Rework typically lowers true throughput in aerospace factories, even when reported output looks stable or improving. It does this by consuming constrained capacity at bottleneck steps, adding variability and queues, and loading quality and MRB workflows that do not show up clearly on standard production dashboards.
1. Why the scoreboard lies: first-pass vs “headline” throughput
Most plants report throughput as total units shipped or hours closed in the MES/ERP. If reworked units are counted the same as first-pass units, the headline metric hides a key fact: more capacity was consumed to achieve the same shipped volume.
- First-pass throughput (true flow): units moving from start to finish without rework.
- Headline throughput: all completions, including units that visited one or more rework loops.
As rework rises, you can see “good” shipment numbers while:
- Cycle times extend.
- WIP grows in front of bottlenecks and MRB.
- Schedule adherence and promise dates degrade.
In regulated aerospace environments, where every rework path adds inspections, signoffs, and documentation, the gap between headline and true throughput can be significant.
2. How rework consumes bottleneck capacity
Throughput in an aerospace factory is governed by the true bottlenecks: specialized processes, limited-qualified operators, key inspection resources, and MRB/engineering decisions. Rework directly competes with new work at these points.
- Critical process centers: NDT, heat treat, composites cure, high-precision machining, final assembly integration, and final test are often capacity-constrained. Sending units back to these centers means:
- Added setups, run time, and changeovers for non-standard operations.
- Interrupt-driven prioritization (e.g., hot AOG or late-line rework).
- Reduced available slots for planned first-pass work.
- Inspection and quality bottlenecks: Every rework loop usually requires additional inspections, verifications, and documentation reviews. This increases queue time in front of CMMs, bench inspection, and quality signoff desks.
- MRB and engineering capacity: Complex rework often needs MRB decisions, concessions, and engineering rework instructions. This can be the true hidden bottleneck: queues of NRB/NCRs waiting for analysis and approval, holding WIP and starving downstream stations of releasable work.
In practice, even a modest increase in rework rate at a bottleneck can have an outsized impact on line-wide throughput, particularly in high-mix, low-volume aerospace programs where setups and configurations are already complex.
3. Effects on flow, WIP, and schedule adherence
Rework does not just consume hours; it disrupts flow and predictability.
- Longer and less predictable lead times: When rework volume is variable, queues fluctuate in front of bottlenecks and MRB. That variability often dominates lead time, not the nominal process duration in the routing.
- WIP inflation: As units loop back for rework, WIP counts rise. True WIP becomes a mix of:
- Normal in-process units.
- Units waiting for MRB/engineering decisions.
- Units waiting for rework capacity or specialized tools/fixtures.
- Schedule fragility: Rework tends to be urgent and late in the build sequence. Expedites for rework often cause resequencing, preemption of planned jobs, and reactive scheduling, which further degrade first-pass throughput.
- Knock-on effects in upstream and downstream areas: When assemblies are held for rework, upstream operations may be starved of kitting space or return loops; downstream lines may run short, then overcompensate with overtime or batching when reworked units finally arrive.
4. Impact on capacity planning and utilization metrics
In brownfield aerospace environments with legacy MES/ERP and manual travelers, capacity models often assume ideal or low rework rates. When actual rework is higher:
- Planned vs actual capacity diverge: Resource models underestimate required hours at constrained steps and MRB. Planners may assume a workstation can support a certain number of units/month, but effective capacity is reduced by unplanned rework.
- Utilization looks high, but throughput stagnates: OEE or utilization can be high because resources are busy doing rework, not because the line is producing more good units. Without separating rework hours, management may misinterpret high utilization as “efficient” rather than a symptom of poor quality.
- Program and ramp-up risk: For new programs, unmodeled rework can consume ramp-up capacity, leading to late deliveries and missed ramp curves despite nominal equipment availability.
In regulated aerospace, changing routings or adding new equipment to relieve these bottlenecks carries qualification and validation burdens, which lengthen the timeline to recover true capacity.
5. Documentation, traceability, and compliance overhead
Every rework loop in aerospace must be traceable and documented. This overhead does not show up in simple throughput counts but directly affects effective throughput.
- NCR/MRB processing: Logged nonconformances, MRB decisions, and concessions take engineering and quality time, and often add days or weeks of queue.
- Configuration and build record complexity: Each rework route may require updated travelers, revised digital work instructions, and additional as-built evidence in the MES/QMS. This is especially heavy in AS9100/AS9102 contexts.
- Audit exposure: Higher rework means more variation in actual vs planned process, more deviations, and a more complex traceability story. While this does not directly change physical throughput, it often triggers more controls and checks that further slow flow.
Because full replacement of legacy MES/QMS is disruptive and risky, many plants layer digital rework/NCR workflows on top of existing systems. How well these workflows are integrated strongly affects how much administrative drag rework adds to throughput.
6. Measuring rework’s true impact on throughput
To understand and control rework’s impact, aerospace factories usually need more granular metrics than “units shipped” or basic OEE.
- First-pass yield (FPY) at each key operation: Track FPY separately from overall scrap. Low FPY at a bottleneck is a direct threat to throughput.
- Rework loops per unit: How many times, on average, does a unit revisit a critical station or re-enter MRB?
- Rework hours as a share of capacity: For constrained resources, track what percent of run hours are spent on rework versus first-pass work.
- NCR/MRB cycle time: Long decision times translate directly into WIP aging and schedule slip.
- Lead time variability: Compare lead times for first-pass units to those that experienced rework. The spread shows the stability impact.
In many brownfield plants, assembling these metrics requires stitching data from MES, QMS/NCR systems, and sometimes paper travelers. Data quality and integration maturity will limit how precise this analysis can be at first.
7. Practical ways to reduce rework’s drag on throughput
Given the qualification burden and long asset lifecycles in aerospace, fully replacing core systems or re-laying out entire lines just to address rework is rarely practical. More common levers include:
- Attack the largest FPY losses at bottlenecks: Use structured problem solving (8D/RCCA) on the operations where rework most constrains throughput.
- Improve operator guidance at critical steps: Digital work instructions, in-process checklists, and mistake-proofing at known high-defect operations can materially reduce rework without major equipment changes.
- Streamline NCR and rework workflows: Digitize and standardize MRB processes to reduce queue time, while still maintaining traceability and approvals. Even when physical rework is required, faster decision-making can unblock WIP and stabilize flow.
- Separate and visualize rework load: Make rework visible in planning and dispatching tools so it does not silently consume prime capacity. Some plants reserve specific time windows or capacity slices for rework to protect first-pass flow.
- Feed learnings back to design and planning: Systematically link recurrent rework issues to design reviews, FAI outcomes, and process capability work so the upstream system reduces the creation of rework in the first place.
8. Summary
Rework in aerospace factories almost always reduces true throughput, even when reported output stays flat or improves. It does this by consuming bottleneck capacity, increasing variability and WIP, loading MRB and inspection resources, and adding traceability overhead. The exact impact depends on product mix, system integration, and process maturity, but separating first-pass throughput from total completions is essential if you want a realistic view of capacity and schedule risk.
