The delivery date was set at the enquiry
Somebody takes a dimension off one of last week's documents and turns it into something to be bought. A pump with a duty, a transformer with a rating, four hundred metres of a particular pipe.
What happens next is a sequence with about a dozen dates in it. The programme carries two of them: the delivery, and the installation that follows. Everything between the drawing and the lorry happens in a system project controls has no reason to open.
And the date that decides when the thing actually arrives is near the front of that sequence rather than the back — which is why an item can be badly late for a reason that occurred in its third week, and why chasing it in its ninth month achieves so little.
A dozen dates, and the programme has two
Set the chain out. It runs roughly like this. A need is identified from a document. An enquiry goes out. Quotations return. There is a technical evaluation and a commercial one. An order is placed. The vendor produces drawings and data. Those are reviewed and released. Manufacture begins. Inspection or testing happens at the works. The item is shipped. It arrives. It is offloaded, stored, and eventually installed.
Twelve events, each with a date, each capable of slipping. Two of them appear on a construction programme.
And they are the last two, which is the part worth sitting with. Every event that can be acted on is upstream of the two that get watched.
The consequence isn't that procurement is unmanaged. There is a whole function managing it, with its own trackers and its own weekly discipline. The consequence is that the two systems meet at one point only, so a problem anywhere in the first ten events stays invisible to planning until it presents as a late delivery with no explanation attached to it.
And when it presents that way, the response it gets is the response a late delivery gets: pressure on the supplier, an expedited freight quote, a request for a recovery plan. All of which address the last three weeks of a problem that has been running for five months.
The second review loop
Last week described a review loop: a document goes to the Engineer, a period runs, and either an objection or a deemed no-objection comes back.
There is a second loop with the same shape running one layer down, and it is the one that matters here. The vendor produces drawings and data sheets. Those go to the contractor's engineering team for review. Where the requirements call for it, they then go to the Engineer as well — which makes three parties and two gates for a single vendor document.
Each gate has laps, and each lap has the property week 15 identified: the clock restarts rather than resuming. But this loop has an additional feature the first one doesn't. Until the vendor's drawings are released, the vendor isn't manufacturing anything.
Which inverts where the pressure should sit. In the upper loop a lap costs review time. In the lower one a lap costs review time and pushes the start of manufacture, so the same three weeks are spent twice over on the same item.
Lead time is quoted from release
This is the sentence worth carrying out of the week, and it is the kind of thing everybody in procurement knows and nobody writes on a programme.
When a vendor quotes a lead time, they are quoting from the point at which they can start cutting metal — which is release of their drawings, not receipt of the order. A quotation of, say, twenty weeks means twenty weeks after approval.
So the order date and the delivery date are separated by two numbers — the quoted lead time, and however long the drawing loop takes — and only the first of them appears anywhere. A programme that shows delivery at order date plus lead time is a programme that assumes the vendor's drawings were approved on the day the order was placed.
Which is the same unmarked assumption as last week, one layer deeper, and it is the reason long-lead items are the ones that fail. They fail at the front, quietly, months before anybody looks.
There is a reason nobody catches it, and it isn't carelessness. The order was placed on time. The vendor was appointed on time. Every visible milestone in the first quarter of that item's life reports green, because the thing that has slipped — a drawing sitting in a review queue — isn't a milestone on anybody's list.
Three clocks: ownership, payment, arrival
The last mechanism is commercial rather than programmatic, and it is where a tender-stage decision comes back.
Three things happen to a bought item and people speak as though they were one. Ownership of plant and materials passes on a defined event. Payment for them is a separate question with its own conditions, and under the standard forms there is a route to being paid for items that have been shipped or delivered but not yet built in. Arrival on site is a third thing again, and none of the three is required to coincide with the others.
The route to early payment has a condition worth knowing before it applies to you: it covers only items listed in the Contract Data for payment when shipped or delivered. Where nothing was listed at tender, the provision doesn't apply, and there is no mechanism to be paid for a transformer sitting in a factory for eight months no matter how much it cost to build.
That listing was a tender-stage act, from week 7, done in the fortnight before a bid by people whose task was to price rather than to finance. Cost & Cash Week 10 takes apart what commitment, accrual and expenditure each mean. What belongs here is the timing: the decision that governs your cash position in year two was taken in a fortnight in year zero, and it was taken by omission.
System design
Row one asks for something an order form has no field for: what the quoted lead time runs from. The number gets recorded and the reference point gets dropped, which is the sixth appearance on this track of a figure travelling without the thing that makes it mean something.
| Record | Produced by | Required quality | Verified against | Feeds |
|---|---|---|---|---|
| Order, with its quoted lead time | Procurement, on placing it | Records what the lead time runs from, not only its length | The quotation it accepted | Every date downstream of it |
| Vendor document register | You, alongside the review register | Lap number and release date per document, same discipline as week 15 | Responses actually issued | When manufacture can start |
| Release date | The reviewing party, or the clock | Captured on the day, because manufacture starts from it | The document it releases | The only defensible delivery forecast |
| Inspection and test records | The works, before shipment | Held before the item leaves, since after shipment they cost a trip | The specification they test | Acceptance · later defect arguments |
| Contract Data list for early payment | The tender, by act or omission | Known in month one, because it can't be added later | The contract as executed | The cash position in year two |
Row five is the only entry in this dictionary whose producer may have been an omission. A list that was never made is still a decision, and it is one that can be discovered in month one and can't be reversed in month twenty.
Practical insight
Build a procurement date line for your long-lead items only. Not everything — the ten or fifteen things whose absence stops your work.
One row each, and columns for the dates that matter rather than all twelve: order placed, vendor drawings received, vendor drawings released, manufacture start, inspection, ship, arrive. Seven columns, and you will be able to fill four of them from documents you already hold.
Then do the one calculation nobody does. Take the quoted lead time, add it to your vendor-drawing release date, and compare the answer to the delivery date in your programme. Where they differ, you have found a delivery that won't happen, and you have found it while the release is still in somebody's in-tray rather than when the crane is booked.
Put your release dates in the weekly report as a column of their own. They are the earliest hard evidence of a late delivery available anywhere on the project, they cost nothing to collect, and they turn a procurement problem into something a programme can respond to while responding is still cheap.
Key takeaways
- A purchase creates about a dozen dated events, and a construction programme carries two of them.
- The two systems meet only at the delivery date, so a problem in the first ten events surfaces as a late delivery with no explanation.
- A second review loop runs between vendor and contractor, and where the requirements call for it the Engineer reviews as well — three parties, two gates, one document.
- Each gate restarts its clock on a resubmission rather than resuming it, exactly as the first loop does.
- Until vendor drawings are released the vendor is manufacturing nothing, which is what makes this loop different from the first.
- A quoted lead time runs from release of the vendor’s drawings, not from receipt of the order.
- So a programme showing delivery at order date plus lead time assumes the drawings were approved on the day of the order.
- Long-lead items fail at the front, quietly, months before the failure becomes visible as a date.
- Ownership, payment and arrival are three separate events on three separate conditions.
- Payment for items shipped or delivered but not built in applies only to items listed in the Contract Data, so an omission at tender removes the route entirely.
- That listing was done by people pricing a bid, and it governs the cash position two years later.
Records born here. The enquiry and the technical evaluation behind the selection · the order, with its date and its quoted lead time · the vendor document register, with laps · the release date of each vendor document · the inspection and test records at the works · the shipping documents · the list of items in the Contract Data for payment when shipped or delivered, or the record that there was none.
What is coming next
Between the order and the delivery there is a job that has no drawing, no work face and no percentage complete, and somebody does it every week.
It consists of ringing a factory and asking whether a thing that was promised is still on course, and of knowing which answers mean yes. It never appears on a programme, produces nothing measurable, and is the difference between finding out about a slipped delivery in March and finding out in September.
Enjoyed this lesson?
Join with Google to get each new lesson the moment it's published — and help me see which topics matter most to you. No spam, one email a week, unsubscribe anytime.
Already following on LinkedIn works too — this is just for the weekly email.