top of page
Search

The Defect Factory Is Eating Your Profit

Chris Butterworth
Nov 9
6 min read

In a typical manufacturing company, the cost of poor quality is larger than the profit.


That is not a figure of speech. It is what happens when you put two well-documented numbers side by side.


Take a manufacturer with $100 million in annual revenue. If it earns the average net margin for its sector, roughly 7%, it keeps about $7 million a year. If its cost of poor quality (COPQ) sits at 10 to 15% of revenue, a range widely reported for ordinary manufacturers, it spends $10 to $15 million a year on scrap, rework, re-inspection, sorting, warranty, returns, expediting and the engineering time burned chasing defects.



The factory inside your factory


Put plainly: this company spends more money making things wrong than it earns making things right. There is a second factory running inside the first one, and its only product is failure. Call it the Defect Factory: all the activity and cost that goes into producing nonconforming product.


Armand Feigenbaum famously named part of it the “hidden plant”: the unrecorded rework that never shows up in the reports. But the Defect Factory is bigger than that. It also includes scrap, warranty claims and returns that are recorded, just scattered across so many cost lines that nobody adds them up.


If that were true, wouldn’t these companies go out of business? No, and the reason matters. The cost of poor quality is not subtracted from the profit you report. It is already buried inside your costs, and your competitors carry a similar burden, so market prices have absorbed it. The 7% is what is left after the Defect Factory has been paid.


The good news is in the same arithmetic. Every dollar of COPQ that is permanently eliminated drops almost straight to the bottom line. Recovering even a third of it would raise profit by roughly 50 to 70%, without a single new customer, product or capital project.


Poor quality does not show up as a loss. It shows up as a profit you never see.

What the research says


The numbers vary by industry and by how honestly costs are counted, but every credible source lands in the same uncomfortable territory.


  • American Society for Quality: quality-related costs reach 15 to 20% of sales revenue in many organizations, and as high as 40% of total operations in some.

  • Crandall & Julien, IISE: COPQ in manufacturing averages about 15% of sales, ranging from 5 to 35% depending on product complexity.

  • Quality Digest: experts estimate COPQ at 5 to 30% of gross sales, yet most executives believe theirs is under 5%, or don’t know it at all.

  • McKinsey, medical devices: even in a heavily regulated industry, direct quality costs run 6.8 to 9.4% of sales, about two-thirds of it from poor quality.

  • Gartner, via Quality Magazine: only 23% of manufacturers actively track their total cost of quality.

  • NYU Stern (Damodaran), Jan 2026: US manufacturing subsectors average a net profit margin of about 7.2%.


Two points matter most for senior leaders. First, COPQ is routinely larger than net margin. Second, most companies are not measuring it, so the people best placed to act on it cannot see it.


One manufacturer described in Quality Digest added up its failure costs and found they equalled 20% of sales: in effect, one day of every five-day week spent producing nothing but waste.


Why the Defect Factory never closes


If manufacturers were good at solving problems, COPQ would not be this high. A defect that is truly understood and fixed at its root does not come back. The fact that failure costs stay at 10% or more of revenue, year after year, tells us that most problems are being contained, not solved.


This is not new. Joseph Juran was writing about the cost of quality in 1951. Three quarters of a century later, we have better machines, better software and better quality systems, and the Defect Factory is still running. What has not improved at the same pace is how companies find root causes.


Walk through almost any plant and you will see the same pattern:


  • A problem is assigned to whoever owns the area: a process engineer, a quality engineer, a production supervisor.

  • That person is talented and busy, and has had little or no training in applied statistics or structured investigation.

  • A team meets, brainstorms possible causes and picks the most plausible ones.

  • Corrective actions are applied against several suspects at once, the problem fades, and the file is closed.

  • Months later the problem returns, or a cousin of it appears, and the cycle starts again.


This is the real reason COPQ is so stubborn. I would argue problem solving is the single most important capability a manufacturer can build, because it is the only lever that recovers money already being lost.


Problem solving is the profit engine of a manufacturing company, and in most companies it is running on guesswork.

Divergent versus convergent: 8D and Investigative Problem Solving


The 8D method is the most common problem-solving framework in manufacturing, and it has real strengths. It gives structure to a team, a format for customer reports and a discipline of containment and follow-up. But 8D is primarily a management and reporting framework. It tells you that you must find the root cause in step D4. It does not give most teams a reliable way to do it.


In practice, D4 is usually filled with brainstorming, fishbone diagrams and 5 Whys. These tools are divergent: they generate a long list of possible causes based on opinion and experience. The team then has to choose among them, often by vote or by whoever is most senior. The result is many candidates for root cause, several corrective actions, and no proof of which one mattered.


Investigative Problem Solving works the other way. It is convergent. Instead of asking people what they think the cause is, it uses data drawn from the process itself, upstream of the defect, to rule out whole families of possible causes at a time. With each step the search space shrinks, until only one explanation is left standing. Applied statistics, sampling and probability make each step trustworthy.



The two approaches side by side



The last row is the real test. If you have found the true root cause, you can make the defect appear and disappear at will. Few 8D reports can make that claim.


In my own career I have used this convergent approach successfully many times, in battery, electronics, plastics and high-tech manufacturing.


“It is a capital mistake to theorize before one has data.” — Sherlock Holmes

Make problem solving a job, not a side task


Most companies treat root cause analysis as something extra that engineers do between their real responsibilities. That is why it rarely gets done well. Nobody becomes skilled at something they do occasionally, under pressure, with half their attention.


I recommend a different model: a small, dedicated problem-solving group whose only job is finding and eliminating root causes. It can start with a single person.


  • Skill compounds. Someone who works problem after problem builds statistical judgment and investigative instinct that no one-off assignment can create.

  • Talent is used properly. Your process and design engineers excel at their own work. Pulling them into investigations they are not trained for wastes their time and the problem.

  • Results become measurable. Track the group by one number: COPQ dollars permanently removed.

  • The business case is easy. In a $100 million company with 10% COPQ, an investigator who recovers just 1% of that $10 million covers much of their own cost. A capable one should recover many times that.


Your engineers still own their processes and implement the fixes. The difference is that the investigation is led by someone trained to converge on the answer.


Where to start


The Defect Factory has been running for decades because companies keep using the same approach and expecting different results. Changing that needs no new quality system or large capital budget. It needs three decisions:


  1. Measure your real COPQ. Count everything: scrap, rework, sorting, re-inspection, warranty, returns, expediting, downtime and engineering hours spent firefighting.

  2. Invest in genuine problem-solving skill. Train engineers in applied statistics, sampling and convergent investigation, not just in filling out a report template.

  3. Dedicate at least one person to problem solving full time, and measure them by COPQ removed.


If your cost of poor quality is larger than your profit, problem solving is not a support function. It is the fastest route to the profit you already earn and then give away.

Training in Investigative Problem Solving


In-house, hands-on training that teaches your engineers to converge on true root causes using process data, sampling and applied statistics. Give me your team, and I will return them ready to take on your most expensive quality problems.


Chris Butterworth, Six Sigma Master Black Beltchris@belfieldacademy.com


References


  1. American Society for Quality. Cost of Quality (COQ). asq.org. ASQ’s 15–20% figure covers total quality-related costs, of which failure costs are the largest part.

  2. Crandall, R. E. and Julien, O. Measuring the cost of quality. Institute of Industrial and Systems Engineers (iise.org).

  3. Archambeau, S. What is Your Company’s Cost of Poor Quality? Quality Digest, 2004.

  4. McKinsey & Company. Capturing the value of good quality in medical devices.

  5. Quality Magazine. Is Your Total Cost of Quality Picture Complete? (citing Gartner).

  6. Manufacturing Financial Benchmarks, compiled from NYU Stern Damodaran industry data, January 2026.

  7. Taylor & Francis. Cost of poor quality (on Juran’s 1951 work and studies placing COPQ up to 40% of revenue).


Prefer a printable copy? Download the PDF version below.



 
 
 

Comments


bottom of page