Little's law calculator: work in process, throughput and lead time

Little's law says that, in a stable system, the average number of items inside equals the throughput times the average time each item spends there: L = λ × W. Enter any two to get the third, such as lead time from work in process and throughput, then check the law on a simulated process.

Work in process 60, 30 per hour120 minW = L ÷ λ = 60 ÷ 30 = 2 hours of lead time
Simulated process: items inside2.877time average, 12 seeded replications
Throughput × time in system2.8771.669 per minute × 1.724 min

How to use Little's law

  1. Choose what you want to find: items in the system, throughput or time in the system.
  2. Enter the other two averages for a period when the system was stable.
  3. Read the result and the time between two finished items.
  4. Optionally enter a process of your own and press the check to see the law hold in a simulation.

Solve L = λ × W

What do you want to find?

Starting values are an example. Use averages over a long period in which the system was not filling up or emptying: Little's law holds for a stable system in steady state.

Result

Items in the system60 L = λ × W
Per item2 min one item leaves every

Units: L in items, λ in items per hour, W in minutes; the formula is applied as L = λ × W ÷ 60.

Check it on a simulated process

Open this exact case in Simflowly

The law and its three forms

L = λ × W W = L ÷ λ (lead time from work in process and throughput) λ = L ÷ W (throughput from work in process and lead time)

L is the average number of items in the system, λ the average rate at which items leave (equal to the rate they arrive when the system is stable) and W the average time an item spends inside, from entry to exit. Use consistent units: this calculator takes λ per hour and W in minutes and divides by 60.

Worked example

A fulfilment process has 60 orders open on average and completes 30 an hour. Lead time is W = L ÷ λ = 60 ÷ 30 = 2 hours, or 120 minutes, and one order leaves every 2 minutes. To halve the lead time at the same throughput, you must halve the work in process. That is the sense in which WIP limits shorten lead time: the law is arithmetic, not a policy.

Little's law in the engine

Simflowly's simulation reports the time-average number in the system, the throughput and the time in the system as separate measurements, so the law can be checked rather than assumed. A queue with 100 random arrivals an hour, 4 servers and a mean service time of 1.5 minutes with spread cv 0.6 (seed 2026, 12 replications of 20,000 minutes) gave:

Little's law measured in an engine run
MeasuredValue
Throughput λ100.13 per hour
Time in the system W1.7 min
λ × W2.877 items
Items in the system L (95% interval 2.862 to 2.892)2.877 items

The law holds for any service-time shape and any arrival pattern, which is why it is the check to run first on any model: if L, λ and W do not agree, the run was too short, it started out of steady state, or the measurements use different boundaries.

When it does not apply

  • Not stable. If WIP grows during the period, the averages of arrivals and departures differ and the law is only approximate.
  • Different boundaries. L, λ and W must describe the same set of items: the same start and end points.
  • It says nothing about variation. The average lead time does not tell you how long the slowest orders wait; for that, you need a simulation or a distribution.

The law was proved in general by John D. C. Little, "A Proof for the Queuing Formula: L = λW", Operations Research 9(3), 1961.

Questions

What is Little’s law in simple terms?

It links three averages: the number of items inside a system, how fast they come out and how long each one stays. If 60 orders are open and 30 are completed an hour, the average order takes 2 hours. It needs no assumptions about arrival or service distributions, only that the system is stable.

How do I calculate lead time from WIP?

Divide work in process by throughput: lead time = WIP ÷ throughput. With 60 items in process and 30 finished an hour the lead time is 2 hours. Use the average WIP over the same period as the throughput and make sure both refer to the same items and the same boundaries.

When does Little’s law not apply?

When the system is filling or emptying during the measurement, when items are counted differently on entry and exit, or when averages come from different periods. It also gives nothing about the spread: two systems with the same L, λ and W can differ greatly in how long the slowest items wait.

Does the law hold in the simulation?

Yes, as it must. In the example, with 4 servers and a service time spread of cv 0.6, the engine measured 2.877 items inside against 2.877 from throughput times time in the system; the largest gap in any single run was 0.02%.

Keep going

Updated 20 November 2026. Free to use, no account. The calculation runs in your browser; nothing you type is sent to us. The engine behind the simulation checks is validated against exact queueing results.