• Mark Meuldijk, Partner |

It's common that there is resistance against Data Governance despite being a K-Enabler for more "trendy" topics. Starting small, thinking big and realizing the IT-connection can make the difference between a success story and a failed project.

In the past few years, more companies started to become data driven. Words like Data Science, AI, Industry 4.0, IoT have earned a regular presence in corporate speeches. But it is not so common to hear Data Governance among them. Even not being the flashiest term, it is instrumental for the success of any of the emerging data driven technologies.

Resistance

In companies that have successfully performed for many years or even decades without focusing on data, it is common to encounter resistance from the various stakeholders. Data governance is usually perceived as something not clearly identifiable, having to do with lots of rules, concepts and definitions. No wonder that in companies where data management has never been a priority, this topic can be instinctively perceived negatively. And, who could blame a colleague who has been doing a certain job well for years, for not being enthusiastic on the upcoming set of new activities which are approaching, and for which it's difficult to perceive the supposed added value?

K-Enabler

However, the world has changed dramatically in the last years. Historical consumer habits changed, and opportunities are coming out of the blue. Companies who cannot keep up with the pace will be inevitably overrun, regardless of their positioning in the market, the strength of their brand and/or the quality of their goods/services.

No company will claim to not be customer focused, never ever. But in concrete, does every company have a structured, modelled, easy-to-present understanding of their single customers? Being able to produce a specific offer or a tailored advertising on expected future needs or wishes by past touchpoints?

Data Governance is about setting the foundations on data, from two sides: conceptual and physical. The first one is about having common definitions across the company, with no reference to any system. If a definition is clear, different actions and processes can be built around it. Back to customer orientation, questioning who or what is a customer? Is it the same as a consumer or a client? Most likely not and having this definition could define the perimeter of interest/areas to target. The same goes for articles, materials, products, goods etc. Furthermore, who has the authority/competence to define such words? This is also part of data governance, to identify the right roles in terms of "who can decide what". 

After, there will be additional steps to identify from the various systems where or how data is stored. If all is well documented, it can give powerful algorithms to analyze and maximize the potential of data. But if we skip the setting up of data governance, can we really expect that an algorithm will understand itself the right boundaries?

Start small, think big

Rolling out data governance in a complex, large organization is far away from being simple and immediate. We personally recommend having an approach with a small start but with big plans. It's quite probable that the organization will be reluctant in taking more tasks and/or allocating FTE. Therefore, showing an added value or providing a concrete tangible benefit, will pay back much more than a long, detailed presentation. 

A common mistake is to think that purchasing an expensive data governance tool at the beginning will make alone/have a major part in the rollout of data governance. This is a mistake just because these tools are extremely complex and will not magically solve the organizational issues, nor provide a universal model for the definitions. This kind of tools will be strong enablers only once there are clear concepts defined and a certain commitment on the part of the company. There is nothing bad in starting with home-made solutions and migrating into the professional ones only when a good degree of maturity has been reached. Having a very simple spreadsheet which is clear, accessible, robust and reliable will help much more in the beginning than a professional solution which nobody in the organization has an idea to what use for.

IT-Connection

As mentioned before, the transferring abstract definitions into real systems, tables and fields is a key aspect of data governance. For this it will be crucial to gain the support and help from the people who know every system very well (IT), for instance which are the key tables and how they are interconnected with each other. Often the same table will contain different conceptual objects and to make a proper split, technical expertise is a must. With this, the circle will be closed: having abstract definitions and where to find them in the system landscape.

Only at this stage can complex algorithms be programmed, and it will be clear in their specifications and in their design where to gather the relevant data in the enterprise landscape.

Conclusion

Please allow us the following comparison to summarize. Think of AI/Data science/Industry 4.0/IoT as having the physical fitness to compete in a professional sport. Therefore, data governance is the regular exercise needed to obtain and maintain it. But in the end, practicing regularly is not only a need, it can also be fun, don't you think?