Workers in firms at the 90th percentile of firm-level labour productivity produced 3.5 times more output than workers in firms at the median in 2023, according to the ONS bulletin on trends in UK business dynamism and productivity. That gap is why a list of HR activity counts won't satisfy a finance director. You need human resource performance indicators that form a linked system, where each measure has a formula, an owner, a threshold and a tested connection to output.
I would start with five indicators, define them tightly, and resist adding more until the data behind those five is trusted. A smaller set that's read the same way by HR, managers and finance beats a long dashboard nobody can reconcile.
Table of Contents
- Five Core HR KPIs and How to Calculate Them
- How to Define an HR KPI So Everyone Reads It the Same
- Sickness Absence as a Measure of Lost Capacity
- Linking HR Indicators to Productivity and Business Results
- Building Live HR Reporting in Power BI
- Common KPI Mistakes and UK and EU Data Rules
- Speak to an HRManagement365 specialist
Five Core HR KPIs and How to Calculate Them
A human resource performance indicator is a measured value that shows whether people processes and workforce capacity are working, set against a stated target for a stated period. If there's no target or no period, you've got a metric, not an indicator.
The five most UK teams adopt first
Employee turnover tells you whether you're losing people faster than planned. Split voluntary and involuntary exits from day one, because they need different responses. A resignation problem and a performance management problem shouldn't sit in one number.
Time-to-hire runs from approved requisition to accepted offer. Report it as a median by role family, since one hard-to-fill specialist role can distort a mean. When managers ask why recruitment feels slow, practical guidance for hiring managers can help them tighten shortlisting and interview decisions, which is often where the delay sits.
Sickness absence should be based on hours, not occasions. Engagement needs its response rate shown next to the score, or you'll celebrate a result based on a fraction of the workforce. Output per employee needs agreement with finance first, including how you'll calculate full-time equivalent values consistently, before you report any trend.
| KPI | Calculation | Reporting unit | Reference point |
|---|---|---|---|
| Employee turnover | Leavers in period ÷ average headcount × 100 | % per period, voluntary and involuntary shown separately | Compare with your own prior periods |
| Time-to-hire | Days from approved requisition to accepted offer | Median days by role family | Compare with your own prior periods |
| Sickness absence rate | Hours lost to sickness ÷ contracted hours × 100 | % of working hours | ONS UK rate 2.0% in 2025 |
| Engagement score | Favourable responses ÷ total responses × 100 on agreed items | % favourable, with response rate | Compare trend and response rate, not single scores |
| Output per employee | Business output ÷ average FTE | Output per FTE in agreed units or currency | Requires an agreed finance or operations source |
These five are descriptive measures. They show what happened, not why. Write down your own five definitions before you build any report, agree them with finance and operations, and don't let the dashboard design run ahead of that conversation.
How to Define an HR KPI So Everyone Reads It the Same
Say a 240-person organisation records 18 leavers in one quarter. That's a quarterly turnover rate of 7.5%. The arithmetic is trivial. Getting HR, finance and managers to produce the same 7.5% next quarter is where the harder work lies.
The visual below sets out the sequence to follow for every indicator.
Seven steps from calculation to trusted indicator
- Write the question: The indicator must answer something specific, such as whether people are leaving faster than planned.
- Set the numerator: Here it's leavers with a termination date in the quarter, so late-processed leavers don't move between periods.
- Set the denominator: Use average headcount across the period, not headcount on the last day, which swings with joiner timing.
- Define the population: For example, permanent employees only, excluding agency workers and contractors who sit outside your establishment.
- Set the refresh frequency: Monthly suits absence, while quarterly suits turnover, where small numbers make monthly movement noisy.
- Name one owner: One person is accountable for the figure and for explaining its movement, not a shared inbox.
- Set an intervention threshold: For example, an early voluntary exit rate that triggers a review of stay interviews and onboarding.
The CIPD people analytics factsheet identifies employee turnover and absence lost-time rates as descriptive people-analytics measures, indicators of what happened rather than explanations of why it happened. That distinction matters. Separate voluntary from involuntary exits, then analyse cohorts by recruitment source, manager, pay band, training completion and review outcome. Cohort analysis can test whether weak onboarding tends to come before early voluntary exits, instead of leaving you guessing from a headline rate.
The short video below shows how structured HR data supports this kind of analysis.
Leading indicators follow the same design. Objective-setting completion, review timeliness, mandatory training completion, Right to Work evidence status and overdue approvals each need a numerator, denominator, population, owner and threshold.
Practical rule: Correlation guides investigation. It doesn't prove cause, and your commentary should never claim otherwise.
Keep the written definition unchanged once it's agreed. If you change the population mid-year, your trend breaks and you'll spend the next meeting arguing about method instead of action.
Sickness Absence as a Measure of Lost Capacity
The ONS sickness absence release for 2025 puts the UK sickness absence rate at 2.0% of working hours in 2025, equivalent to an estimated 148.8 million working days lost, an average of 4.4 days per worker. The ONS defines the rate as the percentage of working hours lost because of sickness or injury, so calculate from hours lost, not from a count of absence occasions.
That headline hides more than it reveals. The illustration below reflects the risk sitting underneath a single percentage.
One percentage, several different risks
| Workforce segment | Sickness absence rate (2025) | How to use it |
|---|---|---|
| Women | 2.4% of working hours | Compare with men and with role mix |
| Men | 1.7% of working hours | Compare with women and with role mix |
| Public sector | 2.9% of working hours | Benchmark against sector peers |
| Private sector | 1.7% of working hours | Benchmark against sector peers |
| Part-time workers | 2.8% of working hours | Test working-pattern effects |
| Full-time workers | 1.9% of working hours | Test working-pattern effects |
| UK overall | 2.0% of working hours | Headline figure only; segment before acting |
The same percentage can come from many short absences or a few long-term cases. Those need different responses, different manager actions and different occupational health involvement. Segment by department, location, role and working pattern before you act.
Frequency, duration and cause
UK government research found 49% of workers had taken sickness absence during the previous year, while 7% had cumulatively taken more than four weeks off in the previous 12 months, in the survey of employees and self-employed workers 2024 to 2025. Track both patterns. Flag cumulative and continuous absence against your own policy thresholds so return-to-work conversations happen on time.
Cause categories need consistent labels. The CIPD Health and Wellbeing at Work report 2025 reported average absence of 9.4 days per employee per year, with mental ill health cited by 41% of respondents within their top three causes of long-term absence and minor illness the leading short-term cause at 78%. If your managers use free-text reasons, you can't reproduce that kind of split.
Presenteeism complicates the picture further. In that government survey, 58% of workers had worked despite not feeling well enough to do so under the same UK government survey findings framework. A falling absence rate can therefore mean work has shifted onto colleagues who are unwell or overloaded, not that capacity has recovered.
Pair the rate with overtime, cover arrangements, repeat episodes and workload signals. Structured records and policy rules, supported by sound leave and absence management processes, let you measure capacity truly recovered rather than absence merely suppressed.
Linking HR Indicators to Productivity and Business Results
A dashboard can show near-complete appraisal returns and no change in output. That's because completion proves process compliance, not business effect. Treating activity as performance is the most common weakness I see in HR reporting packs.
The diagram below shows the connection you need to make explicit.
Why the denominator changes the story
The ONS productivity publications show how sensitive this is. Its flash estimate for Q2 2026 reported output per hour 0.7% higher year-on-year and output per worker 1.4% higher, while the Labour Force Survey-based estimate put output per hour 0.2% lower and output per worker 0.4% higher. Same economy, same quarter, different data source and denominator, different conclusion. Don't claim causation from a single dashboard metric when national statisticians can't agree on direction.
A worked test you can copy
Say an operations team introduces time-to-competence and new starters reach full productivity in nine weeks instead of fourteen. That's promising, not proven. Test it against orders processed per FTE over two quarters, controlling for seasonal volume, training mix and any automation change in the same period. Only then can you say whether the improvement in the HR indicator lines up with an improvement in output.
| HR indicator | Business outcome it may influence | Confounders to test |
|---|---|---|
| Time-to-competence | Output per FTE, service levels | Training mix, automation, seasonal volume |
| Regretted turnover | Revenue per employee, project delivery | Pay changes, hiring mix, restructuring |
| Absence and overtime | Throughput, quality defects | Workload peaks, cover arrangements |
| Skills coverage | On-time delivery of key projects | Role changes, contractor use |
| Internal mobility | Time to fill critical roles | Vacancy policy, manager span |
Each pairing needs a defined population, time window, denominator, data owner and list of confounders. Present results with confidence limits and as hypotheses to test, not as claims that training completion or employee engagement activity automatically causes productivity.
Link each activity measure to an output measure, or stop reporting the activity measure to the board.
The practical hurdle for UK mid-market organisations is rarely the chart. Employee data sits in one system, output and cost data in another, often reconciled in spreadsheets. Linking the two into one governed model is usually the hardest technical step, and the one worth doing properly.
Building Live HR Reporting in Power BI
Power BI is Microsoft's analytics and reporting service. It connects to Dataverse, Dynamics 365 and Microsoft 365 data and can refresh on a schedule, so your figures update without someone rebuilding a spreadsheet each month.
The hierarchy below shows how those sources feed a live dashboard.
Three dashboard tiers
HRManagement365 is built on Microsoft Dynamics 365, Power Platform and Dataverse, so employee, recruitment, leave, time and performance records can sit in one governed data model. Build three views on top of it.
- Executive view: Trend lines against target for the five core indicators, with commentary from the named owner.
- HR operations view: Cases, approvals, overdue reviews and data quality exceptions that need action this week.
- Manager view: Team absence, overdue reviews and upcoming expiries, with row-level security restricting each manager to their own team.
Allow drill-through to individual records only where a decision needs it. Most managers need a team pattern and a prompt to act, not a spreadsheet of personal data.
What is standard and what is configured
| Layer | What it holds | Standard or configured |
|---|---|---|
| Source records | Employees, leave, time, recruitment and performance data in Dataverse and Dynamics 365 | Microsoft platform; HRManagement365 configures the data structure |
| Data model | Relationships, calendar table, headcount and absence measures | Power BI standard; measures written to agreed definitions |
| Business Central or F&O link | Workforce and cost data for productivity analysis | Integration developed per customer |
| Dashboard pages | Executive, HR operations and manager views | Power BI standard; page design configured per customer |
| Access control | Managers see their own team only | Power BI row-level security configured per customer |
Visuals, scheduled refresh, drill-through and row-level security are standard Microsoft functionality. Your KPI definitions, turnover and absence measures, finance links and page design are configured or developed for your organisation, because your policies and populations aren't standard.
Set refresh frequency to match your definitions, monthly for absence and quarterly for turnover and engagement. Refreshing daily data you only review quarterly adds cost and noise. A dashboard can only show what the underlying data and definitions support, so fix gaps in absence reasons or leaver categories before you publish.
Common KPI Mistakes and UK and EU Data Rules
The most damaging mistake is counting activity, such as training completions or form returns, as if it were performance. Close behind are reporting only a company-wide average that hides a department under pressure, and sharing individual health reasons more widely than the decision requires.
Rules differ by jurisdiction
In the UK, UK GDPR and the Data Protection Act 2018 treat health information in absence records as special category data, so access should be limited and purpose-specific. ICO guidance on employment and monitoring should shape any tracking of individuals. Right to Work checks follow Home Office guidance. Structured evidence and expiry tracking can support compliance processes, but they don't guarantee compliance.
In the EU, GDPR Article 9 governs health data and Article 88 allows member states to set employment-specific rules. Germany, for example, gives works councils co-determination rights over monitoring systems. Employment, payroll and tax requirements vary by member state, so check each KPI design locally before rollout.
Before go-live, answer three questions in writing.
- Access: Who can see individual absence data, and for what decision?
- Population: Which worker categories are included in each measure?
- Approvals: Which works council or data protection approvals are needed?
Speak to an HRManagement365 specialist
Start small and prove the approach. Choose five indicators, assign an owner and a threshold to each, and run a 90-day baseline before you set targets. Targets set without a baseline are guesses, and managers will treat them that way.
Your first 90 days
Confirm definitions with HR, finance and operations, using the seven-step design for each indicator. Extract baseline data and check it for gaps, especially leaver categories, absence reasons and FTE values. Review the first results with managers and listen for challenges about population or timing, because those challenges usually reveal definition problems. Then decide which indicators are ready for the live dashboard and which stay back until data quality improves.
Three decisions need explicit answers. Decide when to include contractors, since including them in headcount but not in turnover distorts both measures. Decide when to split by location, which matters where UK and EU sites work under different absence policies. Decide when to hold a metric back, because publishing a figure you don't trust damages confidence in every other figure on the page.
I would rather publish three indicators you can defend in a board meeting than twelve you have to qualify every time.
How HRManagement365 can help
HRManagement365 is a UK and EU HR solution powered by Hubdrive and Microsoft, built on Microsoft Dynamics 365, Power Platform and Dataverse. Hubdrive and Microsoft provide the standard platform. We extend it with our own integrations, customisations, workflows and HR applications, and we also work as a Hubdrive reseller and solution partner.
That means we can support HR implementation and consultancy, Power BI reporting, Business Central and Finance & Operations integration, custom workflows and approval processes, and custom HR applications across the employee lifecycle. Customisation is scoped to your policies and definitions, not sold as standard functionality it isn't.
Discover how HRManagement365, powered by Hubdrive and Microsoft, can help improve, connect and automate your HR processes across the UK and EU. Phone +44 1522 508096 today, or send us a message via our contact form.
HR Management 365 helps UK and EU organisations build defensible human resource performance indicators on Microsoft Dynamics 365, Power Platform and Dataverse, with Power BI reporting and integrations customised to your own definitions and policies. If you'd like help turning five agreed indicators into a live, governed dashboard your managers and finance team can trust, visit HR Management 365 to start the conversation.
