Cost-reduction programmes rarely announce themselves as quality risks. They arrive as engineering change requests, supplier proposals, value-engineering workshops and lean sourcing decisions — each individually defensible, each signed off by people intent on protecting the product. Across two decades in automotive and aerospace, I have learned that quality erosion from cost-down activity is almost never a single bad decision. It is the accumulation of small, unverified departures from the validated baseline, each of which looked harmless on the day it was approved.
The pattern is consistent across industries. A resin is swapped for a cheaper grade with equivalent data-sheet properties. A heat-treatment cycle is shortened. A fastener supplier changes its wire stock. A gasket compound is reformulated. Six months later, field failures appear that nobody can trace to a single root cause, because the product on the shelf no longer matches the product that was originally qualified.
This article covers the three mechanisms — unapproved changes, material substitution, and unearned approval propagation — and, most importantly, how to detect them before your customers detect them for you.
How Unapproved Changes Actually Happen
The majority of unapproved changes are not deliberate deception. They are process failures dressed up as commercial decisions. Purchasing finds an alternative source for a raw material at a better price and issues the change without looping in engineering, because the specification looks the same. A supplier changes its sub-tier — a new plating house, a different compounder — and sees no obligation to notify, since the drawing requirements it holds are unchanged. The change is real, the notification is absent, and the validation status of your product quietly expires.
The second pathway is the informal trial. A production engineer runs a batch with a modified parameter to see whether cycle time improves. The trial works, scrap looks fine, and the new setting becomes the de facto standard without ever passing through change control. Nobody documents it, so nobody tests it. Then a machine is rebuilt, an operator retires, or a different shift encounters a corner of the process window the trial never explored — and the failures begin.
The third pathway is the slow drift of legitimately approved changes. A change is approved with a caveat — approved for the current climate-chamber qualification only — and the caveat evaporates as the change propagates to other part numbers, other plants, other customers. Each downstream application inherits an approval it never earned. When you audit years later, you find the same substitution sitting under part numbers whose requirements are far more demanding than the original case that justified it.

Material Substitution: Where Equivalence Claims Break Down
Equivalence claims are where technical judgement fails most often. Data sheets compare nominal properties — tensile strength, melt flow index, hardness — but the properties that govern field performance are frequently the ones not printed on the sheet. Impact behaviour at low temperature. Fatigue response under vibration. Chemical resistance to the specific fluids the assembly actually sees. A replacement polyamide with identical strength at room temperature can embrittle dramatically at minus thirty, and nothing on the comparative data sheet would tell you.
Metals carry the same trap. A substitution within the same nominal grade can shift cleanliness level, inclusion morphology, grain size after forming, or residual stress distribution after machining. Weldability, hydrogen-embrittlement susceptibility in plated parts, and thread-rolling performance all live in these details. I have seen fastener batches from an alternate mill pass every dimensional check and tensile test, then fail torque-tension consistency so badly that clamp-load scatter doubled in the assembly — a failure mode invisible until you measure the friction coefficient, which nobody had specified.
Elastomers and adhesives are worse still, because their properties depend on cure kinetics, ageing behaviour and compound formulation details that suppliers treat as proprietary. A chemically equivalent seal compound from a second source can differ in plasticiser content, and the difference surfaces two years later as compression set or plasticiser migration into an adjacent plastic part causing environmental stress cracking. Equivalence must be demonstrated against the application's failure modes, not against a data sheet's list of tested properties.
Data-sheet equivalence versus application equivalence
What the data sheet proves
- Nominal tensile strength and hardness at room temperature
- Melt flow index and density within grade limits
- Dimensional conformance to drawing
- Same nominal grade designation
What the application demands
- Impact toughness at service temperature extremes
- Fatigue life under real vibration spectra
- Chemical resistance to actual assembly fluids
- Friction behaviour and clamp-load consistency
Detection: Watching the Physical Baseline
Detection starts with having a baseline to compare against. If you hold no retained physical record of what good material looks like, you cannot recognise when it changes. Retain samples of incoming material lots for safety-critical and field-failure-prone commodities, stored under controlled conditions. When a suspicious failure appears, compare today's material against the qualified reference using the same analytical techniques on both. Without the reference, you are arguing about an unanswerable question.
Analytical techniques matter more than paperwork. Fourier-transform infrared spectroscopy flags a polymer compound change in minutes, even when the grade designation is identical. Differential scanning calorimetry reveals shifts in melting behaviour and crystallinity that accompany resin substitution or regrind abuse. Optical emission spectroscopy confirms alloy identity on incoming metals. Microhardness traverses across case-hardened surfaces expose shortened carburising cycles. Density, melt-flow checks and ash content each catch specific substitution classes. The point is not to test everything constantly — it is to sample intelligently and escalate whenever commercial signals suggest a change is likely.
Those commercial signals are your best early warning. A supplier wins new business at a price nobody can explain. A purchase-price drop arrives with no engineering change notification. Delivery lead times suddenly improve on a material that was previously scarce. Each is a flag to pull a sample and run the comparison suite. Treat purchasing data as a detection input, not just a financial one: the quality department that reads commodity price movements will often know a substitution risk exists before the supplier admits it.
What to Check in the Change Process Itself
Change control needs teeth, and teeth means verification, not approval signatures. Every proposed change touching material, process parameters, sub-tier sources or geometry should carry a verification package proportionate to risk: requalification tests, or a documented technical rationale for why the existing qualification remains valid. The rationale is the weak point. Review a sample of recent changes and count how many justifications say form, fit and function unchanged with no supporting evidence. That phrase, unsupported, is where silent erosion lives.
Audit your bill of materials against physical reality. A disciplined approach is periodic teardown and material verification of current production parts: take finished goods off the line, disassemble, and confirm materials and components match the released specification using the analytical methods above. Discrepancies between the documented BOM and the physical product are direct evidence of unapproved change. I have never run such an exercise on a mature product line and found nothing.
Supplier change-notification obligations belong in the contract, with flow-down to sub-tiers and with consequences. Beyond the contract, maintain an open channel with supplier engineering, not only purchasing, because engineers at the supplier often know about process tweaks the commercial relationship never surfaces. Internally, trend your own parameter records: process monitoring data over months will show the de facto changes — cycle times creeping down, temperatures shifting, torques drifting — long before anyone formally requests approval for them.
Escalating verification when commercial signals appear
- 01Commercial triggerUnexplained price drop, sudden availability, or new low-cost source on a qualified material.
- 02Pull retained referenceRetrieve the qualified baseline sample from controlled storage.
- 03Run comparison suiteFTIR, DSC, OES or hardness traverse — matched techniques on both samples.
- 04Review change recordsCheck whether any ECN covers the difference the lab just found.
- 05Contain or acceptBlock the lot pending requalification, or document evidence-based equivalence.
The Governance Failure Behind Most Erosion
Underneath the technical mechanisms sits a governance problem. Cost-down programmes are measured on savings realised; quality is measured on defects found. The gap between those two measurement systems is precisely where erosion occurs. Savings are booked in the quarter the change is implemented. The failures arrive one, two, three years later, by which time the connection is analytically difficult to prove and politically inconvenient to pursue. Everyone involved in the original decision has moved on to the next programme.
Savings are booked in the quarter the change ships; the failures arrive years later, when the connection is analytically hard to prove and politically inconvenient to pursue.
Closing that gap requires linking every realised saving to its verification record, so that when field failures emerge the first question is: what changed, when, and what was done to confirm it was safe? That question must be answerable within hours, not weeks. An organisation that cannot answer it will spend the weeks anyway — in containment, in customer escalations, and in expensive root-cause investigations that end at the same door.
The practical instruments are familiar: IATF 16949 and AS9100 both require control of changes and verification of their effects, and PPAP exists precisely to force requalification evidence when a process, material or location changes. The recurring failure is not the absence of requirements but the absence of a register that connects each cost saving to its verification artefact. Build that register, audit it quarterly, and the governance gap narrows to a management task.
Treating Every Saving as Guilty Until Qualified
My rule after years of watching this cycle: treat every cost-saving change as guilty until proven qualified. It is a sceptical stance, and suppliers occasionally find it tiresome, but the alternative is discovering in the field, at the customer's expense, that equivalent was a claim rather than a fact. The burden of proof sits with whoever proposes the change, not with whoever finds the failure.
The economics always run in the same direction. The savings arrive first; the failures come later, with interest. Every mechanism in this article — retained baselines, analytical spot checks, commercial-signal monitoring, teardown audits, savings-to-verification linkage — exists to shorten that time window. Shorten it enough, and the substitution is caught at receiving instead of at the customer.
Quality erosion from cost-down activity is not fate. It is the predictable output of a measurement system that books savings immediately and discovers costs later. Change the measurement system, and you keep the savings without funding the interest.
