Implementing People Analytics correctly

By Redaktion techport.ai, HR-Beratung · Last updated on 21 August 2026

The HR dashboard has 24 tiles. It shows turnover, sickness rate, age structure, time-to-hire, training days and gender distribution. It is opened by three people, all from HR. Management saw it once. Their question was why production in plant two could not fill its vacancies, and the dashboard had no answer to that.

People Analytics begins not with data, but with a question that someone wants answered in order to make a decision. Everything else is reporting, and reporting is useful, but it is something different.

How you can tell

There are dashboards, but no decision is based on any of them.
Analyses are requested when a problem is acute and then take weeks.
Data protection and the works council block evaluations because their purpose is unclear.
HR cannot answer management questions, even though the data would be available.

Why this happens

Analytics is treated as a tool question: buy a BI tool, connect data, build tiles. What is missing is the path from question to answer: Who asks, why, what data is needed, what method, who decides what afterwards. Without this path, evaluations are created that are technically correct but practically useless, and data protection and works council concerns remain justified because the purpose is unclear.

Our approach

Collect questions. Together with management and department heads, we identify the decisions due in the next year for which HR data plays a role. This leads to specific questions: Why does recruitment in plant two take twice as long? Is early fluctuation dependent on onboarding?
Define methodology and data. For each question, we determine what data is necessary, where it comes from, at what level it will be evaluated, and which method is appropriate. Usually, clean comparisons across time and groups are sufficient. Statistical models are the exception.
Clarify the framework. Purpose, data basis, aggregation level and access are agreed upon with data protection officers and the works council and incorporated into the Betriebsvereinbarung [works agreement] for the HR system. This ends the blockage because the purpose is now clear.
Evaluate, communicate, decide. The answer is a one-page document with findings, classification and recommendations, not a dashboard. It goes to the person who asked and leads to a decision. Questions that recur are subsequently automated.

What you gain from it

HR answers management questions with data before they are asked.
Data protection and co-determination become partners, because every purpose is named.
Investments in analytics tools follow demand rather than supply.

From our projects

Collecting decision-making questions with management is the step in our projects that changes everything else. Twenty dashboard tiles become five questions, three of which can be answered with existing reporting. Data protection and works council concerns usually resolve themselves once the purpose, aggregation level and access are named for each question, because the concerns were directed against the lack of clarity, not against the evaluation itself.

Good to know

Evaluations of personal data require a purpose and a legal basis. For analyses at group level with minimum group sizes, this is usually the legitimate interest according to Article 6 Paragraph 1 Letter f GDPR, documented in a balancing test. Evaluations that concern the behaviour or performance of individual employees are subject to co-determination. Systems that prepare decisions about employees may fall under the high-risk obligations of the AI Regulation from December 2027.

Frequently asked questions

Which tool do you recommend?

That depends on the question. Many questions are answered by the HR system's reporting. For cross-system connections, a BI tool, often already present in the company, is beneficial. A dedicated People Analytics tool is rarely the first step.

Do we need a Data Scientist in HR?

Not initially. You need someone who can translate questions into data requirements and understands the systems. Statistical depth can be sourced externally if a question demands it.

Let's talk about Implementing People Analytics correctly

In a thirty-minute first call we clarify where your biggest lever is and whether we are the right partner for it.

Book a first call Our software