How to Track HR KPIs in Your HRMS: Metrics and Dashboards
To track HR KPIs in an HRMS, pick five or six metrics tied to real business problems, write down exactly how each one is calculated, make sure the system captures the underlying data cleanly, and then build role-based dashboards with thresholds and scheduled reports. Turnover, time to fill, cost per hire, absenteeism, training completion and engagement are the usual starting set.
This guide is for HR leaders, HRIS analysts and HR business partners who already have an HRMS and want numbers that leaders actually believe. A common situation: the CFO asks for last quarter's turnover, two people pull it, and they get two different answers. Everything below is about preventing that.
Disclosure: The author works at CloudApper, which makes CloudApper AI.
Key takeaways
- A KPI is only as good as its definition. Agree the formula, data source and owner before anyone builds a chart.
- Most "the dashboard is wrong" complaints are data problems: vague termination reasons, missing requisition dates and duplicate records.
- Segment every headline number by department, location, tenure and manager, because averages hide the problem teams.
- Start small, set baselines for a few months, and add predictive measures only once the basics are trusted.
Which HR KPIs should you track in an HRMS?
The right list depends on the problem you are trying to solve. A hospital short on nurses cares most about time to fill and first-year turnover. A distribution center fighting overtime overruns watches absenteeism and schedule adherence. A company opening new sites in a hurry lives and dies by recruiting speed. The table covers the most common starting set.
| KPI | Common formula | HRMS data needed | What it tells you |
|---|---|---|---|
| Turnover rate | Separations in period / average headcount x 100 | Termination dates, reasons, headcount snapshots | Retention health, split into voluntary and involuntary |
| Time to fill | Days from requisition approval to offer acceptance | Requisition and offer dates from the ATS | Where recruiting stalls |
| Cost per hire | (External + internal recruiting costs) / hires | ATS data plus finance cost data | Recruiting efficiency by source |
| Absenteeism rate | Unplanned absence hours / scheduled hours x 100 | Time and attendance records, absence codes | Early signs of burnout, disengagement or scheduling problems |
| Training completion | Assignments completed on time / assignments due x 100 | Learning system records | Compliance exposure and learning habits |
| Engagement score | Average or percent favorable on a consistent survey index | Survey results linked to org structure | Team climate and an early warning for turnover |
For turnover, borrow the categories the Bureau of Labor Statistics uses in its JOLTS survey. Quits are voluntary separations initiated by the employee. Layoffs and discharges are involuntary separations initiated by the employer. Retirements, transfers to other locations, deaths and separations due to disability count as "other separations." If your termination reason codes map cleanly to those three buckets, your trends become easier to explain and to compare with national data. Our HR glossary entry on employee turnover covers the common variations.
Write a one-page definition sheet
Before anything gets built, capture each KPI in a short sheet:
- Name and plain-English meaning. "Voluntary turnover: people who chose to leave."
- Formula, including edge cases. Do contractors count? Are interns in the headcount? Is a rehire within 30 days a separation?
- Source fields. The exact HRMS fields and reports used.
- Refresh frequency and owner. Who answers questions when the number looks odd.
- Target or threshold. What level triggers a conversation.
It feels bureaucratic. It's also the single step that stops the "two answers to one question" problem.
Why are my HRMS KPI numbers wrong?
When leaders distrust an HR dashboard, the cause is almost always the data going in, not the reporting tool. These are the usual suspects:
- Inconsistent reason codes. If managers choose "Other" for half of all terminations, voluntary turnover means nothing.
- Missing dates. Requisitions that sit in email for two weeks before anyone opens them in the ATS make time to fill look better than it is.
- Headcount timing. Dividing by end-of-period headcount instead of an average inflates turnover in shrinking teams and deflates it in growing ones.
- Contractor confusion. Contingent workers counted in some reports but not others.
- Absence coding. Planned PTO coded as unplanned absence, or the reverse.
Fix the inputs first. Audit the fields each KPI depends on, lock down picklists, and make required fields required at the moment of entry rather than chasing them later. A quick test: pick one department and one month, count separations by hand, and compare that to the system report. If they disagree, don't build the dashboard yet.
How do you build an HR KPI dashboard in Workday, UKG or SAP SuccessFactors?
The major HCM suites all ship configurable reporting. Workday, for example, lists Analytics & Reporting and Workforce Management among its HCM products. Menus differ by vendor, but the build process is much the same.
- Start from the definition sheet. Every chart should trace back to one written definition.
- Build and validate the underlying report. Check it against a manual count for one department and one month before you visualize anything.
- Design by audience. Executives need three to five trend lines. HR business partners need drill-downs by segment. Managers need their own team only.
- Set thresholds. For example, voluntary turnover at a location running above its trailing twelve-month average by a margin you agree in advance.
- Apply security. Use role-based access so managers can't see other teams' data or individual survey responses.
- Schedule delivery. A report that lands in inboxes on the first Monday of each month gets read. One that waits to be opened usually doesn't.
If the standard reports can't give you the cuts you need, our guide to custom HRMS reporting walks through the options. For ideas on what to put on a team-level view beyond the classic metrics, see performance dashboards that track burnout and alignment.
How do you turn HR KPIs into decisions?
A number on a screen changes nothing on its own. What changes things is a routine around it.
- Segment before reacting. A stable company-wide turnover rate can hide one location losing a third of its new hires.
- Pair lagging and leading indicators. Turnover tells you what already happened. Engagement, absenteeism trends and 90-day check-in results hint at what's coming.
- Give every red metric an owner. A threshold breach should produce a named person and a date for a response, not just a colored cell.
- Review on a fixed rhythm. A monthly people-metrics review with operations leaders keeps HR data in business decisions instead of in an HR-only deck.
When it doesn't apply: very small teams. With 15 people, one resignation swings turnover by more than six points. Below roughly 50 employees in a segment, look at the individual stories and trends over a longer window instead of reacting to monthly percentages. For a retention-focused playbook, see how to use HR analytics to improve employee retention.
Can AI predict HR KPIs like turnover?
Predictive models can estimate which teams show patterns that often come before people leave, such as falling engagement, rising absenteeism or a long gap since the last promotion. Treat those outputs as prompts for a conversation, not verdicts. Check models for bias across demographic groups, explain to managers what drives a risk flag, and keep a person in the loop for any decision about an individual employee. Our article on measuring engagement with HRMS data and predictive analytics covers the method in more depth.
Some teams add an AI layer on top of the HRMS they already run rather than switching systems. CloudApper AI is one example. Its AI agents cover data processing, natural language understanding and predictive analytics, and it includes configurable dashboards and reports. Integration connectors are listed for Workday, UKG, SAP, Oracle PeopleSoft and Oracle Fusion Applications, and the platform can run in a private cloud such as AWS or Azure, or on-premise. Whatever tool you use, the definition sheet and data audit above still come first.
Frequently asked questions
How do you calculate employee turnover rate?
Divide the number of separations in a period by the average headcount for that period, then multiply by 100. Average headcount is usually beginning plus ending headcount divided by two, or the average of monthly counts. Report voluntary and involuntary turnover separately, because a spike in quits calls for a very different response than a planned restructuring.
How many KPIs should an HR dashboard have?
An executive view works best with three to seven metrics. More than that spreads attention thin and turns reviews into a skim. HR business partners can have deeper drill-downs behind each headline number. Start with five or six KPIs, build baselines over a few months, and only add a metric when a new business question genuinely needs it.
What is the difference between HR metrics and HR KPIs?
A metric is any measurement, such as headcount or the number of applications. A KPI is a metric you've chosen because it shows progress on a specific goal, and it has a target and an owner. Turnover becomes a KPI when leadership commits to reducing it in a particular group by a set amount within a set time.
What to do next
- Pick five or six KPIs that map to this year's top business problems and write the one-page definition sheet.
- Audit the HRMS fields each KPI relies on, starting with termination reasons and requisition dates.
- Build and validate one dashboard per audience, then schedule monthly delivery.
- Set up a monthly people-metrics review with operations leaders, and add predictive measures only after the basics are trusted.
Sources
- JOLTS Definitions, U.S. Bureau of Labor Statistics
- Human Capital Management Software, Workday
- CloudApper AI platform, CloudApper (vendor page)
Last reviewed: October 2026.

Comments
Post a Comment