Choosing the System of Record for Each Scheduling Field

Before your second ERP import, decide which system owns each scheduling field and write it down: fields the ERP owns are mapped in the mask and never edited by hand in EDGEBIC, and fields EDGEBIC owns are left out of the mask entirely so no import can touch them. The rule sounds administrative. It is the difference between a set of machine counts and setups that hold for years and a set that silently reverts every time somebody runs the nightly file.

EDGEBIC by User Solutions has integrated with ERPs through import masks since 1991, and across 35+ years of scheduling for the US Navy, GE, BAE Systems, and Cummins, ownership drift is the failure that costs the most trust. It never announces itself. A value a planner set deliberately is simply not there next month, and the dates that follow from it are wrong in a way that looks like the scheduler's fault.

The mechanism you are governing

An import writes only the fields you mapped. Everything else on the record is untouched, however many times the mask runs. That single behavior is what makes ownership enforceable, because unmapping a column is a permanent decision rather than a habit somebody has to remember.

There is a second, softer layer. On an update run, a blank cell keeps the existing value rather than wiping it, so a partial refresh file is safe. Useful, but do not build ownership on it: a zero is a value, not a blank, and an ERP that emits 0 in a column you meant to ignore will happily write that zero. Blank-cell preservation protects you from gaps in a file. Unmapping protects you from the file.

The three categories

Every field a schedule needs falls into one of three buckets.

ERP-owned. The ERP maintains it as a matter of daily business, so importing it keeps the two systems aligned for free.

EDGEBIC-owned. The ERP does not hold it, or holds it at a grain that cannot drive capacity. It is configured once and excluded from every mask.

Negotiated. Both systems have an opinion and you have to pick, per shop.

Category 1: fields the ERP should own

Map these and leave them alone inside EDGEBIC. The corollary matters: if a routing is ERP-owned, do not hand-edit it locally. A routing import wipes and recreates that product's steps once per run, so a local edit disappears at the next import with no message. The change workflow is in handling engineering changes with an ERP routing re-import.

Category 2: fields EDGEBIC should own outright

These have no export column anywhere, in any ERP, at the grain a finite capacity engine needs.

• Machine instance count. How many identical machines a work center really holds. Most ERPs model a work center as one resource with a capacity figure, which is why this is the highest-value field on the whole load and the most dangerous one to leave mapped.
• Shifts and plant holidays. The calendar every scheduled date is computed against.
• Sequence-dependent setup matrix. The changeover cost that depends on what ran before. No ERP carries a from-product-to-product matrix. See the setup matrix explained.
• Work center groups. A named pool of interchangeable machines, re-shopped on every reschedule, with per-member efficiency factors.
• Operator skills and rosters. No routing export models a certification.
• Bottleneck flags, transfer batches, and queue policy. Planner judgment about how the plan should be built.

The full tiering of what an ERP can and cannot supply is in which ERP fields EDGEBIC needs to schedule.

The instance count, in numbers

This is the field worth being pedantic about, because the failure is invisible and expensive.

A cell of four identical mills, imported as one machine, produces a plan roughly four times too long:

Nothing errors. The dates are simply pessimistic, and the planner who fixed them last month has no reason to check whether they are still fixed. Unmapping the column is a thirty-second job that holds forever.

Category 3: the negotiated fields

Three fields genuinely depend on how your shop works.

Priority. If your ERP carries a meaningful order priority that customer service maintains, import it. If priority is really a conversation the planner has each morning, keep it local and set it in EDGEBIC.

Efficiency and default setup on a work center. Some ERPs hold real numbers here; many hold a default nobody has revisited since implementation. Import them only if you trust them, and unmap them the moment a planner starts correcting them by hand.

Progress. Exactly one system nets completed work: either actual hours reach EDGEBIC and the order import carries the original quantity, or the ERP nets the quantity and no actuals arrive. Doing both subtracts twice. The full case is in reconciling ERP work order quantities after partial completion.

Write the decision down where the mask lives

Name your masks after the routine and the ownership rather than the file: Weekly Work Centers (ERP-owned columns only) outlives wc-final-v3.csv. Then keep a one-page table beside the masks listing each entity, the fields mapped, and the fields deliberately excluded with a short reason.

That page is what makes the decision survive the person who made it. A year later, somebody looking at a work center mask with no capacity column mapped needs to know that was on purpose.

Signs the ownership has drifted

Three symptoms show up before anything looks obviously broken:

• A value you set has reverted. Almost always a mapped column you meant to exclude.
• A high Created count on a routine import of stable master data. Usually identifier drift rather than ownership, and covered in keeping product masters in sync.
• Two people describe the same field differently. The planner says setups are maintained in EDGEBIC, the ERP team says they come from the routing. One of them is about to be surprised.

Catch these on a cadence rather than by accident. The per-import pass is in the import reconciliation checklist, and the slower drift check is in a monthly ERP to EDGEBIC data audit.

What ownership buys you

A settled ownership map is what lets the recurring routine stay short. Export, pick the mask, run, read the counts, schedule. Nobody re-checks instance counts, nobody re-enters setups, and the values that encode your shop's real knowledge accumulate instead of eroding. On top of that stable base the engine does its work: finite capacity across shifts and machine instances, work center groups, sequence-dependent setups, lot streaming with transfer batches, and mathematical optimization with a proven optimality gap. The EDGEBIC product overview maps the engine, and the ERP integration architecture shows the import layer every ERP shares.

Draw your own map

Bring one product export, one work center export, and one routing export to a demo and mark each column ERP-owned, EDGEBIC-owned, or negotiated. It usually takes twenty minutes and it is the cheapest hour of integration design you will spend.