There is a direct relationship between the value your organization gets from its data, the trust your organization has in its data and how formally that data is being governed. This is not new news. In fact, this has always been the case. Value comes from the ability to use the data to make good decisions, to predict behavior and to answer questions that lead to improvements in efficiency and effectiveness. Trust comes from understanding the data, knowing the meaning of the data, where the data came from and knowing how to use the data to improve operations. This is not new.
All we need to do is remember the formal effort we make to assure that the data in our decision support databases and data warehouses (among other places) are of the highest quality … And to remember that the efforts we make to assure quality data in the information systems we are building and the software packages that we are investing in … And to remember the efforts we make to deliver quality data through resources focused on master data management initiatives, improving analytical capabilities and corralling big data resources. It’s all in the data.
Organizations know that the value of these resources will come from the data that is managed through these efforts. Yet, organizations still focus on these efforts independently and with little regard for how this data can be leveraged together to improve the success of our organizations. The same data, defined and produced differently, depending on the needs of sponsor rather than the strategic needs of the organization, results in siloed data resources that are difficult to integrate, share and leverage across the organization.
This is a reason why artificial intelligence and data-centricity are still a dream, or at least significantly far in the future, for many organizations. Organizations continue to spend large amounts of money on their data without formally governing the data as a valued cross-enterprise asset. Governance builds consistency. Governance builds interoperability. Governance build value and trust in the data. And we recognize that the data will not govern itself. It’s all in the data.
I have been known to state repeatedly that “the data will not govern itself.” I hope that the statement is not getting old. It should be stated more often to drill the point home that formal data governance requires that we focus of getting people across the organization to take the “right” actions at the “right” time with the “right” data to get the “right” value out of the data as often as possible.
Data governance is “the execution and enforcement of authority over the definition, production and use of data and data-related assets.” Execution and enforcement of authority sums it up quite well. This definition summarizes the fact that people’s actions are the focal point of data governance. You can even call it “people governance” instead of data governance because it is the behavior of people associated with the definition, production and use of the data that will improve the quality of the data. It is the behavior of people that will lead to improving the value and confidence in the data because we all know that the data, if left to its own accord, will not govern itself.
One way to improve the value people get, and trust people have, in the data is to improve their knowledge about the data that is the focus of enterprise investments. This includes knowledge of the data resources that are available to them and knowledge about the specific data that resides in these data resources. The term for the information used to improve people’s knowledge of the data is called metadata. Metadata is basically “data about data”.
Metadata exists in data catalogs, business glossaries, data dictionaries and repositories, made up of random databases, spreadsheets and documents that describe the data. Data documentation is an afterthought in many system development or acquisition efforts. Metadata is scattered across the organization, siloed from other metadata, and it is not formally managed, kept up to date, or easily accessed by the people. People do not have the metadata they need to improve their efficiency and effectiveness in their job functions. Metadata is often incomplete, inconsistent, inappropriate, or unavailable. Metadata is not governed — meaning that people are not formally being held accountable for the data documentation or metadata. This causes a problem. The reason being, just like the data, the metadata will not govern itself.
Do we see a trend here? Is there something consistent in the message contained in this column? I would hope so. The bottom line is that data does not naturally or automatically increase in value or become more trusted without a resolute effort. That resolute effort must be commanded by the highest level of the organization and orchestrated at the strategic and tactical levels of the organization, to demonstrate value and to gain the trust of the people at the operational level. The effort requires that people be given the time, skills and tools that are needed to deliver value and trust in the data. Metadata is a necessary component of that. In general, organizations must get their people heavily involved and invested in the need for improved data. It will not magically happen. The data (and metadata) will not govern itself. It’s all in the data