Most articles about data governance I’ve read in recent years were terribly boring, so I try to use an analogy to make the dry topic more interesting. I have to admit that I don’t know how and why this idea came to my mind, but let’s try to describe all that can go wrong in data governance along the seven deadly sins. As good governance is not so much about defining fancy dashboards or creating impressive job titles, but rather breaking up silos, we will mainly refer to this aspect.
Maybe the worst an organization could do is not to do anything at all. When talking to customers we are often confronted with a lack of awareness as well as a lack of understanding. Missing or inadequate data governance is not recognized as root cause of various problems. Rather are the issues blamed on an “under-delivering IT” or “irreconcilable differences between business units”.
Even if data governance has been identified as a problem, organizations often shy away from tackling the topic or address it halfheartedly. As long as an organization has not understood and embraced data as one of its most important assets, this is not going to change quickly. Emphasizing the value and importance of data to an organization is a senior management obligation.
In an extraordinarily remarkable conference presentation about a successful data warehouse implementation the business lead of the program compared people who safeguard data like a treasure to Gollum, the fictional character of J.R.R. Tolkien’s legendarium. Almost every organization employs people who see data as “their precious” which might not extend their life beyond natural limits, but – actually or purportedly – protects their position in the organization.
A key aspect of good governance is to overcome silos, which often do not manifest in different technologies, but rather in the hearts and minds of the employees. A key component to break up silos and promote the sharing of data across departmental boundaries is the establishment of a “common language in data”, which can and should become the main instrument of a successful governance organization.
Wrath is what we have often encountered when people were eventually forced to give up their beloved “shadow IT”, and to a lesser extent in heated harmonization discussion. To moderate and accompany such a process you need not only empathy and experience, but also senior executive commitment and a common goal.
In organizations that are perceived leaders in data and analytics, governance objectives have been woven into the MBO system. As soon as bonuses and promotions depend on meeting objectives related to data governance, people show an astonishing willingness to forsake old objections and prejudices.
Initially I thought it might be impossible to stretch the analogy so far as to integrate lust into an article about data governance and breaking up silos. However, it has become much easier when applying the following definition “Lust is a psychological force producing intense wanting or longing for an object, or circumstance fulfilling the emotion.”
No small number of data governance projects has failed or significantly underdelivered, because of wrong priorities. Instead of tackling the hard problems, such initiatives were misused to acquire new tools loaded with unwarranted promises. While tools such as metadata management and data quality solutions can be of great help, no governance project has ever failed or succeeded only because of the tooling.
Rome wasn’t built in a day, and neither is a well-working governance organization. While it is desirable to have ambitious goals, they need to be attainable as well. Hubris can also manifest itself in the belief that it’s “not a big deal” or that external support is just a waste of money.
Organizations that haven’t dealt with governance so far will need support from outside the company: be it by hiring experienced people, sending employees to respective trainings, or tasking consultants with the support in setting up or improving a governance organization.
Gluttony can have many forms in the context of data governance initiatives, when understood as over-consumption, i.e. a situation where resource use has outpaced the sustainable capacity of an ecosystem.
To find a meaningful path between inactivity and exhaustion, you need to understand an organization’s ability to change in the first place. Successful governance project we have supported all had one thing in common: a roadmap gradually introducing new capabilities and familiarizing the people involved with new concepts and ideas.
One may argue that an organization has already achieved a lot when people envy others for being in charge of data governance or data in general. While it is true that you are better of when rather more than less people want to be responsible, an organization need not waste time and money of turf wars.
Nowhere is this more evident as when establishing the position of a Chief Data Officer and the question who he or she should report into. Here you have the choice between “the good, the bad, and the ugly”: in organizations that take the topic seriously, the CDO will be on equal footing with the other C-level executives; a little less good, but still acceptable, is when the CDO reports into let’s say the CFO or CMO, depending on the industry and focus of an organization; it’s only really ugly if data is still perceived as “some IT thing” and the CDO consequently reports into the CIO.
Stefan is a Director in the Finance & Risk practice of TME, focusing on Data & Analytics.
He has supported numerous banks in setting up and implementing successful data governance initiatives.