Metrics That
Tell the truth.
Vanity metrics dress up bad sprints. The four that actually move decisions are velocity (capacity sense), cycle time (flow), throughput (output), and escape rate (quality). Calculate them, then read what to do when they shift.
Why most agile dashboards lie
Three failure modes: optimising for velocity (Goodhart's law), single-point summaries hide tail risk (mean cycle time vs p85), and tracking commit-vs-complete with no second metric to sanity-check it. The toolkit has the calculator for each metric and the post that explains when each one signals what.
The toolkit
1. Frame: Metrics that actually matter
Four agile metrics worth tracking (velocity, cycle time, throughput, escape rate), four that lead you wrong (story-point precision, individual velocity, sprint commitment %, lines-of-code), and Goodhart's law applied.
- 4 useful metrics
- 4 vanity metrics
- Goodhart trap
2. Calc: Velocity Calculator
Paste last 6 sprints, get rolling average, trend, and a planning-range (low/likely/high). Calibrates capacity without anchoring on the most-recent sprint.
- Rolling 6-sprint average
- Planning low/likely/high
- Trend signal
3. Analyse: Cycle Time Calculator
Paste raw cycle-time values, get p50/p85/p95, histogram, and tail-shape diagnosis. The metric that actually predicts when stories will finish, not velocity.
- p50/p85/p95 percentiles
- Histogram + tail shape
- Pattern diagnosis
4. Diagnose: Velocity dropped, here's the playbook
Velocity dropped 30% in two sprints? The calm playbook: when a dip is noise vs signal, the seven real causes (in probability order), the one-meeting fix, and what NOT to do under pressure.
- 7 root causes
- Noise vs signal
- One-meeting fix
Related reading
Sprint velocity benchmarks (and why "industry average" lies)
Why benchmarking against other teams is meaningless, what your own velocity range should look like, and the five contextual factors.
ReadBurndown charts: what they actually tell you
The four shapes (on-track, hockey-stick, late-slip, under-delivery) and the structural cause behind each. Read the chart, not the line.
ReadForecasting sprint dates with p85
Why the mean lies (long tails), what p85 means, the math, and the stakeholder message that earns trust (always share p50 *and* p85).
ReadAgile estimation techniques compared
Story points, planning poker, t-shirts, hours, no-estimates; the trade-offs between each, where each works, and the three honest questions that pick one for your team.
ReadMetrics that ship next sprint, not next quarter
SprintFlint tracks velocity, cycle time, throughput, and escape rate automatically; no spreadsheet rebuild. Plus burndown, sprint goal hit-rate, and the four anti-patterns dashboard tells you when one's drifting.