Dashboards and flow metrics
Design and implement a dashboard, including flow metrics
Lead time versus cycle time
The definitions are precise and are examined directly:
Lead time measures the total time from the creation of work items to their completion.
Cycle time measures the time it takes for a team to complete work items once they start actively working on them.
So lead time includes the queue — the period between someone asking and anyone starting. Cycle time excludes it.
That difference is the diagnosis. A team with short cycle time and long lead time is fast but overloaded: work waits in a backlog before anyone touches it, and hiring more engineers will not help as much as limiting work in progress. Reading only cycle time hides the problem entirely, because by that measure the team looks excellent.
The wider set
| Metric | Signals |
|---|---|
| Lead time | Total responsiveness, queue included |
| Cycle time | Execution speed once started |
| Deployment frequency | How often value reaches users |
| Time to recovery | How fast you recover from failure |
| Change failure rate | Proportion of changes causing failure |
Designing the dashboard
Show what someone will act on. A dashboard nobody acts on is a screen, not a control — and the honest test is whether anyone has changed a decision because of it in the last month.
Primary sources