Let’s face it: data governance is an unfortunate term. For most people, it sounds like something bureaucratic and compliance-heavy – not something designed to support business performance or help people do their jobs better.
If we shifted the language to something like information resource management – similar to human resource management – the idea becomes clearer: you're managing one of your organisation’s most valuable assets.
The reality is, good data governance isn’t about enforcing rules or building frameworks. It’s about enabling people to work more confidently and effectively with data – and doing so in a way that fits naturally into how they already work.
For years, organisations have tried to implement data governance in one of two ways:
- A compliance-driven model, where governance is imposed top-down, often linked to legal requirements like POPIA. It’s enforced, monitored and people are penalised for non-compliance.
- A ‘build it and they’ll follow’ model, where a framework is created and the hope is that people will adopt it voluntarily.
Both approaches eventually lose steam. The first creates resistance, because no one likes being told what to do in an overly rigid way. The second often ends in apathy, because it’s not linked to real value or business outcomes.
IT is only the data custodian – the whole business owns the data.
A more sustainable alternative is a non-invasive, human-centred approach that meets people where they are, focuses on practical value and builds from the inside out.
What non-invasive data governance looks like
Non-invasive data governance is an approach developed by Robert S Seiner. The key insight behind this approach is that data governance is already happening in most organisations. Businesses wouldn’t function if they weren’t governing their data to some extent. It’s just that it is being done informally and inconsistently.
There are always people in the business who know where the data is, how it’s used, what it means and what to watch out for. These are typically your subject matter experts – your informal and unrecognised data stewards. The goal isn’t to add work or give them new titles. It’s to recognise a role they are already performing, support them and create structure around what’s already working.
This is what some call data governance by stealth. Instead of trying to heavy-handedly change everything, it evolves naturally and organically.
To adopt this approach, start with the basics:
Check for readiness: The organisation must want to do it. If there’s no recognition that governance matters – whether for data quality, efficiency, or AI readiness – it’s a non-starter.
Get executive support: Like any meaningful change, governance won’t get far without buy-in from leadership. It needs vocal sponsorship and visible involvement from senior decision-makers.
Do a critical analysis: This involves education and awareness-building – helping the business understand what good data governance looks like, and what’s involved. Then benchmark where it is now, identify gaps and define best practice for its context.
Build a roadmap: The analysis should feed into a clear roadmap, including an operating model, change plan, communications strategy and recommendations for supporting tools. Governance isn’t a tech project – it’s a people one – so change management is critical.
Show it’s already happening: One of the most powerful parts of this approach is pointing out that governance is already being done – just in silos and without coordination. Bringing this to light often leads to “aha” moments that build buy-in and reduce resistance.
Business support is critical
Too often, data governance is seen as IT’s responsibility. But IT is only the data custodian – the whole business owns the data. Like HR, finance or operations, data is driven by business functions and should be treated as such.
The challenge is that, unlike some other business processes, the benefits of data governance can be indirect and hard to quantify. The company may not be able to say exactly how much better the business runs because of improved metadata management or lineage – but the impact is real.
Better governance leads to better data quality. That, in turn, leads to less rework, lower risk, fewer errors and more trust in decision-making.
The need for governance has never been more urgent – especially as organisations move towards AI, automation and self-service analytics.
For AI to deliver value, clean, organised, well-understood data is required. The phrase “garbage in, garbage out” still holds. Governance is what underpins the people and process elements of data quality – beyond just running a few tools.
Likewise, in the world of self-service business intelligence, if people don’t know what data exists, where to find it, or how to interpret it, they’re unlikely to use it effectively. That’s where a good data catalogue, supported by governance, becomes essential. It helps make relationships visible, builds trust, and lets people get on with their work without second-guessing the numbers.
Making it work in the real world
This approach isn’t about creating a new department or forcing change on an already stretched team. It’s about surfacing what’s already happening, giving it structure and aligning it to business goals.
Done well, governance becomes part of the organisation’s culture – embedded in everyday operations and decision-making, not bolted on after the fact. It supports transformation, enables innovation and helps people do their jobs better.
And most importantly, it reminds us that data governance isn’t really about the data. It’s about the people who work with it.
Share