We’ve recently seen a trend in Data Governance with much debate over the differences between Data Governance and Data Management.
Industry expert, Gartner, suggests that in 2025 some 80% of organizations seeking to scale their digital businesses will have failed because they do not take a modern approach to data and analytics governance (Gartner – Turn Data and Analytics Governance Policy into Action Through Well-Defined Processes, August 2021)
So, for those organizations starting on their data governance journey or simply wanting to further leverage the benefits for their business, they could very quickly come to the conclusion that data governance is at best at a crossroads or, at worst, a crisis!
That said, for those having worked in Data Governance in the past decade, this is not surprising as these debates are nothing new and were being had ten years ago! And this is why Gartner’s downbeat assessment is probably not too far away from the future reality, and traditional data governance is indeed not delivering and is in urgent need of a fundamental rethink.
One exciting new emerging area of thinking is the work being collectively labelled under the umbrella term of “data mesh”. The premise of data mesh is that it connects distributed data across different locations and organisations. A data mesh is designed to ensure that data is universally available, easily discoverable, secure, and interoperable with all the applications that need access to it. Does this then afford us a refreshing way forward?
While the thinking around data mesh is still evolving and immature, it still rightly requires Data Governance at its heart. Current work makes it clear having this is critical to a successful data mesh implementation. The crucial difference, though, is data governance in a data mesh solution would be much more decentralised or federated across the organization – or put another way embedded into the business fabric. Intuitively, this makes sense as those who have implemented data governance in the past and who know that the best “foot soldiers” for delivering a successful Data Governance programme need to properly mobilise a network of data owners, data stewards & data custodians throughout the business user end of an organisation. But now, for the bad news, decentralisation also has clear cost implications. Large monolithic data teams at the core of an organisation require big data remediation budgets, typically only available to the largest corporate national, and international enterprises.
Therefore, so good so far. But if we unpack this logic, on what basis will a move to a more decentralised approach for data governance have any more chance of success than having a more centralised one? Undoubtedly, the same pain points to a successful execution will remain in either operating model?
For example, you would struggle to see how the same challenges and issues that Gartner identified as preventing the success of Data Governance initiatives – the lack of key governance skills, the lack of clarity over crucial roles, the lack of skills & experience to drive governance initiatives, ongoing resistance to governance adoption from the wider business & a lack of leadership to support the initiative – will be addressed simply by moving to a decentralised data operating model. In fact, in some cases, it may make it harder to achieve.
Indeed, Gartner reveals that the single biggest reason for the failure of current data governance programmes is their lack of a standardised (data) governance approach. Playing devil’s advocate would this be more difficult to overcome in a decentralised model than a centralised.
So, where does this leave us?
It feels that even if the data mesh approach becomes an increasingly important option to drive a business data strategy forward from a technical perspective, it is unlikely to be a “magic wand” that will either refresh or resolve the ongoing Data Governance challenges Gartner warns of.
It seems the same challenges of fixing Data Governance in an organisation will still require the ongoing appreciation of the underlying issues, winning the hearts and minds of business users, and applying data governance in a structured and actionable manner. And without significant budgets for most organisations, such incremental change will need to be achieved through a stealth approach rather than creating an important data transformation initiative. To this end, the Oakland Group has recently written a useful guide on how this might be achieved (Data Governance by Stealth Lighthouse Paper, September 2021).
Data Governance is not at a crossroads as such; instead, it is still on a (long) and challenging journey regardless of the label being placed upon it.
Andrew Sharp is a Data Governance Lead at The Oakland Group