Try it while you read
Open the free Cost of Poor Quality (COPQ) Calculator tool and click Load sample: it reproduces the case study below.
Theory: the PAF model
The cost of quality is split into prevention, appraisal, internal failure and external failure. COPQ is the failure part. The classic argument, from Feigenbaum and Juran, is that money spent on prevention reduces failure costs by much more, so the total falls when the mix shifts towards prevention.
Key point
Step by step
- Scope: one plant or product family, one period, net sales of the same scope.
- Collect scrap postings, rework orders, inspection hours, downtime logs, customer debit notes, warranty and premium freight.
- Enter each line in the COPQ calculator as an amount or quantity × cost per unit, with the cause and any amount recovered from suppliers.
- Report COPQ % of sales with the benchmark, the structure and the Pareto, and track it monthly in the trend tab.
- Improve: take the top failure item into an 8D or A3 project and value the improvement case with ROI and payback.
Example: plant with 24 M EUR sales
Prevention costs 224 k EUR, appraisal 455 k EUR, internal failure 712 k EUR and external failure 368 k EUR; suppliers paid back 58 k EUR. Net COPQ is 1 021 k EUR, 4.3 % of sales against a 3 % target, a gap of 301 k EUR per year. Internal failure is the largest block, led by scrap and rework. An extra 60 k EUR of prevention that cuts internal failure by 20 % and external by 25 % saves about 223 k EUR per year, a payback of a little over three months.
Common mistakes
- Counting scrap at the material price only.
- Changing definitions between months, so the trend shows accounting changes instead of quality.
- Reporting COPQ without an owner for the largest items.
- Adding hidden-cost estimates to the booked figure.
Frequently asked questions
Who should own the COPQ figure?
Quality prepares it, controlling confirms the numbers. A figure that finance does not recognise will not convince management.
Should I include lost sales?
Report them separately if you can estimate them; mixing estimates into booked costs weakens the credibility of the whole report.
What is a good first target?
A reduction of the current level, for example 20 % in a year, is more useful than an external benchmark that may use different definitions.