A team can have five top priorities and very little finished work. Starting another assignment feels responsive, but it also creates another item waiting for attention. A work-in-progress limit makes that trade-off visible.
Define what counts as started and what counts as finished before choosing a number. Otherwise, people can make the board look healthier simply by renaming unfinished work or moving it into an ambiguous column.
Try a small, observable limit
Imagine a three-person team agrees to have no more than three ordinary tasks actively in progress. This is an illustrative experiment, not a universal ratio. When a fourth request arrives, the team must finish something, stop something explicitly or agree on an exception.
Include work waiting for review when it still consumes the team’s capacity. If every task is parked under “almost done,” the limit is not describing the real queue. Show blocked work and name what it needs to move again.
Discuss the oldest unfinished item first. Perhaps a reviewer needs a smaller package, a decision is missing, or a task was larger than expected. The limit should prompt help and clearer decisions rather than blame for not looking busy.
Make urgency accountable
An urgent item may need to interrupt the plan. Record who made that decision and which work is delayed. An unlimited emergency lane simply recreates the original problem under a more dramatic heading.
Review the experiment after a defined period. Look at completed work, waiting time, quality and the team’s experience. A lower number is not automatically better if it leaves essential work stranded; a higher number is not justified merely because every stakeholder wants immediate starts.
The leadership task is to protect the meaning of a commitment. When the team cannot take on another item without delaying something already promised, make that consequence part of the conversation. A visible limit gives people a shared reason to finish, renegotiate or decline.