Should I add a buffer to Little's law?

No for the prediction itself: the law already reflects whatever WIP the floor carries. A buffer is a separate target decision, so set the tolerance band and compare the count with it rather than inflating the formula.

What counts as current WIP?

Units released to the line and not yet finished, in the scope you use for throughput. Count physical pieces at a moment in time, or open work orders if they close when parts finish, and use one method throughout.

How long a window should the averages use?

Several times longer than the lead time itself, with a stable mix. Shorter windows mostly measure which orders happen to be open today, which is exactly what the tolerance comparison is meant to expose.

What if throughput changes during the window?

The prediction is only as good as the average. If the rate trended up or down, split the window at the change and run each half separately, or wait for a stable stretch before counting.

Lean Manufacturing & Operations calculator

Little's Law WIP Calculator

Check the WIP the floor is holding against the level Little's law predicts from throughput and lead time. You need the rate, the average lead time, the current count and your tolerance.

What this calculator does

Predicted WIP from throughput and lead time, with the flow time current WIP implies and the deviation from the prediction.

Formula used

Predicted WIP = throughput × average lead time (Little's law)
Flow time at current WIP = current WIP ÷ throughput
WIP difference = current WIP − predicted WIP; flow time difference = flow time at current WIP − average lead time
WIP deviation = |WIP difference| ÷ predicted WIP × 100
WIP tolerance band = predicted WIP × tolerance ÷ 100

Inputs explained

Throughput Rate: Good units finished per day over the measurement window.
Average Lead Time: Average time from release to completion for recent units.
Current WIP: Units released and not yet finished, from today's floor count.
WIP Tolerance: Difference from the prediction you will accept before investigating.

How to use the result

Best suited to line count check against flow time, after a batch size cut, job shop flow audit.
Ramp-ups, shutdowns and priority changes break the steady-state assumption. Averages hide the long tail: one stalled order can carry weeks of flow time.

Current U.S. benchmarks

U.S. manufacturing runs at 75.7% of capacity (Federal Reserve, Aug 2026). New factory orders are up 8.5% year over year (Census).

Common questions

Should I add a buffer to Little's law? No for the prediction itself: the law already reflects whatever WIP the floor carries. A buffer is a separate target decision, so set the tolerance band and compare the count with it rather than inflating the formula.
What counts as current WIP? Units released to the line and not yet finished, in the scope you use for throughput. Count physical pieces at a moment in time, or open work orders if they close when parts finish, and use one method throughout.
How long a window should the averages use? Several times longer than the lead time itself, with a stable mix. Shorter windows mostly measure which orders happen to be open today, which is exactly what the tolerance comparison is meant to expose.
What if throughput changes during the window? The prediction is only as good as the average. If the rate trended up or down, split the window at the change and run each half separately, or wait for a stable stretch before counting.

Last reviewed 2026-10-01.