flow-efficiency

Use Kingman's formula: wait time grows non-linearly as utilization approaches 100%, going from 4x at 80% utilization to 99x at 99%. Combined with variation in work items, pushing everyone to be busy creates exponentially longer queues and flow times.

Resource efficiency measures how busy people or machines are; flow efficiency measures how much of a work item's lifetime is spent receiving value versus waiting. Knowledge work typically shows only 5-20% flow efficiency, meaning most time is spent waiting in queues.

Pick a flow unit such as a user story, map its journey from entry to delivery, and divide value-adding time by total time in the system. Comparing active work time against queue and blocked time on a sample of stories usually reveals 80-95% waiting.

Resource efficiency fits when the resource is genuinely scarce and expensive, demand is predictable, flow time matters less than throughput, and switching costs are negligible. Most knowledge work meets none of these conditions, so flow efficiency wins.

Little's Law states throughput time equals work in process divided by throughput rate, so reducing items in flight directly shortens delivery time. WIP limits also keep utilization below the zone where Kingman's formula makes queues explode.

Reduce variation first, since teams usually control story sizing more than staffing pressure. Slicing stories smaller and standardizing handoffs narrows cycle-time spread, which improves flow even at the same utilization level.