TL;DR

People analytics is the practice of using workforce data to make better decisions about hiring, retention, performance, and organisational design. It replaces gut feel with evidence.

What People Analytics Actually Is

People analytics treats the workforce like a dataset - one you can query, model, and act on. At its most basic, it means tracking metrics like time-to-hire or turnover rate. At its most sophisticated, it means building predictive models that flag flight risk before an employee submits their resignation.

The field sits at the intersection of HR, data science, and organisational psychology. Most companies operate somewhere in the middle: they have data (ATS exports, HRIS records, engagement survey scores) but lack the infrastructure to connect it meaningfully. People analytics is the practice of building that connection.

The questions people analytics can answer fall into a few categories:

Descriptive: What does our workforce look like? Where are our attrition rates highest? How long does it take to fill a role by department?

Diagnostic: Why are people leaving the engineering team? Does our interview process predict on-the-job performance?

Predictive: Which employees are most likely to resign in the next 90 days? Which sourcing channels produce candidates who stay longest?

Prescriptive: Given what we know about flight risk, which retention interventions are worth the investment?

Most organisations do a lot of the first, some of the second, and very little of the third and fourth.

Why It Matters for Recruitment

Recruiting generates a disproportionate share of the data that people analytics depends on. Every application, screening decision, interview score, offer letter, and hire outcome is a data point. The ATS is often the richest structured HR dataset in the business - and it's almost always underused.

For talent acquisition teams, people analytics can answer questions that have historically been answered by anecdote:

• Which job boards produce candidates who actually pass the phone screen?
• Which interviewers are calibrated well versus consistently lenient or harsh?
• What does the average candidate journey look like, and where do we lose people?
• Are diverse candidates dropping out at a specific stage?

For HR more broadly, recruitment data feeds workforce planning models. If you know your average tenure by role type and your growth targets by function, you can calculate hiring volume needs 12-18 months out rather than scrambling when headcount opens up.

The challenge is data quality. Garbage in, garbage out. People analytics only works when the underlying data is clean, consistently entered, and connected across systems.

In Practice

A 500-person SaaS company is losing account managers at a rate that's costing them roughly £800,000 per year in replacement costs. The HR team knows turnover is high; they don't know why or where to intervene.

The people analytics team pulls 18 months of data: hire date, department, manager, tenure, performance rating, engagement survey responses, exit interview themes, and time between hire and first promotion. They build a simple attrition model and find three patterns: account managers who were hired into teams with a specific manager churned at 2.3x the average rate, account managers who didn't receive a promotion signal within 14 months left at significantly higher rates, and those who scored below 3.5 on their 6-month engagement survey almost never made it to 18 months.

Three actionable levers, none of which required a new tool. They already had the data. They just hadn't looked at it together.

Key Facts