Turnkey PCBA Material Attrition and Overage: Who Owns the Extra Parts

The first turnkey reconciliation statement is where most buyers discover that they never actually agreed on a quantity.

The quote said 500 assemblies. The invoice says 500 assemblies. But alongside it there is a material line that does not reconcile against the BOM you sent, and a short note about attrition. Somewhere in a warehouse there are parts you paid for that are not in your boxes. Nobody is being dishonest. The contract simply never said which of three possible quantities it was pricing.

Three quantities, none of them equal

A turnkey PCBA job counts the same component three times, and the three numbers are structurally different.

BOM quantity is the theoretical demand: quantity per board, multiplied by the number of boards ordered. It is the only number in your purchase order, and it is the only one of the three that never gets consumed in reality.

Kitted and issued quantity is what actually goes to the line. It is BOM quantity plus the material the process consumes without placing on a shippable board: machine setup and feeder priming, first-article boards, parts lost to pick errors and tape splices, panel units scrapped at any stage after placement, and whatever has to be issued to cover rework. This number is always higher than BOM quantity, and how much higher depends on the component package, the placement count, the panel array format and the batch size, not on anyone's goodwill.

Purchased quantity is what the contract manufacturer actually bought. It is the issued quantity rounded up to how distributors sell: full reels, full trays, minimum order quantities, minimum packaging quantities. A part needed in a quantity of a few hundred may only be purchasable as a reel of several thousand. This number is often dramatically higher than the other two, and the rounding is imposed by the supply chain rather than chosen by the assembler.

Every disagreement about turnkey material cost lives in the two gaps: issued minus BOM, and purchased minus issued. The first gap is attrition. The second gap is overage. They have different causes, different magnitudes and different natural owners, and a contract that treats them as one topic will settle both of them by default in the supplier's favour, because the supplier is the only party holding the numbers.

Gap one: attrition, and what actually drives it

Attrition is consumed material. Once it is gone, the only question is who pays for it, and that question is answerable in advance because the drivers are known in advance.

Setup and priming. Every feeder that gets loaded consumes a leading length of tape before the machine reaches a reliable pick. This cost scales with the number of distinct part numbers and the number of machine setups, not with your order quantity. On a small build with a long BOM, setup attrition can be the dominant component of the whole attrition figure. On a large repeat build it is nearly irrelevant per unit.

First article and process proof. Boards built and then sectioned, tested destructively or simply held as reference. These are real assemblies that consume a full board's worth of every part on the BOM. The quantity is a decision, not an accident, which means it can be stated in the contract.

Placement losses. Pick failures, parts dropped in transit through the machine, tape splice losses at reel changes. These scale with total placement count and with package type. Very small passives in large quantities behave differently from a handful of large connectors.

Post-placement scrap. Any panel unit rejected after parts are on it takes its full component content with it. This is the attrition item most sensitive to your own data quality, because footprint and polarity problems in the BOM and CPL produce scrap that nobody budgeted for.

Rework and replacement. Parts removed during rework are rarely reusable. Reballing or reusing a removed part is a decision that needs explicit permission, and the default answer should be no.

The practical consequence: an attrition allowance stated as one flat percentage across an entire BOM is a fiction that will be wrong in both directions. It overcharges you on the handful of expensive, low-count, hand-placed parts and undercharges on the high-count passives. Ask for attrition to be declared by category, not by one global number.

Gap two: overage, and the four ways it gets declared

Overage is not consumed. It exists, physically, as parts you paid for and did not receive. There are only four honest ways a contract can handle it, and the difference between them is worth more than most unit-price negotiations.

Declaration A: buyer pays full purchased quantity, parts belong to the buyer. You are charged for the whole reel. The remainder is your property, held at the assembler's site or shipped to you with the order. This is the cleanest option and it is the right one when a repeat order is likely, because the remainder is pre-positioned inventory for the next build.

Declaration B: buyer pays only issued quantity, assembler absorbs and retains the remainder. You pay for what went into your boards. The assembler keeps the rest and carries the holding risk, usually reflected in the unit price. Clean, as long as it is written down, and appropriate for commodity parts that the assembler will genuinely consume on other jobs.

Declaration C: buyer pays full purchased quantity, assembler retains the remainder. You pay for the reel, and the reel stays with the assembler as their stock. This is the arrangement that generates every turnkey dispute, and it is almost never written down explicitly. It arises by silence: the quantity gap is billed, ownership is never addressed, and the parts quietly become the assembler's inventory.

Declaration D: remainder is returned or credited at a stated point. Either physically shipped back with the order, or held for a stated period against a named follow-on order and credited if that order does not materialise.

Any of A, B or D can be fair. The failure mode is C by default. The test is simple and you can apply it to a quotation you are holding right now: for a part where the minimum purchasable quantity visibly exceeds your demand, can you point at a sentence that says what happens to the difference? If not, you are in C.

The parts where the gap is structural, not marginal

Not every line on your BOM needs this treatment. Overage exposure concentrates in a predictable set of cases, and those are the lines to argue about.

Parts sold only in full reels when you need a few hundred pieces. Parts with a distributor minimum order quantity far above your demand. Single-source and allocated parts where the assembler has to take whatever quantity is available at the moment of purchase. Parts with a cut-tape premium, where buying the exact quantity costs more per piece than buying the reel, so buying the reel is the rational choice and the remainder is unavoidable. High-value parts, where even a small absolute remainder is a meaningful amount of money. And parts with a shelf-life or moisture-sensitivity constraint, where a remainder held for months may not be usable on your next build at all, which makes Declaration D on those lines close to worthless unless the holding conditions are also stated.

For everything else, a reasonable blanket rule is enough. Spending negotiation time on the remainder of a reel of common passives is not where the money is.

What to fix at RFQ time, before any of this is negotiable

Once the parts are purchased, your leverage is gone, because the money is already spent and the only remaining question is how to split a cost that already exists. Everything below belongs in the RFQ, not in the reconciliation call.

Require the quotation to separate the material subtotal from assembly labour and from any procurement handling charge, rather than presenting one blended per-assembly price. A blended price makes all three quantities invisible.

Require an itemised quantity column showing, per line, BOM quantity, planned issued quantity and planned purchased quantity. The assembler already computes all three to produce the quote. Asking to see them adds no work; refusing to show them is itself informative.

Require an overage declaration code per line, using the A/B/C/D framework above or the assembler's own equivalent. One global sentence in the terms is not sufficient, because the honest answer genuinely differs between a reel of resistors and a single allocated controller.

Require the attrition basis to be stated by category rather than as a single percentage, and to distinguish setup consumption from per-placement consumption, since the two scale completely differently with your order quantity.

Require a named first-article quantity, with a statement of whether those assemblies are shippable, destroyed, or held.

Require a stated disposition point for retained remainder: how long it is held, under what storage conditions, against what follow-on order, and what happens when that window expires. A remainder held indefinitely with no stated endpoint has already been written off.

Require the quote to state the data revision it was built from. Attrition and overage estimates are both derived from the BOM and CPL as submitted; a revision change after quoting invalidates both, and that is a separate conversation from the one about who owns a reel. The ownership of the parts themselves is a distinct clause, covered in our note on material ownership in PCB assembly contract manufacturing — quantity and ownership are two different questions and need two different sentences.

Contract language that closes it

Three provisions, each short. Write them into the purchase order, not the email thread.

Quantity basis. State that material is charged on a declared basis per line, that the three quantities are disclosed with the quotation, and that any purchased quantity exceeding the declared planned figure by more than a stated allowance requires written approval before purchase.

Remainder ownership. State the default declaration code that applies when a line has not been individually coded. Make the default A or B — never silence. A single sentence such as "any material purchased in excess of issued quantity and charged to the buyer remains buyer property" removes the entire Declaration C failure mode in one stroke.

Disposition and reporting. State that retained buyer-owned remainder is reported with the delivery documentation, with quantity, part number, date code and storage location, and that it is applied against the next order for the same part before any new purchase is made. Without the last clause, buyer-owned remainder sits in a bin while the next order buys another reel.

Note that approval of a substitute part interacts with all of this: a mid-build substitution can leave you owning a remainder of a part you no longer use. If your build has substitution risk, the approval route and the quantity consequence should be settled together — see our note on alternate part approval in turnkey PCBA BOM risk.

Failure modes worth recognising

Accepting one blended per-assembly price. It is the single decision that makes every quantity question unanswerable for the rest of the project.

Treating attrition and overage as one topic. They are consumed material and existing material. Conflating them means the existing material gets settled under rules written for the consumed material, which always means you lose it.

Negotiating the attrition percentage down and ignoring overage. On a build with several minimum-order-quantity parts, the overage gap can exceed the entire attrition figure. Shaving a point off attrition while leaving ownership undeclared optimises the smaller number.

Assuming paid-for means received. Payment establishes a charge. It does not establish possession or title. The two must be stated separately.

Letting the remainder be described in email. A remainder acknowledged in a message thread and not in a document has no disposition point and no reporting obligation. In practice it will be consumed, scrapped or forgotten, and no one will have done anything wrong.

Not asking on the first order. The first build is when the reels get bought. Every subsequent order inherits whatever happened the first time, including the inventory you do not know you own.

FAQ

Is it unreasonable for a contract manufacturer to charge me for a full reel when I only need part of it?

No. Buying a full reel is often the cheaper and faster option, and cut-tape sometimes carries a premium that makes partial purchase worse for you. The charge is legitimate. What is not legitimate is charging for the full reel without stating who owns the remainder. The problem is never the purchase decision; it is the silence about title.

What is a normal attrition percentage?

There is no single number that is meaningful, and treating one as a benchmark is how buyers get the comparison wrong. Attrition is driven by the number of distinct part numbers, the number of machine setups, total placement count, package mix, panel array format and batch size. A short-run build with a long BOM behaves nothing like a high-volume build with a short one. Ask for the basis and the categories rather than the percentage, then you can compare two quotations meaningfully instead of comparing two unrelated numbers.

Can I just supply the parts myself and avoid the whole question?

Consigning material moves the question rather than removing it. You then own the overage by definition, but you also own the shortage risk: if attrition exceeds what you supplied, the line stops until you ship more, and the schedule consequence is yours. Consignment also means you absorb the minimum-order-quantity rounding yourself at purchase time. It is a legitimate choice, but it is a different allocation of the same gaps, not an escape from them.

Who owns leftover parts if the contract says nothing?

In practice, whoever is physically holding them, which is the assembler. Absent a clause, there is no reporting obligation, no disposition point and no mechanism to apply the remainder against your next order. This is why the default-ownership sentence matters more than any individually coded line: it converts silence from a loss into a defined outcome.

Should remainder be shipped to me or held at the assembler?

Held, in most cases, provided the holding terms are written. Shipping reels back and forth adds freight, handling and moisture-sensitivity exposure for parts you intend to use at the same facility anyway. Held remainder is only useful if three things are stated: the holding period, the storage condition for any moisture-sensitive or shelf-life-limited part, and the rule that held stock is consumed before a new purchase. Without the third, held remainder has no practical value to you.

Does this change for a repeat order?

Yes, and in your favour if the first order was documented. Setup attrition is largely a per-build cost that does not shrink, but overage on a repeat order should be smaller or zero, because the remainder from the first build should be applied first. If a repeat quotation shows the same purchased quantities as the first one, the remainder from build one has either been consumed elsewhere or was never recorded as yours. That is a question to ask before the second purchase, not after.

What if a part becomes obsolete while my remainder is in storage?

This is the reason a disposition point exists. Buyer-owned remainder with no expiry becomes a liability on someone's shelf, and the loss lands on whoever is deemed to own it. State what happens at the end of the holding window: returned, credited, scrapped with notification, or ownership transferred. Any of those is better than an undefined bin.

---

Related links

• PCB assembly
• quote request