Why Data Governance fails?
This article was originally published in the StrategicRISK magazine [Dec-2017]
https://www.strategicrisk-asiapacific.com
Analytics, AI and machine learning, visualisation and data science are increasingly ‘sexy’ areas to be working in today and all of them make use of large volumes of data. Reliable data is a prerequisite, but sometimes this is assumed or even ignored by the analytics communities.
Proper data management, along with data cleansing to ensure high levels of quality, is often simply too difficult. Coupled with this, risk managers in these fields face an uphill battle to comply with all the business processes and change management that a proper data governance framework would require.
I see this issue in a different, more familiar way: data governance is the Cinderella of the tech world, working hard in the background to keep data clean whilst being overshadowed by her AI, analytics, data science and visualisation ugly stepsisters.
Unlike fairytales, the lack of data governance across the different industries goes beyond what is accepted reality. Data governance is the elephant in the room that no one wants to talk about, much less do anything about.
To add an extra level of complexity, ‘data governance’ means different things to different people. Look at a large corporation as an example: the legal department may think of it in the context of privacy; the IT department considers it as cyber security; compliance might think of processes and workflows, while the majority of employees may not be able to differentiate data governance from data quality. The fact is, in theory, all parties may want good data governance, but in reality, most organisations lack the internal drive to achieve it.
The main reason people see data governance with different lenses is because they are trying to understand the problem in a bottom-up context instead of a top-down approach. Data governance should be part of the culture within an organisation. It is a core building block that should be instantiated from the time the enterprise architecture is drafted.
Data governance education should be an area in large corporates where employees are subject to compliance training. Many employees, especially in large corporations, are obliged to undergo compliance training on subjects like discrimination, money laundering and fraud. I believe data governance compliance training should also be on that list in order to embed it into the firms’ culture.
In my experience, it is the implementation phase where things tend to go wrong; where the data governance champions end up being labelled as boat-rocking troublemakers. Implementation requires big changes on views individuals have about running their business. Change is hard, but implementation of data governance requires genuine support and sponsorship from the top. Without executive support and advocacy, we could end up in another ‘bottom-up’ quest doomed to fail. On the other hand, you do not want morale to suffer as employees feel the burden of more top-down governance pressure.
Perhaps a middle ground is to cement the message at board level, while aiming to create awareness at the employee level. Once everyone has the motivation to solve the problem, the transition to true cultural change has a better chance of success.
Disclaimer: This document reflects the author’s personal views and not the views of the organisations where the author is or has been affiliated.
Director at EY | Finance and Actuarial Transformation
7 年Quote of the week for me: " data governance is the Cinderella of the tech world, working hard in the background to keep data clean whilst being overshadowed by her AI, analytics, data science and visualisation ugly stepsisters".
Data & AI Leader | Co-Founder @ EdgeRed
7 年Good points! ????