There is no way to eliminate all the pain in data governance. It is hard work. But with a good plan, you can minimize the pain and maximize the gain.
Part of a blog series.
Data governance is a stupid name. Data and information are not the same thing. The name “data governance” implies that data (albeit governed data) is the ultimate goal. Data is okay, but information is better.
Today we find ourselves drowning in data. Standing on the corner of a busy intersection, we can look around and see magazines in the stand telling us how to “lose 15 lbs overnight.” A sign in the grocery store across the street reads “Peaches are Half-Priced.” But right now, the information I need to know is, “can I cross Main Street safely?” Fortunately, there is an accurate, clear and convenient information system in place: the cross-walk signal. Is the red hand-lit or the walking green man? Data governance, or as it should be called "Information Governance" is the process of getting the right information to the right people at the right time. [cta bg="dark-blue" button="View Webinar" link="https://go.centricconsulting.com/NAT-FY22-09-WBNR-DataGovernanceHowtoSucceedandWheretoStart_01LandingPage-OnDemand.html"] Join our data and analytics experts to learn how taking an agile, minimally invasive approach to data governance can lead to higher adoption, greater business value, and increased accountability in your organization.[/cta]The Value of Data Information
Information has value. Some companies like Facebook are founded on data as a commodity. But even if your company sells widgets, your data has value. When properly collected and used, data can help a company lower costs, increase revenue, develop a competitive advantage, and reduce risks. The Declaration of Independence states, “We hold these truths to be self-evident, that all men are created equal…” But that truth does not extend to data. Not all data is created equal. My love of peaches is why I cross the street, but the price isn’t going to make my trip across Main Street safer. Raw data, without context, is next to useless. Three properties of the information that have been properly transformed from raw data:- Must be Correct – accurate and complete
- Must be Clear – consistent (correlated, connected, and conformed) and comprehensible
- Must be Convenient
“Data Governance” Isn’t a Four-letter Word
Data governance has been around for many years. Many companies, excited about the promise, started their data governance program. Many never finished. The work was difficult, and the return proved lackluster. Before long “Data Governance” became a dirty word. Too much effort without any benefit led companies to abandon their projects. The goal wasn’t wrong, the approach was wrong, and the pain was real. Where did we go wrong? Several places: we started wrong, we attempted too much, and we took some ill-advised short-cuts.1. The Business Must Lead
Data governance is a business exercise, not a technical one. Many companies have a wall between the business group and the technology group that neither side understands. Too many times reporting projects are thrown over the wall toward the technology group. Many are last-minute and high-priority. Too many times the technology group throws their collective hands into the air because they feel they are asked to do the impossible. This is bad, but it can get worse. Imagine that the technology group simply pushes past their own ignorance and delivers a report that is just wrong? The business thought the request was simple to understand, but they lacked the required technical expertise. They forgot a simple truth. Just like they were not hired for their technical knowledge, the technology group was not hired for their business knowledge. Since the overall usefulness of the information matters to the business, the business must take ownership. The mechanism for taking that responsibility is data governance.2. Don’t Try to Do It All at Once
Data Governance is not a once-and-done exercise. The Big Bang apparently worked for the universe but is a poor choice for Data Governance. Too many companies have tried to develop governance for everything in their data sphere. They get stuck in analysis, with every element uncovered leading to more. Without a defining scope, they get lost in an endless cycle of discovery. You may find mountains of data, but can you safely cross the street? Let’s consider an agile process. Start with the core components of the core systems. Deliver some value, then add more. Continue adding more, over, and over, and over again. How about a more proactive approach? Start looking at what you want to get out of your analytics. What are the critical success factors for your business? Can you identify 10 Key Performance Indicators (KPIs)? Now round that out with 50 or so additional result or performance indicators. What elements are required to perform these measurements? You’ve just scoped your data governance program for the first two or three iterations.3. Know Which Shortcuts to Avoid
Data governance is not a care-free path. There are risks from start to end. It reminds me of a game that I played when Atari first came out. Pitfall. The character, Pitfall Harry, was challenged with finding treasure while avoiding various obstacles. Some obstacles would simply cost points for each second you were in contact, while others would take a life. The same can be said of the obstacles within a data governance program. Finding a new, better formula for key performance indicators will cause a momentary hiccup. What reports have been delivered that may require an update? In heavily regulated industries, like finance or insurance, this may require new updates to previously delivered regulatory reporting. While other industries will simply move forward with the updated values from this day forward. Some pitfalls are more severe. Imagine if your data warehouse design is based on a faulty understanding of the company’s business. There is no way to modify a faulty design. This pitfall requires returning to the earlier stage of business-centric data design. Three properties of a successful information governance program:- Business-centered conceptual data architecture. Not a data exercise, a business exercise.
- Agile, not big-bang delivery. Rinse and repeat.
- Deliberate, diligent technical (logical and physical) data design. No shortcuts. No quick fixes.
