What is people analytics? A practical guide for companies
By Deibson Silva · Reviewed on
People analytics is the structured use of data about people and work (profiles, performance, engagement, hiring, development) to make decisions about teams with more than intuition. It does not replace the judgement of whoever leads: it puts evidence in front of them, within the limits set by data protection law and, in the EU, the AI Act.
What is people analytics in practice?
In practice, people analytics means asking specific questions about your workforce and answering them with data the organisation already holds or can lawfully collect. A typical question: which key roles have nobody ready to step in? Another: which areas take part least in engagement surveys, and why?
The difference from a traditional HR report lies in how the data is used. A report describes what happened (joiners, leavers, training hours). People analytics combines that data with other sources to explain what is happening and to support the next decision: whom to prepare for a role, which team needs a different way of working, what profile to look for in the next vacancy.
Three uses are usually described, from least to most demanding. Descriptive answers "what is happening". Explanatory answers "why it is happening". Predictive tries to anticipate "what may happen". The closer you get to the third, the more care you need with data quality and with the law, because decisions start to affect specific people.
What data does people analytics use?
It draws on several sources, each with its own use and its own precautions. The table sums up the most common ones in a company or in a consultancy that works with client companies.
| Data | Example | What it is for | Handle with care |
|---|---|---|---|
| Workforce | Joiners, leavers, tenure, roles | A picture of the organisation over time | Only the data needed for the purpose |
| Performance | 360 reviews, objectives, OKRs | Deciding where to invest in development | Calibrate as a team before using it |
| Behavioural profile | Behavioural assessment and motivators | Understanding how each person works and decides | Tell the person and explain their report |
| Engagement and culture | Anonymous pulse surveys, organisational DNA | Measuring the lived culture by area | A minimum number of responses per group |
| Hiring | Vacancies, pipeline stages, interviews | Seeing which profiles fit each role | A person makes the hiring decision |
| Development | 70-20-10 plans, training, succession | Preparing successors for key roles | Agree the plan with the person |
Which questions can people analytics answer?
It answers questions that lead to a decision well. Questions asked out of curiosity tend to end up on a dashboard nobody opens.
Questions that do lead to a decision: which key roles have no successor ready, and what each internal candidate still needs; which teams are made up of very similar profiles and miss other points of view; where performance and potential ratings disagree, and why; what behavioural profile a role really calls for before the vacancy is advertised; which areas live a different culture from the one leadership believes it has.
Leave out any question that looks for an answer about someone's private life. If a question cannot be explained out loud to the workforce, it is not a good people analytics question.
How is it different from human capital reporting?
Human capital metrics are the foundation that people analytics builds on. ISO 30414, whose second edition was published on 24 August 2025, sets requirements and recommendations for human capital reporting and disclosure, both internal and external.
According to the ISO technical committee responsible for it, the 2025 edition organises reporting into 11 human capital areas, introduces auditable requirements alongside recommendations and sets a baseline of 14 required metrics for any organisation that applies it. It is a useful reference for deciding what to measure; it does not tell you what to decide.
People analytics starts where the metric ends: when a figure on leavers, training or workforce composition is combined with performance, profile or engagement data to explain something and propose an action.
What limits does the law set?
The law does not forbid analysing people data, but it sets clear conditions. These are the rules any people analytics project in the EU, and in Spain in particular, needs to take into account.
| Rule | What it requires | Since |
|---|---|---|
| GDPR, Article 5 | Data that is adequate, relevant and limited to what is necessary for the purpose (data minimisation) | 2018 |
| GDPR, Article 22 | The right not to be subject to a decision based solely on automated processing that produces legal or similarly significant effects | 2018 |
| GDPR, Article 35, and the Spanish regulator's list | A data protection impact assessment where people are profiled or evaluated, including work performance, personality and behaviour, together with another criterion on the list | Current AEPD list |
| Spanish Workers' Statute, Article 64.4.d | The works council may learn the parameters, rules and instructions of algorithms or AI systems that affect decisions on working conditions and employment, including profiling | In force |
| EU AI Act, Article 5(1)(f) | AI systems may not be used to infer a person's emotions in the workplace, except for medical or safety reasons | 2 February 2025 |
| EU AI Act, Annex III, point 4 | AI systems for hiring, employment decisions, task allocation based on personal traits and performance evaluation are high-risk | 2 December 2027, following Regulation (EU) 2026/1744 |
What changes with the EU AI Act?
Above all, it changes the responsibility of whoever uses the tool. If an AI system takes part in hiring, promotion or performance decisions, the organisation using it has obligations of its own as a deployer, including informing workers' representatives and the people affected before putting it into service.
Regulation (EU) 2026/1744, known as the digital omnibus on AI and published in the Official Journal on 24 July 2026, moved the application of those obligations for Annex III systems to 2 December 2027. The ban on inferring emotions at work did not move: it has applied since 2 February 2025.
If you operate in the United Kingdom, the UK GDPR applies, and the Information Commissioner's Office publishes guidance on employment data and on automated decision-making. The EU AI Act still matters when an AI system's output is used in the EU.
How do you start a people analytics project?
Start with a decision, not with the data. These six steps work for a small business and for a consultancy taking the work to a client company.
First, choose a decision that comes up again and again and is made by gut feeling today, such as who is ready for a key role. Second, list the data that already exists for that decision and the data that is missing. Third, define the legal basis and check with whoever handles data protection whether an impact assessment is needed. Fourth, tell workers' representatives and the workforce what is measured and why. Fifth, try it in one area before extending it. Sixth, review the results as a team and record who decided what.
A small project that reaches a decision is worth more than a complete dashboard nobody uses.
Which mistakes should you avoid?
The most common mistake is using data for something other than the purpose it was collected for. An anonymous engagement survey cannot be used to decide about one person, and a behavioural profile is not a performance rating.
Other common mistakes: showing results for groups so small that you can guess who answered; turning a profile into a fixed label; letting a score decide on its own; measuring what is easy to measure instead of what matters for the decision; and not explaining to the workforce what happens to their data.
Sources
- Regulation (EU) 2016/679 (GDPR), Articles 5, 22, 35 and 88
- AEPD (Spanish Data Protection Agency): list of processing that requires an impact assessment (in Spanish)
- Spanish Workers' Statute, Article 64.4.d (BOE-A-2015-11430, in Spanish)
- Regulation (EU) 2024/1689 (AI Act), Article 5(1)(f) and Annex III
- Regulation (EU) 2026/1744 (digital omnibus on AI), Official Journal of 24 July 2026
- ISO/TC 260: ISO 30414:2025, human capital reporting and disclosure
- ICO: Employment information (UK GDPR guidance)
Frequently asked questions
- Is people analytics the same as HR analytics?
- The terms are used almost interchangeably. HR analytics often refers to the department's processes (hiring, training, payroll), while people analytics focuses on people and teams, but the practical difference is small.
- Can a small business do people analytics?
- Yes. With one well-chosen decision and the data it already has (reviews, profiles, surveys), it can start without a data team.
- Do I need a data protection impact assessment?
- It depends on the processing. The Spanish regulator's list requires one in most cases that meet two or more of its criteria, and the first criterion is profiling or evaluating people, including work performance, personality and behaviour.
- Do I have to inform the works council?
- In Spain, if algorithms or AI systems affect decisions on working conditions or employment, Article 64.4.d of the Workers' Statute gives the works council the right to learn their parameters, rules and instructions.
- Does people analytics decide whom to promote?
- It should not. Data orders and explains; a person makes the decision about each person, and the GDPR protects against decisions based solely on automated processing.