Stop starting, start finishing
October 2026 · agile, flow, complexity
Two teams of three people work through the same endless backlog, item for item. Each item takes between 2 and 8 days of focused work. One team finishes what it has before starting more, so it never has more items in progress than people. The other keeps starting, up to a limit you choose.
Watch the busy team’s first weeks: everyone is working and nothing is done. Then give it a different limit and race again.
Starting more doesn’t make anyone faster. Each person splits their day across more items, so every item waits longer, and every switch between items costs part of the day.
The waiting is the bigger effect. Little’s law says the average time an item spends in progress equals the amount of work in progress divided by how fast the team finishes it. Double the work in progress at the same pace and everything takes twice as long, even before switching costs make the pace worse.
The model is deliberately simple: no meetings, no blockers, no bugs coming back. Real teams have all of those, and each one makes the case for a low limit stronger.