Count One Changeover as One Elapsed Event

A changeover can contain several stops and brief restarts. If a report treats each segment as a complete changeover, its average may be mathematically correct and unusable for scheduling.

The planner needs the elapsed transition between two defined operating states.

Define the parent event

Agree a start and finish with operations. A useful production boundary may be the last acceptable unit of the outgoing run to the first acceptable unit of the next run. Some decisions also need a separate timestamp for sustained production.

Keep cleaning, tooling, material staging, checks, and short restarts as child activities within that event. Preserve their timestamps and reason codes. The goal is to add a reliable transition boundary, not erase diagnostic detail.

Do not automatically add every child duration. Two activities may occur at once.

Reconstruct a small sample

Select transitions with different products and requirements. Compare the event log with a direct observation or another authorized record covering the same interval.

For a hypothetical transition running from 10:00 to 10:20, the machine logs four stops totaling twelve minutes and brief running periods totaling eight. The average stop lasts three minutes. The elapsed transition still lasts twenty.

That difference does not mean twelve minutes were hidden or that operators misreported. It means the report answered a different question.

Document how retries, breaks, interruptions, and abandoned changes belong to the parent event. Keep incomplete events flagged until their boundary can be resolved.

Feed the schedule the right quantity

Calculate transition allowances from complete events for comparable transition types. Retain the distribution and sample count. A sparse category should carry uncertainty rather than a confident average.

Compare the revised allowances with a representative schedule. Longer measured changeovers do not create a new loss; they reveal time the plan may already be consuming.

Improvement work can then target the child activity that drives the elapsed duration. Faster tooling preparation matters if tooling delays completion. It may change nothing if another required activity still finishes later.

The accounting unit is the whole transition. The diagnostic detail is the work inside it. A useful system preserves both.