When better data means helping people better
Leger des Heils wanted numbers it could trust at national scale, where the same word had meant different things across regions. We unified how the business defines its concepts with how IT calculates them, and made every number traceable to source, so trust finally has one place to live.
1The challenge
Leger des Heils exists to help people, across the Netherlands and across the full range of care and welfare. For an organisation like that, data is never the point. The person it serves is the point. The ambition it brought to us was simple to say and hard to do: make its numbers trustworthy enough to act on, at national scale.
Every important number has two sides that rarely meet. The business knows what a concept should mean: what counts as quality, as recovery, as a satisfied client. IT knows how that concept is calculated from the underlying data. When those two are not joined up, you get figures nobody fully trusts, and that was the gap Leger des Heils set out to close. Across regions the same word could mean different things, and national figures took real manual effort to produce, because meaning and calculation had never been unified.
2Our solution
So we built the layer where the two meet, on top of a fully modernised foundation: an upgraded Delta Lakehouse, infrastructure as code in Terraform, dbt, and governance, data quality and a data catalog built in from the start. The business defines the concepts that matter, in its own words and with clear ownership. IT owns how each one is calculated from source. And the governed layer binds them together, so a definition is not just written down somewhere, it is the definition the calculation actually follows. That is where trust gets unified: one agreed meaning, one agreed calculation, owned on both sides and joined in the middle.
With that in place, a number stops being a matter of who produced it. We wired complete lineage across the Power BI environment, more than 1,500 reports and 500 datasets, so any figure can be followed from the report all the way back to its source and shown to be the calculation of the definition the business agreed. The business owns its master data outright, and governance, quality and the catalog keep the link between meaning and number honest over time.
3The results
We did not hand this over and leave. We levelled up the teams to run and extend the platform themselves, developing with CI/CD, notebook engineering and expert Databricks solutions, and we worked closely with the business across HR, healthcare, finance and operations to sharpen the insights that matter. The result is data the organisation defines, owns, and trusts, in service of the mission rather than the other way round: the confidence to see clearly, act well, and keep its focus on the people it is here for.
“The hard part was never the technology. It was agreeing what our numbers mean, and making sure they are calculated that way everywhere. Datahub built the layer where our definitions and IT's calculations finally meet, with lineage back to source across more than 1,500 reports. Now a number means the same thing wherever you see it, and we can trust it enough to act on it.”
