Data Governance: Start with strategy, not with data

“Do you want to become a data-driven company?” The question sounds rhetorical, but it is not. Anyone who answers it seriously has to state which decisions should rest on data in future and which ones will no longer be made on gut feeling. This is where every data governance effort begins, and in practice this is exactly the step that gets skipped.

The typical start looks different: a catalog is procured, a glossary is begun, a role matrix is copied from a template. A few months later there is a set of rules that nobody would have missed. The cause is not a faulty tool but a missing why. Data governance with no link to the corporate strategy has no addressee, and what has no addressee does not get prioritized.

From corporate strategy to data strategy

The workable sequence is: corporate strategy, data strategy, governance measures. Where does the company want to be in a year, where in five? Which of those goals can data actually support? A company that wants to grow needs reliable figures on recruiting, sales, capacity utilization and so on, connected in a way that makes a cause-and-effect chain visible. Only once that chain holds can you justify which data deserve governance at all.

A simple derivation has proven itself here: objective, outcome, benefit, deliverables. Every strategic objective becomes a measurable outcome, every outcome a concrete benefit for the business, and from that follow the deliverables it takes to get there: a data map, defined responsibilities, quality reporting, a data product. This list is not an end in itself but a rationale. It explains to management why this particular measure comes first and another one later.

Two insights belong in the picture from the outset. First: every data strategy is different, because it depends on the company's maturity, its business model and its culture. Templates help with structuring, not with deciding. Second: not all data are equally valuable. Data are capital, and capital is not treated alike. Anyone who tries to regulate everything with the same rigor ends up regulating nothing effectively.

Organizations are vertical, data flow horizontally

This is where the real heart of the matter lies. Departments, budgets and targets are organized vertically, while the data of a customer process run across CRM, ERP, ticketing system and reporting. Every handover is a point at which accountability can get lost. The result is familiar from every steering meeting: “Your reports are wrong.” Answer: “No, your data are wrong, my report is right.”

This exchange does not end with a better report, it ends with an assignment of responsibility. Who supplies the data, who signs off on it, who decides how a field is defined? Data governance is essentially the answer to those three questions, documented and backed by consequences. Hence the second rule of thumb: processes are data and data are processes. Anyone designing a new process should design it from the data perspective, not solely from the roles of the people involved.

The prerequisite for this is transparency. A data map shows which core systems exist, which data sit in them and where those data flow. It is unspectacular and still the basis for almost everything that follows. That applies in particular to security and regulation: how is a company supposed to protect its most important data if it does not know where they are? Experience shows that the initial inventory is the biggest hurdle during implementation, less a technical one than an organizational one.

Data is business, not IT

As long as data counts as an IT topic, governance remains a background service. The mindset has to change all the way up to the executive board, and in this direction: if leaders do not think in terms of data, why would they lead their teams that way? Enablement therefore starts at the top. Training for executives, then an internal format for everyone that explains why the topic matters, and people who carry it forward as ambassadors. A data governance initiative is sustained by multipliers, not by a staff unit.

The benefit can be put refreshingly plainly: you need to be able to make the right decisions faster than your competitors. That is the economic core of governance, and it is the wording that lands in the boardroom. It also helps to keep the goals visible, whether on a board, in the intranet or on the wall. Initiatives rarely lose focus because someone objects; they lose it because people forget.

Case by case instead of big bang

One final point decides between success and failure: there will be no big bang. Data governance takes shape case by case, always with an eye on value creation and always in the order why, how, what. A maturity model helps to establish honestly where the company stands today and where it wants to go. A simple scoring scheme for use cases becomes a communication instrument for management, because it makes prioritization transparent.

The effect shows with every case: what was learned in the first use case accelerates the second. Amazon deliberately concentrated on a single product category at the outset, books, and used that time to lay the foundations for working with data. Focus first, build expertise, then adapt. The uncomfortable part remains: data strategy is hard work, and many organizations no longer take the time for it.

Data Governance: Strategy is key

From corporate strategy to data strategy

Objective

Outcome

Benefit

Deliverables

Organizations are vertical, data flow horizontally

Data is business, not IT

Case by case instead of big bang

Marco Wirtl