The most frequently asked questions about HR KPIs
Anyone who works with personnel data knows that the theory of indicators is simple; the hard part is deciding which ones to measure, how to interpret them, and when to stop using them. These are the questions that come up again and again in HR teams, with concrete answers and no inflated promises.
What are the most commonly used HR KPIs in practice?
The ones that survive in most dashboards are few: voluntary turnover, average time to fill vacancies (time to fill), absenteeism, cost per hire, offer acceptance rate, training participation, and satisfaction or eNPS. Not because they are perfect, but because they are calculated with data that almost every company already has in its ATS or ERP.
An important nuance: eNPS is popular but fragile. With workforces of fewer than 30 people, a single negative response moves the result by several points. In those cases, it is better to read it as a quarterly trend, not as an exact figure.
How often should these indicators be reviewed?
It depends on the nature of the data. Turnover and absenteeism work well with monthly or quarterly review; time to fill and cost per hire are better read per closed process or per quarter. Training participation usually makes sense annually. Reviewing everything every week generates noise and meetings that do not lead to decisions.
How is cost per hire calculated without fooling yourself?
You add up everything invested in filling a vacancy (job postings, interviewer time, agencies, assessments, referral bonuses) and divide it by the effective hires in that period. The usual mistake is counting only external spend and forgetting internal hours. According to industry studies, the real cost is usually considerably higher than the agency invoice, and that difference is what explains why some “cheap” processes end up being expensive.
A clear limitation: comparing this KPI across departments with very different profiles (a specialized technician versus a customer service position) leads to false conclusions. It is better to compare similar processes with each other.
Is turnover always bad?
No. There is healthy turnover: people who leave after a complete cycle and are replaced without losing critical knowledge. The useful indicator is not the gross percentage, but early voluntary turnover (departures before 12 months) and the turnover of hard-to-replace profiles. A case that is seen frequently: a company celebrates lowering its overall turnover from 18% to 12%, but when breaking it down discovers that early departures went up. The aggregate figure was hiding the real problem.
Which KPI measures climate better than an annual survey?
None replaces it, but simple combinations work better than a single metric: survey participation rate, eNPS evolution by team, and the ratio of internal applications to open positions. If people do not move internally and also do not respond to surveys, climate is not being measured: it is being assumed.
How many KPIs should an HR dashboard have?
Between five and eight at most. A dashboard with twenty indicators is not read; it is skimmed. Practical experience says that the panels that are maintained over time are the ones that fit on one screen and have an assigned owner per metric. If no one knows who acts when a KPI deviates, the indicator is decorative.
What should not be measured with KPIs?
The quality of leadership or culture cannot be reduced to a number. Proxies can be used (team turnover, participation, internal promotions), but treating them as a definitive measure distorts decisions. According to documentation from people measurement organizations, indicators are signals for asking questions, not verdicts.
How to start if the company has never measured anything?
Choose three indicators, define their formula in writing, and set a quarterly review. Consistency of calculation matters more than sophistication. Changing the turnover formula halfway through the year invalidates any comparison and erodes trust in the data.