Thijs van der Feltz, MBA, is an enterprise data management specialist and information architect and has been devoted to data management for over 30 years. He has hands-on experience in numerous facets of this field. With a long history in financial services, he is currently in the middle of implementing data governance at an insurance company in the Netherlands. He is one of the founding members of the recently established DAMA-Netherlands chapter.
How long have you been working in Data Governance?
If we define it loosely, I started doing Data Governance about 30 years ago, long before that name emerged. It was called data administration then and was taken very seriously at the company I worked for. Things were different at that time, of course. With the mainframe being king of the jungle, there was already some serious governance going on, except that IT was in charge instead of the business. Over the years the business gained more control of IT, and rightly so, but governance has lagged behind. The king is practically gone and what we have left is a rapidly expanding jungle of disparate data that is cultivated by innovation and savvy users. Hence the need for data governance is now greater than ever.
How did you start working in Data Governance?
In the 1980s I worked for a consultancy firm that specialised in information strategy, databases and data modelling. After 5 years, I joined a merchant bank to set up their data administration. In time, this field of work evolved into data governance, data architecture/modelling and metadata management, which I have all embraced during my career. In fact, in a recent article Robert Seiner suggests that data modelling is data governance; not literally of course, but in terms of discipline, rigor and effect, the analogy is a good one.
What were your initial thoughts when you first fully understood what you had got into?
When this ‘new’ field called data governance emerged, it did not come as a surprise to me because I had already been doing data governance stuff for years. It added some missing parts to data management as we knew it, and it had a nice ring to it that appealed to business people.
Data management and data governance are intrinsically connected, but are also distinctly different. One can argue that we have a chicken-and-egg discussion as to ‘who is on top’ but ultimately, data management is the task that needs to be done. Data governance is needed to make sure that this is properly embedded in terms of roles and responsibilities. Semantically, the term data governance is sort of a misnomer, because we cannot govern data (the inanimate ‘thing’) itself. But hey, what’s in a name; if it works, don’t fix it. What can and must be governed, however, is the behaviour of people, and data management processes. That’s what data governance is about.
Are there any particular resources that you found useful support when you were starting out?
There were very few resources back then, but I found several books that were particularly influential and supportive to my efforts: “Data administration” (William Durell, 1985), “Fourth Generation Data” (Dan Tasker, 1989) and “Case*Method – Entity Relationship Modelling” (Richard Barker, 1990).
At that time, I was consultant with a company that was a pioneer in data centric solutions, with data management as a core competence. I had the good fortune of having access to a wealth of expertise provided by all my colleagues.
What characteristics do you have that make you successful at Data Governance and why?
Data governance is a complex field, so my multi-disciplinary educational background (engineering, economics, MBA), which spans the business-IT gap, has given me an advantage in comprehending the big picture.
What also helps is that I’m passionate about data and relentless in seeking answers and finding ways to convince people of the importance of data governance. I have experienced first hand the advantages of mature data management and understand the root causes of data quality issues. I am very critical about superficial approaches to fixing data quality, which is not always appreciated by some of the people I work with. Diplomatic skills should not be underestimated when trying to achieve data governance success.
What is the biggest Data Governance challenge you have faced so far?
One challenge is the knowledge gap between the business and IT. While the business clearly wants high quality information and more and more uses for it, they generally don’t want to be bothered with the details of achieving data quality. They just want to plug their (information) appliance into an outlet in the wall, but don’t care what’s behind that wall. This leads to another challenge: obtaining management buy-in.
In the other corner is an IT organisation that is primarily focussed on technology and not so much on informational aspects such as semantics, purpose, and taxonomy. The result of this polarisation is often poorly defined business and data requirements. This is a root cause of cost overruns and mutual frustration between business and IT. Data governance plays an important role in bridging this gap, by ensuring that the business takes control and responsibility of its own data. Data governance sets the rules and policies, while IT is responsible for the data management execution that correctly implements them in a way that is transparent and verifiable by the business.
Data belongs to everyone, but nobody wants to own it. Business engagement, responsibility and accountability are key, but that is much easier said than done. Strong leadership and executive sponsorship are prerequisites but actually getting an organisation to embrace a ‘data as an asset’ culture is the ultimate challenge and very hard work.
What have you implemented or solved so far that you are particularly proud of?
Having designed and built a custom built enterprise metadata repository system for data modelling at the previously mentioned merchant bank has given me great satisfaction and pride. While there are many good tools on the market, none offered a complete solution that we were looking for, so rather than trying to synchronise multiple overlapping tools, we decided to build an integrated solution ourselves. It was used from 1990 to 2010 to serve the needs of hundreds of application developers and database designers and it was our metadata integration hub. We incrementally developed an enterprise data model (ca. 4000 entities) that was mapped to our relational databases across some 250 systems. The data modelling process was incorporated in the application development life cycle, and certified data models helped to create accurate business requirements. The system provided us with considerable control of our (master) data, as well as reliable impact analysis that enabled exceptional agility. Eventually this system faded into obscurity, after a takeover that resulted in our systems succumbing to the new organisation. Nonetheless, it has been very inspirational to have worked in such a mature data management environment, with highly capable team members and supportive management. It has opened my eyes to the potential of data management done right.
I am also proud of my recent progress in establishing and promoting a corporate business glossary. A seemingly simple list of words and definitions can be an endless source of discussion, so it’s important to be firm at times and to keep moving. Two factors have proven to be successful: (1) a comprehensive set of guidelines & standards to improve the quality and overall consistency and (2) embedding the glossary in existing policies and procedures, to encourage its use, to enforce consistent use of terminology, and to leverage reusability by averting the need to redefine terms in multiple documents.
What single piece of advice would you give someone just starting out in Data Governance?
Get involved in teams of like-minded and experienced people and start doing it. There is no shortage of good books and blogs in the internet to help learn the trade, but compare sources to separate truth from fantasy and hypes. Some of my favourite sources of knowledge are the TDAN, B-Eye-Network and DataVersity. Also, the recent Data Governance books by John Ladley and Robert Seiner are very good. To get a broader overview, become familiar with some general management and quality assurance books such as those by W. Edwards Deming, Peter Drucker and Danette McGilvray.
Data governance is business responsibility, but it is important to realise that the success of data governance strongly depends on the maturity of the data management function. To this effect and in the spirit of the old adage “you can’t manage what you can’t measure (or define)” I would to add “you can’t govern what you can’t manage”. One should therefore understand the essentials of data management and business-IT interaction. Data governance is a multi-disciplinary function and requires a certain level of seniority to get things done, so prior experience in different data-related fields is highly recommended before seriously getting involved in data governance.
The rewards of data governance success are considerable, but passion, patience and perseverance are indispensable while pursuing this goal.
Finally, what do you wish you had known or done differently when you were just starting out in Data Governance?
For a long time, I believed that it was sufficient to enforce a solid data management discipline using advanced tooling for data modelling and metadata management. This idea was further reinforced by, as mentioned earlier, positive experiences with a mature data management environment. In recent years, it has become very obvious to me that the biggest challenge is people, culture and change. No surprise here, just confirming the findings of Nicola’s previous interviews. While I have no regrets about my career path, I do wish I had become aware of this earlier, to have gained more experience in tackling the ‘sociological’ aspects of data governance.
I wondered if you could share a memorable data governance experience?
During my years at the bank, we were involved in several successive mergers and acquisitions. Part of the process was systems integration. During these sessions, our new colleagues were surprised that we had only a single customer database, while to us, this seemed like such an obvious design choice. To make a long story short, our IT system was chosen several times to be the leading environment for the new organisation. In time, I also witnessed that such decisions are more often politically motivated, with data being of lesser concern. Could this be a reason why the anticipated benefits of mergers often fall short of expectations?
What are your final thoughts?
I would like to thank you, Nicola, for giving me the opportunity to share my views and join the ranks of data governance experts that you have interviewed over the years. Their views and public contributions have been a valuable source of knowledge and inspiration for me in my daily work.
Having read my interview with Thijs you can also read my free report which reveals why companies struggle to successfully implement data governance.