Rework, Scrap, or Concession? — A Five-Step Cost Decision Method for Nonconforming Products

A 40-minute MRB meeting at an automotive parts company was held to discuss a batch of 2000 injection-molded parts with surface shrinkage marks. The production department suggested rework: "A bit of polishing and they'll be good to use, the cost is the lowest." The quality department advocated for scrapping: "No one can guarantee the performance and appearance consistency of reworked parts, and if the customer finds out, it could lead to a batch complaint." The sales department proposed a concession: "There are only five days left until the delivery date, and scrapping and re-producing will definitely not meet the deadline. Negotiating a price reduction for release will result in the least loss."

Each department had a valid point, but no one could convince the others. The vice president finally made the decision: "Calculate the cost, and go with the cheapest option." — But the problem lies precisely here: no one in the meeting room could calculate the cost of these three options. How long would rework take, how much production would need to be made up for scrapping, and how much subsequent risk would a concession carry? These were all "intuitive" judgments. Thus, the decision became a game of positional bargaining: the production department always leans towards rework, the quality department always towards scrapping, and the sales department always towards release.

The disposal of nonconforming products is the most routine and costly decision point in quality costs, yet most companies lack a standardized approach. This article provides that standard: it places the costs of the three disposal paths on the same table, compares them using "risk-adjusted relevant costs," and reaches a conclusion within five minutes, leaving a traceable record of the calculations.

1. Why "Choose the One with the Least Loss" is a Pitfall

The costs of the disposal paths may appear as three different types of losses, but their structures are entirely different and cannot be simply compared by amount.

The cost of rework and repair is "additional investment." It includes rework labor hours, auxiliary materials, re-inspection after rework, and the time occupied by equipment and production lines. The larger the batch, the more the cost grows linearly. The hidden cost lies in the fact that rework is a second processing, which can alter the dimensional chain, surface condition, and assembly stress. The failure rate of reworked parts is typically higher than that of parts produced in one go.

The cost of scrapping is "replacement." Besides the loss of already invested materials and labor hours, the real cost is re-production—re-production will occupy production capacity, require rescheduling, and may delay delivery. In a factory with saturated capacity, this opportunity cost is often much larger than the material loss.

The cost of concession is "risk exposure." The direct cost is only the customer discount, price reduction, and the labor hours for review and communication, making it appear the cheapest. However, it transfers the risk to the future: the probability of this batch being detected by the customer, and the external failure cost per piece (return logistics, production line stoppage claims, fines, expedited re-supply, loss of trust) should all be included based on their expected values.

Therefore, the comparison should not be about "which path costs the least," but rather "which path results in the lowest risk-adjusted relevant cost for the company." There are two key terms here.

One is relevant cost. Only compare the costs that will change due to this decision. The materials and labor hours already invested cannot be recovered regardless of the chosen path and do not serve as a basis for decision-making—unless scrapping requires re-production, in which case the consumed production capacity is a relevant cost. One of the most common errors in quality cost calculations is mixing sunk costs with relevant costs.

The other is risk adjustment. The probability and cost of reworked parts and concessioned parts being detected by the customer differ. A classic empirical ratio is 1-10-100: it costs 1 yuan to resolve a defect in the prevention stage, 10 yuan to detect and handle it within the factory, and 100 yuan if it reaches the customer. Using this scale, the seemingly cheapest part of a concession is often severely underestimated.

2. Five-Step Method: Calculate the Three Paths on the Same Table

Step One: Define the Decision Unit and Boundaries.

First, clearly describe what this batch of goods is: batch number, quantity, current status (completed and awaiting inspection, work-in-progress, finished goods in inventory), materials and labor hours already invested, remaining time until delivery, and whether it involves regulatory or safety characteristics. This step prevents miscalculations—reworking a semi-finished product and reworking a finished product have entirely different cost structures; a delivery date 20 days away and one with only 5 days left have entirely different costs for scrapping and re-production.

At the same time, establish a compliance red line: items involving safety characteristics, regulatory characteristics, or customer-mandated prohibition of concessions should be immediately removed from the candidate options and not enter the cost comparison. Compliance issues are not subject to economic discussions; they are a prerequisite, not a variable.

Step Two: List Cost Elements for Each Option, Retaining Only Relevant Costs.

Create a column for each of the three options and go through the following checklist. Fill in the amount if applicable, write zero if not, and mark "to be confirmed" if uncertain. Do not leave any blanks.

When listing costs, follow this principle: only include expenses that will occur because "this option is chosen." The material costs already incurred are not a basis for decision-making, but the new material required for re-production is.

Step Three: Convert Hidden Costs into Money.

This is the key to making the table useful. Without converting hidden costs, the table will only show labor hours and materials, and the decision will revert to intuition.

• Labor Rate = (Direct labor + social security and welfare + workshop indirect costs) ÷ effective labor hours. Do not use only the wage rate. Using the wage rate alone can make rework appear 30% cheaper.
• Capacity Opportunity Cost = occupied machine hours × marginal contribution of those hours. When capacity is available, it can start from zero; when capacity is saturated, it must be included, otherwise scrapping will always seem more expensive than rework.
• Expected External Failure Cost = exposure probability × external failure cost per piece × quantity. The exposure probability is estimated using historical data (detection rate of similar defects at the customer's site), and the external failure cost per piece is the average of return logistics, rework or replacement, customer production line stoppage claims, and fines.
• Concession Discount = negotiated price reduction per piece × quantity, plus the implicit impact of extended payment terms and future order shares.

Step Four: Calculate the Total and Use Break-Even Points for Quick Judgments.

Sum up each column to get three "risk-adjusted total costs," and sort them from smallest to largest. The lowest one is the recommended option. Also, calculate two critical values, which are more valuable for decision-making than the total:

• Break-even Point for Scrapping = total relevant cost for scrapping this batch
• Break-even Point for Concession = total relevant cost for conceding this batch

Take the smaller of the two (i.e., the total cost of the second-best option), divide it by the batch quantity, to get the "maximum rework cost per piece": as long as the actual rework cost per piece is below this value, rework is the optimal solution. This way, the site does not need to fill out the entire table each time; a simple division and comparison can make the judgment.

Calculation Example: A batch of 2000 injection-molded parts with surface shrinkage marks from an appliance production line, with a unit selling price of 65 yuan, material cost already invested at 18 yuan/piece, re-production of 2000 pieces requiring 8 hours of production line time (with a marginal contribution of 6000 yuan/hour for this line), rework labor hours at 0.12 hours/piece, labor rate at 45 yuan/hour, and a customer discount negotiation space of 12 yuan/piece.

The conclusion is clear: rework is the cheapest, followed by concession, and scrapping is the most expensive. When converted to a per-piece cost, rework is 13.75 yuan/piece, below the 15.55 yuan/piece limit calculated for the concession option, so rework is the right choice. If the rework labor hours increase to 0.18 hours/piece (raising the rework cost per piece to 16.45 yuan), the conclusion would flip to a concession release—the decision boundary is not in the total cost, but in the changes of a few key unit prices.

Step Five: Execute, Document, Update Records, and Trigger Improvements.

Once the decision is made, the action is not just "rework." Three things must be documented:

1. The "Cost Calculation Table for Nonconforming Product Disposal," archiving the data for the three options, assumptions, and sources of exposure probability, so that the next similar issue can directly use the data.
2. The actual disposal costs should be recorded in the "internal failure cost" section of the quality cost ledger, ensuring that the disposal losses are reflected in the monthly quality cost report.
3. The issues triggered by this disposal should be converted into improvement items—a batch of 2000 reworked parts costing 2.75 million yuan, if similar shrinkage marks occur twice a month, the annual cost is 66 million yuan, enough to invest in a mold temperature monitoring and process parameter alarm system. Disposal decisions address the current batch, but root cause improvements address the next batch.

3. Six Common Pitfalls

Pitfall One: Only Calculating Labor Hours and Materials, Not Capacity Opportunity Cost. In a factory with saturated capacity, the true cost of scrapping and re-producing is the displaced orders. If this is recorded as zero, scrapping will seem abnormally cheap, and rework will seem abnormally expensive, leading to the opposite conclusion.

Pitfall Two: Using Sunk Costs as a Basis for Decision-Making. "The materials are already invested, scrapping is too wasteful" is an emotional judgment, not a cost judgment. The materials already invested cannot be recovered on any of the three paths; they only influence whether to re-produce, not whether this batch can be used.

Pitfall Three: Assuming Rework Equals Restoration. Rework is a second processing, which can alter the dimensional chain, stress state, and surface quality. The residual failure probability of reworked parts must be included in the expected loss, otherwise, the cost of rework will always be underestimated. The compliance issues of reworked parts being shipped as new parts also need to be separately confirmed.

Pitfall Four: Only Calculating the Price Reduction for Concession, Not the Risk Exposure. The visible cost is the 12 yuan/piece discount, but the invisible cost is the 8% probability of being detected by the customer and the 320 yuan external failure cost per piece. Converting the invisible part into an expected value often significantly reduces the apparent advantage of the concession option.

Pitfall Five: Using Wage Rate Instead of Labor Rate. Omitting social security, welfare, and workshop indirect costs artificially lowers the costs of rework and re-production, leading to a systematic bias towards "using more labor, scrapping less," which is often the most expensive path.

Pitfall Six: Not Documenting Decisions, Rehashing the Same Issue Each Month. Without a record of cost calculations, the disposal conclusions cannot be reused, and the MRB meeting each month will be the same. The actual disposal costs also cannot be recorded in the ledger, leaving the internal failure cost in the quality cost report incomplete.

4. In Summary

The disposal of nonconforming products should not be decided by who has the loudest voice in the meeting room, but by the risk-adjusted relevant costs—once calculated clearly, most disputes are no longer valid.

Use cost calculations, not positional arguments

Knowledge code: 4.3.1

Version: v20261012

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.