IIIOrganization & Project ManagementGuide
What Errors Cost: Measuring Rework and Complaints in Euros
Almost every company knows how many complaints it had. Hardly any knows what they cost. Yet a simple calculation is enough to turn an annoyance into a figure you can base decisions on.
Key takeaways
- The cost of errors comprises all costs incurred because something was not right the first time: rework, corrections, follow-up deliveries, complaints, and credit notes.
- The classic model of quality costs distinguishes four groups: prevention, appraisal, internal failure, and external failure. The last two are the cost of errors in the narrower sense.
- The cost of errors is calculated like process costs: minutes per case times the cost rate, plus material, freight, and credit notes. A tally sheet over four weeks is enough for a first figure.
- Whether every preventive measure pays off is disputed in research. What is documented is that companies that measure quality costs are more likely to reduce them.
Why does hardly any company know its cost of errors?
A fitter drives to the customer a second time because a part was missing. Accounting rewrites an invoice because the price was wrong. In the workshop, a component that was already finished is reworked. Each of these cases is small, and none appears in the accounts as an error. It is hidden in wages, travel costs, and material, where the correct work is recorded too.
That is why companies almost always underestimate their cost of errors. This leads to the guiding question of this guide: How does a small company measure what rework and complaints cost, and how does it use the figure to eliminate the costliest errors first? To answer it, we first clarify the terms and introduce the quality cost model. Then we review the research, weigh two perspectives, and work through a complete example.
An error that is not measured in euros counts as bad luck. One that is measured becomes an investment decision.
Terms: error, rework, cost of errors
An error here is any deviation from what was agreed or intended, whether in the product, the service, or a document. Rework is the work needed to fix an error before or after it reaches the customer. A complaint in the sense used here is a grievance that refers to a specific defect and is usually tied to a demand, as the guide to handling complaints defines it.
The cost of errors comprises all costs incurred because something was not right the first time. This includes working time for rework and correction, additional material, freight and trips, credit notes, and price reductions. Harder to grasp are the consequential costs: an order that does not come back, or an appointment that falls through for another customer because of the rework. This guide names them but does not include them in the calculation, so that the figure remains robust.
The thinking model: prevention, appraisal, failure
In 1956, the American quality expert Armand Feigenbaum proposed a classification that is still the most widely used model of quality costs. In the literature it is called the PAF model, for prevention, appraisal, and failure. It assigns all costs related to quality to four groups (Feigenbaum 1956; Schiffauerova and Thomson 2006).
| Group | What belongs to it | Example |
|---|---|---|
| Prevention | costs to keep errors from arising at all | checklist before sending a quote, instruction, work instruction |
| Appraisal | costs to find errors before delivery | final inspection, four-eyes check of an invoice |
| Internal failure | costs of errors discovered within the company | rework in the workshop, invoice corrected before sending |
| External failure | costs of errors that reach the customer | complaint, second trip, credit note, follow-up delivery |
Source: after Feigenbaum (1956) and Schiffauerova and Thomson (2006)
The idea can be compared to a leaking roof. You can check the tiles regularly, you can set out buckets, or you can repaint the ceiling after every rain. All three cost money. But only the first addresses the place where the water comes in. The model shows how much a company spends on buckets and paint, and whether it would pay to put more into the tiles.
What research shows about quality costs
Schiffauerova and Thomson (2006) reviewed the published literature on quality cost models and on reports of their success. According to their review, the PAF model is the most frequently used. Companies that systematically record quality costs report that they reduce them and improve quality. A model only works, however, if it is tailored to the situation, purpose, and needs of the individual company.
For small companies there is less data. Teplická and Hurná (2021) analyzed the quality costs of 50 small industrial companies with 10 to 40 employees and at most 3 million euros in annual turnover in engineering, metalworking, and automotive supply over several years using the PAF model. Total quality costs fell over the period, mainly through lower costs of internal and external failures. According to the authors, fewer complaints led directly to savings, while the companies’ overall costs rose for other reasons. The study describes trends; it does not test causes.
A study by Omurgonulsen (2009) of seven large Turkish food manufacturers over the years 2000 to 2005 supports the model’s basic assumption with figures. When spending on prevention and appraisal rose by one percent, the cost of failures fell by 0.83 percent. According to the author, the relationship was driven mainly by external failures, that is, errors at the customer.
Is there an economic optimum? Two perspectives
The first view is the classic one. It says: every euro spent on prevention lowers the cost of failures, but with diminishing effect. At some point, the next inspection costs more than it saves in errors. So there is an economic optimum, and zero defects are too expensive. This view is supported by the measured trade-off between the two types of cost in Omurgonulsen (2009).
The second view goes back to Philip Crosby (1979) and is summed up in the title of his book: quality is free. Prevention that eliminates an error at its cause works not once but on every future order. The costs of missing quality are therefore almost always underestimated, and the optimum lies much closer to zero defects than the classic curve suggests.
For a small company, I consider a distinction more useful than choosing between the two. For appraisal, the classic view applies: an additional check costs time on every order and finds errors instead of preventing them. For prevention at the cause, Crosby is closer to the mark: a changed template or a mandatory field costs once and works permanently. The often-quoted rule of ten, according to which an error becomes ten times more expensive with each stage on the way to the customer, is a rule of thumb. There is no robust evidence for the factor of ten, but the direction matches the finding that external failures are the most expensive.
Inspect more
- costs on every order
- finds errors before the customer sees them
- turns external into internal failures
- pays up to an optimum
Prevent at the cause
- usually costs once
- keeps the error from arising
- reduces internal and external failures
- pays as long as the cause recurs
The cost of errors in euros: the calculation in four steps
The calculation works as for process costs. The guide to what a process costs describes the core formula: minutes per run times the cost per minute of practical capacity. For errors, material, freight, trips, and credit notes are added. The tool for this is called the error cost list here.
Define error types
Write down the five to eight errors that recur in the company, each with a short label. A category for other errors is allowed, but must not become the largest group.
Count for four weeks
Whoever fixes an error makes a tally mark for the error type and notes the minutes. For complaints, the complaint register supplies the cases.
Convert to euros
Minutes times the area’s cost rate, plus material, freight, travel, and credit note per case. Times the number of cases gives the cost per error type.
Rank by cost
Sort the error types by their cost, not by their frequency. Often the most frequent error is not the most expensive.
| Error type | Cases | Cost per case | Cost per month |
|---|---|---|---|
| correct an invoice | 10 | 20 min. office: €17 | €170 |
| deliver missing parts | 6 | 40 min. office: €34, freight €25 | €354 |
| rework in the workshop | 8 | 1.5 hrs. workshop: €98, material €30 | €1,024 |
| complaint at the customer | 3 | 75 min. office: €64, 4 hrs. installation: €260, travel €60, credit note €150 | €1,602 |
| total | 27 | · | €3,150 |
Worked example
Worked example with freely chosen values. Office at 0.85 euros per minute as in the article on calculating process costs, workshop and installation at an assumed 65 euros per hour. Extrapolated, about 37,800 euros a year.
The example shows what the list achieves. Correcting invoices is the most frequent error and the cheapest. The three complaints are the rarest and cost almost half. Whoever only counts would start with the invoices. Whoever measures in euros starts with the complaints.
What the figure should trigger in the company
An error cost list is not a report but a ranking for decisions. For the most expensive error type, the cause is sought, and it is sought in the process. The essay on errors as process signals describes the substitution test for this: could another capable person in the same position have made the same error? If so, the measure belongs in the process and not in a conversation with the person.
With the figure in euros, every measure can be evaluated. If a checklist before delivery takes two minutes per order at 150 orders a month, that is 300 minutes or about 255 euros. If it halves follow-up deliveries, it saves 177 euros and does not pay for itself on that alone. If it also prevents one complaint a month, it saves more than 700 euros. What good checklists look like is shown in the article on checklists in the company. How to find errors before they happen for the first time is described in the article on finding errors in advance with FMEA.
The list therefore belongs in a rhythm. Once a month it is updated and compared with the previous month, as lean quality management provides for anyway. One limit should be kept in mind: the list only measures what is reported. Whoever uses errors as an occasion to assign blame gets a shorter list, not fewer errors.
Conclusion: how does a company measure what errors cost?
We set out to ask a small company measures what rework and complaints cost and how it uses the figure. The answer is an error cost list based on the PAF model: a few error types, four weeks of tally marks, minutes times cost rate plus material, freight, and credit note, ranked by euros instead of frequency. Consequential costs such as lost orders are left out so that the figure is robust, and they are still one more reason to start with external failures.
As the studies show, companies that measure quality costs are more likely to reduce their cost of errors, and more prevention goes along with fewer failures. Whether there is an optimum depends on whether you inspect or prevent at the cause. The sober observation is: the most expensive errors are rarely the most frequent ones, and without a figure in euros you usually work on the wrong one first.
What you can do this week
- With two employees, write down the five to eight error types that occur most often.
- Put up a tally sheet with error type and minutes wherever rework takes place.
- Take the cost rate per minute from your process cost calculation or estimate it from salaries and practical working time.
- After four weeks, calculate the cost per error type and sort by euros.
- For the most expensive error type, look for the cause in the process and set a measure with a deadline.
Frequently asked questions
What is the cost of errors?
The cost of errors comprises all costs incurred because something was not right the first time: rework, corrections, follow-up deliveries, complaints, and credit notes. In the quality cost model, it is divided into internal and external failure costs.
How do you calculate the cost of errors?
For each error type, multiply the minutes needed to fix it by the cost rate per minute and add material, freight, travel, and credit notes. That result times the number of cases gives the cost for the period. Four weeks of counting is enough for a first figure.
What is the PAF model?
The PAF model goes back to Armand Feigenbaum and divides quality costs into four groups: prevention, appraisal, internal failure, and external failure. It is the most widely used model of quality costs.
Is the rule of ten for error costs true?
The rule of ten states that an error becomes ten times more expensive with each stage on the way to the customer. It is a rule of thumb without robust evidence for the factor of ten. The direction is right: errors at the customer are usually the most expensive.
Does more quality control pay off?
Additional inspections pay off up to a point because they cost time on every order. Prevention at the cause, such as a changed template, usually costs once and works permanently. The error cost list shows where each measure pays.
Sources
- Feigenbaum, A. V. (1956): Total Quality Control. Harvard Business Review, 34(6), 93-101.
- Schiffauerova, A.; Thomson, V. (2006): A review of research on cost of quality models and best practices. International Journal of Quality & Reliability Management, 23(6), 647-669.
- Teplická, K.; Hurná, S. (2021): New Approach of Costs of Quality According Their Trend of During Long Period in Industrial Enterprises in SMEs. Management Systems in Production Engineering, 29(1), 20-26.
- Omurgonulsen, M. (2009): A research on the measurement of quality costs in the Turkish food manufacturing industry. Total Quality Management & Business Excellence, 20(5), 547-562.
- Crosby, P. B. (1979): Quality Is Free. New York: McGraw-Hill.
