Metrics
M01
Days from start date to first PR merged. The single most-tracked onboarding metric. DORA elite teams measure 1-2 days; low performers see 2-4 weeks. Lagging indicator of dev environment friction.
Measure: Git history; first non-revert commit by author email. | Tools: Jellyfish, LinearB, Swarmia, internal Git script.
Elite
1-2 days
Low
2-4 weeks
M02
First merge to production environment. Highly correlated with platform-engineering investment. Mature golden paths produce day-one production deploys.
Measure: CI/CD logs filtered by author. | Tools: GitHub Actions, GitLab CI, internal deploy logs.
Elite
Day 1-3
Low
Month 1+
M03
Weeks until story-point output is at 80-90% of team average. The ground truth ramp metric. Sensitive to team size and ceremony quality.
Measure: Sprint velocity divided by team-average velocity. | Tools: Jira, Linear, Shortcut, Swarmia.
Elite
6-8 weeks
Low
5-8 months
M04
Total onboarding cost divided by annual salary. Below 25% is healthy; above 35% suggests programme drag or codebase complexity issues.
Measure: Output of the calculator on this site. | Tools: Calculator on this site / internal finance spreadsheet.
Elite
15-22%
Low
30-45%
M05
Hours per week of senior engineer time absorbed. Should peak in week two and decline rapidly. Sustained mentor load past week six is a strong signal of programme failure.
Measure: Calendar audits or pair-programming sessions logged. | Tools: Loom history, calendar tags, mentor self-report.
Elite
Peaks 8-12 hrs/wk, < 3 hrs by W08
Low
20+ hrs/wk sustained
M06
Survey score from new hires rating their onboarding experience. Run at Day 30, Day 60, and Day 90. Trends matter more than absolute numbers.
Measure: Post-onboarding survey, 0-10 scale. | Tools: Lattice, CultureAmp, Officevibe, Google Form.
Elite
70+
Low
< 30
M07
Percentage of engineers still employed at 90 days. Industry average sits at ~78% (DevPath). Best-in-class programmes report 95%+. The most expensive metric to miss.
Measure: HR system, cohort by start month. | Tools: BambooHR, Workday, internal HRIS.
Elite
97%+
Low
70-78%
M08
Weeks for surrounding team velocity to return to pre-hire baseline. The team-side cost of onboarding. Hides in 'we are slower this sprint' if not measured.
Measure: Team velocity before vs after new joiner. | Tools: Jira velocity reports, Linear analytics.
Elite
2-3 weeks
Low
8+ weeks
Sources
DORA Accelerate State of DevOps 2024-2025, Stack Overflow Developer Survey 2025, daily.dev First-90-Days Playbook, GitHub Octoverse 2025, BCG Software Productivity Index, Brandon Hall Group Onboarding ROI study, DevPath / Zartis early-attrition research, SHRM replacement-cost framework.
FAQ
Time to first commit. It is cheap to measure (Git history), already understood by engineering leaders, and tightly correlated with dev-environment quality. Once it is in place, layer on time to first deploy and time to full velocity.
The team-level metrics, yes. The per-person metrics, sparingly. Time-to-first-commit shared with the new hire on Day 1 builds focus. Sprint velocity recovery shared with the new hire builds anxiety.
Time to first commit and time to first deploy are onboarding-specific instances of the DORA 'lead time for changes' family. Teams already tracking DORA can plug onboarding metrics into the same dashboard with little extra work.
It usually points to codebase complexity or unclear scope. People are staying because the team is good, but they are slow to contribute because the system is complex. Invest in dev environment automation and ADRs.
Updated 2026-04-28