Data is the lifeblood of any organisation today. When treated well, it informs decisions, fuels innovation and keeps operations running smoothly. When treated poorly, it causes frustration, slows progress, and wastes money. Most organisations have tried fixing data after the fact. They spend months untangling old systems, cleaning up errors, and enforcing rules nobody remembers. It rarely works. The cost is high, the impact limited, and momentum is quickly lost.

The smarter path is to look forward. New systems, processes, and modernisation initiatives should be built with governance baked in, not patched on later. Data governance by design ensures that from day one, the right rules, roles, and quality checks are part of how work is done. This approach embeds governance in the very DNA of the transformation lifecycle. Here’s how it can be done.

The problem with retrofitting governance

Most headaches come from legacy systems and processes that were never designed with consistent data practices. The bigger and older the organisation, the worse it gets.

The symptoms are familiar: data that’s hard to find, definitions that change depending on who you ask, quality that can’t be trusted, and processes that block sharing or reuse. Security and regulatory risks pile on top. The result is mistrust, inefficiency, and a constant scramble to fix the mess.

Organisations often try retrofitting governance onto these systems. It means untangling complex data flows, wrestling with unsupported technology, and imposing rules on systems that were never built to follow them. Engineers and consultants who knew the ropes have long left. Even when inconsistencies are spotted, systems like SAP S/4 resist fundamental changes while transactions are ongoing, making quick fixes impossible. Trying to fix old data in a live system is like trying to rebuild a ship while it’s sailing—expensive, slow, and fraught with risk.

I’ve spent years in this battle. Tracking down owners, deciphering old ETL scripts, reconstructing intent from fragmented documentation. Generative AI might help uncover tribal knowledge in the future, but today, it’s mostly manual, time-consuming work. Few organisations see a meaningful return, and many abandon the effort, leaving the same issues for the next wave of projects.

Look forward, not backward

Data leaders who achieve impact focus on prevention. Instead of repairing every mistake, they design governance into new initiatives from the start. This shifts the approach from reactive to proactive, from remediation to prevention.

Transformation projects are unique windows of opportunity. Teams of architects, engineers, analysts, and domain experts come together with deep understanding of sources, quality, lineage, and intended use. But this knowledge is fleeting. Once projects end, people move on, contractors leave, and the metadata begins to decay. The moment attention shifts away, fields start to deteriorate. Data that was once clean goes sour. Capturing this knowledge while the eye is still on the field is critical. Business glossaries, data catalogues, and dictionaries are not just tools, they are lifeboats for knowledge that would otherwise be lost.

Governance in the existing lifecycle

The solution is not a separate governance process. Start with what already exists. Most larger organisations already have a transformation lifecycle, whether called project governance, solution delivery, or portfolio management. These frameworks define scope, decisions, and responsibilities. Embed governance expectations into this lifecycle, and any project of sufficient size automatically falls under governance.

Where no lifecycle exists, don’t invent one solely for governance. Create a lightweight, practical framework that serves multiple purposes, with governance as a built-in element.

Managing change

Introducing governance into a lifecycle won’t happen overnight. It may seem heavy at first, but most of these activities are not new, they are simply applied consistently and earlier.

Data quality rules should be part of technical specifications. Reusing enterprise data models saves time and provides a shared language. Identifying domain owners gives development teams a place to raise concerns and avoid workarounds. Central teams can provide modelling and data definition as a service. Even with all this, systems like SAP S/4 may resist fundamental changes while active transactions exist. This reinforces why governance must be addressed at design time—not later.

Teams are happier when they enable new projects rather than cleaning up old ones. Focusing on the lifecycle reduces risk, saves cost, and allows business users to work with trustworthy data from day one.

That’s a Wrap

Data quality by design is not theoretical. It is practical. It is achievable. Embedding governance into the transformation lifecycle turns data from a recurring headache into a strategic asset. Decisions made at the right time reduce risk, improve quality, and ensure the organisation can rely on its data to drive insight and innovation every time.