From 300 days to 30: the unglamorous work of fixing delivery

Delivery organisations rarely fail loudly. They fail slowly, through a backlog that stops being triaged, releases that become risky enough to postpone, and a team that quietly reorganises itself around avoiding blame. By the time the metrics are bad enough to escalate, the underlying problem is cultural rather than technical.

The first move was to make the truth visible without making it personal. One board, every defect, an owner and an age against each. No new tool, no new process document, just the honest picture in one place that the whole team and I looked at together every morning.

The second move was to give things back their owners. A defect without a named owner is not a defect, it is a wish. We closed or consciously parked everything nobody would own, which removed a surprising share of the backlog on day one.

The third was cadence. We committed to a release rhythm and protected it, which forced work to be sliced small enough to actually finish. Automated pipelines and test coverage followed, because by then the team wanted them, rather than being handed them as a mandate.

Within a year, average resolution moved from over 300 days to under 30, monthly defect volume from over 300 to single digits, and release cadence from twice a year to more than fourteen times. None of those numbers came from buying anything.

Teams do not need to be told to care. They need the fog cleared, an owner named, and a cadence they are allowed to trust.

What I would do again