What is an appropriate level of concession volume for a mature program?

There is no single acceptable concession volume that applies across all mature programs.

In general, a mature program should not rely on concessions as a routine operating mechanism. Concession volume should be low, stable, and trending downward or at least remaining within a clearly understood control band. If concessions become a normal path to ship product, that usually indicates unresolved process capability issues, design tolerance problems, supplier variation, inspection escape patterns, or weak change control.

That said, the right threshold depends on context, including:

• product criticality and risk tolerance
• customer and contractual requirements
• whether the program is production, sustainment, repair, or spares-heavy
• process capability and measurement system quality
• supplier quality performance
• how a concession is defined, classified, and counted in your quality system

So the practical answer is not a benchmark percentage by itself. The better question is whether concession volume is expected, exceptional, and explainable.

What good looks like

For a mature program, concession activity is usually appropriate when all of the following are true:

• most concessions are isolated rather than repetitive
• repeat concessions on the same part family, operation, tool, or supplier are rare
• each concession has clear technical rationale and disposition traceability
• there is visible linkage to corrective action when recurrence crosses a defined threshold
• the business is not using concessions to mask chronic scrap, rework, planning instability, or schedule pressure

If recurring concessions are common, the program may still be shipping product, but it is not operating in a mature state in any meaningful quality or execution sense.

What to measure instead of asking for a single number

A raw concession count can be misleading. A better assessment uses a small set of normalized measures such as:

• concessions per 100 units, lot, or work orders
• concessions by defect type, process step, and part family
• repeat concession rate within a defined time window
• concessions tied to supplier-caused versus internal-caused nonconformances
• cost, cycle time, and queue impact of concession processing
• share of concessions closed with permanent corrective action versus administrative closure only

This matters because a low count can still hide significant risk if the same issue recurs on critical hardware, while a higher count in a complex sustainment environment may be understandable if tightly controlled and non-repetitive.

When concession volume is too high

Concession volume is probably too high for a mature program if any of these patterns appear:

• the rate is flat or increasing after process stabilization
• the same dimensions, features, documents, or suppliers drive repeated concessions
• MRB capacity is overloaded and becomes a production bottleneck
• engineering spends material time reviewing predictable issues that should have been designed out or controlled out
• operators or supervisors expect concessions as part of normal completion
• schedule recovery depends on concession approval speed

At that point, the issue is not just quality. It is also capacity, cost of poor quality, and evidence that process learning is not being converted into standard work, tooling changes, supplier controls, or design updates.

Brownfield reality

In many regulated plants, concession volume is distorted by system fragmentation. The same event may touch MES, ERP, QMS, and MRB workflows differently, and some plants still carry part of the record in paper or spreadsheets. That means your concession baseline may be unreliable until definitions, data capture, and workflow ownership are cleaned up.

Because of that, full system replacement is usually not the first answer. In long lifecycle, qualified environments, replacement programs often fail or stall due to validation effort, downtime risk, integration complexity, and the burden of preserving traceability across legacy records. A more realistic path is often to standardize definitions, improve routing and evidence capture, and integrate existing systems well enough to separate true concession demand from administrative noise.

Practical guidance

An appropriate level for a mature program is the lowest level that is sustainable without pushing risk downstream, while keeping each concession exceptional, justified, and traceable.

If you need a management rule, use this one: concessions in a mature program should be rare enough that recurrence stands out quickly, not common enough that people stop noticing. Set internal thresholds by product family and risk class, then review trend, recurrence, and operational impact together rather than relying on a generic industry target.

No single concession percentage can tell you whether the program is healthy. Recurrence, concentration, and dependence on concessions for routine output are usually the more important signals.