A factory that promises a slot without owning a machine is making a promise it cannot enforce. The slot exists only if the machine is available, the operator is present, and no higher-priority job arrives. All three can change after the promise is made.
This is not dishonesty. It is a structural weakness in how capacity is committed. Most shops sell time they do not control, and then manage the gap by reshuffling. The reshuffle works often enough to keep the system going, and fails often enough to make delivery unpredictable.
Reserving capacity is the attempt to fix that. It means setting aside a defined window on a defined resource, for a defined period, before the work arrives. It is a commitment on both sides, and it only works if both sides treat it as one. The terms for standing volume arrangements are set out on the dealership application page.
There are three practical arrangements, and they differ in who carries the risk.
The first is a forward block: a named window on a named operation, held for a fixed period, with a minimum quantity to justify the setup. The buyer commits volume; the factory commits the window. If the buyer underfills, the factory still ran the setup, so the shortfall has a cost.
The second is a rolling reservation: a standing share of a recurring window, reviewed each cycle. It suits programmes with steady demand and low variation. It gives less protection against a spike, because the window is sized to the average.
The third is a priority agreement: no specific window, but a defined position in the queue when work is released. This is the weakest of the three and the most common, because it costs nothing to grant and nothing to prove.
A priority agreement fails because it has no physical referent. Nothing on the floor changes when it is signed. The planner knows about it, but the planner also knows about every other priority agreement, and they cannot all be honoured in the same week.
When two priority jobs collide, the tie is broken by whoever escalates loudest, which rewards noise rather than planning. Buyers learn this quickly and start escalating pre-emptively, which consumes the planner's time and degrades the schedule further.
A reservation avoids this by being specific. A window on a named operation is a physical fact. Two buyers cannot both hold the same window, so the conflict surfaces at the point of sale rather than at the point of production.
Reserving capacity costs the factory something real: the option to sell that window to someone else. In a busy month, that option has value, and a reservation forgoes it. This is why reservations are usually priced, and why a factory that offers them freely is probably not maintaining them.
There is also an internal cost. A reserved window that goes unused is visible waste, and it takes discipline to leave it empty rather than fill it with a job that happens to be ready. That discipline is the whole point, and it is the first thing to erode when the shop is under pressure.
A credible reservation therefore needs an owner. Someone has to be accountable for the window being held, and that person has to have the authority to refuse work that would consume it.
The buyer gives up flexibility. A reservation commits volume, and volume that does not arrive is still charged for the setup it displaced. It also removes the option of shopping the next order around, because the value of the reservation depends on continuity.
This is a real trade and it should be made deliberately. A buyer with volatile demand should not reserve, because the shortfall cost will exceed the reliability gain. A buyer with steady demand and a hard downstream commitment should reserve, because the alternative is a queue position that moves.
A reservation that cannot be checked is a priority agreement with better language. Four things make it checkable.
With those five in place, a reservation becomes an operational document rather than a gesture. It can be audited, which means it can be trusted.
Work spread across four facilities in Shanghai complicates reservations, because a window on one operation is only useful if the upstream and downstream operations can feed it. A held slot in welding is worthless if the forming department cannot deliver to it.
The fix is to reserve along the chain rather than at a single point. That is harder, because it requires the planner to hold windows at several operations at once, and it reduces the total work the shop can accept. Shops that do it well tend to be the ones with a single planning authority across all sites.
A buyer can test this by asking whether the reservation covers the finishing and inspection steps as well as the fabrication. If the answer stops at the machine, the reservation is partial and the queue will reappear downstream.
A reservation holds when three conditions are true: the resource is named, the window is owned by someone with authority to defend it, and both sides have accepted a cost for breaking it. Remove any one and the arrangement drifts back into priority language.
For buyers, the practical question is not whether a factory offers reservations but whether it can describe one precisely. A shop that can explain the volume floor, the review point, and the person who owns the window is a shop that has done this before. The same discipline shows in how the operation is organised, in the way custom work is scoped, and in the parts listed under other parts, stainless exhaust systems and titanium components.
