Metrics for automatic testing
Flow metrics measure how work actually moves through a process: cycle time (how long one item takes end to end), lead time, throughput, and work in progress. The point is measuring what's actually moving, not how busy people look.
Here's why that belongs on a testing page: "time from a bug being reported to a regression test existing for it" is a concrete, trackable cycle time. And a growing pile of skipped, ignored, or flaky tests is itself a flow-metric signal — work in progress that isn't actually moving — meaning testing debt is piling up faster than it's being paid down, whether or not anyone's tracking it on a board.
Learn more at Atlassian's guide to Kanban, or watch ProKanban's "The Kanban Guide - Kanban Metrics" for a video introduction to the metrics themselves.
Examples you can ask AI
- What do 'cycle time' and 'throughput' actually mean, in plain terms, for a team's day-to-day work?
- How would I start tracking cycle time for bug fixes on my own team, without buying new tooling?
- How do I use flow metrics to make a concrete case that a growing pile of flaky tests is real, measurable technical debt?