Flowline

FlowlineLearn › Little’s Law

Little’s Law: WIP, throughput, and cycle time

Little’s Law is the closest thing operations has to a conservation law. The average amount of work in a system equals the rate work arrives multiplied by the average time each piece spends inside. It holds for any stable system, with no assumption about how anything is distributed — which is exactly why it is so useful and so easy to misuse.

L = λ W — or, on a factory floor: WIP = throughput × cycle time

What it tells you that you did not already know

Rearranged, the law says cycle time = WIP ÷ throughput. On a line already running at its bottleneck rate, throughput cannot rise — so every extra unit of work-in-process converts directly into lead time. That is the quantitative case for pull systems: the WIP you release beyond what the constraint can absorb does not become output, it becomes waiting.

Measured on Flowline’s canonical line, going from 6 to 24 Kanban cards per edge raises jobs in the system by 53% and measured flow time by 44%, while throughput moves 7%. Extra WIP buys time in system, not output.

The trap: measuring the law with the law

Most dashboards compute cycle time as WIP divided by throughput. Check Little’s Law against that number and you have proved nothing — you restated a definition. For most of this project’s life the law was on the explicitly-not-claimed list for exactly that reason.

What makes it evidence is a second, independent measurement. Flowline now stamps each unit when a chain-head station releases it, has it inherit the latest stamp among its inputs at every station after that, and measures it when it ships. That number contains no WIP term anywhere, so comparing it against a time-averaged job count and a counted throughput is arithmetic that can actually come out wrong.

Measured flow time ÷ (WIP ÷ throughput), mean of three seeds, 3 h each
LineRatio
~4% cumulative scrap0.975
~34% cumulative scrap0.776

The second row is not an error. The law relates WIP to all departures; flow tracking measures only units that ship. A scrapped unit is work-in-process while it lives and then leaves unmeasured, so the ratio is pulled down by roughly the scrap fraction — certified on the low-scrap line, with the bias asserted separately rather than absorbed into a wider tolerance.

What this does not claimA cycle-time readout defined as WIP ÷ throughput is a definition, not a measurement. Flowline labels its in-game readout as exactly that, and reserves the claim for the independently tracked flow time above.

Try it on a real line

Each of these is a fixed-seed puzzle factory in Flowline’s Challenge Mode. Every one is verified before it ships: the intended fix must win, doing nothing must fail, and a plausible wrong fix must also fail — so the puzzle isolates the idea it claims to teach. They run in your browser.

Open Challenge Mode

Keep reading