For consumer hardware companies
At a hundred units a cent per part is noise. At two hundred thousand it is a salary. Consumer hardware is the segment where engineering decisions and cost are the same conversation.
Cost visibility where the decision happens
The design decisions that determine unit cost are made months before anybody negotiates a price, by engineers choosing parts. By the time procurement is quoting, the cost is largely fixed and the remaining work is trimming a few percent off a number somebody else set.
So cost belongs on the screen where the part is chosen. Manufacturing PLM shows catalogue price at quantity beside a component during selection, and your ERP's negotiated price where one exists — labelled by source, never averaged, because those are different facts and an engineer needs to know which one they are looking at.
A rolled-up estimate on a resolved structure is the useful artefact. Not a costed BOM maintained in a spreadsheet, which is stale the week it is written, but an estimate resolved through the same engine as the structure, at the volume you specify, labelled as an estimate.
The reason it must be labelled as an estimate is that a number carrying more authority than it earned is worse than no number. Catalogue price at ten thousand is not what you will pay, and a system implying otherwise will be believed once and distrusted permanently.
The contract manufacturer relationship
Consumer hardware is built by somebody else, usually far away, frequently across a time zone gap that makes every clarification cost a day. The quality of the handoff is the quality of the programme.
What travels badly is everything that is not in the released package: the note about which side of the tolerance to favour, the alternate that was approved verbally, the reason a component cannot be substituted despite looking substitutable. Every one of those becomes an email, and emails do not reach the next person to do the job.
The supplier portal carries the released package — structure, drawings, specifications, AML with substitution rules — as controlled copies that supersede themselves. The CM sees what is current because what they see is what is current, rather than because somebody remembered to send an update.
Deviations matter more here than teams expect. A CM finding an out-of-tolerance batch at two in the morning their time will make a decision; the question is whether that decision is recorded as a deviation with a bounded quantity or lives in a message thread that nobody keeps.
The failure it prevents
A product goes to volume with a cost model built during design. Six weeks in, actual cost is eleven percent above model and margin is gone.
The cause is four components substituted during the ramp for availability, each individually approved, each slightly more expensive, none of them reflected in the model — because the model was a spreadsheet built once from a BOM snapshot and never resolved again.
No single decision was wrong. Four reasonable substitutions compounded past the point where the product made money, and the mechanism that would have caught it — a cost estimate resolved from the current structure rather than from a six-month-old copy — did not exist.
How it meets the rest of the product
Part reuse search is worth more in this segment than any other. A component already qualified, already in the AML and already bought at volume is cheaper than an equivalent new one before anybody negotiates, and finding it requires classification-based search rather than description matching.
The change calendar carries the NPI cadence, which in consumer hardware is unusually date-driven — EVT, DVT, PVT and a ship date that does not move. Changes cluster around those gates, and seeing them against the gate they belong to is how a programme stays legible.
Compliance rollup handles the certifications volume products need — CE, FCC, UKCA, RoHS, REACH — resolved through the structure rather than tracked per product in a document, so a substitution that affects a declaration is visible when it is made rather than when a customer asks.
How it meets your ERP
The ERP owns cost and always should. Manufacturing PLM reads it, shows it beside the engineering decision, and never authors it — because a cost figure engineering maintains is a cost figure that disagrees with the one finance reports on.
Where an ERP purchase history read exists, actual paid price is the most valuable number on the screen. An engineer choosing between two components with a catalogue price difference of four cents and an actual paid difference of eighteen is making a materially different decision with the second number.
Publication timing matters at NPI pace. Structures change fast between gates, and publishing every change to the ERP produces churn nobody wants — so publication is tied to gate releases rather than to every revision, which is a configuration decision made once per programme.
Where the boundary is
Manufacturing PLM does not cost your product. It shows catalogue and ERP-sourced values, rolls them up through a resolved structure and labels the result an estimate. Standard costing, landed cost, overhead allocation and margin analysis are your ERP's and finance's work.
It also does not manage your CM relationship commercially. Quoting, purchase orders, capacity commitments and payment are elsewhere; what lives here is the definition the CM builds to and the record of what was agreed about it.
Facts
| Where cost is decided | Part selection, months before anybody negotiates |
| Shown at selection | Catalogue price at quantity · ERP negotiated price, labelled by source |
| Never | Averaged into one number |
| Rollup | Resolved through the structure, labelled an estimate |
| CM handoff | Controlled copies that supersede themselves |
| Deviations | Bounded and recorded, not a message thread |
| Highest-value read | Actual paid price from your ERP |
| Publication | Tied to gate releases, not to every revision |
Frequently asked
Why show cost during part selection?
Because that is when unit cost is decided. By the time procurement is quoting, cost is largely fixed and the remaining work is trimming a few percent off a number engineers set months earlier without seeing it at the moment they set it.
Which price do engineers see?
Both catalogue price at quantity and your ERP's negotiated price where one exists, labelled by source and never averaged. They are different facts, and an engineer choosing between components needs to know which of the two they are actually looking at.
Is the rolled-up cost reliable?
It is an estimate and labelled as one, resolved through the same engine as the structure at the volume you specify. Catalogue price at ten thousand is not what you will pay, and a system implying otherwise gets believed once and distrusted permanently afterwards.
What goes wrong during a ramp?
Compounding substitutions. Four components swapped out for availability, each individually approved and each slightly more expensive, none of them reflected in a cost model that was a spreadsheet built once from a snapshot. No single decision is wrong and the margin disappears anyway.
How does this help with our contract manufacturer?
The released package — structure, drawings, specifications, AML with substitution rules — reaches them as controlled copies that supersede themselves. They see what is current because what they see is current, rather than because somebody remembered to send an update.
Why do deviations matter so much here?
Because a CM finding an out-of-tolerance batch at two in the morning their time will make a decision regardless. The question is only whether it is recorded as a deviation with a bounded quantity or lives in a message thread that nobody keeps afterwards.
Does Manufacturing PLM replace our costing?
No. It shows catalogue and ERP-sourced values and rolls them up through a resolved structure as a clearly labelled estimate. Standard costing, landed cost, overhead allocation and margin analysis all remain the work of your ERP and of your finance team entirely.