Supply on site: why the material does not arrive and how to stop losing days
Site supply is the process that runs from spotting the need for a material to having it available at the work face where it will be used. It sounds linear and it is not: it has between five and eight links, each with its own lead time, and one of them failing is enough to leave a whole crew standing.
It is the most cited cause of days lost on site and, oddly enough, one of the worst designed processes in the whole operation.
The full chain
| Link | Who controls it | Where time is lost |
|---|---|---|
| Spotting the need | Site engineer | Spotted late, once it is already short |
| Requisition | Site engineer | Held back to order “all at once” |
| Approval | Head office | Waiting on a signature with no deadline |
| Quoting and purchase | Procurement | Chasing price instead of availability |
| Manufacture or preparation | Supplier | Real lead time differs from the promised one |
| Transport | Supplier | No visibility until it turns up |
| Receipt and unloading | Site | No space, no equipment, nobody to receive it |
| Haul to the work face | Site | The material is on site but not where it is used |
The last link is the most underrated. Material on site is not material available. A pallet of block at the gate, when the crew is working on level 4, is no use until somebody lifts it.
Where it really breaks
Spotting it late. The root cause. When the need is spotted by seeing an empty work face, the clock on every link that follows starts late by definition. No amount of fast purchasing makes up for a late detection.
Promised lead time versus real lead time. Every supplier promises a date and every experienced buyer knows it is optimistic. If nobody records each supplier’s actual delivery time, planning is done on false data, project after project.
Approval with no deadline. A requisition waiting for a signature is pure time. And because no deadline was ever committed to, nobody counts it as a delay.
Price ahead of availability. Procurement optimises what it is measured on, which is normally savings. A supplier 8 % cheaper who delivers two weeks later can cost three days of a crew standing idle, which is worth more than the saving. That calculation is almost never done.
Improvised receipt. The truck arrives, there is nobody to receive it, no storage space, no unloading equipment. The truck waits, or it leaves.
How to work out when to order
The calculation is simple and almost nobody makes it explicit:
Order date = Date of use
− supplier's real lead time
− approval time
− receipt and haul time
− safety buffer
With example figures for a material with a real lead time of 15 days:
| Component | Days |
|---|---|
| Supplier’s real lead time | 15 |
| Internal approval | 3 |
| Receipt and haul to the work face | 1 |
| Buffer | 5 |
| Total lead time | 24 days |
In other words: to use the material on day 60 of the programme, it has to be ordered on day 36. And to order it on day 36 you have to know on day 30 that it will be needed, which requires somebody to be looking at the programme ahead of time.
The buffer is not pessimism, it is the variability of the lead time. If the supplier delivers in anything from 12 to 20 days depending on the month, planning on 15 means failing half the time.
The three records to keep
You do not need a purchasing system. You need three records.
Real lead time by supplier and by material. Not the promised one: the measured one. Order date and actual arrival date, accumulated. After six months you hold the most valuable information in your procurement department.
Critical materials with their lead times. The five or six materials that stop the project if they are missing. With their lead time calculated and checked against the programme.
Days lost for want of material, with the material and the cause. It is what turns “we are behind on supply” into a number, and —this matters— it is what supports a delay that is not yours when the material is supplied by the client.
The record that protects you
There is a consequence here that gets underrated.
When the material is supplied by the client and does not arrive, you have a delay that is not attributable to you. But only if you recorded it on the day it happened, with the cause, the material, and how many people were left standing.
Without that contemporaneous record, the delay is presumed to be yours. And since price adjustment mechanisms are calculated against the programme, a delay you could not prove was somebody else’s also moves your price settlement date against you.
In other words: the record of a day lost to material is worth far more than a day.
What can be fixed without spending
Five concrete things, none of which needs software or extra headcount.
A committed deadline for internal approvals. Give a requisition 48 hours to be approved, and record it when that is missed.
Measure procurement on availability, not only on price. Adding an on-time delivery indicator changes behaviour within a month.
A weekly look at the programme four weeks out. Check what will be needed in a month and whether it has already been ordered. Fifteen minutes a week.
A single receiving point with a named person responsible. So the truck knows who to look for and there is somebody to receive and sign.
Record the lost day on the day it happens, with the cause and the material. It is the record that is worth money later.
Frequently asked questions
- Is it better to buy everything up front?
- It reduces the risk of running short and increases the risk of damage, theft, obsolescence and tied-up capital. The reasonable answer is to stagger it: long lead times on the critical materials, scheduled deliveries on the ones consumed continuously.
- How do I negotiate real lead times with a supplier?
- With their own history. A supplier you can show that their last six deliveries averaged 19 days when they promised 12 has little left to argue. Without that figure, the conversation is about perceptions.
- What do I do if the material is supplied by the client?
- Record it with the same rigour, or more. Date requested, date committed, date actually delivered, and crew days standing idle. It is the evidence that supports an extension of time.
- How much buffer is reasonable?
- It depends on the supplier's variability, not on the lead time. A consistent supplier justifies a short buffer; an erratic one needs a buffer proportional to their spread. If you do not measure the variability, you cannot size it.
- Is a dedicated storekeeper worth it?
- On mid-sized projects, almost always. The cost of the position is usually less than the days lost to improvised deliveries and to material that is on site but not at the work face.
Chilean, designing for Latin America. Field research surfaces what actually matters to clients, and that becomes products non-technical people adopt on their own — legal, education, accounting — and that show up in productivity from week one.
See all their articles →
