Interloop Editorial Team
Ask a finance head about revenue, margins or cash flow, and you'll get an answer in seconds. Ask an HR head how many people left last quarter, from which teams, and why, and you'll often hear: "Let me get back to you."
That's not because HR people don't care about numbers. It's because people data in most growing companies is scattered across spreadsheets, payroll files, attendance systems and inboxes. Pulling a simple attrition figure can take days.
The opposite problem exists too. Some companies discover HR dashboards and start tracking everything: 40 metrics, colour-coded, reviewed monthly by nobody.
This piece is about the middle path, and a practical approach to KPI implementation for HR. A short list of people numbers that actually help leaders make decisions, how to calculate them, and how to start tracking them without building a data team.
Key insights
- Leaders know it matters. In Deloitte's 2018 Global Human Capital Trends survey of more than 11,000 business and HR leaders, 84% rated people analytics as important or very important.
- But few trust their data. In Bersin by Deloitte research, only 39% of organisations said they had good or very good quality data to support people decisions.
- Engagement is low and costly. Gallup's State of the Global Workplace 2025 report found global engagement fell to 21% in 2024, with lost productivity costing the world economy an estimated $438 billion.
- Managers move the numbers. Gallup estimates that around 70% of a team's engagement is down to its manager, which is why people metrics should be read team by team, not just company-wide.
- Hiring takes time. SHRM's 2025 benchmarking puts the median time to fill a non-executive role at 44 days. Every day a key role sits empty has a cost.
Start with questions, not metrics
The most common mistake with people metrics is starting from a list of what can be measured. The better starting point is: what do our leaders need to decide, and what would help them decide better?
In a growing company, the questions usually sound like this:
- Are we keeping the people we most need to keep?
- Can we hire fast enough to meet the growth plan?
- Are our new hires working out?
- Are our managers doing a good job of leading their teams?
- Is our people cost growing in line with the business?
- Are we treating people fairly?
Each question points to a few numbers. Each number should point to an action. If a metric goes up or down and nobody would do anything differently, it probably doesn't need to be on the dashboard.
Two more rules help. First, always look at a metric by team, location or level, not just as a company average. A healthy 12% attrition rate can hide one team losing 40% of its people. Second, look at trends. One month's number tells you very little. Six months tells you a story.
The short list
These ten metrics cover most of what a growing company needs. You don't need all of them on day one. Pick the four or five that match your biggest questions and add others as your data improves.
| Metric | How to calculate it | What it tells you |
|---|---|---|
| Voluntary attrition rate | Voluntary leavers ÷ average headcount × 100, for the period | Whether you're keeping people. Track it annualised and by team |
| Regretted attrition | Leavers you wanted to keep ÷ total leavers × 100 | Whether you're losing the right people or the wrong ones |
| First-year attrition | Hires who left within 12 months ÷ hires in that cohort × 100 | How well hiring and onboarding are working |
| Time to fill | Days from role approved to offer accepted | Whether hiring can keep pace with the growth plan |
| Offer acceptance rate | Offers accepted ÷ offers made × 100 | How competitive your offer and candidate experience are |
| Internal fill rate | Roles filled internally ÷ total roles filled × 100 | Whether people can grow inside the company |
| Engagement score | Share of favourable answers to a short, regular survey | How people feel about their work, team and manager |
| Absenteeism rate | Unplanned absence days ÷ scheduled working days × 100 | Early warning of burnout, disengagement or team problems |
| Training hours per employee | Total training hours ÷ headcount | Whether you're investing in skills |
| People cost as % of revenue | Total employee cost ÷ revenue × 100 | Whether headcount growth is sustainable |
A note on fairness: once you have these basics, it's worth splitting a few of them, such as pay, promotions and ratings, by gender, location and level. Gaps there are often invisible until someone looks.
Getting the data in one place
Most of these metrics need data from more than one source. Attrition needs joining and exit dates. Regretted attrition needs performance ratings. Time to fill needs recruitment dates. People cost needs payroll. When each lives in a different spreadsheet, every report becomes a manual project, and numbers from different teams rarely match.
The practical answer is a single system of record for employee data. It's the step most HR KPI implementation projects skip, and the one that decides whether the numbers can be trusted. Zoho People software, for example, part of the wider Zoho HRM suite, holds the core employee record alongside attendance, leave, performance and engagement surveys. Its reports can show headcount, joiners and leavers by department or location, and its analytics can trend attrition and absenteeism over time. Connected to Zoho Recruit and Zoho Payroll, the same data covers hiring timelines and people costs too.
A single system won't fix everything by itself. You still need clean data and consistent definitions, which we cover in the steps below. But it turns a three-day reporting exercise into something a manager can look at on a Monday morning.
How we suggest starting
- Ask leaders for their top three people questions. Use these to pick your first four or five metrics.
- Write a one-line definition for each metric. Agree, for example, whether interns count in headcount, or whether a transfer counts as a leaver. Inconsistent definitions are the most common reason HR numbers get questioned.
- Check your data quality. Make sure joining dates, exit dates, exit reasons, departments and managers are filled in correctly for everyone.
- Set a simple monthly view. One page, one trend line per metric, split by team. Share it with the leadership team.
- Agree what triggers action. For instance, any team with annualised attrition above 20%, or any critical role open for more than 60 days, gets a conversation.
- Add metrics slowly. Once the first set is trusted and used, add the next few. A good KPI system implementation grows in stages.
Common traps
Averages that hide problems. A company-wide number can look fine while one team is in trouble. Always slice by team.
Metrics without context. An attrition rate of 15% means different things in a call centre and in a research lab. Compare against your own trend first, and against industry benchmarks second.
Measuring people, not systems. Metrics should help you find and fix process problems, not single out individuals. If people feel watched, the data gets worse, not better.
Dashboards nobody reads. If a metric hasn't led to a decision or a conversation in six months, either it isn't useful or it isn't reaching the right people.
Waiting for perfect data. Start with what you have, be open about its limits, and improve it as you go. A roughly right number that's used beats a precise number that's late.
Interloop. Orchestrating business — today, and tomorrow.
