The Restart Clock Ends at Good Output

Cleaning complete is a useful milestone. It does not necessarily mean the next product is ready to run at its planned rate.

A bakery may still need equipment assembly, approved setup checks, product availability, or first-production acceptance. Treating these as separate reporting categories can obscure the total production effect of a transition.

I would measure the complete restart before deciding which part to improve.

Give the clock explicit endpoints

Start with the last good output of the previous run. Record cleaning completion, mechanical readiness, any required release, first good output, and the point at which the new run reaches its agreed operating condition.

Define those events with the operating and quality owners. The definition must fit the line and product. There is no universal number of minutes or temperature band that establishes readiness.

Keep any production awaiting a release decision separate from accepted output. A running machine and an approved unit answer different questions.

Draw the actual dependencies

Some work must follow other work. Some can run in parallel. Some shares the same people and therefore cannot overlap in practice.

Use the observed timeline to find which unfinished activity actually delayed the next milestone. Adding every task duration will overstate elapsed loss when tasks overlap. Looking only at the longest task will understate it when the sequence contains several dependencies.

For a hypothetical restart, mechanical setup and material preparation might finish together, while a required check starts only after setup. Moving material preparation earlier would not shorten this particular restart. Making that change may still help elsewhere, but the evidence does not support a capacity claim here.

Measure output recovery without counting it twice

During restart, the line may produce at a reduced rate or create rejected units. Compare accepted output with a relevant production rate for that product and configuration over the same interval.

If a rate already counts only good units, do not subtract rejects again. If it counts all units, reconcile rejected and held quantities before valuing the result.

Use a representative run segment as the comparison, not the single fastest minute. Record unusual disturbances separately so they do not silently become the permanent startup standard.

Choose the next intervention

A readiness checklist helps when missing preparation delays restart. Resource coverage helps when a required reviewer is unavailable. Equipment or process work may be appropriate when a repeatable physical recovery limits the sequence.

The measurements should decide which explanation deserves a trial. They should not be used to pressure a team into declaring readiness earlier.

Finally, check whether the recovered interval can complete useful work in the schedule. A shorter restart may reduce paid hours or create room for an order. It may also leave spare minutes that cannot yet be used.

The value claim belongs to the actual conversion. The measurement improvement belongs to knowing when production truly resumed.