How it's calculated

Expected failures m = units × hours per year × (lead time ÷ 12) ÷ MTBF Hold the smallest k where Σ(e^−m × mⁱ ÷ i!) for i = 0…k ≥ target

This suits parts that fail randomly across a fleet, such as motors, drives, sensors and valves. It answers "how many do we need on the shelf to be 95% sure we won't run out before a replacement arrives?"

For a single critical part with a long lead time, holding one spare may be justified by downtime cost alone. See the downtime cost calculator.