The Yield Looks Good. The Rework Cost Says Otherwise.
A final pass rate can look healthy while repeated touch-up, retest, diagnosis and component replacement quietly consume margin and capacity.
The short answer
Final yield can look healthy even when the factory spends heavily to recover failed units. Electronics rework cost tracking should connect each defect and repair to labour, replaced material, retest, equipment time, queue delay and shipment impact.
The right question is not only “Did the unit eventually pass?” It is “How much avoidable work was required to make it pass, and what decision will prevent recurrence?”
Ninety-eight percent passed eventually
The monthly dashboard shows a strong final yield. Almost every unit shipped. The quality team solved difficult problems, production met the quantity, and customer returns remained low.
At the same time, the rework bench worked overtime. Test queues grew. Engineers investigated recurring faults. Extra components were consumed without clear job cost. Urgent orders waited behind units that had already been processed once.
The outcome metric is positive. The process economics are not.
Separate first-pass and final yield
First-pass yield measures units that complete the defined process without repair or repeat. Final yield measures units ultimately accepted.
Both are useful. First-pass yield reflects process stability and direct flow. Final yield reflects the factory’s ability to recover acceptable output. Reporting only final yield hides the distance between them.
Define the denominator and stage consistently. A rolled first-pass measure across several tests can behave differently from a product-level measure. Management should know exactly what the number represents.
A defect creates several forms of cost
Direct repair labour is the most visible cost, but rarely the largest complete consequence.
The unit may consume diagnostic engineering, component replacement, solder materials, repeated test time, additional inspection, handling and documentation. It occupies work in progress, competes for constrained equipment and may delay the order.
Some failures damage panels or associated units. Others cause line stops, extra setup or containment of a wider lot. Customer notification and expedited freight can follow.
Cost tracking should capture useful categories without turning every technician into a cost accountant.
Begin with the defect event
The original failure should identify product, revision, serial or batch, process, failure code, time, equipment and work order. Diagnosis adds suspected or confirmed cause. Repair records the action, labour and material. Retest records the outcome.
These relationships allow cost and recurrence analysis by the condition that created the work. A generic monthly “rework expense” cannot show which product or decision should change.
The article Does the Serial Number Survive the Rework Bench? explains how the unit history should remain connected through repair.
Use practical labour capture
Minute-perfect time entry can create more burden than insight. Define a method appropriate to the operation: start-stop scanning, standard time by repair code, bench attendance allocation or a hybrid for complex cases.
Accuracy should be sufficient to distinguish recurring high-cost problems and understand capacity. Review obviously missing or extreme entries, but do not encourage technicians to optimise the time record at the expense of diagnosis quality.
Separate hands-on repair from waiting where possible. A unit can spend two days in rework while receiving twenty minutes of labour. Both flow time and labour matter.
Replaced material must reach the job cost
Components used during repair often come from bench stock or miscellaneous issue. If consumption is not connected to the serial and work order, inventory and cost both drift.
Capture the replacement component, quantity and lot at the required traceability level. Record scrapped or returned removed material where meaningful.
High-value components, repeated replacements and parts associated with safety or reliability deserve stronger control. Low-value consumables can use simpler policies, but their aggregate cost should not disappear.
Retest consumes constrained capacity
A repaired unit often returns to equipment already planned for new production. Retest can become an invisible demand stream.
Measure test attempts, duration and queue by product and failure. If a product routinely requires repeat cycles, nominal test capacity understates the true load.
Scheduling and quotation should use evidence from actual repair demand. Otherwise the factory keeps promising capacity that quality recovery has already consumed.
Delay is a commercial cost
Not every hour of rework changes shipment, but unresolved failures can block lot release or consume the buffer protecting delivery.
Connect rework status to order milestone and promise. Show whether the affected quantity can be substituted, partially released or must hold the shipment. This helps customer teams communicate evidence rather than wait for a final surprise.
Where rework causes expediting, premium freight or customer concession, attach that consequence to the underlying event.
Distinguish recovery from correction
Repairing a unit restores product value. Corrective action prevents recurrence. A factory can become highly skilled at recovery while leaving the original process unchanged.
Use thresholds by recurrence, severity, cost and customer risk to trigger structured investigation. The trigger should consider aggregate pattern, not only dramatic individual failures.
Link corrective action to the relevant product, component, programme, process or instruction. Then verify whether first-pass performance improves after implementation.
Avoid blaming the detection point
The station that records a defect may not be the process that created it. Test and inspection often reveal upstream conditions.
Analysis should distinguish origin, detection and responsibility. Rewarding low defect counts at one station can discourage honest reporting or push problems downstream.
Cross-functional review should ask what condition allowed the defect, why it escaped earlier and what control would prevent or detect it more economically.
Add rework to product profitability
Standard cost may include an assumed scrap or efficiency factor, but actual rework can vary significantly by product, revision, launch stage and customer requirement.
Connect repair labour, material and test consumption to job or product margin. This reveals a bestseller or strategic programme whose revenue looks healthy while recovery erodes profitability.
Quotation can then use realistic yield and test demand. Engineering can compare redesign cost with recurring repair. Operations can prioritise improvement by total business consequence.
Build a small useful dashboard
Start with first-pass yield, final yield, repaired units, average repair labour, replacement material, repeat test attempts, rework queue age and estimated total cost.
Allow users to filter by product, revision, customer, line, defect and cause. Show trend and recurrence rather than a static monthly total.
Do not present cost as false precision when capture is still maturing. Label estimated and actual components clearly. Consistent directional evidence is valuable if its limitations are understood.
The wider operational context appears in One Product, Hundreds of Decisions. Can Your Factory See Them All?.
Run a ten-unit cost study
Select ten recently repaired units from one recurring defect family. Reconstruct failure, diagnosis, labour, material, retest and elapsed time. Estimate downstream delay and expediting where relevant.
Compare the total with the value shown in the current job-cost report. The difference defines the hidden-cost problem more convincingly than a general discussion.
Then select one preventive action and measure the next production run. Improvement should reduce both failure occurrence and recovery demand.
Set an improvement threshold
Not every individual repair needs a formal project, but repeated cost should not remain invisible because each event seems small. Define thresholds using recurrence, severity, cumulative labour, replacement material, test demand and shipment risk.
A low-value defect repeated hundreds of times may deserve earlier action than one dramatic but isolated failure. Review the threshold monthly as product mix changes. Assign an owner, expected benefit and verification date to every selected action, then compare the next run with the original baseline.
Frequently asked questions
Is first-pass yield always the best quality measure?
No single measure is sufficient. First-pass yield reveals direct flow, while severity, customer impact, defect recurrence and final acceptance add necessary context.
Should rework labour be charged to production or quality?
Accounting treatment varies, but operational analysis should connect the cost to the product, process and cause that created it. Department allocation should not erase responsibility for improvement.
How accurate must repair time be?
Accurate enough to support capacity, costing and prioritisation. Use a simple repeatable method before pursuing minute-level precision.
Does successful repair mean the process is effective?
It means recovery was effective. Process effectiveness requires fewer defects and less recovery demand over time.
Can rework data improve quotation?
Yes. Actual yield, repair labour, material replacement and retest demand help estimators price the real production route, especially for complex or immature products.
Count the work behind the pass
A green final result is important, but it should not hide the path taken to reach it. Every repeat test, repair touch and replaced component consumes capacity and margin.
Connect that work to the unit, product and cause. Then quality achievement and commercial performance can improve together instead of one masking the other.