the operational claim — a system moves at the speed of its bottleneck, and improving anything that is not the bottleneck changes nothing at all.
Which is why the useful question is never all five. It is which one is binding today.
Roth and Shapley (2012) and Milgrom and Wilson (2020) established market design for markets that already have participants — matching, auctions, allocation.
BJ Fogg's behaviour model underlies constraints one to four at the level of the individual. Goldratt's Theory of Constraints supplies the binding-constraint logic, applied here to demand rather than production.
The extension is the move from individual to market, and the addition of Coordination.
Whether a person can comprehend what is on offer — what it is, how it works, why it matters, and whether it is for them. Not awareness, which measures whether they have heard of it.
Whether the rails exist and this person can reach them. Physical and infrastructural; legal, since an unlicensed rail is not a rail; and economic, since a price beyond them is a barrier like any other.
Whether they believe participation is safe. Perceived risk rather than brand affinity: will I lose money, will I be defrauded, is there recourse, will this work at all.
Two mechanisms, one constraint. Ignition converts intention into action today. The loop brings people back without a new prompt each time. Ignition without a loop produces a spike; a loop without ignition never starts.
Whether participation becomes more valuable as others participate. It cannot be assessed by testing one person, and no single organisation can relieve it alone — which makes it the constraint that requires convening rather than campaigning.
Bilateral trust does not scale. If every participant must individually assess every other participant, transaction costs rise faster than the market can grow. Coordinating institutions convert expensive person-to-person trust into cheap institutional trust.
It is also fragile in proportion to its efficiency. Concentrate trust in one institution and you create a single point of failure — and when it fails, trust does not degrade gradually. It collapses across the whole market at once.
Supply is endogenous, not external. It responds to demand it can see, it responds with a lag, and it does not respond at all when it is blocked rather than absent. But the vacuum does not stay empty: where legitimate supply is missing, informal or fraudulent supply arrives instead — which then binds Trust for the legitimate market that follows, often for a generation.
The framework addresses latent demand, not manufactured demand. A considered decision not to take part is not a constraint, and treating it as one produces a discipline that assumes everyone would participate if only the friction were removed. We would rather not build on that assumption.
Worth being plain about, since the two are easy to confuse from the outside.
Both have a scored instrument. Understanding is measured by assessment rather than recall — whether someone can explain how a thing works, not whether they have heard of it. Access resolves to binary checks: does the rail exist, is it permitted, is it affordable, how many steps stand in the way. Two people running these on the same population should get the same answer.
These are established through evidence and judgement: incident history, recourse audits, return-rate curves, ecosystem and stakeholder mapping. Rigorous, but not yet reducible to a score. Instruments for Trust and Momentum are being built. Coordination may never have one, since it is a property of a market rather than of a person.
We say which is which in every diagnostic, per check. A finding that rests on judgement is labelled as one.
A diagnosis that cannot be wrong is not much of a diagnosis, so each call is recorded in advance and states five things: the binding constraint, the intervention, a quantified expected change against a named metric, the window, and the result that would count as disconfirming.
The fifth is the one most often left out, and without it the exercise is a forecast rather than a method. The record is kept whether or not it flatters us, and it is kept by constraint — because being reliable about access and unreliable about coordination is a useful thing for a client to know before commissioning either.
Calibration data will be published once the sample is large enough to mean anything. It is not yet.
Campaigns influence decisions.
Systems influence behaviour.
Markets influence culture.