top of page

MEP First-Pass Yield: Budget Repair and Repeat Inspection Separately

Oct 25
4 min read
A 120-kit batch splits into 102 first-pass successes and 18 corrected kits.

Original illustrative engineering schematic. The worked calculation below uses explicit example inputs.


A completed quantity register can hide the work that happened between installation and acceptance. If every assembly eventually passes, final acceptance may be 100%, even though many assemblies required correction and another inspection. An estimator who uses only the final accepted count cannot see that extra effort. Conversely, adding replacement quantities for every rejected inspection can overstate materials when the original assembly was repaired rather than discarded.


This article models a hypothetical, non-safety-critical accessory assembly process. It is not a welding acceptance rule, pressure-test procedure or permission to accept defective work. The project quality plan determines the required inspection and disposition. The calculation simply keeps accepted units, inspection events and repair labour on separate bases so an allowance can be checked.


1. Fix the measurement boundary


ASQ defines first-pass yield around units that complete the defined process without rework, rerun, retest or other corrective diversion. Decide which process is being measured before recording a percentage. In this example, the boundary begins when a completed kit enters its first inspection and ends when that first inspection gives a result. A later successful inspection does not retrospectively turn an initial failure into a first-pass success.


Keep a unique identifier for each kit. Count the number inspected for the first time, the number accepted on that first inspection, the number sent for correction, and the number eventually accepted. A record of inspection events alone is insufficient: one kit inspected three times must not appear as three separate kits in the denominator. Record incomplete inspections separately rather than silently treating them as passes.


2. Reconcile an original 120-kit example


Assume 120 kits are assembled. At the first inspection, 102 are accepted and 18 require a minor correction. All 18 are corrected once and accepted at the second inspection. No kit is scrapped, no replacement kit is made, and no further correction is required. These are invented planning inputs, not Nexora production data or a quality benchmark for a contractor.


First-pass yield = 102 / 120 = 85%.


Final accepted quantity = 102 + 18 = 120 kits.


Final acceptance ratio for this closed batch = 120 / 120 = 100%.


Inspection events = 120 initial + 18 repeat = 138 events.


The two percentages answer different questions. The first shows how many kits cleared the stated boundary without correction. The second shows the eventual disposition of this completed batch. Neither establishes compliance by itself: the inspection criteria, evidence and authorised acceptance still matter. Never reduce mandatory checks simply to improve the reported first-pass percentage.


120 initial inspections plus 18 repeats equal 138 events; 120 kits are ultimately accepted and total labour is 90.30 person-hours.

Original illustrative estimating graphic. Inputs are hypothetical and do not establish project performance.


3. Build labour from activities, not one blanket multiplier


Assume initial assembly requires 0.60 person-hours per kit, each inspection requires 0.10 person-hours, and each affected kit requires 0.25 person-hours of correction. The inspection figure is the combined labour for the defined event, not elapsed waiting time for an inspector. Assume the initial assembly allowance excludes inspection and correction. If the historical allowance already includes them, this build-up would double-count effort.


Initial assembly = 120 × 0.60 = 72.00 person-hours.


Initial inspection = 120 × 0.10 = 12.00 person-hours.


Correction = 18 × 0.25 = 4.50 person-hours.


Repeat inspection = 18 × 0.10 = 1.80 person-hours.


Total modelled effort = 90.30 person-hours; extra correction/reinspection = 6.30.


Dividing the clean baseline of 84 hours by 0.85 gives about 98.82 hours, which does not match this model. That shortcut treats the entire baseline as though it must be repeated for failures. Here only a limited correction and another inspection recur; the original assembly is not repeated. A yield multiplier can be valid for a different process, but its assumptions must match what actually happens.


4. Test sensitivity without declaring a forecast


For a planning comparison, keep quantity and activity times fixed while changing only the assumed first-pass yield. At 90%, twelve kits require correction, giving 4.20 extra person-hours and 88.20 total. At 80%, twenty-four kits require correction, giving 8.40 extra person-hours and 92.40 total. This is sensitivity arithmetic, not evidence that a training initiative will achieve either result or deliver a financial saving.


The model assumes every initially rejected kit succeeds after one correction. If repeat failures occur, record the additional repair and inspection cycles explicitly. If some kits must be scrapped, add the appropriate replacement materials and fabrication activities using the approved disposition. Do not count a repaired kit as both a replacement assembly and an accepted original. Separate defect types when their correction effort differs substantially.


5. Carry the evidence into the estimate


A usable estimating record contains the process boundary, inspected population, observation dates, defect categories, correction route, repeat-event count and labour basis. Distinguish person-hours from elapsed programme delay: six hours of repair effort does not necessarily mean a six-hour project extension. Access, inspector availability and sequence can govern elapsed time. A small convenience sample from an easy location may not represent the next package.


Before adopting an allowance, ask the quality and delivery teams whether the observed batch had comparable components, access and procedures. Retain the original records and show a range when the evidence is weak. The practical next step is to add separate correction and repeat-inspection lines to the labour build-up while leaving the accepted quantity unchanged. The MEP Estimators Handbook provides a broader quantity-takeoff and labour-productivity reference; this example does not imply a dedicated yield calculator is included.



Continue the calculation review



Sources and calculation basis


Primary references checked on 25 October 2026. The worked examples and diagrams are original illustrative calculations, not manufacturer selections or project approvals. Apply the stated limitations and the requirements governing the actual project.



Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page