TL;DR: HR analytics exists on a maturity ladder of four types—descriptive, diagnostic, predictive, and prescriptive—and most UK HR teams are already doing foundational analytics without realizing it. You don't need enterprise-level resources or data scientists to start extracting real business value from your workforce data.
Introduction: Most HR Teams Are Already Doing Analytics — They Just Don't Know It
If you've ever pulled a turnover report, tracked absence rates in a spreadsheet, or counted how many days it took to fill your last vacancy, congratulations — you've already done HR analytics. You just did it at the most foundational level.
There's a persistent myth in mid-sized UK businesses that "HR analytics" is something reserved for large enterprises with Workday implementations, dedicated people analytics teams, and data scientists on the payroll. The reality is quite different. The four types of HR analytics form a maturity ladder, and most UK HR teams are already standing on the first rung without realising it.
The CIPD has long emphasised foundational metrics as the starting point for data-driven HR practice — not because the basics are all you need, but because they're the bedrock everything else is built on. This article explains each of the four types, what data you actually need to do them, and how to progress up the ladder without a significant budget. Whether you're a solo HR manager in a 60-person business or leading a team of five in a 400-person organisation, there's a practical starting point here for you.
The 4 Types of HR Analytics: A Quick Overview
The four types of HR analytics aren't four separate disciplines — they're a progression. Think of them as a maturity ladder: descriptive → diagnostic → predictive → prescriptive. Each level builds on the one before it, and each delivers genuine business value in its own right.
A useful analogy: imagine your car dashboard. Descriptive analytics tells you your current speed and how many miles you've covered. Diagnostic analytics explains why the engine warning light just came on. Predictive analytics warns you that, based on tyre wear patterns, you're likely to have a blowout in the next 50 miles. Prescriptive analytics tells you to pull over at the next services, get the tyre changed, and take the A-road rather than the motorway for the rest of the journey.
In brief: descriptive tells you what happened, diagnostic tells you why it happened, predictive tells you what's likely to happen next, and prescriptive tells you what to do about it.
Crucially, you don't need to reach prescriptive analytics to get real value. Each level delivers meaningful insight on its own. Most mid-sized UK companies — those with 50 to 500 employees — operate primarily at the descriptive level, and that's a perfectly valid and genuinely useful place to be.
Type 1: Descriptive Analytics — What Happened?
Descriptive analytics is the most accessible and most commonly used type of HR analytics. It summarises historical workforce data to show what has happened over a given period — last month, last quarter, last financial year.
The metrics UK HR teams are already tracking in this category include headcount by department, voluntary turnover rate, time-to-hire, absenteeism rate, and gender pay gap figures. That last one is mandatory for UK employers with 250 or more employees under the Equality Act 2010, which means a significant proportion of HR teams are already producing descriptive analytics whether they call it that or not.
The data sources available to most mid-sized UK teams don't require enterprise software. Payroll exports — including HMRC RTI data — give you headcount, salary bands, tenure, and employment type. HRIS platforms like BambooHR, Breathe HR, and Personio generate starters and leavers reports, job change histories, and absence logs. Recruitment ATS systems provide time-to-hire and source-of-hire data. You probably have more raw material than you think.
A practical tip worth emphasising: a well-structured Excel or Google Sheets dashboard tracking five or six core metrics, updated consistently every month, is genuine descriptive analytics. Consistency matters far more than sophistication at this stage.
For UK context, CIPD guidance highlights that voluntary turnover rates vary significantly by sector. Retail and hospitality often exceed 30%, while professional services typically sit closer to 10–15%. Simply knowing where your organisation stands relative to your sector benchmark is a meaningful, actionable insight — and that's descriptive analytics doing its job.
Type 2: Diagnostic Analytics — Why Did It Happen?
Diagnostic analytics goes a layer deeper. Once your descriptive data surfaces a pattern — turnover is up, absence is spiking, time-to-hire has stretched — diagnostic analytics asks the more important question: why?
Here's a concrete example. Your descriptive data shows that voluntary turnover spiked in Q3. Diagnostic analytics investigates the cause. Was the attrition concentrated in one team or one office location? Did it cluster around a particular manager? Was it skewed towards employees in their first 18 months of tenure? Did it correlate with a change in your flexible working policy or a restructure announcement? The answers to these questions are what allow you to act, rather than simply observe.
The diagnostic methods available to lean UK HR teams are more accessible than most realise. Exit interview analysis — when structured consistently and reviewed in aggregate — is a powerful diagnostic tool. Pulse survey cross-tabulation (comparing engagement scores by team, tenure band, or line manager) can reveal patterns invisible in company-wide averages. Absence pattern analysis by department or manager often surfaces issues long before they escalate into formal grievances or tribunal claims.
In terms of tools, you can get surprisingly far with Excel pivot tables or Google Sheets. Teams with access to Microsoft 365 licences — which many UK businesses already have — can use Power BI at no additional cost. Tableau Public offers a free tier for those who want more visualisation capability.
One important nuance: diagnostic analytics requires asking better questions of your data, not necessarily collecting more data. A structured exit interview template, analysed consistently across 20 leavers, is worth far more than ad hoc conversations that produce no comparable data.
A UK GDPR note worth flagging: when drilling into individual-level data for diagnostic purposes, anonymisation thresholds matter. Avoid analysis on groups smaller than approximately eight to ten people to protect individual privacy under UK GDPR. This is particularly relevant when segmenting by team or manager in smaller organisations.
Type 3: Predictive Analytics — What Is Likely to Happen?
Predictive analytics uses historical patterns and statistical modelling to forecast future workforce outcomes. In practical HR terms, this means moving from reactive to proactive — intervening before problems occur rather than analysing them after the fact.
Common HR use cases include attrition risk scoring (identifying which employees are most likely to leave in the next six months), time-to-hire forecasting for workforce planning, absence trend forecasting ahead of peak periods, and skills gap projection as the business evolves. The value proposition is straightforward: if you can identify that a particular cohort of employees is at elevated flight risk three months before they hand in their notice, you have time to do something about it.
For mid-sized UK teams, an honest assessment is warranted here. True predictive modelling requires sufficient historical data — typically two to three years minimum — clean data hygiene, and some degree of statistical literacy. It's not out of reach, but it does require investment of time and, in some cases, tooling.
That said, accessible starting points exist at non-enterprise price points. Some modern HRIS platforms — including Personio and HiBob — include basic predictive features within their mid-market tiers. People analytics tools like Visier offer SME-oriented options. And a mid-sized HR team of two or three people can begin with simple trend extrapolation — projecting headcount needs based on historical growth rates, for instance — before moving to more sophisticated regression-based models.
It's also worth noting that AI-powered HR tools are increasingly making predictive analytics accessible without requiring a data scientist on the team. This is a natural bridge: as AI becomes embedded in HR workflows, it can surface predictive signals from data that previously required specialist analysis. We'll come back to this point shortly.
Type 4: Prescriptive Analytics — What Should We Do About It?
Prescriptive analytics is the most advanced of the four types. It doesn't just predict outcomes — it recommends specific actions to achieve a desired result. Think of it as the difference between a weather forecast that says "70% chance of rain tomorrow" and a travel app that says "leave 20 minutes earlier, take the M6 rather than the M1, and pack an umbrella."
In HR terms, prescriptive analytics might surface insights like: "If you increase the frequency of manager one-to-ones in this team, modelling suggests attrition risk drops by a meaningful margin." Or: "Adjusting your hiring timeline by two weeks at the offer stage reduces offer decline rates based on historical patterns." Scenario modelling for restructuring or headcount planning — where you can test different assumptions and see projected outcomes — is another prescriptive application.
An honest framing is important here. Prescriptive analytics is largely the domain of enterprise platforms and specialist people analytics teams. Most mid-sized UK HR teams are not operating at this level yet — and that's entirely fine. The goal isn't to reach prescriptive analytics immediately; it's to build the data foundations that make prescriptive insights trustworthy when you do get there.
AI is beginning to democratise prescriptive capability, with tools that surface recommended actions based on workforce data patterns increasingly available at non-enterprise price points. But there's a critical caveat: prescriptive recommendations are only as good as the data and assumptions behind them. Garbage in, garbage out. This is precisely why investing in clean, consistent descriptive data now pays dividends later — it's the foundation that makes everything above it reliable.
What Data Do UK HR Teams Actually Have Access To?
Before investing in any analytics tooling, it's worth taking stock of what data you're already sitting on. Most mid-sized UK HR teams are surprised by how much they have — and how underused it is.
Payroll data, via HMRC RTI submissions or payroll software, gives you headcount, salary bands, tenure, employment type, and absence-related SSP claims. HRIS data, if you're using platforms like BambooHR, Breathe, Personio, or Ciphr, provides starters and leavers records, job changes, performance review scores, and training completion rates. Recruitment ATS data covers time-to-hire, source of hire, offer acceptance rates, and cost-per-hire. Engagement and pulse survey data — even a simple quarterly eNPS — provides employee sentiment segmented by team or tenure band.
For UK employers with 250 or more employees, mandatory reporting data is already being produced: gender pay gap figures under the Equality Act 2010, and increasingly, voluntary ethnicity pay gap reporting as stakeholder expectations evolve. Apprenticeship levy returns provide additional workforce data points.
What most mid-sized teams are missing is integration across these systems — payroll, HRIS, and engagement data sitting in one place rather than three separate exports. Clean historical data going back three or more years is another common gap, as is consistent data definitions across the business (does "turnover" mean all leavers, or just voluntary leavers?).
The practical insight: before spending a penny on analytics tools, audit what you already have. Map every system that holds workforce data, assess what's clean and what's inconsistent, and identify the gaps. You'll likely find you're closer to a solid descriptive analytics capability than you thought.
How to Start Building an HR Analytics Capability on a Limited Budget
The question HR teams in smaller organisations ask most often is: how do we actually get started? Here's a practical, sequenced approach that doesn't require enterprise software or a data science background.
Step 1 — Audit your data. List every system that holds workforce data. Assess what's clean, what's inconsistent, and what's missing. This is the unglamorous but essential first step.
Step 2 — Pick three to five core metrics and track them consistently. Don't try to measure everything at once. Start with voluntary turnover rate, time-to-hire, absence rate, headcount by department, and cost-per-hire. These five metrics alone will tell you a great deal about your workforce.
Step 3 — Build a simple monthly dashboard. Excel or Google Sheets is entirely adequate at this stage. A dashboard you update every month for 12 months is worth infinitely more than a sophisticated tool you abandon after three. Consistency over complexity.
Step 4 — Add diagnostic questions once you have six to twelve months of descriptive data. Start asking "why" questions. Use structured exit interviews, manager surveys, and absence pattern analysis by department. This is where the real insight starts to emerge.
Step 5 — Evaluate tools when your data foundation is solid. Only once you have clean, consistent descriptive data should you consider investing in a dedicated people analytics platform or AI-powered HR tool. The tools will only be as good as the data you feed them.
On budget: the barrier to starting is lower than most HR teams assume. CIPD offers free metrics guidance and templates. Microsoft Power BI is included in many Microsoft 365 Business licences. Google Looker Studio is free. The investment required to build a genuine descriptive analytics capability is primarily time and discipline, not software spend.
As teams progress from descriptive to diagnostic and predictive analytics, AI-powered tools like Aura can add a dimension that traditional analytics misses entirely. Think of Aura as a layer that sits on top of your HR knowledge base — answering employee questions instantly while simultaneously surfacing patterns that feed your analytics. When 40 employees in your Manchester office ask about flexible working policy in the same month, that's a diagnostic signal worth investigating, and it shows up in Aura's data before it ever appears in your turnover figures.
The HR Metrics Every UK HR Director Should Be Tracking
A practical, UK-specific list of the metrics that matter most — not an exhaustive catalogue, but the ones that give you the clearest picture of workforce health.
Voluntary turnover rate is the foundational metric. Track it monthly and segment by department, tenure band, and line manager. The CIPD's annual surveys provide sector benchmarks to contextualise your figures.
Time-to-hire — from job approval to offer acceptance — is a critical operational metric. CIPD data suggests the UK average is four to six weeks depending on role level, though this varies considerably by sector and seniority.
Absence rate is both a wellbeing indicator and a cost metric. The UK average sits at approximately 5.7 days per employee per year according to the CIPD Health and Wellbeing at Work survey. Segmenting by team or manager often reveals hotspots that company-wide averages obscure.
Cost-per-hire — including advertising costs, agency fees, internal recruiter time, and onboarding costs — helps HR make the business case for investment in employer brand and retention.
Gender pay gap is mandatory for employers with 250 or more employees, and increasingly scrutinised by candidates, investors, and the press. Tracking it year-on-year, not just at the annual reporting deadline, gives you a clearer picture of progress.
Employee Net Promoter Score (eNPS) is a simple, consistent measure of employee sentiment that's easy to track quarterly without a complex survey programme.
Training completion rate and L&D spend per employee are becoming more important as skills gaps widen and the link between development investment and retention becomes clearer.
Revenue per employee — or an equivalent productivity metric relevant to your sector — helps HR professionals speak the language of the board and demonstrate workforce ROI.
The goal isn't to track all of these simultaneously from day one. Pick the three to five most relevant to your current business challenges and build from there.
Where AI Fits Into Your HR Analytics Journey
AI is making analytics more accessible to lean HR functions — tools that previously required a data science team are increasingly available at mid-market price points. But it's worth being clear-eyed about where AI genuinely adds value and where it's marketing noise.
The honest answer is that AI analytics is only as good as the data it's grounded in. An AI tool fed inconsistent, incomplete, or poorly defined data will produce unreliable outputs. This is why building descriptive and diagnostic foundations first isn't just good practice — it's a prerequisite for getting value from AI-powered analytics.
One data source that's consistently underappreciated in HR analytics discussions is employee questions. The volume, topics, and patterns of questions employees ask HR every day are a rich, real-time signal of workforce sentiment, confusion, and emerging issues. Traditional analytics captures this signal only after it's manifested in turnover or absence data — often months later.
This is where a tool like Aura offers something different. Think of Aura as a layer that sits on top of your HR knowledge base — answering employee questions instantly, in any language, at any hour, while simultaneously surfacing patterns that feed directly into your analytics picture. A spike in questions about redundancy processes, flexible working entitlements, or parental leave policy can signal an emerging workforce concern weeks or months before it shows up in your turnover data. That's diagnostic analytics in near real-time.
The key message is this: AI augments your analytics capability — it doesn't replace the need for HR judgment in interpreting and acting on what the data shows. The most valuable analytics insight is only useful if an HR professional knows what to do with it. Data informs; humans decide.
Conclusion: Start Where You Are, Build From There
The four types of HR analytics — descriptive (what happened), diagnostic (why it happened), predictive (what's likely to happen), and prescriptive (what to do about it) — form a progression, not a checklist. You don't need to be at the top of the ladder to get value from the journey.
Most mid-sized UK HR teams can begin with descriptive analytics today, using tools they already have. A consistent monthly dashboard tracking five core metrics, built in Excel or Google Sheets, is a genuine starting point. The teams that eventually reach predictive and prescriptive analytics are, almost without exception, the ones that invested in clean, consistent data at the descriptive level first.
The maturity ladder is a journey measured in months and years, not weeks. Start where you are, pick your five metrics, and build from there. If you're curious about how AI can help surface workforce insights from your day-to-day HR interactions — and feed those signals into your analytics picture — explore how Aura works and see what patterns your employee questions are already telling you.
