Menu
Book a demo
The Cost of Poor Quality (COPQ) in Manufacturing: 2026 Guide

The Cost of Poor Quality (COPQ) in Manufacturing: 2026 Guide

Cost of poor quality (COPQ) explained: the formula, the four PAF categories, and a brake pad line priced in OEE points and euros per shift, month and year.
The Cost of Poor Quality (COPQ) in Manufacturing: 2026 Guide

Key takeaways

  • The cost of poor quality (COPQ) is what a plant loses because products or processes fail their requirements, inside the plant and at the customer.
  • COPQ formula: COPQ = internal failure costs + external failure costs. Add prevention and appraisal and the total is the cost of quality (COQ).
  • Company-wide estimates run high: up to 20% of sales (Juran) and 25% or more (Feigenbaum, 1994), most of it failure.
  • Worked example (sample figures): on a brake pad finishing line, poor quality costs 8.13 OEE points a shift, and only 1.68 of them show in the quality factor.
  • In money: EUR 803.98 a shift, EUR 607,808.88 a year on a sold-out line, against EUR 273.52 a shift on the scrap report. Quality holds and restarts, with the slow running, startup rejects and sorting that follow them, are about two thirds of it.

The cost of poor quality (COPQ) is every cost a manufacturer carries because a product or process fails its requirements. Inside the plant that means scrap, rework, re-checks, sorting and the line time lost to quality holds; outside it, returns, claims, sorting at the customer and replacement freight. The standard COPQ formula is: COPQ = internal failure costs + external failure costs.

In quality cost terms, COPQ is the cost of nonconformance; prevention and appraisal, the cost of conformance, sit outside it. Armand Feigenbaum first presented quality costs in detail in 1956, in four categories (prevention, appraisal, internal failure, external failure), as a 2020 ASEE paper by Mustafa Shraim of Ohio University sets out: the PAF model.

The first part of this guide covers the formula, the categories and the questions plant teams ask. The second works one brake pad line through OEE and euros, and shows that most of the line time it loses to poor quality sits in availability and performance, outside a scrap report.

How Do You Calculate the Cost of Poor Quality (COPQ)?

Internal failures are defects caught before the product leaves the plant; external failures are defects the customer finds. The money comes from records a plant already keeps: the scrap log, rework hours, sorting time, complaint files and freight invoices.

To calculate COPQ on a real line, in five steps:

  1. Pull scrap, rework and re-check counts from the quality records, or derive them from first pass yield and rolled throughput yield data.
  2. Assign each failure a fully loaded cost: the material, labour, machine time and energy already invested, plus the time spent reworking or replacing the part.
  3. Add external failure costs: sorting at the customer, returns, claims, complaint handling and express replacement shipments.
  4. Express the total as a share of revenue or of production cost, so it can be tracked over time and compared across lines and sites.
  5. Add the line time poor quality takes: quality holds and restarts (availability), slow running after a restart (performance) and the good units that scrapped units displaced (quality). Price that time at the margin of a unit only if the line is sold out.

Worked example, one shift (sample figures): scrap EUR 273.52 (52 brake pads at EUR 5.26 cost of goods) + rework EUR 35.00 + sorting EUR 52.50 + a shift's share of one customer return a month, EUR 43.00 = EUR 404.02 a shift. The second part builds the example in full, with the line time the same defects take.

How large COPQ is across a whole company depends on how much gets counted. Juran estimated that unplanned quality-related costs could be as high as 20% of sales, as cited in Shraim's paper. Feigenbaum told IndustryWeek in July 1994 that the cost of quality in major American companies can be 25% or more of sales, most of it failure. A measured case shows the split: in a 2022 case study of an export apparel factory in Addis Ababa, the cost of quality came to 23% of sales revenue, with prevention at 4.8%, appraisal at 10.2%, internal failure at 7.4% and external failure at 0.54%. Scrap and rework also take line time, and the cost of downtime calculator prices lost output.

The line-time part needs stop data. OEE software records stops with their reasons, so a quality hold and its restart stay apart from breakdowns and changeovers. In Fabrico, run time and counts come from the machine's PLC or an IoT device with sensors on the machines you connect, and operators give stop and reject reasons at the machine. Stations without a usable machine signal get their output counted by the camera add-on. Where the camera add-on covers the station, the video for the minute of a stop can be opened from the shift timeline, streamed on request and not stored in Fabrico. With the selling price and margin of each SKU entered, losses are priced in money. So a quality hold shows up in euros beside the scrap, and the short video below shows those views.

What you see: a shift timeline, unclassified time waiting for a reason, and the operator's line-down screen, on sample lines. Book a demo.
Plays from YouTube. Google may set cookies when you press play. Privacy policy.

The four quality cost categories and where each shows up in OEE

The cost of quality (COQ) adds the cost of good quality, prevention and appraisal spent on purpose, to the cost of poor quality, internal and external failure lost as a consequence. COPQ is therefore a subset of COQ, and the full comparison is in cost of quality vs cost of poor quality. A cost-of-quality report lists the four categories in money; the table adds where each one shows up in OEE, who owns it and how to measure it.

CategoryExamples on a lineWhere it shows up in OEEWho owns itHow to measure it
Prevention (cost of good quality)Training, error-proofing, process design, supplier qualification, a grinding wheel dressing intervalNot a loss but planned spend; it shows as fewer quality stops and rejectsPlant and quality managers (budget)Spend per project against failure costs before and after
Appraisal (cost of good quality)Incoming, in-process and final checks, testing, calibration, auditsA check that stops the line: availability; one that slows it: performanceQuality managerChecking hours a shift, test and calibration costs
Internal failure (COPQ)Scrap, rework, re-checks, downgrading, sorting at the line, quality holds and restartsScrap and rework: quality factor; holds and restarts: availability; slow restarts: performanceProduction manager, with qualityRejects by reason, quality stop minutes by reason, rework hours, scrap at cost of goods
External failure (COPQ)Returns, claims, sorting at the customer, replacement freight, complaint handling, lost ordersOutside the plant's OEE; it comes back as sorting, rework or a quality holdQuality manager, with salesCost per complaint: invoices, freight, credit notes, hours

The split matters for budgets. Shraim notes that conformance costs come from activities that are normally planned, while failure costs are not, and that the costs of failure, internal and external, could reach as much as 70% of the total cost of quality. How prevention and appraisal trade off is covered in prevention vs appraisal costs, and the line between a defect caught in the plant and one the customer finds in internal vs external failure costs. Root cause analysis then confirms that each prevention step removes a real failure mode.

Frequently asked questions

What is the Cost of Poor Quality (COPQ) in manufacturing?

The cost of poor quality (COPQ) is the money a plant loses because products or processes fail their requirements: internal failure costs such as scrap, rework, sorting and quality holds, and external failure costs such as returns, claims, sorting at the customer and replacement freight. It is the failure half of the cost of quality.

What are the four categories of quality costs (the PAF model)?

Prevention, appraisal, internal failure and external failure. Prevention stops defects (training, error-proofing, process design), appraisal finds them (checks, testing, calibration, audits), internal failure is the cost of defects found before shipment and external failure the cost of defects the customer finds. The first two are the cost of good quality; the last two are COPQ.

What are internal failure costs in COPQ?

The costs of defects found before the product leaves the plant: scrapped material and the work in it, rework labour, re-checks, downgrading, sorting suspect stock, and the line time lost to quality holds and restarts. On the sample brake pad line they come to EUR 361.02 a shift in cash, before the lost line time is priced.

How do appraisal and prevention costs relate to COPQ?

They sit outside COPQ but drive it. Prevention removes the causes of defects, so it lowers internal and external failure costs together. Appraisal finds defects earlier, which moves cost from external to internal failure and shortens quality holds, but leaves the cause in place. Raising appraisal without prevention buys more checks and keeps the failure modes behind them.

Where does the cost of poor quality show up in OEE?

In all three factors. Scrap and rework lower the quality factor. A quality hold and a restart with first-off checks are stops, so they lower availability. Slow running while a restarted line proves it is stable lowers performance. On the sample brake pad line, poor quality costs 8.13 OEE points a shift, and only 1.68 of them sit in the quality factor.

How is the cost of poor quality calculated?

Add internal failure costs (scrap at cost of goods, rework and sorting hours at a labour rate) and external failure costs (sorting at the customer, returns, freight, complaint hours) for one period, then divide by sales for the share. Add the line time quality takes, at the unit margin, if the line is sold out. The sample line comes to EUR 404.02 a shift in cash and EUR 803.98 with the lost margin.

What percentage of revenue does poor quality cost?

It depends on how much is counted. Juran estimated unplanned quality-related costs at up to 20% of sales, and Feigenbaum in 1994 put the cost of quality at 25% or more of sales in major American companies, most of it failure. A 2022 apparel case measured 23% of sales, of which failure was 7.94%. The sample brake pad line, counted at line level only, comes to 1.49% of its sales in cash and 2.97% with the lost margin: a floor, not a company total.

What is an example of the cost of poor quality?

On a brake pad finishing line (sample figures), a thickness drift causes a 16-minute quality hold, a restart with first-off checks and slow running while the setting is proven. The shift scraps 52 pads, re-coats 40, sorts suspect stock and carries its share of one customer return a month: EUR 404.02 in cash, EUR 803.98 with the margin of 404 lost pads.

Why is COPQ larger than the scrap report shows?

A scrap report counts scrapped material. It leaves out rework and sorting hours, the stops and slow running after a quality problem, complaints and returns, and costs booked elsewhere: Shraim notes that hidden quality costs such as corrective action time and premium freight end up in indirect labour, fixed and administrative costs. On the sample line the scrap report shows EUR 273.52 of the EUR 803.98 a shift.

How can a plant reduce its cost of poor quality?

Rank the costs by line time and money, then remove causes rather than add checks. On the sample line, quality holds and restarts, with the slow running, startup rejects and sorting that follow them, are about two thirds of the cost, so the cause of the thickness drift and a standard restart come first. Give quality holds their own stop reason, give rejects a reason at the machine, and track the cost by line and shift every week.

Where COPQ hides on the line: availability, performance and quality

Ask what poor quality costs a line and the first answer is the scrap figure. That figure is real, but it is one line of the internal failure costs, and it is the only one a scrap report collects.

Feigenbaum had a name for the rest. In the same IndustryWeek interview he called it the hidden organization or hidden factory, the part of an organization that exists to do bad work, tying up about 20% to 40% of the total capacity of many American companies. On a line, OEE shows where that capacity goes.

The Lean Enterprise Institute's lexicon entry on OEE sets out what each factor holds: availability covers the downtime from equipment failures and adjustments, performance covers running below design speed and stoppages of a few seconds, and quality covers scrap and rework as a percentage of all parts run. Only the third looks like a quality cost, yet a quality problem on a line does not end at the reject bin.

When a check finds a drift, the line stops for a quality hold while the cause is found: an availability loss, and so is the restart with its first-off checks. A line that restarts at reduced speed loses performance. Rejects come in two kinds: a 2019 paper on overall equipment cost loss in the Jordan Journal of Mechanical and Industrial Engineering describes reject parts from steady-state production (process defects) or on startup after a stop event (reduced yield), and prices the losses of all three OEE factors in money.

So poor quality costs a line in all three factors. On the sample brake pad line worked through below, 444 good pads a shift are lost to poor quality, and 352 of them sit in availability and performance.

Where poor quality hides in OEEBrake pad line, one shift, sample figures444 good pads lost to poor qualityAvailability: 252 pads, 4.62 pointsQuality hold and restart, 21 minutesPerformance: 100 pads, 1.83 pointsSlow running after the restartQuality factor: 92 pads, 1.68 points52 scrapped, 40 reworkedPoor quality costs 8.13 OEE pointsThe quality factor shows 1.68 of them6.45 sit in availability and performanceOEE this shift: 78.57%Sample figures for illustration only.
Good pads a sample brake pad line loses to poor quality in one shift, by OEE factor: the quality factor holds a fifth of the loss. Sample figures for illustration only.

Below, the split is priced in euros and ranked, and three manager roles get their first steps.

Worked example: the cost of poor quality on a brake pad line, per shift, month and year

Sample figures for illustration only. An automotive supplier finishes disc brake pads for passenger cars on one line: grinding, slotting and chamfering, powder coating, shim fitting, marking and packing. The finishing line is the plant's constraint. It works 3 shifts on each of 21 working days, 63 shifts a month, and each shift has 455 planned minutes: 8 hours less 25 minutes of breaks. The ideal rate is 12 pads a minute, so a perfect shift would make 5,460 good pads.

On an average shift the end-of-line thickness check finds one drift. The line stops for a 16-minute quality hold while the setter dresses the grinding wheel; the restart and first-pad re-check take 5 minutes, and the line then runs at 8 pads a minute instead of 12 for 25 minutes. Beside the line, 2 people spend 45 minutes each sorting the pads made since the last good check.

Carry the shift through OEE:

  1. Availability. Quality hold and restart take 21 minutes, and other stops (a conveyor jam, waiting for cured pads, a planned product change) 48 minutes, so run time is 455 minus 69 = 386 minutes. Availability = 386 ÷ 455 = 84.84%.
  2. Performance. At 12 a minute, 386 minutes would make 4,632 pads. Slow running after the restart costs 100 pads (25 minutes x 4 a minute) and other short stops and slow cycles 150, so the count is 4,382. Performance = 4,382 ÷ 4,632 = 94.60%.
  3. Quality. 52 pads are scrapped: 20 startup rejects in the first pads after the restart and 32 with chipped friction material. 40 pads fail the coating check and are re-coated off the line. Good first time: 4,290. Quality = 4,290 ÷ 4,382 = 97.90%.
  4. OEE = 84.84% x 94.60% x 97.90% = 78.57%, the same as 4,290 good pads ÷ 5,460 possible.

Now sort the losses by cause. Measured in good pads against the 5,460 possible, poor quality takes 444 pads a shift, 8.13 OEE points, and the quality factor holds only 92 of them:

Quality loss, one shiftOEE factorGood pads lostOEE points
Quality hold and restart (21 minutes x 12 a minute)Availability2524.62
Slow running after the restartPerformance1001.83
Scrap (52) and rework (40)Quality921.68
Total lost to poor qualityAll three4448.13

About four fifths of the loss, 352 of the 444 pads or 6.45 of the 8.13 points, sits outside the quality factor. A plant that reads COPQ from the quality rate alone would see about a fifth of it.

Then price it. Take a sample selling price of EUR 6.25 a pad and the 15.84% gross margin NYU Stern reports for US auto parts companies (35 firms; the data are as of January 2026). Each pad then carries EUR 6.25 x 15.84% = EUR 0.99 of gross margin and EUR 5.26 of cost of goods. Labour is priced at EUR 35.0 an hour, Eurostat's 2025 labour cost level for manufacturing (NACE C) in the EU27, dataset lc_lci_lev. A sector average of 35 US companies stands in for real figures: use your own SKU margin when you have it, and your own labour rate.

Cost, one shiftHow it is pricedEUR a shiftView
Scrap52 pads x EUR 5.26 cost of goodsEUR 273.52Finance (cash)
Rework40 pads re-coated, 1.0 hour x EUR 35.0EUR 35.00Finance (cash)
Sorting at the line2 people x 45 minutes, 1.5 hours x EUR 35.0EUR 52.50Finance (cash)
Customer returnEUR 2,709.00 a month ÷ 63 shiftsEUR 43.00Finance (cash)
Cash COPQ, internal + externalSum of the four rows aboveEUR 404.02Finance (cash)
Lost margin404 good pads not made (252 + 100 + 52 scrapped) x EUR 0.99EUR 399.96Line time, if sold out
Total a shiftCash COPQ + lost marginEUR 803.98Both

The customer return is one complaint a month: a vehicle plant finds pads with a missing shim and blocks the delivery. Third-party sorting at the customer costs EUR 1,395.00, express freight EUR 544.00, sorting the returned lot 12 hours (EUR 420.00) and the complaint report 10 hours (EUR 350.00): EUR 2,709.00 a month. Good pads from the returned lot go back to stock.

Scrapped pads carry both cost and margin, and that is not a double count: the cost of goods is cash spent, and on a sold-out line their 52 slots could have made pads that sell. The 40 reworked pads still sell, so they carry only the labour.

A shift loses EUR 803.98, of which EUR 404.02 is cash and EUR 399.96 lost margin. A month is 63 shifts, EUR 50,650.74, and a year is 12 x 63 = 756 shifts, EUR 607,808.88. The scrap report alone would show EUR 273.52 a shift, EUR 206,781.12 a year.

What poor quality costs, in eurosBrake pad line, one shift, sample figuresScrap, 52 pads x EUR 5.26: EUR 273.52Rework, 1.0 hour x EUR 35.0: EUR 35.00Sorting at the line, 1.5 hours: EUR 52.50Customer return, a shift's share: EUR 43.00Lost margin, 404 pads x EUR 0.99: EUR 399.96Cash (finance view)Line time, if sold outA shift: EUR 803.98A month (63 shifts): EUR 50,650.74A year (756 = 12 x 63): EUR 607,808.88Scrap report alone: EUR 273.52 a shiftSample figures for illustration only.
What poor quality costs the sample brake pad line in one shift, in cash and in lost margin, with the month and year. Sample figures for illustration only.

Against the line's sales, 4,290 good first time plus 40 reworked pads a shift at EUR 6.25, EUR 27,062.50, the cash COPQ is 1.49% and the total 2.97%. That is far below the company-wide estimates, because one line leaves out engineering time, warranty provisions and lost customers. The finding holds either way: the euros follow the line time, and the line time sits mostly in availability and performance.

Which COPQ cost to attack first

Four failure costs compete for the same improvement hours on the sample line, plus the option of checking more. What each returns a year (756 shifts) differs by more than ten times:

OptionBack a shift (example)A year, sold-out lineEffort (rough estimate)OwnerRisk to watch
A. Cut quality holds and restarts: the cause of the drift (wheel dressing interval) and a standard restartEUR 525.98 (21 minutes, 100 slow pads, 20 startup rejects, the sort)EUR 397,640.8840 to 80 hours of engineer and setter time (EUR 1,400 to 2,800)CI or process engineerA dressing interval set by the calendar, not by measured drift
B. Cut process defects: chipped pads in handling and transferEUR 200.00 (32 pads at EUR 6.25)EUR 151,200.0016 to 40 hours (EUR 560 to 1,400), plus partsProduction managerFixing the easiest transfer, not the one that chips pads
C. Cut coating rework: booth settingsEUR 35.00 (one hour of labour)EUR 26,460.008 to 24 hours (EUR 280 to 840)Process engineerRework that reaches the customer as a complaint
D. Stop escapes: error-proof the shim fitting, close the complaintEUR 43.00 (EUR 2,709.00 a month)EUR 32,508.0024 to 60 hours (EUR 840 to 2,100), plus any sensorQuality managerThe customer's next order, which no shift figure prices
E. Check more (appraisal): a thickness check at the grinder every hourShorter holds, fewer pads to sortCosts EUR 10,584.00 a year (8 checks of 3 minutes a shift)Operator time every shiftQuality managerPaying for checks while the drift keeps coming

The money uses the example's EUR 0.99 margin a pad and counts only on a sold-out line; effort is an estimate for one line at EUR 35.0 an hour, without software. Options A to D add up to the full EUR 803.98 a shift, because each removes a different part of it.

Who owns each option. A belongs to the CI or process engineer, because the cause sits in the process setting; B to the production manager; C to the process engineer; D to the quality manager, with the customer; E to the quality manager as a stopgap. A comes first: it returns about two thirds of the total, more than B, C and D together, for 40 to 80 hours.

Questions to answer before choosing:

  1. How many minutes a shift did the line stand in quality holds and restarts last month, counted from the stop log and not from memory?
  2. How long does the line run slow after a restart, and how many rejects come in the first pads?
  3. Is the line sold out? If not, price the time as capacity, not margin.
  4. Which complaint cost the most last year, counting sorting at the customer, freight and hours?
  5. Which option removes a cause, and which only adds a check?
  6. To an OEE software vendor: can a quality hold get its own stop reason, can operators give rejects a reason at the machine, and can scrap and rework be reported apart from good output by line and shift?

More checking is not prevention. In the apparel case cited earlier, appraisal cost 10.2% of sales against 7.94% for all failure: the factory spent more on finding defects than it lost to them. Checks shorten a quality hold; only a removed cause ends it.

Sold out or not: how to price the line time poor quality takes

The EUR 399.96 of lost margin a shift is money only on a constraint line whose extra pads would find buyers. That condition decides how to price the line time.

Two cases, on the same 404 pads a shift:

  • Sold out: every pad not made is a sale lost, so the time is worth the unit margin, EUR 0.99 a pad on the example line, and the scrapped pads cost their full selling price.

  • Not sold out: the same 404 pads are about 34 minutes of capacity a shift (404 ÷ 12 a minute = 33.7), worth what that capacity saves: overtime avoided, a second line not bought, shorter lead times. The cash COPQ, EUR 404.02 a shift, stays either way.

To see how much line time scrap and rework take, compare production monitoring software for live stops and output by line with MES software, which adds order and job context.

In a business case: count the cash saving and, separately, the capacity a project returns, and say which case the line is in. Reclaimed capacity priced at full margin on a line that is not sold out overstates the return.

Machine condition, restarts and the first good part

Some quality losses start with the machine's condition rather than with the operator, as in these cases:

  • A vibrating spindle creates surface chatter.

  • A clogged filter causes painting defects.

  • A drifting temperature sensor causes under-cooked food.

On the example line the drift came from the grinding wheel, and the cost came from what followed: the hold, the restart and the slow running. To see that chain, put quality stops and rejects on the same time axis as the other stops. Real-time OEE (availability x performance x quality) per line and shift shows it when three habits hold:

  • Give a quality hold and the restart after it their own stop reasons, apart from breakdowns and changeovers.

  • Compare rejects in the first pads after a restart with rejects in steady running; the gap is the startup loss, explained in startup loss vs reduced yield loss.

  • Link a recurring cause to the people who fix it. In Fabrico, maintenance is built in: when a stop reason is set up for it, logging that reason opens a maintenance task for the machine. A repeated drift on the grinder then reaches the people who can remove it.

Checks at the source: appraisal that shortens quality holds

Appraisal adds no value to the product, but where a check sits decides how much a defect costs once it is found.

On the example line the thickness check is at the end of the line, so a drift means sorting every pad made since the last good check. A check at the grinder finds the same drift with fewer pads behind it.

Three ways to move it to the source:

  • First and last part checks: the operator checks the first pads after a restart and the last pads before a planned stop, so a drift is caught before a full batch is made.

  • Checks where the cause sits: measure the characteristic the machine controls at that machine, thickness at the grinder rather than at packing, and record the result with the time and shift.

  • Result: a defect caught at the step that made it costs that step only; caught at the end of the line, it has taken the time of every step after it, and the sort covers more pads.

COPQ actions for the plant manager, the quality manager and the CI engineer

Each role owns one part of the cost of poor quality. One week of stop data measures it on a single line; one quarter of projects cuts its largest part.

RoleIn the first weekOver the quarter
Plant managerGet two COPQ figures for the constraint line: cash a shift, and line time lost to quality in all three OEE factors.Fund the option with the largest return for its effort and report COPQ in euros at the monthly review.
Quality managerList last year's complaints with every cost attached: sorting, freight, credit notes, hours.Remove the cause of the largest complaint and move one end-of-line check to the step that causes the defect.
CI or process engineerPull a month of quality holds, restarts and slow restarts from the stop log, with the rejects after each restart.Fix the cause of the largest hold, write a standard restart, and compare hold minutes and rejects before and after.

Five steps for this week, on one line:

  1. Pick the constraint line: the one the rest of the plant waits for.
  2. Check that quality holds and restarts have their own stop reasons, and that rejects carry a reason.
  3. Add up last month's scrap at cost of goods, rework and sorting hours, and external failure costs.
  4. Count the good units lost to quality in all three OEE factors, priced at the unit margin if the line is sold out.
  5. Price a shift, a month (its shifts) and a year (12 x the month's shifts), and bring the ranking to the next production meeting.

Conclusion

The cost of poor quality is larger than the scrap figure because quality problems take line time as well as material.

Measured in OEE points and euros a shift, it shows where to start: on the example line, with the holds and restarts in availability and performance.

Shifting spend to prevention pays when the step removes a cause the line data has priced: a drift that triggers holds, a transfer that chips pads, a missing shim.

Count the scrap, then count the time. A light MES with OEE by line and shift keeps both visible from one shift to the next.

To see quality holds, restarts and rejects in OEE and in money on sample lines, book a demo.

Track OEE and downtime in real time
See where lines lose time
Book a live demo.
See where lines lose time
See where lines lose time

Real Fabrico tiles, sample data.

By clicking the Accept button, you are giving your consent to the use of cookies when accessing this website and utilizing our services. To learn more about how cookies are used and managed, please refer to our Privacy Policy and Cookies Declaration