Often marred with bureaucracy and red tape; Data Governance is a challenging feat to achieve in any organisation.
Unlike a data science or engineering initiative, Data Governance is effectively implementing business change.
Reluctance to change is not a new phenomenon. A journal published in 1948 entitled “Overcoming Resistance to Change” highlights how the different participation rates from factory workers created drastically different results in acceptance of the said change.
Even though the research goes back more than 70 years, the fundamental psychological traits of human beings stay the same.
So why does Data Governance often end up failing?
#1 When Data Governance is treated as a limited-time project
#2 When Data Governance does not have Executive Sponsorship
#3 When Data Governance is confused with other analytics initiatives
#4 When fixation is on Enterprise-wide governance rather than an executable strategy
Let’s dive into it!
Too often, Data Governance is treated like a limited-time project. This puts undue pressure to achieve unrealistic goals. Engaging with the business stakeholders, explain the importance of data governance, get their buy-in, and rolling out training takes time.
Takes a long time!
Also, the return-on-investment trickles in slowly, by which leaders run out of patience.
If this is treated as a project, there is an expectation of the project coming to an end. Instead, governance activities need to be embedded across the business and made business as usual.
Merely getting people to realise governance is their responsibility and not IT’s, takes months of consistent effort.
Avoid this error by having a dedicated governance team leading the business change with set goals and metrics. Involve experts as required, however, build your own business capability to lead the function. Treat Data Governance as a new unit that has the accountability of enforcing governance with buy-in across all the various business areas.
Where do I get started with this one?
People are either fearful of their leaders or are inspired by them. Which one is more suited can be debated another time!
The point is, if the leaders and executives are not behind this business change of embedding Data Governance, this is a sure-fire way of sabotaging the entire initiative.
Executives would also understand how best to implement change within their respective areas, they would know how to rally their troops and they’re able to deal with resistance that inevitably arises.
Avoid this error by having the main executive sponsor for the initiative, usually the CDO (Chief Data Officer), who can influence other executives and senior leadership teams to get behind the idea.
This seems unusual, however, the difference between data management and analytics is not well-understood.
I explain here how there is a lack of understanding of data management: learn why data scientists are failing?
Someone who has spent a significant part of their life working on an analytics project can’t become an expert in data management, overnight. So, if you’re thinking of restructuring your team and handing over the Data Governance baton to your best analytics colleague; I’d urge you to think again!
A team led by someone who doesn’t have a passion for this area will inevitably lead to failure.
Unlike an analytics initiative of implementing tools, metrics or dashboards; Data Governance embedding requires strategy building and execution. It’s not uncommon to hire external help to kick-off the process whilst using this person to help build internal capability.
Avoid this error by using the most trusted influencer in the team to lead the initiative. Feel free to run Data Governance embedding loosely in line with the largest analytics initiative, however, do not join them up. Otherwise, you risk derailing your goal.
This is when the strategy is thrown out of the window and fixation is on working across the whole business in the shortest possible time.
The (proverbial) elephant must be eaten one bite at a time. But to extend on this analogy, the elephant must also be cooked and prepared one piece at a time.
Starting on an enterprise-wide rollout must start from one or two business areas. Especially areas that have some loyalty for Data Governance and its influential leadership. The reason for this is to make sure the initiative doesn’t hit roadblocks in its first few weeks. This is also the time when the new team and strategy is most vulnerable to sabotaging from detractors.
Avoid this error by using the success in a few areas as an example for getting other areas involved in the Data Governance roll-out. Human beings suffer from FOMO (fear of missing out) excessively, so having embedded Data Governance in a few areas will mean other areas will want to get involved.
This, by no means, is an exhaustive list; I could easily add many more points on poor strategy alone. However, I hope this succinct post has helped you understand typical Data Governance pitfalls and how to avoid them.
Do you agree with what I’ve said above, what are your thoughts? Feel free to reach out to me via my email at [email protected], if you have some feedback or if you just want to say hello!
If you’re still reading this, I hope you’ve found some value in this blog post.
If you’d like to be kept informed of more content like this, subscribe to my newsletter.
Also, check out my other blog on Data Engineer Jargon