The Savings Your Plants Will Politely Decline

The number you agreed to is two numbers

A supplier comes back on a packaging category with a concession split in two. Part of it starts now. The rest arrives after your plants run line trials and qualify the new spec. The headline reads as one percentage. Procurement books it as one percentage. The agreement is signed as one percentage, usually against a multi-year term.

It is two numbers, and only one of them is yours yet.

The second number is an option the supplier wrote and your plants decide whether to exercise. The strike price is line time, changeover minutes, QA attention, and a slot on a project list that was full before your agreement existed. Nobody at the table in the negotiation pays that price. The plant pays it, and the plant did not sign.

Why a thin contingent rate guarantees the trial never happens

Plant project lists are ranked. A plant manager running a facility with a known constraint has a short list of things that will move output this quarter, and a longer list of things somebody else wants. A packaging trial that requires pulling a line, running an unproven film or format, holding product for QA, and absorbing the risk of a bad run has to beat everything on that first list to get scheduled.

So the rate on the contingent tranche is not a price term. It is a priority term.

Lower the contingent rate and you make plant trialling look less attractive, which makes getting it prioritized harder, which means the saving never gets captured. The negotiation that looked like it was about percentage points was actually setting the probability that any of it happens. A supplier who routes value behind a gate they do not control, and you cannot schedule, has quoted you a number without committing to it.

This is also where the term commitment turns against you. You trade four or five years of volume for a headline rate, and the term runs on day one regardless of how many trials ever get scheduled. The contingent half can convert at zero and the commitment still binds. That asymmetry is the whole trade, and it is almost never priced as one.

The discipline here is to keep three states separate, because they get collapsed constantly. Negotiated savings is a sentence in an agreement. An implemented change is a qualified spec running in production. Realized margin is money that showed up. A savings report that treats the first as the third is not a report; it is a forecast wearing a ledger's clothes.

Negotiate the total, let the supplier pick the instrument

My rule in these negotiations is that I will not trade term for a number the client's plants still have to earn. If the supplier wants the long agreement, the value comes up front, and trial work gets priced separately as work.

That usually means striking the contingency and asking for the full rate against everything, then letting the supplier decide how to deliver it. Price, rebate tier, allocation across years: that is their internal problem. If they cannot show the number as price and need to put part of it in a rebate, it is worth asking whether you actually care. Often you do not, and saying so gives the supplier a path to yes that costs you nothing.

Where you do care, name why, because the instrument changes the economics. Price lands on every invoice from the effective date. A rebate lands later, sometimes a year later, and sits in a different place in the P&L. A tiered rebate only pays if volume cooperates. Those are not interchangeable, and a sourcing team that treats them as interchangeable will sign a worse deal while believing it held the line.

Then split the work you actually will do from the work you will not. Some trials are genuinely low effort and low risk, and those are worth committing to by name. Others require real line disruption and are not worth the squeeze; be direct that those will not get scheduled, and decline to accept a rate that assumes they will. A supplier's own effort and risk ranking is a useful starting point for that split, and it is also a negotiating artifact: if they called something easy, they have already conceded it is cheap to qualify.

One more thing to fix before any of this is measurable. Check which way your savings number moves when volume moves. If a report shows less savings in a month where the plant bought more, it is measuring usage, not the thing the agreement changed. Measure rate against rate, forecast price minus actual price, applied to actual volume. Volume is not what procurement controls, and a method that punishes the business for growing will get argued with instead of acted on.

What a well-run category agreement reads like

Every line in the savings schedule is in one of two columns, granted or contingent, and the columns are never added into a single headline. Each contingent line names the plant that owns it, the trial date inside the current quarter, and the line hours the trial consumes. Term length is quoted against granted value only. The monthly report shows rate against rate with volume held out of the variance, and it shows implemented separately from negotiated. When a trial slips, the line moves back to contingent that month rather than staying in the total.

The next test

Open the current savings schedule and circle every line that still needs a plant to do something. Add those up. That sum is the portion of your agreement that is a forecast, and if it is more than a small share of your headline number, you negotiated a term commitment against work nobody has scheduled.