Sourcing Books the Savings, Operations Pays the Bill

The Number Was Real on Paper

A mid-market food brand ran a packaging savings program with roughly $1M projected. The realized number came in a fraction of that. The folding-carton work was, in the team's own words, a debacle. The corrugated savings never landed at all. The savings were booked, forecast, and reported. The floor simply never produced them.

Two hundred miles away, a Midwest protein processor was running film and corrugated optimization at the same time, and hitting the same wall from the other direction. Sourcing wanted alternate specs across suppliers, three to seven alt specs to take to market. Operations was waiting on validation protocols before anything could move. Two halves of one decision, each waiting on the other, no shared model between them. The sourcing side and the optimization side were, plainly, not aligned internally.

Two Owners, One Number, No Bridge

Here is the mechanism underneath both. A material savings number has two owners who almost never sit in the same meeting. Sourcing owns piece price. They can negotiate a thinner film or a cheaper board grade and book the delta the day the quote comes back. Operations owns run rate, scrap, and changeover. They find out the new spec seals slower, jams more, or fails a drop test after the savings are already in the forecast.

The projected number is real inside the sourcing model. The realized number lives on the floor, in run-rate and scrap and rework. With no shared model spanning both, the constraint is not the price and it is not the line. It is the unowned seam between the two. That is what constraint alignment means at the organizational level: the binding constraint has migrated to a handoff, and the org is still optimizing the two ends.

Build One Model Across the Seam

Put both numbers on every spec change before you book a dollar. A spec change should carry the piece-price delta and the modeled run-rate and scrap impact from day one. If you cannot state the floor impact, you do not have a savings; you have a quote. Run alt specs through a validation gate before they enter the forecast, not after, so the number that gets committed is the number the line can actually hold.

Then name an owner for the seam. Someone has to be accountable for realized savings, not booked savings, and that person has to be able to see both the sourcing model and the floor data. The protein processor's instinct was right: develop the alt spec with each supplier, go to market like-for-like and on the alternate, then qualify. The failure mode is doing the qualification after the savings are in the plan instead of before.

What a Well-Run Program Looks Like

A well-run program reconciles booked savings to realized savings every month, and the gap stays inside roughly 10%. Every spec change carries a piece-price number and a run-rate and scrap number before it enters the forecast. Alt specs pass a validation gate before they are quoted as savings. One named person owns realized savings across sourcing and operations, and the validation protocol exists before the negotiation starts, not after the carton fails on the floor.

A savings program does not fail at the negotiation. It fails at the seam nobody was asked to own.