What is not measured is not improved. Quality metrics help you understand software health and guide decisions. This guide introduces you to the most important metrics and how to use them.
Why Measure Quality
Visibility
Understand current state.
Trends
Getting better or worse?
Decisions
Based on data, not intuition.
Communication
Demonstrate quality to stakeholders.
Code Metrics
Code Coverage
Percentage of code covered by tests.
Cyclomatic Complexity
Complexity of paths in the code.
Duplication
Repeated code.
Technical Debt
Cleanup work estimate.
Maintainability Index
Maintenance ease score.
Bug Metrics
Defect Density
Bugs by code size (KLOC).
Bug Escape Rate
Bugs found in production vs full.
###MTBF
Mean Time Between Failures.
MTTR
Mean Time To Repair.
Severity Distribution
Proportion by severity.
Testing Metrics
Test Coverage
Unit, integration, E2E.
Test Pass Rate
Percentage that passes.
Flaky Tests
Unstable tests.
Test Execution Time
How long to run suite.
Process Metrics
Lead Time
From commit to deploy.
Cycle Time
From the beginning to the end of a feature.
Deployment Frequency
How many deploys per period.
Change Failure Rate
Percentage of deployments that cause problems.
DORA Metrics
What are they
DevOps Research and Assessment metrics.
Four Metrics
- Deployment Frequency
- Lead Time for Changes
- Change Failure Rate
- Time to Restore Service
Classification
Elite, High, Medium, Low performers.
Code Review
Time to Review
How long for code to be reviewed.
Review Thoroughness
Substantive comments.
Approval Rate
Pass rate.
Production Metrics
Uptime
System availability.
Error Rate
Error rate in production.
Response Time
Request latency (p50, p95, p99).
Throughput
Requests per second.
User Satisfaction
###NPS
Net Promoter Score.
App Store Rating
Evaluation in stores.
Bug Reports
Volume of user reports.
Support Tickets
Bug-related tickets.
##Security
Vulnerabilities
Number and severity.
Time to Patch
How long to fix.
Security Incidents
Security incidents.
Tools
SonarQube
Static analysis, code metrics.
Codecov/Coveralls
Coverage tracking.
DORA/Sleuth
Deployment metrics.
Datadog/New Relic
Production metrics.
Dashboards
Visibility
Metrics accessible to the team.
Update
In real time when possible.
###Trends
View evolution.
Alerts
Notifications for anomalies.
Using Metrics
Don't Gamify
Metrics for insight, not punishment.
Context
Numbers without context are deceiving.
Trends > Absolutes
Direction matters more than point value.
###Action
Metrics that don’t lead to action are useless.
Metrics by Phase
Development
Complexity, coverage, review time.
Release
Lead time, deployment frequency.
Production
Error rate, uptime, response time.
Post-Release
Defect density, customer satisfaction.
Continuous Improvement
Baseline
Establish starting point.
Goals
Set realistic targets.
Retrospectives
Discuss metrics regularly.
###Experiments
Test interventions.
Common Errors
Metrics as Objective
100% coverage does not mean quality.
Many Metrics
Focus on a few important ones.
Ignore Context
Different teams and projects.
Punishment by Metrics
It causes perverse behavior.
Conclusion
Quality metrics are tools for understanding and improving software. Choose the right ones for your context, monitor trends and take data-driven action. Quality is a continuous journey, not a destination.
##FAQs
1) What coverage is sufficient? 70-80% is a good target. But it depends on what is tested.
2) What metrics are most important? DORA metrics for DevOps. Defect density and uptime for quality.
3) Do metrics replace good judgment? No. They inform decisions, they do not decide.
4) How to convince the team to adopt metrics? Show value. Don't use it to punish.
5) Are metrics valid for a small team? Yes. Simplify, but measure the essentials.