Material loss in production is not a defect, it is physics and process capability. Sheet metal has offcuts, chemical conversion has losses, electronics assembly has test failures, machining generates chips. The commercial question is not how to eliminate loss but who pays for it, and that question is answered by a clause most agreements handle in two sentences: an expected yield percentage and a statement that loss above it is the converter’s responsibility.
Two sentences is not enough, and the reason becomes clear at the first reconciliation. Expected yield measured against what: material issued, material theoretically required, or good units delivered? Over what period, given that a single batch can swing well outside a monthly average? At what value, given that material cost moves and the loss might be valued at contract price, replacement cost or standard cost? And what happens to the scrap itself, which in many processes has real recovery value that somebody is currently keeping?
Where the clause is vague, the practical outcome is that allocation is settled by negotiation each period, weighted toward whichever party has better data and more leverage. That is an expensive way to run a recurring commercial calculation, and it obscures the more useful signal, which is whether actual yield is trending against expectation in a way that should change the price, the process or the supplier. Index-linked price adjustment clauses fail the same way when the formula is left implicit.
What the clause has to specify
The elements below are the minimum for a clause that produces the same answer regardless of who calculates it. Each maps to a dispute that arises when it is absent.
- Expected yield and measurement basis, the yield figure plus exactly what it is measured against, stated as a formula with named inputs rather than as a percentage alone.
- Reconciliation period and cadence, the settlement window, whether it is a rolling average, and who produces the reconciliation and by when.
- Treatment above and below expectation, who bears excess loss, whether better-than-expected yield is shared, and any dead band where no adjustment occurs.
- Valuation method, contract price, standard cost, replacement cost or landed cost, and the date at which the value is fixed.
- Scrap revenue and reclaimed material, who owns scrap and recovered material, whether proceeds offset allocated loss, and the reporting required.
Those five turn allocation into a calculation both parties can run independently and get the same number from. That is the practical test of the clause: if two competent people with the same records produce different answers, the clause is underspecified. The same test applies to any formula-driven contract term, including the evidence expectations in an IATF 16949 contract records review.
Applying a steady-state yield expectation during launch or after a significant engineering change allocates loss the converter had no realistic way to avoid, which either gets absorbed into an inflated steady-state number or becomes a claim. Agree a stepped yield curve with dated milestones for ramp-up and post-change periods, and track the milestones as contract obligations.
Where yield allocation goes wrong
These are the recurring problems, and all of them are contract design issues that surface as commercial disputes.
- Measurement basis is implicit. Buyer calculates against theoretical requirement, converter against material issued, and both believe the contract supports them. The difference can be several percentage points.
- Reconciliation never happens on schedule. Without a dated obligation and a named producer, reconciliation slips until enough periods accumulate that reconstruction is impractical.
- Scrap revenue is unaddressed. In processes with meaningful recovery value, silence means the converter keeps it, which effectively reduces the loss it is bearing without the buyer knowing.
- Engineering changes reset yield without a contract step. A process or design change alters achievable yield, but the contracted expectation stays, so the number becomes fictional until renegotiated.
- Buyer-caused loss is not carved out. Material supplied out of specification, late design freeze or incorrect tooling cause loss the converter should not bear, and the carve-out is usually missing.
A five-step roadmap to make yield allocation work
Step 1, Write the yield formula, not the percentage
Express expected yield as a formula with named inputs and their sources, so the calculation is reproducible. Include worked examples in the schedule, which resolves more future argument than additional clause language.
Step 2, Set the reconciliation as a dated obligation
Define who produces the reconciliation, by when after period end, with what supporting records, and who reviews it. Track that as a recurring contract obligation with an owner rather than as a finance routine.
Step 3, Define the carve-outs for buyer-caused loss
Enumerate the causes that shift loss back to the buyer: out-of-specification supplied material, late design or tooling changes, specification ambiguity. Require the converter to notify and quantify within a window, or the carve-out lapses.
Step 4, Settle scrap and recovery rights explicitly
State who owns scrap and reclaimed material, whether proceeds offset allocated loss, and what reporting is required. In precious and critical metals this term is frequently worth more than the yield percentage itself.
Step 5, Trigger a yield review on change
Make any engineering change, material substitution or tooling change a trigger for reviewing the yield expectation, with a defined process and window. Otherwise the contracted number drifts away from process reality and stops being used.
Keep yield terms measurable and reconciled on time
Aavenir tracks reconciliation duties, yield review triggers and ramp-up milestones as owned obligations against the supply agreement.
Why the clause is a performance signal, not just a cost allocation
Treated well, yield allocation does more than assign cost. A reproducible yield calculation run on a fixed cadence produces a time series that shows whether a converter’s process capability is improving, stable or degrading, and whether the contracted expectation still reflects reality. That is useful in three places: supplier performance review, where yield trend is a harder measure than most scorecard metrics; price negotiation, where a converter consistently beating expectation is earning a margin the buyer is funding; and sourcing decisions, where yield differences between suppliers on comparable work are a genuine capability signal. None of that is available when allocation is settled by negotiation each period, because negotiated outcomes reflect leverage rather than process. Buyers running contract lifecycle management for manufacturing usually feed the reconciliation output straight into supplier review.
Yield clause components and controls
The table sets out each component with the specification needed and the control that keeps it current. Appropriate terms depend heavily on the process, the material value and the commercial structure.
| Component | What to specify | Control | Review trigger |
|---|---|---|---|
| Expected yield | Formula with named inputs and worked example | Schedule referenced by the agreement | Engineering or material change |
| Reconciliation | Producer, deadline, supporting records, reviewer | Recurring dated obligation with owner | Missed period or disputed result |
| Loss above expectation | Bearing party, dead band, cap if any | Calculated in reconciliation | Sustained variance beyond dead band |
| Valuation | Cost basis and valuation date | Fixed in schedule, sourced from price list | Material price movement beyond threshold |
| Scrap and recovery | Ownership, proceeds treatment, reporting | Periodic recovery report as obligation | Recovery value change or process change |
Yield loss allocation is a recurring, calculable commercial outcome that most contracts leave as a recurring negotiation. The fix is specification rather than length: a formula instead of a percentage, a dated reconciliation with a named producer, explicit carve-outs for buyer-caused loss, and a settled position on scrap recovery. Once the calculation is reproducible, the argument mostly stops and a useful performance signal appears in its place. That signal is worth more than the disputed percentage points.
For related guidance, see How to audit sub-tier supplier flowdown requirements, IATF 16949 supplier quality agreements, a control guide and Supplier quality agreement and PPAP commitment tracking.
