This post is written by a human; all errors and omissions are the author’s responsibility.
Data is not a cliché. We’ve all heard from technical vendors that “Data is the new oil”, and “Data is our strategic Asset”. While those statements have a ring of truth, they lull us into a false sense of security. They emphasize only the positive aspects of creating, obtaining, storing and processing data. They omit the fact that data without governance isn’t an asset; it can be a liability waiting to be litigated. Technology leaders tell us that yes, data is important to the company. However, “importance” in most things implies value, but in the world of data, importance is neutral. It can be an asset if used properly, and if not governed can be a liability. Think about a massive oil spill – that is pretty “important,” but it is not positive. Data is not inherently good or bad. It is powerful. But without the guardrails of governance, it is merely noise at best and a toxic liability at worst.
The takeaway is that data needs governance. Martin Treder, in his book “The Chief Data Officer Management Handbook: Set Up and Run an Organization’s Data Supply Chain”, explains that the role of managing data should be treated like a logistics function, a supply chain. I’ve worked in manufacturing companies as an IT leader, and this is a great way to think about the discipline of Data Governance.
Treder explains that data is unique because it is intangible, easily replicated, and non-depletable. However, because of these traits, it is prone to entropy (disorder). Like some physical goods, data tends to degrade in quality and increase in volume automatically if not managed. Unlike physical goods, an increase in quantity does not mean an increase in value. In data, that increase can actually decrease the value. So Governance is not bureaucracy; it is the force that counteracts the natural entropy of data. That’s an essential function.
To expand a bit on Treder’s metaphor of the “Data Supply Chain”, we can compare it to a physical supply chain. If a car manufacturer receives rusty parts to work with, the assembly line stops, or else they will create a faulty product. In the same way, if a Marketing algorithm receives “rusty parts” (dirty customer data), the process might also stop, but worse, might continue with flawed output (spend and targeting decisions).
This is why data is important. Not just because it helps a company or function to be more effective and efficient, but because ungoverned data silently poisons the decision-making source. Data is the base material for the rest of the supply chain. As someone in charge of data at my company, my job isn’t just to provide data; it is to ensure the purity of the data through the supply chain to all our organizations and data users.
I like Treder’s metaphor not just as a mental framework, but also because he uses that framework to break down “Importance” into actions I can take on our business data. You might also find this breakdown interesting:
Treder’s framework starts with Relevance, meaning obtaining the right materials to begin with. Ungoverned data can become a swamp of irrelevant information. Governance ensures we are transporting the right goods (decisions made from data), not just any goods.
The next part of the framework is Trust. This is the quality check on the integrity of the data. For almost all data implementations this is important, but for predictive and relativistic uses like AI this is essential. Never has the saying “Garbage In, Garbage Out” been more true. Predictive and prescriptive analytics are often based on non-deterministic algorithms that include statistics. Even small perturbations in the base data can make for wildly inaccurate results. Treder calls the governance function the “ Quality Assurance (QA) department of the data factory”.
The final part of Treder’s framework is Security. This is the safety of data transport. Data should always be treated as a toxic liquid. In physical goods, this means special planning, transport containers, monitoring and careful movement. The in the data world, the Governance function covers these aspects.
But there’s a common complaint: “Governance slows us down.”
My friend Brent Ozar likes cars. Some of them are really fast. Not once have I heard him say “There were brakes on this car, so I took them off so it would go faster.” Brakes are there for safety and control. Without them, my friend wouldn’t be able to drive as quickly, because there would be no control over that speed. In fact, as he tells me, the faster the car, the bigger the brakes have to be to control it.
Data Governance can also serve as guide rails and brakes, and actually improve speed. Governance streamlines your data supply chain. By standardizing definitions and flows, I show my fellow C-suite that we speed up time-to-insight. For example, if Finance and Sales have a base definition and trusted data on what a “Qualified Customer” is (because of governance), they stop duplicating spreadsheets and start making cohesive decisions (increased efficiency and effectiveness).
My guiding principle for any new constraint or change is that “Process should serve Productivity”. Yes, I do put in more processes to implement Data Governance, but only where it serves the needs of the company, and keeps us safe and compliant.
Governance isn’t the red tape that stops the business; it’s the paved road that allows it to move fast.
If you would like to speak to us about getting Chief Data Officer-level advice and guidance, you can contact us here.
