In the Age of Big Data, many people might think that the practice of Data Governance is a thing of the past – nothing could be further from the truth. Data Governance has often been misunderstood or underappreciated and relatively few organizations have taken the time and made the investment to integrate it into their enterprise processes. So, there are actually several questions that need to be answered here:
- Does the de-normalization of data through exploitation of Big Data technologies discount the need for Data Governance?
- Why isn’t Data Governance more widespread if it indeed still has value and
- What is the value proposition behind Data Governance? (what are the 7 reasons why you need it)
We’ll tackle these questions one at a time.
1 - Does Big Data Require Governance?
The immediate expectation in response to this question might be – well no - as Governance seems to represent the formal and complex approach used for both RDBMS and OLAP data structures. But does this make sense per se? Classifying a data model as normalized, star schema or a de-normalized Big Table doesn’t necessarily impact the nature of the data attributes themselves. In other words, we still need to understand that data regardless of where or how it housed – we still need to know where it comes from, who owns it, where it goes, how it is transformed and so on. If we want the data to be valid, accurate and managed across a lifecycle, Governance is still needed. The technology itself does nothing to prevent us from experiencing a ‘Garbage in / Garbage out’ situation. The adoption of new technology doesn’t imply the need to discard common sense.
2 – Why isn’t Data Governance More Accepted?
This is a tougher question and in fact can’t easily be broken down into a single reason. Some of the most common reasons include:
- To govern data you first have to understand it holistically, and that initial assessment / analysis is a generally the hardest part – and is often why things don’t progress beyond that point (as many of these assessments simply never get completed)
- Often times, all Governance within an organization may be lacking because of the perception that some processes can’t be Agile and just hold things back or slow them down too much. While there is some truth to that, there is also truth in the lesson learned innumerable times that bypassing that Governance causes tremendous impacts later (to cost, efficiency and the ability to deliver and maintain capability).
- Because sometimes the tools get confused with the practice and while there are a number of great data governance tools available – sometimes they become an obstacle in themselves (e.g. some may be considered too expensive, others too complicated or perhaps there might be too many in the mix). The reality is that a lot of Data Governance can occur before or even sometimes without making that investment. It is the practice and not the software used to facilitate the practice that really matters.
3 – Why do Most Enterprises Need Data Governance? Here are 7 good reasons that tend to represent the more or less universal value proposition:
- Data Governance reduces enterprise complexity. At first, as I alluded to earlier, the impression here might be the opposite. But one only needs consider a highly typical data Use Case to see how Governance cuts right through complexity. Perhaps the number one integration issue I’ve seen faced over the past twenty years pretty much everywhere is the proliferation of similar or even the same data across multiple systems (this can include both multiple databases and reporting platforms). This quickly leads to all sorts of confusion and ultimately costs more to manage as long as it stays, well, confused. Governance tackles this type of problem at its core, by first designating authoritative systems and then more strictly controlling the use or reuse of such data. This can translate into business rules across the stack and often results in the elimination of both redundant data elements as well as duplicate systems.
- Data Governance Enhances Security – How one might ask, does it do that? Well, precisely through some of what I’ve already mentioned; including an assessment and classification of what data assets your enterprise has as well as determination of rules and architectural requirements for safeguarding both Data at Rest and Data in Motion. All of this starts with and becomes part of Data Governance. And if we think a little deeper about it, this is only logical when we consider that Data Assets are in fact the number one target of every major Cyber-Attack ever launched. To protect your enterprise, you must first know what’s in it and secondly you must have the ability to control the flow of that information.
- Data Governance is the best 1st Step for Integration – Almost every integration challenge is at its heart a data challenge. How we transport data, transform data and keep everything aligned is to a large part dependent on how well we understand that data. Messaging / Middleware / API Frameworks / EDI / SOA /EAI – you name it - it’s all about the data. Once integration is place, it must be governed – data interfaces (through messaging or other similar mechanisms) – is actually one of the most pragmatic initial places where Data Governance can be instituted.
- Data Governance Enables more Sophisticated Capabilities – such as MDM – Master Data Management is an example of a valuable enterprise capability that simply couldn’t exist if some level of Governance weren’t in place. To deploy MDM, an organization has to understand its core business entities and how they relate to attributes and be able to control them in a consistent manner. Every MDM solution I’ve ever seen either has Data Governance built in or relies on some other existing Data Governance process. MDM is not the only capability dependent on Governance though.
- Data Governance is Critical to Achieving an Effective Analytics Solution – The last thing any organization wants to be getting different answers to the same or similar questions. Data Governance not only helps to de-conflict issues at the data level – it can be used to de-conflict entire solutions. In other words, data governance helps drive consolidation of reporting and reporting architectures as well as the source systems underneath them.
- Data Governance can Impact the Bottom Line – Having Data Governance can make your enterprise more effective, not just from an IT perspective, but also the Business perspective as well. I’ve seen many organizations reduce duplicate systems and eliminate conflicting data and experience immediate results. The amount of benefit is dependent on how many systems can be consolidated or turned off and how improving data accuracy will impact whatever the business mission of the organization may be – but in almost every case – these types of benefits will be realized to some degree.
- Data Governance is often the Keystone upon which more Effective Enterprise Governance is Built – It is a great place to start if no Governance is in place or an even better place to expand if perhaps there are already some pockets of Governance already deployed. Since Data tends to be a cross-cutter, both organizationally and architecturally – it can become the foundation for a wider Governance framework.
In my experience, even in the organizations that didn’t fully implement Data Governance, the elements which were deployed provided obvious and immediate value. The current technology trends tend to point to a heightened need for Governance rather than the other way around, especially with the massive levels of adoption of Hybrid Cloud capability. I’ll talk about that in an upcoming post.
Copyright 2016, Stephen Lahanas
