Replenishment

Keep the next queue actionable

Replenishment is not the same as filling every future slot. It is the policy-driven act of selecting enough prepared work for the delivery system to pull while preserving the ability to respond to new information. The right queue size depends on flow time, uncertainty, and the cost of stale priority.

A concrete example

An illustrative team has capacity to pull four items next week. During replenishment, it confirms three ready items and keeps one slot for a reliability request likely to arrive. The team avoids both an empty queue and a large pile of speculative commitments.

Use evidence

Review backlog age, arrival rate, service classes, and cycle-time distributions. Prepare the next item before capacity becomes idle, but do not turn preparation into a large batch of work that will never be selected.

Limitations

Replenishment cannot remove uncertainty from discovery work. It also depends on product and operational decisions outside engineering analytics. Keep exploratory work visible under an appropriate policy rather than forcing false readiness.