The challenges of data governance are well documented; from fostering collaboration among different lines of business to gaining buy-in from upper management, data governance is difficult at best. Yet it’s crucial to help organizations manage important data assets if they are going to be leveraged for strategic business decisions, especially in big data environments. Data governance is critical to ensuring the health of a data ecosystem as information flows and functions across an organization’s data supply chain, but even in 2018 organizations have a host of reasons for denying the efficacy of efforts to govern data, or delaying attempts to implement initiatives.
In a recent article from the Forbes Technology Council, titled “Data Governance Is Dead — Meetings And Spreadsheets Killed It” the author says, “despite the real value it provides, organizations shy away from implementing governance because of past, misguided experiences.” The article goes on to discuss a scenario in which an organization is implementing a new data lake, the project group hesitates to execute data governance because of organizational misconceptions around what data governance means, and the lake becomes polluted with “ill-defined, poor quality data.” When they finally do apply data governance concepts through a poorly conceived plan consisting of spreadsheets, wikis and monthly meetings, nobody takes their responsibilities seriously, and the “governance” result is frustratingly complex and untrustworthy. The data further degrades, turning the data lake into a mud pit of untrustworthy, inaccessible data, ultimately wasting time and money.
In the example, the project’s data governance advocate eventually takes a different approach that meets with approval and success—wrapping a data catalog around the data lake to “help identify what data is in the lake, what it means, who owns it, who is using it and maybe even where it came from.” The data governance lesson being, ostensibly, that taking a back-end approach and dipping a toe in data governance to gain incremental support to prove value may be the better approach than to wield the “data governance” message like an organizational hammer.
As much as we all might like to wield a hammer from time to time in our professional lives, this is sound advice for a number of other reasons. Anyone who has tried to gather budget, enterprise or executive buy-in, or resource backing for data governance knows how difficult it can be, and that there are many preconceived notions about the costs and challenges of data governance—particularly if early attempts were ill-conceived or ill-supported. Taking an incremental approach towards implementation is an effective way to build support and bolster advocacy across the enterprise, but that advice must come with a caveat: it needs to be part of a comprehensive data governance strategy.
The takeaway, then, is that spreadsheets and wikis don’t represent the death of data governance; they simply characterize an early iteration of its rapid evolution that’s no longer an effective strategy. And as the scenario from the article illustrates, an effective strategy must be firmly rooted in a foundation of data quality.
The most important first step when considering how to implement—or how to redefine—data governance in the age of big data is to consider what you want to accomplish in terms of short-term goals, but also build a framework that is scalable and workable for long-term strategies. One problem with the scenario outlined in the article is that it was being considered and deployed within the confines of a single project, albeit a large one. Within that scenario, the short-sighted attempts at data governance that failed were doomed to do so, because they relied on manual, labor-intensive practices that didn’t benefit the enterprise as a whole. Outside of that big data environment, those governance initiatives were neither relevant nor integrated. No one beyond the scope of the project had any involvement, and their vested interest—the ability to use that data as an enterprise asset—wasn’t dependent upon any actions or accountability outside of the project group. Even the “successful” data governance initiative, the implementation of a data catalog, was narrow in scope and wasn’t presented as a good first step in a larger strategy to bring enterprise-wide, integrated governance to the organization.
When developing a data governance program, you need to first develop a strategy based on needs and priorities across the wider organization. Whether an individual or a committee takes the lead, it is important to identify not just isolated projects, but the broad vision and goals for data governance within the organization. This way, you are more likely to build support that can encourage both accountability and collaboration down the road—two key ingredients for a successful data governance framework.
In the article’s scenario, a lack of data governance led to lack of data quality and a lack of trust in insights from data, which are of fundamental importance to organizations today. In a 2017 CIO WaterCooler survey of senior IT leadership, when asked why they were implementing data governance initiatives, 63% of respondents reported that it was to ensure good quality data, and 89% selected “better decisions, based on better quality data” as the main benefit they hoped to achieve.
With data quality as a fundamental goal of data governance, embedded data quality is a critical ingredient in a solid governance framework. Ongoing data quality monitoring not only ensures the accuracy, consistency and reliability of data, but it also combats one of the main points of resistance to data governance—the misconception that data governance will inevitably be a grueling, time-consuming endeavor. Analytics-enabled data governance is a long way from spreadsheets and wikis, and enables automation of data quality initiatives. Look for a data governance solution that provides a platform with both quality and governance, and doesn’t cobble together “integration” through partnerships necessitated by product gaps.
A successful data governance program requires a comprehensive solution suite to promote a community approach that puts IT and various lines of business on the same page, across all projects and lines of business. Organizations may very well find that a phased approach to an enterprise-wide solution is required to build buy-in, belief, and the requisite collaboration, but it needs to be approached as an organizational initiative and not a department project.
When fully implemented, the solution suite should deliver a complete view into an organization’s data landscape, from the data available, its owner/steward, lineage and usage, to its associated definitions, synonyms and business attributes. It should allow all users to easily define, track, and manage all aspects of their data assets, enabling collaboration, knowledge-sharing, and user empowerment through clarity across the entire organization.
With a complete solution suite, organizations can ditch the spreadsheets and tackle data governance once and for all. Data users will rejoice when they have the trustworthy data they need to do their job right at their fingertips.
If you would like to learn more about data governance, download the eBook below.
For a deeper dive into this topic, visit our resource center. Here you will find a broad selection of content that represents the compiled wisdom, experience, and advice of our seasoned data experts and thought leaders.