Quick answer: Rejection analysis is the systematic process of tracking every rejected part, classifying it by defect type and root cause, and converting the data into corrective actions. Indian MSME factories lose 5 to 15 percent of revenue to poor quality, including scrap, rework, customer returns, and hidden capacity loss. Structured rejection tracking with a cloud ERP, combined with Pareto-driven root cause analysis and closed-loop corrective actions, can reduce rejection rates by 30 to 50 percent within the first year.
Why rejection analysis matters more than rejection rate
Every Indian factory owner knows their rejection rate. It is the number they report in management review meetings and the number customers ask about during audits. But the rejection rate alone tells you almost nothing. A 2 percent rejection rate sounds good until you realize that 80 percent of those rejections come from a single defect on a single machine, and fixing that one issue would cut your rejection rate in half.
Rejection analysis is the discipline of going beyond the number. It answers: what exactly is being rejected, at which operation, on which machine, by which operator, during which shift, and most importantly, why? Without these answers, your quality team fights fires instead of preventing them.
The difference between a factory with 5 percent rejection and one with 1 percent rejection is rarely better machines or better operators. It is better rejection data. The low-rejection factory knows exactly where its defects originate, has eliminated the top three causes, and catches new trends within days instead of months. The high-rejection factory has a notebook of rejection entries that nobody analyses, and the same defects repeat quarter after quarter.
The real cost of rejections: what most factories miss
Most Indian MSME factories calculate rejection cost as the material value of scrapped parts. This captures roughly 20 to 30 percent of the actual cost. The full cost of poor quality (COPQ) includes four layers:
Internal failure costs. These are the costs incurred when a defect is caught inside the factory. Scrap material cost is the obvious one, but add rework labour (the operator who re-machines or re-works the part), re-inspection time (the QC inspector who checks the reworked part again), machine time consumed producing defective parts (this capacity could have produced good parts), and the sorting cost when a batch has mixed good and bad parts and every piece needs individual re-inspection.
External failure costs. These hit when a defect escapes to the customer. Customer return freight (both inbound and outbound), replacement production on an urgent basis (often at overtime rates), credit notes and debit notes, penalty deductions that OEM customers apply for PPM breaches, and the cost of your quality engineer travelling to the customer site for sorting and root cause meetings. For auto component suppliers, a single line stoppage caused by your defective part can result in a penalty of several lakhs.
Appraisal costs. Inspection labour, gauges, CMM time, calibration, and testing. These costs increase when rejection rates are high because more inspection is needed to catch defects. A factory with a 1 percent rejection rate can do sampling inspection. A factory with a 5 percent rejection rate ends up doing 100 percent inspection, which doubles or triples the inspection cost per piece.
Prevention costs. Training, process improvement projects, FMEA development, and quality planning. These are investments that reduce the other three cost categories. Most Indian MSME factories spend too little on prevention and too much on appraisal and failure costs.
Across implementations in Indian manufacturing, we consistently see COPQ between 5 and 15 percent of annual revenue. The government's ZED (Zero Defect Zero Effect) certification scheme has reported that participating MSMEs achieved an 11 percent reduction in COPQ and a 10 percent decline in rework after structured implementation. That is not a marginal improvement. For a factory with Rs. 10 crore annual revenue, a 10 percent reduction in COPQ is Rs. 50 to 75 lakhs saved per year.
Internal rejection vs customer rejection: why the ratio matters
Every defect in your factory will be caught at one of two places: inside your factory (internal rejection) or at the customer's incoming inspection (customer rejection). Where it gets caught determines how much it costs you.
An internal rejection typically costs you the material value plus rework or scrap cost. A customer rejection costs 5 to 10 times more because it adds freight, urgent replacement production, credit notes, customer audit costs, and the intangible cost of eroded trust.
The ratio of internal rejections to customer rejections is a direct measure of how effective your quality system is. If you produce 100 defective parts per month and your internal inspection catches 95 of them, your internal rejection rate is high but your customer rejection rate is low. That is a manageable situation because the defects are not reaching the customer. If your internal inspection catches only 60 out of 100 defective parts, 40 defective parts reach the customer every month. That is a quality crisis waiting to happen.
The target ratio: for every 1 customer rejection, you should be catching at least 10 internally. If the ratio is lower than 10:1, your inspection process has gaps, and defects are leaking through. If the ratio is higher than 20:1, your inspection might be overly conservative (catching too many borderline parts as rejections when they are actually within tolerance), which inflates your internal scrap cost unnecessarily.
The five most common rejection reasons in Indian factories
Across 60+ MSME factory implementations in auto parts, precision machining, sheet metal, and plastics, these five defect categories consistently account for 80 to 85 percent of all rejections:
1. Dimensional out-of-tolerance (35 to 40 percent of rejections). The part does not meet the drawing dimensions. Bore diameter is oversized, length is undersized, hole position is shifted. Root causes: tool wear not monitored (the tool gradually drifts out of tolerance over 50 to 100 pieces), incorrect offset setting after tool change, fixture wear or looseness, and thermal expansion during long machining runs. This is the single largest rejection category in every machining factory.
2. Surface finish defects (15 to 20 percent). Scratches, dents, burrs, chatter marks, and rough finish. Root causes: worn inserts running beyond their tool life, improper handling between operations (parts dumped into bins without protection), incorrect cutting parameters (feed rate too high, spindle speed too low), and contaminated coolant. Surface defects are particularly costly because they are often caught only at final inspection, after all machining operations are complete.
3. Material defects (10 to 15 percent). Wrong material grade, porosity, inclusions, hard spots, and surface cracks in raw material. Root causes: supplier quality issues, inadequate incoming material inspection, mixed lots in the raw material store, and missing material test certificates. Material defects are frustrating because the factory did everything right in production, but the input material was defective.
4. Assembly and process errors (10 to 12 percent). Wrong orientation, missing components, incorrect torque, skipped operations, and wrong sequence. Root causes: missing or unclear work instructions, operator fatigue during repetitive tasks, untrained operators on new products, and no poka-yoke (mistake-proofing) mechanisms. These are purely preventable defects that reflect process discipline rather than technical capability.
5. Cosmetic and visual defects (8 to 10 percent). Paint defects, plating issues, colour mismatch, label errors, and packing damage. Root causes: environmental contamination in paint shops, inconsistent plating bath chemistry, supplier labelling errors, and rough handling during packing and dispatch. Cosmetic defects are subjective, which makes them harder to standardize across inspectors.
How to set up a rejection tracking system
A rejection tracking system does not require expensive quality management software. It requires structured data capture at every inspection point. Here is the minimum data you need to collect for every rejection:
- Part number and description: what was rejected
- Work order or batch number: which production batch it belongs to
- Operation and machine: where the defect was created (not where it was detected, but where it originated)
- Operator and shift: who was running the machine when the defect was produced
- Rejection reason code: a standardised defect code from a predefined list (not free text, because free text cannot be analysed systematically)
- Quantity rejected: how many pieces in this rejection entry
- Disposition: what happens next (scrap, rework, use-as-is with concession, or return to supplier)
- Detected by: which inspection point caught it (in-process, final, or customer)
The critical detail is the rejection reason code. Free-text rejection descriptions like "part not OK" or "dimension problem" are useless for analysis. You need a structured code list. A good starting set for a machining factory includes 15 to 25 codes grouped by category: dimensional defects (5 to 8 codes), surface defects (3 to 5 codes), material defects (3 to 4 codes), process errors (3 to 5 codes), and handling or packing damage (2 to 3 codes). Too few codes and you lose detail. Too many codes and operators default to "other."
Pareto analysis: finding the 20 percent that causes 80 percent of your rejections
Once you have 30 to 60 days of structured rejection data, run a Pareto analysis. Sort all rejection reason codes by quantity (or cost) in descending order and calculate the cumulative percentage. The top 3 to 5 reason codes will account for 60 to 80 percent of total rejections. These are your priority targets.
The power of Pareto analysis is focus. Instead of trying to fix everything at once, you attack the top 3 causes. Fixing 3 issues out of 20 can eliminate 60 to 70 percent of your total rejections. This is where most Indian factories fail: they know their overall rejection rate, but they have never done a Pareto analysis to identify which specific defects are driving it.
Run Pareto at three levels: by rejection reason code (what defects are most common), by machine (which machines produce the most rejections), and by part number (which products have the highest rejection rate). The intersection of these three views pinpoints exactly where to focus. If 40 percent of your rejections are "bore oversized" and 70 percent of "bore oversized" rejections come from CNC Lathe #3, and 90 percent of those happen on Part A, you have a very specific problem with a very specific solution.
Rework management: the hidden cost multiplier
Not all rejections are scrap. Many are reworkable: the dimension is slightly oversize and can be re-machined, the surface has a burr that can be removed, or the assembly is missing a component that can be added. Rework feels like a save because you recover the material value. But rework has its own costs that most factories underestimate.
Double machine time. The reworked part goes through the machine a second time, consuming capacity that could have produced a new good part. If your bottleneck machine is at 90 percent utilization and 5 percent of its output needs rework, you have effectively lost 4.5 percent of your bottleneck capacity to rework. That directly reduces your daily output.
Double inspection. Every reworked part needs re-inspection. The quality inspector checks it once (and rejects it), then checks it again after rework (to confirm it is now within tolerance). Two inspection cycles for one part.
Scheduling disruption. Rework jobs interrupt the production schedule. The machine is set up for the current production order, and a rework batch arrives that requires a different setup. You either do the rework immediately (disrupting the current order) or park it for later (risking that it gets lost or delayed). Either way, the production plan takes a hit.
Traceability risk. Reworked parts can get mixed with original good parts if not tracked carefully. In industries with traceability requirements (auto components, aerospace, medical devices), a reworked part that enters the good stock without proper documentation is a compliance violation that can trigger a customer audit finding.
The correct approach to rework management involves three elements: a separate rework work order that tracks the rework operation, time, and cost independently from the original production order; a mandatory re-inspection step before the reworked part can re-enter good stock; and a rework cost report that shows the true cost of rework by part number and defect type so management can decide whether rework or scrap is the better economic choice.
Root cause analysis: moving from "what" to "why"
Rejection tracking tells you what is going wrong. Root cause analysis (RCA) tells you why. Without RCA, you fix symptoms. With RCA, you fix causes.
The simplest and most effective RCA technique for shop floor defects is the 5 Why analysis. Start with the defect and ask "why?" five times until you reach the root cause. Example: bore diameter is oversize. Why? Tool insert was worn beyond limit. Why? Operator did not change the insert at the prescribed tool life. Why? There is no tool life tracking system, and the operator judges tool wear by feel. Why? Tool life monitoring was never set up because it was considered unnecessary. Why? Management assumed that experienced operators would know when to change tools. The root cause is not the worn insert. The root cause is the absence of a tool life tracking system. Fix the system, and the defect stops recurring.
For recurring or high-impact defects, use the 8D methodology: define the problem, form a team, contain the defect (immediately stop defective parts from reaching the customer), identify root cause, develop corrective actions, implement and verify corrective actions, prevent recurrence, and close the report. OEM customers in the auto sector often mandate 8D reports for every customer complaint. Having the rejection data already captured in your ERP makes the 8D process faster because you can pull the history of that defect type, that machine, and that part number in minutes instead of digging through paper registers.
Customer rejection handling: the process that protects your business
A customer rejection is a quality event that requires a structured response. Handled well, it strengthens the customer relationship. Handled poorly, it costs you the account. Here is the process that works for Indian manufacturers supplying to OEMs and distributors:
Step 1: Acknowledge within 24 hours. Confirm receipt of the rejection notification, assign a quality engineer, and communicate the timeline for your response. Do not debate the rejection at this stage. Acknowledge it.
Step 2: Contain immediately. Check your finished goods inventory and in-transit stock for the same batch. If the defect is batch-specific, quarantine all remaining pieces from that batch. If the defect is systemic, hold all production of that part until the cause is identified. Containment prevents additional defective parts from reaching the customer while you investigate.
Step 3: Investigate and respond. Perform 5 Why or 8D root cause analysis. Provide the customer with a written response that includes: defect description, root cause, immediate corrective action taken, long-term preventive action with timeline, and evidence that the corrective action is effective (data from subsequent production batches). Most OEM customers expect this response within 7 to 10 working days.
Step 4: Track corrective action effectiveness. Monitor the defect for the next 3 to 6 months. If it recurs, the corrective action was insufficient. If it does not recur, close the CAPA (Corrective and Preventive Action) record. Your ERP should flag if the same defect code reappears on the same part number after a CAPA was closed, which indicates the corrective action failed.
Step 5: Update the credit note and quality record. Process the credit note or replacement for the rejected quantity. Update the customer's quality scorecard with the rejection data. If the customer tracks supplier PPM, update your own PPM dashboard to reflect the current status.
How ERPDrive helps manage rejections and rework
ERPDrive is a cloud ERP built for Indian manufacturers, and rejection tracking is integrated into the production and quality workflow rather than being a separate quality module:
Rejection logging at every inspection point. During in-process inspection, the quality inspector logs rejections directly against the work order and operation. They select a rejection reason code from the predefined list, enter the quantity, and choose the disposition (scrap, rework, use-as-is, or return to supplier). This data is immediately visible to the production manager without waiting for a daily or weekly quality report.
Automatic Pareto and trend reports. ERPDrive generates Pareto charts of rejection reasons by part number, by machine, by operator, and by time period. The production manager can see at a glance which defects are driving the most rejections this month, whether rejection trends are improving or worsening, and which machines need maintenance attention. These reports update in real time as rejection data is entered.
Rework work orders with cost tracking. When a rejected part is marked for rework, ERPDrive creates a linked rework work order. This tracks the rework operation, machine time, operator, and re-inspection result separately from the original production order. The rework cost (labour, machine time, and any additional material) is calculated and attached to the original part number, giving you a true cost-of-quality view.
Customer rejection and CAPA workflow. When a customer rejection is received, it is logged in ERPDrive with the customer name, part number, rejected quantity, defect description, and batch or lot number. The system creates a CAPA record linked to the customer rejection. Corrective actions are assigned with due dates and responsible persons. The system sends reminders for overdue CAPAs and flags recurring defects where a previous CAPA was closed but the same defect has reappeared.
Supplier rejection tracking. When incoming raw material fails inspection, the rejection is logged against the supplier and the purchase order. Over time, ERPDrive builds a supplier quality scorecard showing rejection rate by supplier, by material, and by defect type. This data supports vendor evaluation decisions: should you continue with a supplier whose incoming rejection rate is 8 percent, or switch to a more expensive supplier whose rejection rate is 0.5 percent? The ERP gives you the data to make that calculation.
PPM dashboard. For auto component suppliers, ERPDrive calculates and displays your customer PPM (parts per million defective) in real time. You can track PPM by customer, by part number, and over time. When a customer asks "what is your PPM for the last quarter?" during an audit, the answer is one click away instead of a two-day Excel exercise.
A step-by-step rejection reduction plan for your factory
If your rejection rate is above 3 percent and you want to bring it below 1.5 percent, here is a 6-month plan that works consistently across Indian MSME factories:
Month 1: Set up structured rejection tracking. Define your rejection reason codes (15 to 25 codes). Train operators and inspectors on how to log rejections with the correct code. Start capturing data digitally instead of in notebooks. Do not try to fix anything yet. Just collect clean data.
Month 2: Run your first Pareto analysis. With 30 days of data, identify the top 3 rejection reasons by quantity and by cost. Assign a root cause investigation team for each of the top 3. Use 5 Why analysis. Document the findings.
Month 3: Implement corrective actions for the top 3. These are usually straightforward: set up tool life monitoring, fix a worn fixture, add a poka-yoke gauge, update work instructions, or improve incoming material inspection. The fixes for the top 3 defects are rarely expensive. They are usually process changes, not capital investments.
Month 4: Verify corrective action effectiveness. Compare rejection data from Month 3 and Month 4 against the baseline from Month 1. If the top 3 defects have reduced by 50 percent or more, the corrective actions are working. If not, the root cause analysis was wrong, and you need to dig deeper.
Month 5: Attack the next tier. With the top 3 fixed, the Pareto chart shifts. New defects rise to the top. Run a fresh Pareto analysis and attack the new top 3. Also start tracking rework cost separately to identify parts where scrap is cheaper than rework.
Month 6: Establish the routine. Monthly Pareto review, weekly rejection trend review, and CAPA closure tracking become standard practice. By this point, rejection rate should be 30 to 50 percent lower than the Month 1 baseline. The system is now self-sustaining because the data drives the improvement cycle.
Common mistakes in rejection analysis
Using free-text rejection descriptions instead of codes. "Part NG," "not OK," "dimension problem," and "rejected" are entries that cannot be analysed. If you cannot sort and count rejections by reason, you cannot identify patterns. Standardised codes are non-negotiable.
Tracking where the defect was detected, not where it was created. If a bore diameter issue is created at CNC Turning (Operation 2) but caught at Final Inspection (Operation 8), the rejection must be attributed to Operation 2 and the turning machine, not to the inspection station. Otherwise, your Pareto chart shows "Final Inspection" as the biggest problem area, which is misleading because inspection is not creating the defects.
Blaming the operator instead of fixing the system. When 80 percent of bore oversized rejections happen on Machine 3, the problem is Machine 3 (spindle wear, fixture looseness, or thermal drift), not the operator. Penalising operators for rejections creates an incentive to hide defects rather than report them. Fix the machine, fix the process, and fix the training, in that order.
Closing CAPAs without verifying effectiveness. A CAPA that says "operator has been instructed to be more careful" is not a corrective action. It is a hope. Effective corrective actions change the process, the tooling, or the system so that the defect cannot recur even if the operator is not careful. Verify with data: did the defect frequency actually decrease in the 3 months after the CAPA was implemented?
Ignoring rework cost because "we saved the part." Rework is not free. It consumes machine time, operator time, and inspection time. If the rework cost per piece is more than 60 to 70 percent of the manufacturing cost per piece, scrapping and re-making is usually cheaper and faster. Track rework cost per piece to make this decision with data, not intuition.
Frequently Asked Questions
What is rejection analysis in manufacturing?
Rejection analysis is the systematic process of tracking every rejected part, classifying it by defect type and cause, and identifying patterns that point to the root cause. It involves recording what was rejected, at which operation, by which operator, on which machine, and why. The goal is to convert rejection data into corrective actions that prevent the same defect from recurring. Without structured rejection analysis, factories fix symptoms instead of causes, and the same defects keep appearing month after month.
What is the difference between internal rejection and customer rejection?
Internal rejection is when a defect is caught inside the factory during in-process inspection or final inspection before dispatch. Customer rejection is when the defect reaches the customer and is returned or flagged in an incoming quality check. Internal rejections cost you material and rework time. Customer rejections cost 5 to 10 times more because they include freight, replacement production, inspection at customer site, credit notes, and potential loss of future orders. The ratio of internal to customer rejections tells you how effective your quality system is at catching defects before they leave the factory.
How do you calculate cost of poor quality (COPQ) in a manufacturing factory?
Cost of Poor Quality (COPQ) includes four components: internal failure costs (scrap, rework, re-inspection, downgrading), external failure costs (customer returns, warranty claims, credit notes, penalty deductions), appraisal costs (inspection labour, testing equipment, calibration), and prevention costs (training, process improvement, quality planning). For most Indian MSME factories, COPQ runs between 5 and 15 percent of annual revenue. The visible part, which is scrap and rework, is typically only 30 percent of the total. The hidden 70 percent includes lost capacity, expedited shipping to replace rejected lots, and customer trust erosion.
What are the most common rejection reasons in Indian manufacturing?
Across Indian MSME factories, the top five rejection reasons by frequency are: dimensional out-of-tolerance (35 to 40 percent of all rejections), surface finish defects such as scratches, dents, and burrs (15 to 20 percent), material defects including wrong grade, porosity, and inclusions (10 to 15 percent), assembly errors like wrong orientation, missing components, and incorrect torque (10 to 12 percent), and cosmetic or visual defects such as paint defects, plating issues, and colour mismatch (8 to 10 percent). The specific breakdown varies by industry segment, but dimensional and surface issues consistently account for over half of all rejections.
What is a good rejection rate for an Indian MSME factory?
For general engineering and auto component MSME factories, an internal rejection rate below 2 percent and a customer rejection rate below 500 PPM (parts per million) is considered acceptable. Top-performing factories targeting OEM supply achieve below 1 percent internal rejection and below 100 PPM customer rejection. If your internal rejection rate is above 5 percent, you are likely losing 3 to 8 percent of your revenue to scrap and rework. The first goal should be to get below 3 percent through structured rejection analysis and corrective actions, which most factories can achieve within 6 months of implementing systematic tracking.
Can rejection tracking be done without expensive quality software?
Yes. A cloud ERP with built-in quality inspection and rejection logging is sufficient for most MSME factories. Operators log rejections with a reason code, quantity, and operation at each inspection point. The ERP generates Pareto charts of rejection reasons, trend reports by machine and operator, and COPQ calculations automatically. This replaces the manual rejection registers that most Indian factories maintain in notebooks or loose Excel sheets. The key advantage is that digital rejection data can be sliced by time period, part number, machine, operator, and defect type, which is impossible with paper records.
Sources and References
This article cites the following primary sources. Click through for the official source documents: