Cycle Times

Little's Law in Real Life

Here’s a Little formula:

average number of things in queue = average processing rate x average cycle time

No, that’s actually Little’s Law, named after a MIT professor named John Little. It’s often written as (l = λw).

I first learned about Little’s Law at 18F because it’s a major feature of agile methodology. Specifically, agile practitioners have recognized that if you’re focused on optimizing “cycle time,” and you can’t force software engineers to work faster, the lever available to you is to reduce the number of things in queue. So, agile practitioners focus on managing Works in Progress or “WIP” and establish “WIP limits” (pronounced “whip limits”) as a means of controlling and improving cycle-time performance.

But Little’s Law has applications far beyond software development. Managing cycle time is critical to everything from highway tolling, retail-customer transactions, and more. It turns out that, perhaps, based on recent research, cycle time may also be empirically correlated with success across science, startups, and security. As one of the researchers explained:

But when you look at a success group, with each failure, their efficiency systematically improves the inter-event time between two consecutive attempts systematically decreases. So this means they start to fail faster and faster eventually to succeed.

In other words, as the cycle time between attempts goes down, success increases. This finding certainly could have broad implications about how people and organizations should approach their work. Perhaps a Little goes a long way. :rimshot: