Production Data Breaks the Moment Work Leaves the Plant
On most shop floors, plating, heat treatment, surface finishing, and painting are outsourced. Given the capital cost of the equipment and the expense of meeting environmental regulations, that is a sound decision. The problem is that production data goes dark across those steps.
The typical break looks like this. Outbound shipments to the subcontractor are handled as ERP issue slips, while returns are logged by hand or in a spreadsheet on the floor. When defects appear, someone calls the subcontractor, and nothing from that call is recorded anywhere. The result is that nobody knows, in real time, how much WIP is sitting outside the plant or how far along it is.
That gap shows up first in delivery management. If heat treatment has a contractual lead time of three days but actually takes five to seven, and all you have is a ship date and a receipt date, you cannot determine after the fact whether the delay came from the subcontractor's loading schedule or from your own late shipment. One precision-parts manufacturer we worked with was classifying more than half of its average twelve monthly delivery delays as "cause unknown." Only after a tracking layer went in did it become clear that seven of them traced back to one subcontractor's day-of-week loading imbalance.
Where Consigned Material Accounting and Inventory Come Apart
Consigned material is where outsourced processing goes wrong most often.
Ownership versus inventory recognition is the first trap. With free-issue consignment, the buyer retains title to the material, so it must remain on the books as an asset even while it sits offsite. Yet many plants simply deduct it from inventory at the moment of shipment. A large share of the book-versus-physical discrepancies found during semiannual counts originates right here. Paid consignment works the opposite way — it generates a sale and a purchase, so the tax treatment differs. If the system does not distinguish the two types, accounting and production end up looking at different numbers.
Processing loss and scrap allowances are another item that often lives only in the contract and never in the system. If 1,000 pieces go into plating and 985 come back, the actual loss rate is 1.5% — within a 1.5% contractual allowance. Let the same order come back at 975 pieces, and the 2.5% loss rate becomes a claim. If someone is running that judgment on a calculator every time, disputes are only a matter of time. Loss rates should be maintained as master data keyed by item, operation, and subcontractor, with automatic evaluation at the moment of return receipt.
Varied settlement bases also have to be reflected in the data model. Plating is commonly billed by area (dm²), heat treatment by weight (kg), and assembly subcontracting by piece count (EA). Build the price table around piece count alone and everything else leaks back into spreadsheets. At minimum, the price master needs a billing UOM column linked to the unit conversion data on the item master.
Bringing Outsourced Operations into the MES Routing
The governing principle is that an outsourced step is modeled as an operation like any other. Leave it out of the routing and manage it separately, and your progress calculations will always be wrong.
In practice, three transactions carry the load.
Preserving lot traceability is the second challenge. If you cannot connect the lot that left with the lot that came back, backward traceability breaks the next time a customer files a claim. A workable approach is to use the subcontract order number issued at shipment as the lot key, and to record parent-child relationships at return whenever the subcontractor merged or split lots. In processes like heat treatment, where multiple orders routinely go into a single batch, merging is the norm — so the table design has to permit N:M relationships from day one.
The subcontractor interface calls for pragmatism. Demanding dedicated terminals from small shops with no appetite for new systems will stall the project. We recommend preparing three tiers.
A mixed model — tier two or three for high-volume partners, tier one for low-volume, high-mix shops — is the realistic outcome.
Turning Quality Accountability and Claims into Data
Using incoming inspection results only for a single pass/fail decision wastes them. That data should accumulate by subcontractor and become the basis for reorder decisions.
Assigning accountability for defects found at downstream operations requires a minimum evidence set: the in-house inspection record at shipment, the subcontractor's own inspection record, the incoming inspection record, and lot photos or measurements from each point. When those four are linked by lot key, "it was already defective when you shipped it" and "it happened during processing" become claims you can settle with data.
Design subcontractor scorecards along three axes.
| Metric | Formula | Use |
|--------|---------|-----|
| On-time rate | Returns within promised date / total returns | Adjusting order allocation ratios |
| Incoming defect rate | Rejected quantity / returned quantity | Price negotiation, inspection intensity |
| Claim resolution lead time | Days from claim notice to rework return | Assigning expedited volume |
Grading subcontractors A/B/C on these three metrics each quarter puts evidence behind allocation decisions that were previously made on instinct.
Rollout Sequence and the Investment Case
Attempts to onboard every subcontractor at once usually fail. Since the top 20% by volume typically accounts for 70–80% of spend, starting the pilot with the top three to five partners is the sensible move.
Approach matters too. Leaving the existing ERP purchase and settlement process untouched and layering tracking on top of it carries far less risk. If the design reads ERP purchase data to generate subcontract orders and writes return performance back as ERP receipt slips, you gain visibility without touching accounting logic. Full replacement is a decision that can wait until the pilot has proven itself.
Build the investment case backward from recoverable losses. At a company spending ₩200 million a month on outsourced processing, leaking just 0.5% to missed loss evaluations and settlement errors costs ₩1 million a month — ₩12 million a year. Add expedited freight and line idle time caused by delivery delays, and payback generally lands inside a year.
POLYGLOTSOFT Is Here to Help
POLYGLOTSOFT builds outsourced processing traceability step by step — from MES routing design to subcontractor portals and ERP integration. We do not ask you to specify a finished system up front. Under our subscription development model, we recommend running a pilot with your top subcontractors first, then widening the partner and feature scope as real operating requirements surface. Because a dedicated development team is attached for a flat monthly fee, improvement requests from the floor can go into the next sprint immediately. If losing sight of WIP once it leaves the plant is a concern, please get in touch. We will start with a review of your current process.