Agile Metrics That Actually Matter

The Problem With Agile Metrics

Most teams track velocity and think they're measuring performance. They're not. Velocity measures output, not outcomes. A team can have high velocity and still deliver nothing of value.

Good metrics answer one question: Are we getting better at delivering value?

The 5 Metric Categories

1. Flow Metrics — How fast does work move?

Cycle Time: Time from work starting to work finishing.

• Measure per story/ticket
• Track the trend, not individual data points
• Target: decreasing or stable over time

Throughput: Number of items completed per sprint/week.

• More reliable than velocity (not affected by estimation accuracy)
• Use for forecasting: "We complete 8-12 items per sprint"

Work In Progress (WIP): Items currently in progress.

• High WIP = context switching = slower delivery
• Target: WIP should be ≤ team size

Blocked Time: How long items sit in "blocked" status.

• Track average blocked time per sprint
• Target: <10% of total cycle time

2. Predictability Metrics — Can we deliver what we promise?

Sprint Goal Success Rate: % of sprints where the goal was met.

• Target: >80%
• If below 60%, your planning process needs work

Commitment Reliability: Committed points vs completed points.

• Target: 85-100% (not 100% — that means you're sandbagging)
• Track trend over 6+ sprints

Carryover Rate: % of stories carried from one sprint to the next.

• Target: <15%
• High carryover = poor estimation or too much WIP

3. Quality Metrics — Are we building it right?

Escaped Defects: Bugs found in production (not in testing).

• Target: decreasing trend
• Track severity: critical bugs matter more than cosmetic ones

Defect Density: Bugs per story/feature delivered.

• Target: <0.5 bugs per story

Rework Rate: % of stories that come back for fixes after "done."

• Target: <10%
• High rework = weak Definition of Done

4. Team Health Metrics — Is the team sustainable?

Team Satisfaction: Anonymous quarterly survey (1-5 scale).

• Track: psychological safety, workload, growth, collaboration
• Target: average >3.5

Retrospective Action Completion: % of retro actions done before next retro.

• Target: >70%
• If below 50%, retros are theatre

Unplanned Work %: How much of the sprint is consumed by unplanned work.

• Target: <20%
• High unplanned work = poor backlog management or unstable systems

5. Business Outcome Metrics — Are we delivering value?

Feature Usage: % of delivered features actually used by customers.

• Target: >60%
• If below 40%, you're building the wrong things

Time to Value: Time from idea to customer using the feature.

• Track end-to-end, not just development time
• Includes discovery, design, build, test, release, adoption

Customer Satisfaction: NPS, CSAT, or support ticket trends.

• The ultimate measure: are customers happier?

How to Use Metrics Without Destroying Trust

Rule 1: Never use metrics to compare teams. Teams have different contexts, different work types, different constraints. Comparing velocity across teams is meaningless and harmful.

Rule 2: Trends matter more than absolutes. A velocity of 30 means nothing. A velocity that's been declining for 4 sprints means something.

Rule 3: Let teams own their metrics. The team decides what to track and what to improve. Metrics imposed from above become targets to game.

Rule 4: Pair metrics. Never optimise one metric in isolation. If you push for speed, quality drops. Track both together.

Rule 5: Review quarterly, not daily. Metrics are for strategic decisions, not micromanagement. Review trends quarterly. Act on patterns, not noise.

The Minimum Viable Dashboard

If you can only track 5 things, track these:

1. Throughput (items per sprint) 2. Cycle time (days from start to done) 3. Sprint goal success rate (% of sprints goal met) 4. Escaped defects (bugs in production) 5. Team satisfaction (quarterly score)

Everything else is optional until these 5 are healthy.

---

Download the [Sprint Velocity Tracker template](/templates) to start tracking these metrics today.