Data Governance at Exabyte Scale
DAMA-framed Data Governance Office established; full GDPR + EU AI Act compliance; data democratisation as the operating norm.
Theme Risk, Compliance & Governance · Also Scaling
In brief
Situation. The company managed exabytes of data. Trust in that data, across the C-suite, across business units, across regulators, was the prerequisite for everything else.
Complication. At that scale, governance is not a side project. Without unified data concepts, defined ownership, and clear quality standards, every business unit ends up with its own dialect of the same data, and senior leadership ends up making decisions based on inconsistent numbers.
Resolution. I led a comprehensive Data Governance initiative, inspired by the DAMA framework, in partnership with Accenture. We built data dictionaries, defined quality standards, set clear roles and responsibilities, and created the company’s Data Governance Office, chaired by the CTO and operationally managed by me. The Office became the central authority on data-related decisions, including monetisation. Databricks Unity Catalog and related tooling were expanded. Extensive internal training programmes turned employees into data-literate practitioners.
Impact. Departments gained increased autonomy from the centralised data function. They could self-serve while being accountable for the quality of what they produced. More agile decision-making, better regulatory compliance, improved data accuracy. Foundation laid for the company’s long-term ambition of becoming a fully data-driven organisation, trusted by stakeholders at every level: CEO, CFO, CTO, Director of Compliance.
The longer story
Most companies treat data governance the way medieval kingdoms treated tax collection: nobody likes it, but everyone agrees somebody else should be doing it.
The trick to making data governance work is to stop calling it governance. Call it democratisation. Frame it as: “we are giving every team the right to use the data they need, in exchange for the responsibility of keeping the data they generate clean.” Suddenly governance stops sounding like a tax and starts sounding like citizenship.
Once teams felt they had been given autonomy rather than had restrictions imposed on them, adoption of the governance framework accelerated dramatically. Compliance officers stopped being the villain. Data quality became a source of pride.
The cultural shift was the project. The data dictionaries and the Unity Catalog were the tools. The training was the bridge. Without all three you do not have governance, you have paperwork.