The owner bought a way of buying.

Three things the choice fixes

Set the standard comparison out and three columns matter.

How many contracts the owner holds. One, with a single firm carrying design and construction together. Or several — a designer, a constructor, sometimes a manager, sometimes the vendors directly.

What the owner can select on. Competitively bid work is awarded on lowest construction cost. Design-build and construction management are awarded against performance expectations, and the literature is blunt that the winner isn't necessarily the cheapest.

Whether phased construction is possible. Under design-bid-build it isn't. Under design-build and both forms of construction management it is.

The first two columns get argued about in the industry constantly. The third is stated flatly in the standard comparison, as a property of the method, and then largely ignored — possibly because it reads as a technical footnote rather than as the constraint on the calendar that it is.

That third column is the one nobody reads and the one this track cares about, because it decides whether the overlap between engineering, procurement and construction — the thing that shapes every programme in this industry — is even available.

Why these are one choice rather than three

The three columns are locked together by a single mechanism, and once you see it the whole table collapses into one decision.

To select on lowest price you have to be able to price. To price you need a complete set of documents, because a price against an incomplete design is a guess dressed as an offer. And a complete set of documents means the design finished before anybody started building — which is exactly what forbids the phases from overlapping.

So the owner is choosing between two packages, not picking from a menu of independent options. Either a competitive price against a finished design and a sequential programme, or an earlier start with overlap and a selection made on something other than price.

There is no third option where you get the cheapest price and the earliest finish, and an organisation that has committed to a target date assuming both has to give one of them up later, at a worse rate.

Where the overlap was bought

Week 2 said the overlap between phases is a decision rather than a fact about building. This is the meeting where it was made.

Notice what is being traded. Overlap buys calendar time and pays for it in rework risk, because construction starts against a design that is still developing and some of what develops contradicts what is built. That is a real exchange with a real rate, and the rate depends on how much of the design was actually stable when the first concrete went in.

By the time you inherit an overlapping programme, the trade has been priced and accepted. Your job is to administer it. Which is very different from believing you chose it, and the difference shows up in how you argue about rework.

Every contract the owner didn't sign is an interface it owns anyway

A job split into six packages doesn't have less coordination than a job let as one. It has the same coordination, held by a different party.

Under a single contract the coordination between disciplines is inside one firm, which is why it never generates a notice. Split the same scope and each of those seams becomes a boundary between organisations, with a contract on each side and a party in the middle who now owns the sequencing.

That owner is the client, and nothing in the procurement process makes the transfer visible at the time it happens. Interfaces Week 1 takes apart what comes next, when the Engineer named in your contract turns out not to exist and the party controlling your access has no contract with you at all.

The point for this week is earlier. The interface was created here, in a strategy paper, by somebody optimising the procurement. Splitting a scope into packages does exactly what it is meant to do — it lets each part go to the firm best placed to price it, and it can save real money. The saving is banked at award. The cost turns up two years later as a sequencing argument between two contractors who both have a defensible position, and it turns up on a different line in a different budget.

What arrives with your tender documents

A form of contract, a scope, and a set of answers to questions you haven't asked yet.

Whether the design will be finished when you start. Whether somebody else is on the same site under a separate contract. Whether your programme is required to fit around a package let before yours. Whether you are being selected on price or on something you will have to demonstrate. All of it decided, none of it explained, and none of it open.

And one thing that is recoverable, because it changes what you should worry about from day one. If you won on lowest price against a complete design, the design is complete and your exposure is quantity and productivity. If you won against an incomplete design on some other basis, the design will develop during your contract, and how that development is paid for is the commercial question of your whole job.

System design

The fourth row is the only one here you can build yourself, from documents already in your possession, in an afternoon. Every other row is somebody else’s record that you either receive or reconstruct.

Record	Produced by	Required quality	Verified against	Feeds
Delivery strategy paper	The owner, at the final gate	States what was traded, not only what was chosen	The target date and budget it served	Contract form · packaging · the programme shape
Package breakdown	The same paper	Every seam named, with the party who will hold it	The scope it divides	The interface matrix · sequencing obligations
Selection criterion	Procurement, per package	Recorded per package, since packages differ	The design completeness available	What your exposure actually is
Design completeness at tender	The tender pack itself	Issue status per drawing, counted by discipline	The drawing register	Where development will come from, and when
Sequencing obligation	The contract, or its silence	Names who must coordinate whom, and what happens when nobody does	The package breakdown	Every access and interface argument

The last row deserves the attention, because a contract can define it by saying nothing. A contract that says nothing about who sequences two packages has still allocated the obligation — to whoever can't avoid it, which is the party on site.

Practical insight

Establish, in your first fortnight, how complete the design was at tender — and be specific, because the answer varies by discipline.

Take the drawing register from the tender pack and mark each drawing with its issue status. Preliminary, for approval, for construction. Then count them by discipline. What you get is a picture nobody assembles and everybody needs: architecture might be at construction status while the mechanical services are preliminary, which tells you exactly where the development is going to come from and roughly when.

Then find out whether anybody else is working under a separate contract on your site, and ask for their programme rather than their contract. The contract is confidential. The programme is a coordination document, it has to be shared to be useful, and it is the thing that will constrain you.

Both of these take an afternoon and neither of them is anybody's job, which is why you should expect to do them yourself and expect nobody to ask for them.

The register mark-up is the one that pays for itself twice. The first time is now, when it tells you where your risk actually sits. The second is in eighteen months, when you need to show what the design looked like at tender and everybody else is reconstructing it from memory while you have it dated.

Key takeaways

• An owner buys a method of buying before it buys a building, and the method is fixed in a front-end paper that never forms part of the tender.
• The delivery methods aren't a list of contract shapes. They are a list of ways of arranging time.
• The choice fixes three things: how many contracts the owner holds, what it can select on, and whether the phases may overlap.
• Selecting on lowest price requires a complete design; a complete design means a sequential programme.
• So there is no option that gives both the cheapest price and the earliest finish, and a target date can be set as though there were, because the date is fixed in the business case before the strategy paper is written.
• Overlap buys calendar time and pays in rework risk, at a rate set by how much design was stable when construction started.
• By the time you hold an overlapping programme the trade is priced and accepted, and administering it is a different position from having chosen it.
• Splitting a job doesn't reduce coordination. It moves the same coordination across an organisational boundary, and the client keeps it.
• Your tender documents answer questions you haven't asked yet, and none of the answers are open.
• Whether you were selected on price against a finished design, or on something else against an unfinished one, decides what your commercial exposure actually is.

Records born here. The delivery strategy paper · the package breakdown and the seams it creates · the selection criterion for each package · the design completeness assumed at tender · the interface matrix the split implies · the sequencing obligation and which party carries it.

What is coming next

The strategy said how the work would be bought. Somebody then had to go and buy it.

An enquiry goes out, questions come back, answers are issued to everybody or to nobody, prices arrive, and one of them is accepted — sometimes after a negotiation that changes what was asked for. Everything a contractor is later deemed to have known was fixed in those weeks, by a process the delivery team doesn't attend, in documents that stop being read once the contract is signed.