Data-Informed School Development: Who, If Not School Leaders?

In Germany, education is organized by the federal states; VERA, Startchancen, and LiGa are German assessment or school-development initiatives referenced below.

Why data-informed school development is less a question of tools and much more a question of strategy, structure, and the courage to take the first step.

85 percent of school leaders want to develop their school using data. 5 percent do so consistently. This gap is one of the most striking findings from the current Wübben survey "Schools Under Pressure 2025" (Schule im Brennpunkt 2025). It is also the reason I am writing this article.

An honest diagnosis

Anyone who talks to school leaders about this topic tends to hear the same mixture of interest and exhaustion. Data-informed school development sounds sensible, sometimes even appealing — but in practice it collides with staff shortages, substitute teacher scheduling, and what feels like twenty other priorities. The question, then, is not whether school leaders find the topic important. The question is what causes good intentions not to translate into action.

Three answers have crystallized from research in recent years. First, many schools lack a shared language for data, leaving analysis the responsibility of individually motivated people. Second, data is often associated with oversight rather than with learning or pedagogical freedom, which triggers defensive reactions in the teaching staff. Third, there are rarely fixed structures in which data work can actually take place. Staff meetings are full, subject departments are full, professional development days are scheduled up — and trying to build a routine alongside all other demands quickly fails for lack of time.

Data-informed does not mean data-driven. Data are decision aids, not substitutes for decisions. Pedagogical interpretation and the professional work of teachers remain central.

Who bears the responsibility?

Education ministries set frameworks, but they do not shape school practice. State institutes provide dashboards, but not a data culture. Teachers can work with data, but they need institutional backing. External programs like Startchancen or LiGa can provide momentum, but they cannot replace the process within the school.

So who takes responsibility for turning VERA feedback, learning-level assessments, and student surveys into lived school development?

My thesis: school leaders. And their responsibility is threefold.

Enabling

Creating the conditions in which data work can take place. This means fixed time slots in the school calendar, clear rules for handling data, and a small, voluntarily composed data team that does not run alongside other duties but has genuine meeting time. Schildkamp and others have shown for years that this structural step is the most important of all.

Steering

Anchoring data work systematically in school development. This means establishing a cycle in which questions are formulated, data is interpreted, measures are implemented, and results are evaluated. It is precisely this last step that is almost everywhere missing today. Only one in three schools checks whether its measures have actually had an effect. Without this step, the cycle becomes a one-way street.

Working with data

Shaping the values behind data use. Data must not be an instrument of oversight, nor a covert ranking among teachers. Data is a tool of shared responsibility for learners. This attitude is not reflected in mission statements — it shows up in how school leaders themselves work with data: whether they support their own decisions with evidence, and whether they do not push aside uncomfortable findings.

Three levels, one shared path

Data-informed school development rarely succeeds as a grand stroke. It grows from many coordinated steps. It helps considerably to assess honestly where a school actually stands. From my advisory experience and the relevant literature, three levels can be distinguished.

Level 1, Entry. Data is collected, but not used systematically. There is no data team, little data literacy, and in some cases reservations among the teaching staff. At this level it is not about grand concepts, but about trust and initial experience. Take stock, choose a manageable topic, form a small data team, run a short training session, make a quick win visible. Nothing more is needed at the start.

Level 2, Building. Initial analyses are taking place — often triggered by specific events, such as VERA results. A steering group exists, but the work has not yet been established as a regular practice. This is where the greatest leverage for good school leadership lies. Fixed time slots, a cyclical approach such as the data team model developed by Schildkamp and Ehren, a professional development concept, and above all an actively cultivated culture of trust.

Level 3, Deepening. Data-informed school development is embedded in the school program. Data teams work regularly, cycles are documented, and the teaching staff shares a common language. At this level the focus shifts to expansion. Student feedback complements attainment data, AI-supported diagnostics are explored as a new data source, and new assessment formats are accompanied by evidence. Those who reach this level have a responsibility to share what they have learned.

Most schools are moving at the transition between Level 1 and Level 2. That is precisely where the impact of strong school leadership is greatest.

What AI and assessment culture are changing

Data-informed school development no longer stands alone. It intersects with two debates that are currently reshaping school practice — generative AI and changing assessment culture.

Joscha Falck describes four areas of development in the AI competency model: understanding, applying, reflecting, and co-creating. Translated to data work, this means school leaders need not just a tool, but a language with which to place the topic in their staff community. Doris Weßels repeatedly points out that legal uncertainty is the biggest brake on innovation in schools. Data protection law, the AI Act, and state school legislation are complex — but those who do nothing out of fear of violations lose more than those who move forward with proper advice. Hendrik Haverkamp shows that a contemporary assessment culture cannot be conceived without evidence-based reflection. When new assessment formats are piloted, they need accompanying evaluation — otherwise they remain anecdotes.

These three voices enrich the classic discourse on data-informed school development at important points. They make clear that data work today is more than VERA analysis. It is part of a school development that is adjusting to a world with AI and changed approaches to measuring achievement.

Culture beats tool

A fully stocked dashboard is not a concept. A VERA feedback report filed in a binder is not a development process. Anyone who thinks only in terms of tools will fail, because tools without culture have no effect.

The most important levers lie elsewhere. School leaders who are themselves visibly working with data automatically create ripple effects in the teaching staff. A culture of psychological safety allows uncomfortable data to be shared openly. Data storytelling ensures that numbers do not remain abstract but are embedded in a meaningful narrative. And the willingness to evaluate effectiveness closes the cycle that otherwise stays open.

All of this is demanding. It requires a long-term commitment, as Martina Diedrich of the IfBQ Hamburg rightly emphasizes. But it is also what makes the difference between schools that document data and schools that learn with data.