People Centricity case study

Early Leavers Are Warranty Claims on Your Hiring

69sites in the analysis
12months: the window that defines an early leaver
2lenses: exit records and the engagement survey
1definition for every country, tracked monthly

The situation

The question behind the question.

This company had put new-hire attrition on its group scorecard, one of only 2 people measures there, and its new talent acquisition strategy made it one of its first initiatives. Each early leaver is a hire paid for twice: the search, then the replacement, and the months before a newcomer is fully productive. The company had a rate. It did not yet know where early leavers came from, when in their first year they left, or why, nor whether the problem was hiring, onboarding or local.

As it was asked

“Tell us why our new hires are leaving.”

As we reframed it

Treat first-year attrition as a result of how people are hired and welcomed: find where and when it happens, test whether the reasons recorded can be acted on, and turn the answer into a measure the business tracks.

How we worked

How the work unfolded.

01

Separate the signal

We started by separating early leavers from everyone else. 3½ years of terminations from the HR system, across 69 sites, were split by type, tenure, month, site and reason, so first-year departures could be read against all voluntary leavers rather than buried inside them. The comparison changed the question: a turnover rate that looked stable overall can hide a first year that is not.

02

Find where and when

Then the concentration. Early leavers were mapped by site and by the point in their first year at which they left, and profiled by role for each reason they gave. The aim was to turn one company-wide rate into local questions for each site leader to own, and to see whether people left in the first weeks, when onboarding is at stake, or later, when the job itself is.

03

Test the reasons

Then the reasons, tested for one thing: could anyone act on them? Each exit reason was read by site and role, and catch-all categories that point to no fix were set apart from those that do. The engagement survey’s view of new joiners was laid alongside. An annual survey is too slow for a 12-month window, so the leading signal had to come from somewhere else: conversations with new joiners while they are still deciding.

04

Make it a measure

Then the measure. Leavers were counted against the new hires of the period, not against average headcount, so the rate shows what share of a hiring class did not stay rather than moving with hiring volume. One formula, a 12-month window, monthly on the HR scorecard and the recruiting dashboard, owned by talent acquisition. We recommended exit reason codes that point to a cause, check-ins at 30 and 90 days as the leading signal, and reading early attrition as the quality test of the new assessment approach.

The instrument

A first-year leaver, read 3 ways

Against the rest. First-year leavers as a share of all voluntary leavers, with their own trend, so a stable overall rate cannot hide a rising early one.

Where and when. By site, and by the point in the first year when people leave: the first weeks point to onboarding, later months to the job.

Why, if actionable. Each recorded reason tested for whether it points to a fix. A reason nobody can act on is a gap in the data, not an answer.

Every hire comes with a 12-month warranty; early attrition is the claim rate. Read it as one, and it tells you what to fix before the next hire.

What moved

What changed, and what didn't.

1 in 5voluntary leavers had left within their first year, while overall attrition stayed flat
Evidence Decision taken Adopted in practice Measured impact
What moved

1 in 5 voluntary leavers had left within their first year, while overall attrition stayed flat. New-hire attrition is now tracked monthly on the HR scorecard and the recruiting dashboard, defined once for every country, and read alongside how people were assessed and welcomed.

What didn't move

The analysis was descriptive and covered voluntary leavers only: failed probations, the signal of a selection problem, stayed in the backlog, and recruiter, hiring source and hiring manager were not yet linked to who left. Reasons recorded before the codes were redesigned could not be recovered, and whether the check-ins lowered early attrition was beyond what we saw.

What we'd say to a peer

Every hire comes with a 12-month warranty, and early attrition is the claim rate. Separate it from the rest of your turnover, count it against the hiring class, find where and when it happens, and test whether the reasons you record could ever be acted on. Then pair it with an early signal, and read it next to how those people were hired.

Let's talk Discover People Centricity Read more cases
Method note:

Own approach in 4 steps; 3½ years of HR records across 69 sites; tenure, site, month and reason breakdowns; engagement cross-check; cohort-based KPI definition.

WeAreHuman@Work

One piece of evidence on sustainable work, every Sunday. Free, and one click to leave.

Get the weekly edition