Flow Metrics
Where a team's time actually goes: work versus waiting, by stage — for the current sprint, or the most recently finished one if none is running.
What you see
- A "Time spent working" figure — the share of an item's total time, from start to finish, that was spent actively worked rather than waiting. Shown as a percentage, measured across however many items delivered this sprint.
- A standing note: "Most teams are lower than they expect. The number is a map of where the waiting is, not a score."
- "Average days per stage" — a bar chart across the team's own workflow stages, with a note naming which of the stages shown count as waiting rather than working. A table below repeats the same figures, and each stage's name is clickable.
- An average blocked-days figure, shown only when at least one item was blocked at some point.
The stages, in flow order, are: New, Queued, In Progress, Test, Review, Waiting.
How we calculate it
- Active vs. waiting
- In Progress, Test and Review count as active (working) time. New, Queued and Waiting count as waiting time. Time before an item's real start (its first mapped in-progress transition) is excluded entirely from both sides — it isn't counted as waiting.
- Flow efficiency %
- Active time ÷ total tracked time, computed per item, then reported across the sprint as a median and an 85th percentile. This figure is read from the same shared calculation Sprint Summary and Dashboard use for the same sprint, specifically so the same team never sees two different flow-efficiency numbers on two different screens.
- Minimum sample
- Needs at least 10 completed items with usable stage history; below that, the percentage is withheld and the page explains why instead of showing an unreliable figure.
- Average days per stage
- These are averages (means), not medians — a deliberate exception to how most other figures in Agile Gauge are computed, made explicit rather than left implicit. They are whole calendar days, weekends included, not configurable to working days on this view.
- Time in a stage counts every visit
- An item that went Active → Review → Active again shows both Active stretches added together, not just the most recent one.
- Clicking into a stage
- Shows only items currently sitting in that stage right now (a live snapshot, not history), with time in that stage shown to one decimal place — a different, finer precision than the whole-day averages in the chart above it, so a list's total is not expected to match its bar exactly.
- Blocked days
- Averaged only across items that were blocked at some point — items never blocked are excluded from the average, not counted as zero.
What you can click through to
- A stage's name opens the items currently sitting in it, each with its own time-in-stage.
- Each item opens further into its full journey through every raw Azure DevOps state it actually passed through (a finer-grained view than the six stage buckets above).
There is no CSV export on this view.
A different percentile method from Cycle Time
Flow Metrics' headline percentage uses an interpolated median/85th-percentile across items' individual flow-efficiency percentages. Cycle Time's P50/P85 use a different method (nearest-rank, always a real observed value) over raw day counts. The two are not directly comparable calculations, even though both use the word "percentile."
Known limits
- Single-sprint scope only — the current sprint, or the last finished one. There's no multi-sprint trend or window selector here (compare Cycle Time, which has one).
- Below 10 completed items with usable history, the flow-efficiency percentage isn't shown at all.
- Which stages count as "waiting" is stated in words next to the chart, not shown by colouring the bars.