My issue-tracker workflow
Issues, cycles, and how I ship calmly every week without a burndown chart.
26 articles
The tracker is not the process. The process is deciding what not to do, and the tracker is where that decision becomes visible to other people.
Three rules
- Anything without an owner is not a task, it is a wish
- A cycle that is full is a cycle that is finished being planned
- If something rolls over twice it gets closed, not moved
The third rule was the one that changed everything. A backlog you never close is a list of things you have decided to feel bad about.
Cycles are a forcing function
A cycle is not a deadline. It is a container with a fixed size, and its value is entirely in what it makes you refuse. A backlog is infinite and therefore free; a two-week container costs something to put things into, which is the only mechanism I have found that produces honest prioritisation.
Estimating badly on purpose
We estimate in three sizes: a day, a few days, and too big. Anything in the third bucket gets broken up before it can be scheduled. That is the entire estimation process and it has been more accurate than the two point-based systems it replaced.
The accuracy comes from the third bucket, not the first two. Most schedule failures are not small mistakes accumulating; they are one item that was secretly four items.
What I stopped doing
- Burndown charts — they measured whether we guessed well, not whether we shipped
- Carrying items forward silently
- Statuses between “in progress” and “done”
- Estimating anything in the backlog before it was scheduled