Cost of poor quality: building a number that survives a finance review

Four buckets, and only two of them get measured

The conventional split is prevention, appraisal, internal failure and external failure. Prevention covers design reviews, capability studies, training and process improvement. Appraisal is inspection, test and calibration. Internal failure is what you find before shipping: scrap, rework, sorting, downgrade, retest. External failure is what the customer finds: returns, credits, warranty, sorting at their site, and the recovery activity that follows. Most plants can produce appraisal and scrap figures because both sit in a cost centre. Prevention is rarely tracked at all, and external failure is systematically understated, which is exactly backwards for decision-making.

The costs that never appear in the ledger

Capacity lost to running defective material through a constrained machine does not show as an expense, yet it may be the largest single item in a busy plant. Neither does engineering time consumed by containment instead of improvement, the premium freight booked to recover a delivery, the extra inventory held because yield is unreliable, the management hours spent in customer meetings, or the business not quoted because the customer moved. Estimating these roughly and labelling them as estimates is far more useful than omitting them, provided the estimated portion is shown separately from the ledger-based portion.

Agreeing the method with finance before publishing anything

A quality-produced cost figure that finance has not endorsed will be disputed the moment it is used to request money, and the argument will be about method rather than about the problem. Settle in advance which accounts are in scope, how scrap is valued, whether internal labour counts at full cost or at marginal cost, how lost capacity is treated, and what is deliberately excluded. Then keep the method fixed. A stable, conservative and jointly owned figure that moves with reality beats a larger figure recalculated each time somebody needs a bigger number.

The ratio that actually drives decisions

The total is a headline; the useful analysis is the balance between the buckets. Heavy appraisal spending with high external failure says the inspection is not catching what matters, so more of it will not help. High internal failure with low external says the containment works but the process does not, and the money is being burned in scrap and sorting. Low spending everywhere with rising complaints usually means the failures are being absorbed informally and never recorded. Each pattern points at a different investment, which is the whole reason for building the number.

Using it to win an argument, and its limits

The strongest use is a specific case: this defect cost this much last year across these accounts, this change costs less, here is the evidence it works. That is far more persuasive than an aggregate quality cost presented annually, which tends to prompt a target rather than a decision. Be aware of the perverse incentive too. Once the figure becomes a performance measure, the reliable way to reduce it is to record less of it, so any downward movement should be checked against complaint volume and scrap quantity before it is celebrated.

Frequently asked questions

How should scrapped work in progress be valued?

At the cost accumulated up to the operation where it was scrapped, including material, processing and any purchased content, rather than at raw material value or at selling price. Material-only valuation makes late-stage scrap look trivial and hides the argument for detecting defects earlier. Selling price inflates the figure and invites finance to reject the whole exercise. Where the plant is capacity-constrained, the lost contribution from the occupied machine time deserves a separate line.

Is there a benchmark level of quality cost we should aim for?

Published comparisons are unreliable because the scope of what each organisation counts varies enormously, and a low reported figure often reflects narrow accounting rather than good performance. Your own trend on a fixed method is the only comparison worth making. If an external reference is needed, compare the shape — the proportion in external failure versus prevention — rather than the absolute level against turnover.

Who should own the calculation?

Finance should own the numbers and quality should own the categorisation, meeting on a fixed cycle to produce one agreed statement. Quality-owned figures are dismissed as advocacy; finance-owned figures miss the operational categories that make the data actionable. The joint version also survives a change of quality manager, which purely departmental analyses rarely do, and it puts the discussion in front of the leadership team as a business number rather than a departmental report.