QM Management Depth (10) | COPQ: How to Calculate, Who to Present It To, and What Level of Granularity to Use

A certain electronics manufacturing company, with annual revenue of 860 million yuan, operates three product lines—industrial power supplies, automotive modules, and consumer adapters, with a net profit margin of less than 6%. In March 2025, the quality manager presented the COPQ report at the management meeting for the first time: the cost of poor quality accounted for 9.7% of revenue, approximately 83 million yuan. The financial director immediately pushed the report back: "According to your calculation, the company not only made no profit last year but also lost three percentage points. Which accounting item does this money fall under?"

The next forty minutes were entirely spent arguing over whether "rework hours should be multiplied by the workshop hourly cost" and whether "customer complaint handling hours should be counted as a loss." The real issue—where the money is leaking from and who should be moved next month—was not discussed at all. The quality manager later reflected: "I spent three weeks calculating this number, and it was debunked in just three sentences."

The second time he presented, the situation was reversed: he brought three tables, and the boss only took one A4 sheet, immediately deciding to abandon orders for a low-priced consumer product line. What made this happen was not more precise numbers, but rather clarity on three questions—whom the report is for, what level of granularity to use, and which costs should not be included in the main table.

1. Judgment Framework: Three Commonly Confused Questions

Perspective One: Who the Report is For Determines the Tone and the Severity of the Language

COPQ is not just one report, but three. The report for the boss should only answer "which product line and type of failure is the money leaking from, and where should we make changes to be most cost-effective"; the report for the plant manager and process engineers should only answer "which process has the greatest loss, and which equipment should be improved first"; the report for sales and customer quality should only answer "which customer's service cost is so high that we need to renegotiate prices and boundaries."

The same 83 million yuan should be sliced differently for each audience. A poor quality manager often combines all three into one "comprehensive table": the page for the boss is filled with thirty rows of process details, making it difficult for the boss to see the decision options; the page for the workshop is filled with "estimated loss of customer loyalty at 12 million yuan," which the workshop cannot verify or act upon. A table that everyone can read often ends up being a table that no one uses.

The judgment standard is simple: after the table is handed over, can the recipient immediately state what action they will take next? If they cannot, it is likely that the wrong table was given to the wrong person, rather than the numbers being insufficiently accurate.

Perspective Two: Explicit Costs in the Main Table, Implicit Costs in a Separate Appendix—Mixing Them Ruins the Credibility of the Entire Report

Rework hours, scrapped materials, claims, and freight have documents and can be verified against the accounting records, making them verifiable explicit costs, which should be included in the main table. Capacity occupied by rework, work-in-progress and safety stock accumulation, expedited air freight, bulk discounts to customers, and lost quotation opportunities due to quality fluctuations are implicit costs. These costs are often as significant as or even greater than the explicit costs, but they rely on assumptions.

Placing numbers that rely on assumptions on the same table as verifiable numbers will inevitably lead to the first question in the review focusing on the weakest assumption, and the entire report will be dismissed as "numbers fabricated by the quality department." The correct approach is to separate them: the main table should only contain verifiable items, even if it is only 4%; the appendix should contain implicit items, with each item clearly stating the calculation assumptions, data sources, and ranges. The main table answers "how much money is definitely lost," while the appendix answers "how much more might be lost and in which direction." The boss will seriously discuss the "definite 4%," but will be defensive about the "uncertain 9.7%."

Perspective Three: Granularity is Not a Pursuit of Precision, but an Investment

Breaking down by product line, process, customer, and failure mode, each additional dimension requires someone to fill in the numbers, someone to verify them, and someone to interpret them, consuming the recording hours of work teams, the analysis hours of engineers, and the attention of a presentation. The only reason to add a dimension is if it can point to a comparable action in terms of amount.

Companies that stop at the third level of breakdown are often not lazy, but rather assume "the finer, the more professional." In reality, it is a curve of diminishing returns: breaking down from the total table to the product line yields the highest marginal benefit, often identifying over 70% of the losses; breaking down further to the process yields a lower benefit but is still valuable; breaking down to the work team or individual usually yields a benefit lower than the cost of filling in the data and can induce data gaming. The appropriateness of the granularity is judged by whether the breakdown table can tell the reader "which line to move first" within five seconds.

2. Implementation Steps: Five Actions

Action One: Define the Audience and Table List Before Calculating the Numbers. Before starting the calculation, write a one-page "Usage and Delivery Table" listing three to four audiences, which table each person corresponds to, what question each table answers, and the maximum number of pages. The criterion is that each table should have no more than six columns, and any excess should be cut. If it cannot be cut, it indicates that the audience has not been clearly defined. The quality manager should draft this, with confirmation from finance and production.

Action Two: Two Separate Tables for Explicit and Implicit Costs, with the Implicit Table Including an Assumption List. Explicit items should be aggregated based on documents, with responsibility assigned to the workshop and business positions; implicit items should have an assumption list—capacity occupation calculated by actual rework hours multiplied by marginal contribution, inventory calculated by the holding cost rate multiplied by the occupied amount (the holding cost rate should be agreed upon with finance, typically ranging from 12% to 20%), discounts calculated by actual invoices, and customer loss estimated by the historical range of quotation conversion rates. The criterion is that each item in the appendix can answer "what is the assumption, who provided it, and what is the range," and three random items should be answerable upon questioning.

Action Three: Set Subdivision Thresholds According to the 80/20 Rule. Items with an annual amount below 2% of the total or a process loss below 5% of the product line loss should not be further subdivided and should be merged into "other." The criterion is that the breakdown table should cover over 80% of the losses while keeping the number of rows to ten or fewer. Adjust the thresholds quarterly based on actual distribution.

Action Four: Create a Translation Table to Convert Quality Language into the Recipient's Decision Variables. The same loss, for the boss, is "eroding the operating profit margin by 0.9 percentage points"; for procurement, it is "the supplier's incoming quality control (IQC) nonconforming rate translates to 0.13 yuan per part"; for sales, it is "the service cost for this customer has exceeded half of their gross margin, and we need to renegotiate service terms or adjust prices." The translation table should be created once and reused over time. The criterion is that within 24 hours after the presentation, at least one business department should proactively request detailed data—no requests indicate that the translation was not effective.

Action Five: Report Both Management and Accounting Perspectives, and the Difference Must Be Explainable. The quality department should lead the calculation of COPQ (management perspective, including implicit costs), while the finance department should simultaneously produce quality-related expenses (accounting perspective, only including recorded items). Both tables should be presented at the meeting, and the source of the difference should be explained in no more than three sentences, such as the rework pricing method and inventory holding cost. The criterion is that for two consecutive quarters, the explanation of the difference can be independently verified by finance. This step is crucial for gaining credibility—the numbers do not need to be fully agreed upon, but the process must be trusted.

3. Case Development: Three Tables and an Abandoned Order

Returning to the electronics company. At the second meeting, the quality manager brought three tables: a one-page main table for product lines (only containing verifiable losses), a one-page appendix for implicit costs (with an assumption list), and a one-page cross-table for product lines and customers (including service costs).

He only spoke for five minutes, concluding that the consumer adapter line, with annual revenue of 240 million yuan, accounting for 28% of the company's total, consumed 52% of the internal failure costs and 61% of the customer quality service hours. The main table showed a definite loss of 21 million yuan, and the appendix estimated an implicit loss range of 14 million to 26 million yuan, while the gross margin was only 9%, leaving little profit after deducting quality losses. The boss immediately asked: "If we don't accept low-price orders for this line, what will we do with the idle production line?" The quality manager provided a second set of numbers: this line occupies two SMT lines and 60% of the final inspection manpower, which are the most needed resources for expanding the automotive module line, with a gross margin of 24% and a quality loss rate of only one-third of the consumer adapter line.

The decision was made on the same day: within three months, gradually exit from two low-price customers in this product line and reallocate resources to the automotive module line. An additional condition was attached—no delivery breaches during the exit period, and the quality department must assist sales in smoothly transitioning the customers. This became the largest cost item later.

Nine months later, a review was conducted: the company's revenue decreased by 5.8% (approximately 50 million yuan), but internal failure costs decreased by 41%, and external quality service hours decreased by about half. The gross margin increased from 14.2% to 16.5%, and operating profit increased by about 11 million yuan. The quality department did not expand its headcount, only adding one and a half positions for cost analysis, costing about 180,000 yuan per year.

The costs were also real. During the exit process, two customers turned to competitors, one of which was later used as a benchmark case at an industry exhibition, and salespeople still have grievances, with relationships remaining cold. The implicit cost assumptions were questioned more than once, with the most intense debate being the production director's belief that the inventory holding cost rate of 18% was too high. They eventually agreed on 15%, and the appendix was adjusted accordingly, but the conclusion direction did not change—this concession ensured that the reports were not questioned for the next two years. The 83 million yuan figure was recalculated and corrected to 71 million yuan in the second year, and the quality manager proactively explained the correction at the management meeting: "It is better to reduce the number than to soften the criteria."

4. Self-Inspection Checklist

• The report is versioned for different audiences (boss/production/sales), each version answering one question, with no more than six columns.
• The main table only includes verifiable losses, while implicit losses are listed in a separate appendix, with each item clearly stating the assumptions, sources, and ranges.
• Subdivision thresholds have been set, ensuring that the breakdown table covers over 80% of the losses while keeping the number of rows to ten or fewer.
• A translation table has been established, and within 24 hours after the presentation, at least one business department proactively requests detailed data.
• Both management and accounting perspectives are presented at the meeting, and the difference can be independently verified by finance; the calculation results are not tied to the department's current bonus.

If three of the above five criteria are not met, the problem usually lies not in the calculation ability, but in the lack of clarity on "what decision this report is intended to support." The value of COPQ is not in the total amount being scary, but in being able to provide a data-supported answer within five minutes when the boss asks "where should we make changes to be most cost-effective."

Numbers can be corrected, but criteria cannot be softened.

Knowledge code: 4.3.1

Version: v20260920

Author: QTank QTank is dedicated to providing systematic professional knowledge, methodologies, and practical tools for quality management practitioners, helping companies continuously improve their quality capabilities.