What Expediting A Frozen Order Actually Costs You

The earliest ship date is not a line question

A production planner at a mid market frozen prepared foods plant takes the same call every week. A customer wants the earliest date a new item can run. The planner knows the honest answer is weeks out, not days. The person asking assumes that answer lives in line availability, because line availability is the part of the schedule everyone can see. It does not live there. In this plant the binding step is freeze time. Blending and forming fit inside a week. Bagging cannot start for roughly a week after that, because product that has not fully set does not pack efficiently.

Nobody is idle during that wait. No line is occupied. No labor is consumed. That is precisely why the step vanishes from every conversation about capacity, and why the plant keeps agreeing to dates its own process cannot support.

Why a dwell step hides from the capacity conversation

Capacity discussions are organized around assets that cost money while they run. Lines, crews, ovens, baggers. A dwell step costs money while nothing runs, so it has no owner, no utilization number, and no slot on the board. It shows up only as a planner's reluctance, which reads to the commercial side as conservatism rather than as physics.

The dwell step also interacts with the rest of the schedule in ways a line rate never does. In a plant running allergen bearing and tested items, a product containing milk, or one awaiting a gluten result, blocks whatever would run immediately before or after it. The sequence is not a preference. It constrains which orders can be neighbors, which means the dwell step and the allergen rule together decide how much of the week is actually available. Neither one appears in a line speed number. Together they can take more of the week than any changeover does.

This is the general shape: when a system has a step that consumes time but no visible resource, throughput is governed by the interaction between that step and the sequencing rules around it, not by the rate of the fastest or slowest machine. Thermal processing plants know this from retorts and pasteurizers. Coatings and adhesives know it from cure. Frozen foods know it from set time. In every case the plant that manages only the visible asset is managing the wrong thing.

What the rescue actually costs

Every plant with a dwell step has an escape hatch. Here it is accelerated freezing, which can replace most of the wait on demand. The hatch works. It also has two prices, and only one of them gets discussed.

The first price is technical. Accelerating the freeze changes product density, which means the item that comes out is not quite the item the standard process makes. On an established product with run history, a planner can predict how it will behave in the bagger. On a brand new item that has never had a trial, nobody can. So the accelerated path is used most often on exactly the products where its effect is least known, which is the opposite of how a process control would choose.

The second price is capacity, and it is the larger one. The accelerated freezing asset only earns its keep at volume; it needs a high daily throughput to be economic. A rescue run for a single late order can be a small fraction of that volume. Running it means the asset spends a production day producing a fraction of what it produces on a normal day. The order gets protected. The displaced pounds are never charged to it. They are absorbed as a bad day on an asset nobody is holding to a number, so the plant pays in the one currency it is shortest of and never sees the invoice.

Stack a few of those into a month and the pattern inverts. The plant is using its scarcest capacity mostly to repair promises it should not have made, and the capacity it thought it had for growth was spent on rescues.

The decision, and who has to own it

This is a reallocation decision, not an improvement project, and it belongs above the planner. A planner asked to protect a customer date will always use the hatch; that is the correct local choice, and it is the wrong system choice. The decision that has to move upward is simple to state and uncomfortable to hold: accelerated freezing is a reallocation of constraint capacity, so it requires the same approval as any other reallocation of constraint capacity.

Three changes make that real. First, give the accelerated path a minimum volume per run, below which it does not get used without an explicit exception. That single rule prices the hatch. Second, log every use with the order it protected and the pounds the asset did not make that day, so the cost lands somewhere visible. Third, and this is where most plants find their real gain, stop issuing dates that depend on the hatch. A plant that refuses to schedule an item until the materials are physically in house and the full process time, dwell included, fits in the calendar removes most of the demand for rescues at the source. That is not a scheduling refinement. It is the difference between a plan and a wish.

Verification is straightforward. Count rescue runs per month and the displaced volume. If the count falls and the schedule still holds, the capacity was real. If the count falls and the schedule starts missing, the rescues were covering a promise date problem the plant has not fixed yet, and now it knows.

What a well run plan looks like here

The planner can state the full process time for any item, dwell included, and the commercial team quotes from that number rather than from line availability. Accelerated freezing has a named owner, a minimum volume per run, and a count that someone reports monthly alongside the pounds it displaced. No job is released to the schedule unless its materials are in house. Any new item has a trial on record before it is allowed on the accelerated path. The allergen and test sequence is a published rule, so the planner builds the week against it rather than discovering it.

The next test

Pull last month's schedule changes and mark every one that was absorbed by accelerated freezing. Then ask the one question that settles it: on each of those days, how many pounds did that asset not make? If nobody can answer, that is the number your capacity plan is missing.