Your Earliest Date Is Route Math, Not Line Time

The earliest date goes to the wrong person

Ask a plant scheduler for the earliest date you can have an item and you will get a number several weeks out. Ask the commercial team the same question and you will get a number inside the month. Both people are reading real constraints. They are reading different ones, and only one of those readings has to survive contact with the floor.

The scheduler is adding up a route. The item gets blended, moved to a second operation, held, then packed, and several of those steps cannot start until the one before it has finished and sat for a required period. The commercial answer is reading the production schedule and finding open hours on the line. Those open hours are real. They are also the one term in the calculation that was never binding.

When I ask a scheduler how they arrive at a promise date, the answer almost always comes back as a sequence rather than a lookup. They walk the route out loud, step by step, and add the waits. That is the correct calculation. The problem is that it lives in a person rather than in a field, so the organization has one number it can query and one number that is right, and they are not the same number.

Lead time is a sum, and the line is one term

A promise date has four terms. Material has to be in the building. Each step has a run time. Several steps have a mandatory wait after them before the next one can start. And the binding step has to have uncommitted hours in some future window. Most plants can produce the fourth term on demand and cannot produce the first three without asking someone.

That asymmetry is the whole mechanism. Routes usually exist in the system, but the required hold between steps often does not; it lives in a quality specification, or in the head of the person who has run the item for six years, and it is never joined to the schedule. Material lead time lives with purchasing, in a different system, on a different cadence. The line calendar is the only term with a clean owner and a clean query, so it becomes the working basis for every date the business commits to. Each of the four terms has a competent owner. The sum has none.

The failure mode follows directly. A promise shorter than the route total cannot be met by working harder, because none of the missing time is effort. It can only be met by skipping a wait or splitting a run, and both of those move the cost rather than removing it. Compressing a required hold changes the state of the material that the next step depends on, so the packing operation runs below its standard rate and the hours you saved come back with interest at the step that is actually constrained. Splitting a run buys the date and spends a changeover, which takes capacity out of the week to pay for a commitment made outside it. Neither consequence shows up anywhere near the quote that caused it.

This is not a food manufacturing peculiarity. It is the same structure as cure time in construction or soak time in heat treat. A mandatory wait is a physical property of the process, not a priority setting. You can schedule around it, you can sometimes engineer it shorter, and you cannot compress it with urgency. The organizations that get this right treat the wait as a hard input to the date. The ones that do not treat it as a negotiation, and discover every quarter that it does not negotiate.

Price the route before you promise the date

The decision this changes is whether a lead time complaint is a capacity problem. It usually arrives looking like one. Customers want product sooner, the plant cannot deliver sooner, and the proposed answer is more hours, another shift, or another line. My rule is that I do not accept a lead time complaint as a capacity problem until someone shows me the route total for the items in question. Most of the time the line was never the binding term, and a capital request built on that complaint buys hours the item structurally cannot use.

Start by writing the route down for the items that generate the complaints, not for the whole catalog. Step count, run time per step, and required hold time after each step, as fields rather than as convention. Join material lead time per item to the front of it. That gives you an earliest start that is a date instead of an estimate, and a route total that anyone can read without finding the one person who knows. The work is bounded, because the items that cause the trouble are a short list in every plant I have worked in, and the list is already known to the scheduler.

Then decide who owns the quoted date. Not the commercial team working from the line calendar, and not the scheduler answering one order at a time by hand. One calculation, generated from the route total plus the first window with uncommitted hours on the binding step, and one number that both sides read. The organizational change is smaller than it sounds, and it is the part that sticks, because once the quote reads from the route, the route stops being optional documentation.

Be precise about what the value is and when it exists. A documented route is not a changed quote. A changed quote is not a reduced expedite spend. The value shows up as fewer premium freight events, fewer partial runs, and fewer orders late on items that had line hours available, and you only see it by counting those three things before and after over a comparable season and mix. Keep them as three separate states in three separate columns. Collapsing them is how a scheduling improvement becomes a savings number that cannot be defended nine months later.

What a well-run promise date reads like

Every active item has a route in the system carrying step count, run time per step, and required hold time between steps, with the hold as a field rather than a convention. Material lead time is joined to the front of that route, so the earliest start is a date. The quoted date is generated from the route total plus the first window with uncommitted hours on the binding step, and the commercial team quotes that number rather than maintaining a parallel one. Expedite events, split runs, and late orders are counted monthly by item, and any item that repeatedly generates them has its route checked before its capacity does. When someone asks for a date inside the route total, the answer names which wait would have to be skipped and what that costs at the next step.

If you want the check today, take the three items your plant is most often late on and write out their routes by hand, including every required wait. Put the date that was promised next to the total. If the promise is earlier than the total before the schedule enters the calculation at all, the lateness was decided at the quote, and no amount of line time will fix it.