Where you enter.

The economic exposure you intend to hold. The asset, or the class of asset. The fund and the mandate it runs under. The jurisdictions already in play, whether chosen or inherited. The financing you expect to raise against it. The constraints you already know about, including the ones your investment committee has accepted informally and never minuted. What you do not have is a view on which of those is the one that binds.

What is the structural problem? Not which structure to use. Managers arrive routinely with a well-formed transaction and a misidentified difficulty: the constraint named in the room is the jurisdiction, and the constraint that actually binds is a clause in the fund constitution nobody has read against this asset. Both are real. Only one of them decides what can be built, and the other absorbs the attention because it is the one everybody can discuss.

The problem stated structurally, and stated in one place. The constraint that binds, cited to the instrument that creates it. The constraints that were assumed to bind and do not. That is usually the more surprising half. And the questions that have to be answered before any route can be drawn, set out in the order they have to be answered in.

When the difficulty has a name and an owner. It does not end with a proposed structure, and a diagnostic that arrives carrying one has quietly done the second rung in a meeting and skipped the record of it.

Structural Diagnostic, Structural Pre-Feasibility and Regulatory-Structural Mapping all open here. The three differ in what exists when you arrive: a transaction, a strategy, or a regulatory question raised by a change you are contemplating. All three are set out under the ten engagements.

A structural problem already stated, whether it was stated here or elsewhere. The map of what exists: the fund, the vehicle, the asset, the jurisdictions, the instruments, the appointments already made. And, almost always, one candidate route, which somebody in the house has been carrying since the first meeting and which nobody has yet had a reason to test.

What routes are open, and what does each one do to the rest of the transaction? The second half is the whole of the work. A route is never local: move the holding tier and the security package moves, the flows move, the forum that would enforce the security moves, and the administrator has to be asked a question it has already answered about a different structure.

Two, three or four routes drawn to the same depth and set against thirteen structural axes, with every rejected route recorded alongside the constraint that removed it. The axes themselves are at how routes are compared. What is particular to this rung is that the rejected routes are the durable half of the output. The surviving route is redrawn at every rung above; the record of what was ruled out, and why, is what stops a route being reopened by whoever joins the transaction next.

When the routes are distinguished on the axes your committee actually decides on. It does not end with a route selected. Selection is yours, and the rung is finished at the point the choice becomes available to take rather than at the point it is taken.

Structural Options Analysis, which the ladder shortens to Structural Options. One name here, against five on rung three. That reflects how many ways a structure which already exists can be tested, not how often the work opens on either.

A route the house has settled on, or a structure designed elsewhere and now in front of an investment committee. A term sheet drafted or close to it. Frequently counsel already appointed and drafting against the shape, and a target whose own holding chain is about to become yours.

What does the structure actually look like? Which entities, seated where, holding what, under which instruments, with which flows, which security, which governance, which regulatory interfaces and which adviser interfaces. Drawn to one scale, so that the tier your counsel is drafting, the tier your lender is securing against and the tier your administrator has to operate are demonstrably the same tier.

The blueprint: nine layers of one transaction held at one scale, set out at what it contains. The drawing is the visible part. Underneath it is a coherent structural model of how the transaction could work, with the failure points named against the condition that would trigger each one, and the dependencies mapped into the sequence the next rung is built from.

When every tier has an owner, an instrument and a jurisdiction, no tier is described as still to be confirmed, and every remaining question is written as a question and routed to the professional who owns it rather than carried as a shared assumption.

Transaction Architecture and Cross-Border Architecture are drawn here. Structural Feasibility Review, Structural Risk Review and Structural Second Opinion enter here as well, and each of those three tests a structure that already exists rather than drawing one. Five of the ten names sit on this rung, and three of the five arrive only once a shape already exists.

A structure your committee has accepted and your counsel is content to document. The advisers appointed or about to be. And the dates that actually bind, which are the ones set by somebody outside your house rather than the ones written in your own plan.

What has to happen, in what order, and who owns each step? Sequence is the reading of which steps cannot be taken until another has closed, which cannot be taken twice, and which have to be put to a party whose answer, if it is not the expected one, reopens steps already taken.

The implementation framework, built backwards from whichever date is real, with every step carrying the role that owns it. What this rung adds to the drawing is attribution: a dependency the third rung recorded as a fact about the structure becomes here a step somebody has to take, on a date, having first asked somebody else. The steps that cannot be taken twice are marked, and the classes they fall into are at the one-way doors.

At execution-readiness. That is a state with seven conditions rather than a view about a transaction. The framework then goes to the parties you have appointed, and each executes the part that belongs to it, in its own name and under its own authority.

Implementation Architecture. One name, and it is the last one before the work leaves this side of the line.

A second transaction, and a first one already behind you. Nothing else is required, and there is no threshold to clear. The rung is reached by returning rather than by qualifying for it.

What did we settle last time, and what does it bind now? A manager that has done one thing three times holds three structures, and the third was usually drawn without anybody reading the first two. The consequence is that the same question gets a different answer in each, and no document anywhere records that the answers differ.

Structural memory held outside your own house and kept current: your fund architecture, your vehicles as they actually stand rather than as they were designed, the centres you use, your operating model, your governance, the asset classes that recur, the counterparties you appoint, and the reason each earlier structure was drawn the way it was. A new transaction then opens against a map rather than against a blank page, which moves the first rung from weeks of reading to a conversation about what has changed.

It does not. It is the only rung with no closing artefact, and the only one that changes what the first rung takes as its input.

None of the ten. It is what the first four become when a manager comes back, and what it is made of is set out at structural intelligence rather than named as an engagement.