HR: KPIs
Hiring, retention and capability measures, the quarterly bias audit that is mandatory rather than optional, and why individual productivity is not measured here.
Markdown. No sign-up, no email.
The six that matter#
| Measure | Definition | Target | How it gets gamed |
|---|---|---|---|
| Time to hire | Role opened to offer accepted | Under 35 days | Opening the role late, so the clock starts after the search did |
| Quality of hire | Still performing well at 12 months | Above 85% | Measuring at 3 months, when almost everyone still looks fine |
| Regretted attrition | Leavers we wanted to keep | Under 8% | Counting all attrition together, which hides the number that matters |
| Skills coverage | Critical skills with more than one person | 100% | Counting people who did the training rather than people who can do the work |
| Internal fill rate | Roles filled from inside | Above 30% | Reclassifying a sideways move as a promotion |
| Bias audit outcome | Selection rates by group at each stage | No unexplained gap | Not running it, which is by far the most common failure |
Quality of hire is measured at twelve months. Three months measures onboarding. A year measures the decision.
Regretted attrition is the retention number. Total attrition includes people whose leaving was fine for everyone, and averaging the two hides whether we are losing the people we need.
The bias audit is not optional and not annual. Quarterly, by group, at every stage: applied, screened, interviewed, offered, accepted. A gap that appears at one stage and not others tells you exactly where to look.
Two counter-metrics#
| Counter-metric | Catches |
|---|---|
| Offer acceptance rate | A fast, efficient process that candidates decline. Speed achieved by being unpleasant to experience |
| Manager span of control | Growing spans mean people are managed in name only, which shows up in attrition two quarters later |
Workforce planning, quarterly#
| Question | Bad answer |
|---|---|
| Which critical skills sit with one person? | Any, and this is the most common structural risk in a small company |
| What will the roadmap need in 6 months that we do not have? | "We'll hire when we get there" |
| Who has not had a development conversation this quarter? | Anyone |
| Where is capacity above 90% sustained? | Anywhere. That is not efficiency, it is a system with no slack |
What is deliberately not measured#
- Individual productivity. Measures what is easy to count, which is rarely what matters about someone's work, and turns analytics into surveillance.
- Hours worked or logged in. Attendance, not contribution.
- Training hours completed. Attendance again. Measure whether the skill exists.
- CVs screened. Volume.
- Engagement score alone. Useful only alongside attrition; on its own it measures who answered the survey.
Cadence#
| Weekly | Open roles, candidates waiting on us |
| Monthly | Time to hire, offer acceptance, attrition |
| Quarterly | Bias audit, skills coverage, workforce plan, development conversations held |
| Annually | Quality of hire for last year's cohort, pay equity review |
The quarterly bias audit is the entry on this page that must never be deferred. It is the one that catches the failure nobody sees, and its absence is not neutral: it means the failure is happening unmeasured.