How To Define Data Governance And Data Catalog Use Cases
Although Data Governance is a fundamental part of an overall data management strategy, for a governance program to be successful, businesses need to concentrate on the benefits anticipated to result from implementing the program. The same goes for projects aiming to implement a first Data Catalog.
Eliminating data silos within an organization is one of the primary aims of Data Governance. However, while this will help, it is not a direct business benefit.
A further objective of Data Governance is to guarantee that data is utilized appropriately. This serves two purposes: first, it prevents introducing data errors into systems, and second, it prevents the possible exploitation of sensitive information and personal data concerning customers.
In addition, Data Governance can assist in striking a balance between the various mandates for gathering data and the multiple data collection procedures.
Improved data quality, reduced costs associated with data management, and increased access to necessary data for data scientists, other analysts, and business users are some of the benefits resulting from better Data Governance and the Data Catalog technology enabling it. Other benefits include increased accuracy in analytics and improved regulatory compliance.
In the end, Data Governance can improve company decision-making by providing executives with more accurate information. This will result in competitive advantages, more revenue, decreased risks, and improved profits in an ideal world.
For the time being, let's zero in on this particular aspect.
How does Data Governance create value anyway?
Managing the data used in operational systems and the data used in BI and analytics applications that data warehouses, data marts, and data lakes feed require effective Data Governance as a foundation. It is also a vital component of digital transformation initiatives, and it can assist in other corporate processes, such as risk management, business process management, and mergers and acquisitions.
Data Governance will likely find greater application as the range of uses for data continues to broaden and as new technologies are developed. For instance, work is currently being done to incorporate Data Governance processes into machine learning algorithms and other forms of artificial intelligence software. In addition, high-profile data breaches and laws such as GDPR and CCPA have made it necessary to incorporate privacy protections into Data Governance policies. This has made the process of governance a primary focus.
But like for any investment for which we measure return, if we want to do this for Data Governance, we have to understand precisely how, where, and when it will happen for our organization and measure it. In other words, how will we use Data Governance, and in what case, so we can solve a real problem we have? (see how we've weaved in "use" and "case" in that last phrase?)
Focus On The Applications Of Use Cases
The following diagram shows the typical use cases that combine the pillars of Data Governance: data stewardship, data quality, and master data.
Diagram 1 - List of typical Data Catalog use cases
These are general use cases, but many organizations start with "ingesting the metadata for everything." Unfortunately, that is very hard to do and creates an artificial pre-condition to delivering value that is untrue in many organizations and will bog down and sabotage the project.
We favor looking at the list above and peeking around the organization where we could apply the use case to help with a real problem. For instance, one use case is Streamline Trusted Business Reporting. We don't want to create generic functionality that can help with ANY reporting project but focus on a specific one where a lack of trust in a set of enterprise reports is blatant and in urgent need of a fix.
We first document it thoroughly as in the following diagram:
Diagram 2 - A detailed description of a use case.
Even in Diagram 2, the use case is not yet applied. Who cares about this? Which reports? For which reason? How do we save money, make more, or reduce risks? How would we measure this?
Then, we implement an MVP and iterate. And prioritizing use cases by their application to specific data domains, specific users, specific projects, or specific user communities (those that have pressure to solve the problem) will help quite a bit. Also, remember that every separate user community will bring its own flavor of requirements for the same need; focusing on just one for the MVP will save time to value.
Data Governance Minimum Viable Product & Iterate
We sometimes forget that agile is actually ideal in particular circumstances where the business has to learn new ways of doing things (data stewardship) with new technology, and IT also need to get familiar with this new beast they will have to support and evolve, as well as roll-out in the organization.
A clear focus on an MVP that will deliver value (win #1), not carry on for a year (win #2), allows people to get familiar with how Data Governance works (win #3), and if successful, will recruit vocal business supporters of this critical component of Data Culture (win #4), is the perfect way to reduce risk.
Now That You Know
Now that you know, it makes total sense. But you would be surprised how far ahead you are in reading this: many clients pay Incept advisory fees to help convince them that this is the right approach. And if you are already convinced, you can get started.
