Reconcile Downtime Before Changing the Rate

A low downtime percentage is not proof of a measurement fault. A disagreement between the report and independently observed events is a reason to investigate.

Start by making the two records describe the same thing.

Align the boundary

Choose one line and a defined observation window. State which equipment the report covers, when its production clock begins and ends, and how it treats breaks, setup, waiting and short stops.

Observe events over that same boundary. Record start and end times and what the observer can establish about the cause. Preserve unknown causes as unknown.

A stationary machine may be planned idle, starved, blocked or stopped by a fault. Those states can have different reporting treatment while all affecting the schedule.

Reconcile events before percentages

Match observed intervals to recorded events. Look for missing intervals, time-zone or clock differences, threshold filters, overlaps and events assigned to another asset.

Only then compare totals. Ten minutes missing from a two-hour production window and ten minutes missing from a full shift do not produce the same percentage.

Do not force every loss into availability. A reduced rate may appear in a performance measure, while a setup interval may be treated separately under the plant's definitions. The schedule still needs visibility into the time consumed.

Validate the reference rate separately

A rate used for performance comparison needs an explicit basis for the product and configuration. A machine nameplate is not automatically the right baseline for every application.

If the reference changes, preserve the old basis and distinguish a recalculated metric from a physical improvement. Correcting a dashboard cannot by itself create throughput.

Close the measurement change

Assign each discrepancy to its capture, classification or reporting cause. After correction, repeat an appropriate matched observation and retain the reconciliation.

Use the trustworthy record to test the next operating decision. It may support a maintenance action, a different release schedule or further measurement. It does not automatically establish excess staffing or a case for new equipment.