The floor reports what it made, and the inventory is already done
An order becomes a bill of material, the bill of material becomes work on a line, and the line becomes an ordered queue an operator can work from. When they save what they ran, finished parts go into stock and the components they consumed come out — in one transaction, at the moment they said it happened. There is no production paperwork waiting to be entered afterward, because entering it was the reporting.
-
A sequence, not a calendar
A scheduled job has a position, not a timeslot. That is deliberate. A line supervisor reorders a queue dozens of times a shift, and a system that demands a start and an end time for every move turns a thirty-second decision into a planning exercise nobody actually does — which is how scheduling systems end up abandoned while the real schedule lives on a whiteboard.
Dates still exist where they mean something: the due date on the tree node and the production due date on the order line, worked backward through your operating calendar and your shutdown periods.
-
Routing is where you assigned the part, not a second document
You describe your lines once — a schedule is a production line at one of your locations. Assigning a part to a line at an ordinal position is what creates its routing: the set of distinct ordinals is the routing, and two lines at the same ordinal are alternates, different ways to do the same step.
Nothing has to be maintained in parallel. A routing document kept alongside the assignments is a routing document that will disagree with them within a month, and the floor will believe the assignments.
-
One Save moves stock in both directions
An operator opens a run batch, records what they finished and what material it took, and saves once. Run results, material lines and both stock postings either all land or none of them do — it is a single transaction.
Both postings are stamped with the run batch’s start time, not whatever time the operator got round to saving, and are made at the line’s company location, so a company running more than one plant gets stock in the right building without anyone choosing it.
-
The operator’s number is the number
Material is pre-filled from the order’s bill of material and then written exactly as the operator left it. Backbone MRP proposes what the build should have taken and then trusts the person who was standing there, with good and scrapped quantities recorded as a breakdown of what was used rather than additions to it.
A job is finished when the operator says it is finished. Completion is never inferred from quantity, because the quantity being short is very often the whole story — and a system that closes the job on your behalf is a system that hides it.
-
A saved ticket is final, and corrections are visible
Because saving moved real stock, the ticket will not let you go back and rewrite it. A mistake is corrected the way a mistake in any ledger is corrected — with an adjustment that is itself visible — rather than by editing history until the balance looks right.
Confirmed and unconfirmed production both write forward-looking expected inventory changes as well, so what is coming is visible next to what is on hand.
The reason production reporting fails in most factories is not that the software is hard; it is that reporting is a second job, done later, by someone who was not there. Making one Save at the line do the reporting, the finished-goods posting and the component consumption at once removes the second job entirely — and sequencing rather than time-slotting is what keeps the queue matching what the line is actually running, which is the only condition under which any of it stays true.
Worth knowing: this is sequencing, not finite capacity planning. There are no work centers or machines as separate records, no capacity model and no Gantt chart. Backbone MRP does not capture labor hours, setup time or run time; labor and overhead exist as cost lines on a bill of material, not as clocked time.