People Analytics
People Analytics
What is people analytics?
- Make people decisions with HR and organizational data
- Use evidence to check training impact, turnover signals, and engagement
- Privacy and ethics are essential
The first thing people analytics changes is who has to prove the value in the meeting room. Before, HRD had to answer the question 'why spend this much on training?' with qualitative arguments. Now it sits at the same table as line executives, speaking a common language of turnover, performance distribution, and engagement scores. What matters is not the analytics tool but first asking a question you are prepared to act on once you have the answer. Data collected without a question only adds more dashboard screens.
In practice, it helps to think in three layers. Descriptive analytics counts what happened (turnover rate, training completion rate), diagnostic analytics asks why (differences in conditions between leavers and stayers), and predictive analytics looks at what will happen next. Most Korean companies stay in the first layer, but reporting the current state is not analytics. The former shows numbers; the latter tests hypotheses. Also, the smaller the unit of analysis, the shakier the conclusion. Applying a trend from 1,000 company-wide employees to a 12-person team gives you impressions, not statistics.
There is one more trap when measuring training effectiveness. People who sign up for training voluntarily often have higher motivation and performance to begin with, so better results among completers may reflect a selection effect, not the training itself. That is why you need a design that compares completers with a matched group of non-completers, or tracks the same people’s metrics before and after training. What turns correlation into causation is not data but design.
Practice usually goes wrong in two places. First, adopting a solution before the groundwork is laid. At a company where 80% of performance ratings cluster at B and the job description hasn't changed in five years, even the best analytics tool just neatly summarizes garbage. Second, passing attrition-risk scores straight down to team leads. The moment a number like '68% likelihood of leaving' is attached, that employee is treated as someone on the way out, gets left off important work, and the prediction fulfills itself. Combining data from internal systems without checking the purpose of collection and the scope of consent can also create legal problems. Use analysis results as clues for fixing policies, not as labels to stick on people.
Hwi KimCEO of HRDIST & Talent Development Systems Consultant
John KimJohn Kim, author of B2B Marketing Design and CEO of Nanum ConsultingIt spread in the 2010s as HR teams began applying data analysis, and went mainstream as Google and other companies’ examples of 'running HR on data' became widely known.
Consider a hypothetical case. Company A, a mid-sized Korean chemical materials firm, has 1,400 employees and annual revenue of KRW 420 billion (≈ US$314,230,100). Sales new-hire turnover had climbed to 34% for three years running, and hiring and onboarding costs were leaking over KRW 600 million (≈ US$448,900) a year. Management said 'let’s add more onboarding training,' but the HRD team wasn’t even sure training was the real cause.
- The HRD lead suggested, 'Before expanding training, let's look at what separates those who leave from those who stay,' so the team started by pulling together the records of 412 people hired in the last three years.
- HR and IT worked together for two months and linked 11 data points under anonymous IDs, including hiring channel, onboarding completion, first assigned department, number of check-ins at month six, and direct manager tenure.
- The analysis showed onboarding completion made little difference to turnover. Instead, the group with fewer than three manager check-ins within six months of joining had a turnover rate 2.4 times higher.
- When an HR executive said, 'Then we just need to have them do interviews,' the HR manager replied, 'The busier the team, the fewer the interviews. We need to look at workload too,' and added deals closed per person by department.
- Company A cut onboarding classroom training by 16 hours, added three new-hire check-ins over six months to the managers’ mandatory checklist, and ran 1-on-1 coaching practice for 62 managers.
- A year later, sales new-hire turnover fell from 34% to 21%, saving KRW 230 million (≈ US$172,100) in re-posting and retraining costs. The biggest change was that the executive meeting agenda shifted from 'let’s add more training' to 'let’s design the check-in conversations.'