How should aerospace teams measure hold time?

Aerospace teams should measure hold time as elapsed time between a clearly defined hold-start event and a clearly defined release event, tied to the part, lot, serial number, work order, operation, and reason for the hold. There is no useful universal number unless the team separates planned process holds from unplanned delays, captures the responsible function, and preserves traceable timestamps.

The most common mistake is treating hold time as a single average. An average can hide expired material exposure, long-tail MRB delays, queue buildup before inspection, or parts waiting on engineering disposition. Use distributions, aging, and reason-code Pareto views, not just a monthly mean.

Define what kind of hold is being measured

“Hold time” can mean different things in aerospace operations. The measurement method should match the control objective.

• Planned process holds: Required waiting periods such as cure, dry, dwell, stabilization, or inspection timing windows. These are usually controlled by the work instruction, specification, or routing.
• Material or environmental holds: Time-sensitive constraints such as shelf life, out-time, pot life, temperature exposure, humidity exposure, or open-container limits. These often require tighter controls than ordinary queue time.
• Quality holds: Parts stopped for inspection, nonconformance review, MRB, containment, rework planning, or customer approval.
• Operational holds: Parts waiting for labor, tooling, machine availability, fixtures, NC programs, paperwork, planning correction, or material replenishment.

These categories should not be blended without context. A required cure hold and a three-day wait for disposition are both “time,” but they do not have the same meaning or risk.

Use event-based measurement

A practical definition is:

Hold time = release timestamp minus hold-start timestamp, adjusted only by approved exclusions.

The hold-start event may be a nonconformance creation, inspection hold status, material quarantine transaction, operation completion followed by required dwell, or manual hold placement. The release event may be MRB disposition, quality release, next-operation start, material release, or completion of a required process window.

Those events should be defined in procedure or controlled work instructions. If each cell, planner, or inspector starts the clock differently, the metric will not survive serious operational review.

Measure more than elapsed time

Useful hold-time reporting usually includes:

• current WIP on hold by age band;
• median, 80th, 90th, or 95th percentile hold time;
• oldest open holds by program, part family, customer, or value stream;
• reason-code Pareto by hold hours, not just hold count;
• owner or queue responsible for the next action;
• holds approaching technical or material limits;
• MRB and engineering disposition aging;
• repeated holds by part number, operation, supplier, or work center.

For regulated or customer-controlled work, teams should also preserve who placed the hold, who released it, when the status changed, what authority was used, and what record supports the decision.

Be careful with exclusions

Calendar time and business time answer different questions. If the concern is customer flow, capacity, or WIP aging, calendar time is usually the safer default. If the concern is department responsiveness, business hours may be useful as a secondary measure.

Do not subtract weekends, holidays, queue time, or waiting-for-customer time unless the exclusion is intentional and consistently applied. For shelf life, out-time, cure windows, and environmental exposure, exclusions may be inappropriate because the material or process condition does not pause when the plant is closed.

Brownfield systems make this harder

In many aerospace plants, hold data is split across MES, ERP, QMS, PLM, inspection systems, spreadsheets, and paper travelers. A clean hold-time metric usually requires integration or at least a controlled reconciliation method. Full system replacement is often unrealistic because of qualification burden, validation cost, downtime risk, traceability obligations, and long equipment lifecycles.

A more practical approach is to define the authoritative event for each hold type, then map the necessary timestamps from existing systems. For example, a QMS may own nonconformance creation and MRB disposition, while the MES owns operation start, operation complete, and shop-floor hold status. ERP may show material quarantine or inventory status, but not the real shop-floor reason for delay.

Common failure modes

• Missing scans: Operators release or move parts physically before the system status changes.
• Weak reason codes: “Other” becomes the largest category and the metric stops being actionable.
• Overlapping holds: A part may be waiting on quality and tooling at the same time, causing double counting unless rules are defined.
• Partial lots and split orders: Some units move while others remain held, which requires serial- or lot-level handling.
• Timezone and daylight saving errors: Multi-site systems can distort elapsed time if timestamps are not normalized.
• Manual workarounds: Paper tags, emails, and informal quarantines create real holds that never appear in the system.
• Metric gaming: Teams may release and re-hold items to reset aging unless audit trails and rules prevent it.

What good looks like

A credible aerospace hold-time metric has controlled definitions, traceable timestamps, reason codes with ownership, aging visibility, and exception handling for technical limits. It should support operational decisions without implying that a dashboard alone controls compliance, safety, or customer approval.

The metric also needs periodic review. If the data is not used for disposition aging, bottleneck removal, staffing decisions, supplier follow-up, or process correction, it will usually degrade into another report that people no longer trust.