Why it felt slower and was faster
With forty tickets in flight, every ticket waited behind the others on the same desk. With two per engineer, each ticket got continuous attention until it was finished. Total throughput rose because waiting fell, not because anyone worked harder.
Work becomes visible as cards on a board whose columns are the stages it passes through: to do, in analysis, in build, in test, done, or whatever the real stages are. That alone helps, because it shows where work piles up. The method begins with the second step, which is writing a number above each column and refusing to let more cards than that sit in it. When the build column is full, nobody starts building; they help finish something, or they wait, or they fix whatever is blocking the column downstream. Limiting work in progress feels slower and is faster, because items that are started and not finished are the main reason work takes long, and the limit forces finishing.
- Visualise the work: every item, every stage, on one board everybody can see.
- Limit work in progress: a number per stage, enforced, revised deliberately.
- Manage flow: measure how long items take end to end and how many finish per week, and act on the numbers.
- Make policies explicit: what it means for an item to move from one column to the next, written on the board.
- Improve collaboratively: change the system when the measurements say so, with the people who do the work.
Two measurements carry the method. Cycle time is how long an item takes from starting to finishing; throughput is how many finish per period. Both improve when the limits are lowered, up to a point, because less work in progress means less waiting and less switching. Neither needs an estimate, which is why teams tired of estimating often move to Kanban: a team that knows its cycle time distribution can say when an item will probably finish without sizing it. The method has no fixed iterations, no mandated roles and no events beyond what the team chooses, and it fits work that arrives continuously and unpredictably, such as support, operations and maintenance, better than a sprint does.