How to calculate cost of poor quality without double counting

Calculate cost of poor quality by adding the costs of failures found before delivery to the costs of failures found after delivery. Count each underlying cost once. Show supplier recoveries separately, and keep routine prevention and inspection spending outside that failure-cost subtotal.

The difficult part is usually deciding whether two records describe two costs or the same cost twice. A scrap report and its accounting entry, for example, can both be correct without being two separate losses.

This guide is for a manufacturing quality manager building a monthly report with a finance colleague. The worked example uses a fictional UK workshop and pounds sterling. It is a proposed management-reporting method, not a statutory accounting, tax or certification rule. Use your organisation’s approved cost basis and investigate significant uncertainty before making financial decisions.

The American Society for Quality distinguishes prevention and appraisal from internal and external failure. Internal failures are found before the customer receives the output; external failures are found afterwards. Its examples include scrap, rework, repair, complaint handling and returns. This article uses that failure-cost definition of COPQ. ASQ cost of quality

For this method, keep four categories visible. Planned training to prevent errors goes under prevention. Routine inspection goes under appraisal. Rework before delivery is an internal failure. Repairing a defect discovered by a customer is an external failure. A label such as “quality department” is not enough to classify a cost: look at why the activity happened.

Agree how to classify exceptional sorting, repeat inspection and investigation. In the example below, a repeat inspection required because a batch failed belongs to that failure event; routine inspection remains appraisal. Record the convention so another person can apply it consistently.

Start with one site, one reporting month and one currency. Write down whether amounts come from posted costs, approved estimates or a mixture. Give estimates a separate flag and replace them with actual amounts through a visible adjustment. A missing invoice is an unresolved amount, not automatically zero.

Choose a consistent recognition rule with finance. The example uses costs assigned to September 2026 under the fictional workshop’s reporting policy; it is not a lifetime total for every defect discovered that month. Keep the event date as well as the recognition period so a late warranty cost can be linked to an earlier event without being added to both months.

For scrapped stock, record what the unit cost already includes. If the value includes original material, normal production labour and allocated overhead, adding those same components again inflates the loss. Additional rework can be a separate cost, but only when it is outside the scrap valuation and supported by a distinct record.

Every quantity, rate, event and result in this example is fictional. No client records or operational test results are used. The scope is one UK workshop for September 2026. Amounts exclude tax; there are no currency conversions. Within the fictional scenario, finance has approved a £28 hourly resource-cost rate for the separately logged work below. The time entries do not overlap. The six reworked units and the customer repair are separate from the 60 scrapped units.

Internal failure cost is £1,440 + £504 + £156 + £112 = £2,212. External failure cost is £180 + £168 + £90 + £56 = £494. Gross COPQ is therefore £2,706.

The workshop also records a £300 supplier credit, CN-108, relating to S1 and recognised in the same period. Keep the gross failure cost visible, show the recovery as a separate £300 line, then report net COPQ of £2,406. In this example the £1,440 scrap value is before the credit. If an actual export already nets a credit against scrap, restore the gross amount or use the net figure consistently; do not subtract the credit twice.

The same workshop spends £500 on planned prevention and £400 on routine appraisal. Those £900 belong in the wider cost-of-quality report. On this example’s gross basis, total cost of quality is £900 + £2,706 = £3,606. Calling £3,606 the COPQ would mix the categories used here.

Suppose the same site and month have fictional net sales of £120,000, measured before any supplier recovery and on the workshop’s agreed sales basis. Gross COPQ as a percentage of those sales is £2,706 ÷ £120,000 × 100 = 2.26%, rounded from 2.255%. The corresponding net measure is 2.01%, rounded from 2.005%. If the chosen sales denominator is zero, report the percentage as unavailable rather than zero. Label which one you use and check the denominator before comparing periods.

Imagine that a colleague adds a second £1,440 row from the nonconformance log. It refers to the same S1 units and the same write-off as WO-091. The apparent gross COPQ becomes £4,146. Removing the duplicate restores £2,706; no operational improvement occurred. Preserve both source references, but give them one counted cost line.

Do not deduplicate on event ID alone. R1 legitimately contains three different costs. Use an event ID to group the problem, a cost-line ID to identify each distinct amount, and a source reference to show where it came from. Two matching amounts can still be different costs; two different amounts can be an estimate and its replacement. Resolve the relationship before summing.

The same discipline applies to customer compensation. A repair, replacement, refund and credit are not automatically four additive losses. Map what actually happened and what each amount contains. If the records do not establish whether a refund replaces another amount or is an additional concession, hold that component for review rather than inventing a treatment.

Use one row per distinct cost or recovery. The completed S1 record below shows the minimum information behind the first row of the calculation.

For your own ledger, also retain event and posting dates, the source-system name, cost owner, currency, quantity, unit, rate, amount, estimate status and inclusion decision. Keep excluded rows with a reason. A cost list containing only included rows cannot explain why another report has a different total.

In a spreadsheet, use separate amount columns or clearly filtered categories for internal failure, external failure, recovery, prevention and appraisal. Check these identities before release: gross COPQ equals internal plus external failure; net COPQ equals gross minus included recoveries; gross cost of quality equals prevention plus appraisal plus gross COPQ. Never add the gross and net COPQ figures together.

Before accepting the month, trace the scrap total to stock adjustments, rework time to distinct time entries, and customer costs to repair, transport and complaint records. Reconcile differences and record unresolved amounts. A numerical match alone does not prove the source records are complete.

In this example, S1 is the largest recorded event at £1,440 before recovery. That makes it a sensible starting point for a cost investigation. It does not make it the highest safety, customer or compliance priority. Those consequences need a separate assessment even when the recorded cost is small.

Nor is £2,706 a cash-saving promise. The ledger includes £840 of separately logged labour resource cost, and the scrap value contains allocated production costs. Preventing the work may release capacity without reducing payroll or overhead spending. Ask which costs can actually be avoided, over what period, and what the proposed change itself will cost. Do not annualise one month as a forecast without checking volume, mix and recurrence.

Keep uncertain lost-sales or reputation estimates outside the recorded-cost subtotal. If they matter to a decision, show them as a separate scenario with an explicit method and assumptions. Their omission from the subtotal does not mean the consequences are absent.

Turn the chosen investigation into an owned action with evidence of the cause, a proposed change and a test of its result. Use evidence of corrective-action effectiveness before closing the action. Then compare later periods on the same cost basis, explaining changes in output, prices, reporting coverage and recoveries. A lower total caused by missing records is not a successful quality improvement.

ASQ Cost of Quality provides the classification framework. Its public resource was checked on 15 October 2026; a publication or effective date is not stated. It is professional guidance, not legislation. The ledger, calculations, reconciliation controls and fictional workshop scenario are QSE Advisory’ proposed practical method.

QSE Advisory is published by QSE Advisory. Preparation included AI-assisted research and drafting. Reviewer and Assessor: Laraib Naveed Ahmad.

This article is general information. Verify current requirements and seek qualified advice for your specific circumstances.

Spotted an issue? Contact our editorial desk.