Listen to this article:
When it comes to data, there are two types of companies: those that are aware of their data quality issues and those that aren’t looking. Does your CEO know how bad your data quality is? Hopefully not, because that means that you haven’t had a negative event large enough or impactful enough to warrant attention at that level. Yet.
AI is exposing data quality issues, and as a result, data is starting to get more executive attention.
Don’t mistake executive attention for executive interest.
Their interest is in the business drivers. It’s in the increased revenue or customer satisfaction. It’s in the claim that the company is successfully deploying AI and to a lesser degree, analytics. (Although, if analytics was really a data quality driver, we would have seen greater engagement decades ago.) In their most recent annual reports or 10-K filings, every one of the Fortune 50 companies acknowledge the role of AI in their operations through applications, investments, or as part of strategic initiatives. Every one of those Fortune 50 companies also highlight the use of data analytics or data-driven decision making.
Only a couple mention data quality. Maybe it’s implied in the others. Maybe it’s not.
As AI exploded into the mainstream, the widespread assumption was that the data used to power AI and analytics was high quality. Or at least sufficient quality. And why wouldn’t it be? Product is moving. Money is flowing. The job is getting done. The occasional red alert is handled by the Ops Control folks that monitor the applications.
Do your Ops Control folks monitor data in the same way they monitor applications?
If not, then somebody is monitoring data quality, aren’t they? After all, it would be negligent to say that we’re not going to care whether our data is right or not, wouldn’t it? The CEO assumes that the CIO is doing it. Or maybe the CDO. They assume that the DBAs or enterprise analytics team or maybe the application development teams or their business partners are doing it. Experience suggests probably not. When I played volleyball we’d call this a “campfire ball.” Everyone stands around and watches the ball fall in between them, waiting for somebody else to handle it.
When it comes to data, everything is OK until it isn’t. And then the CEO gets involved. That’s not the way we want that to happen. Your CEO is going to care about data either intentionally or accidentally. You would prefer intentionally. If your company’s data is bad, then you are likely to have problems implementing AI. Then the CEO is going to care. I think we’ve established that most CEOs neither understand nor care about data. That’s a problem and we need to deal with it, but we want that caring to be on our terms.
The phrase “a mile wide and an inch deep” has been used for more than a century to describe both fear and superficiality. But it also a metaphor for the required breadth of a CEO’s understanding and awareness. When it comes to data, my Rock Bottom Data Feed podcast partner, John Ladley, has often said that a CEO’s understanding is a quarter of an inch deep if that. I once heard an executive say that talking about data was like looking into the sun: he could only do it for a very short time before it started to hurt.
This is a very real but unnecessary mental block.
Executives deal with data all the time. I’ve seen many executives at many companies poring over pages filled with columns of numbers. They’re just called financial statements, sales reports, or workforce plans. But it’s just data. They assume that it is correct, and that there are processes in place to ensure correctness. And for maybe a couple domains that’s true. Like Finance.
The production, management, and reporting of financial data is governed by GAAP (Generally Accepted Accounting Principles) and other regulatory requirements. Every department has one or more employees that manage budgets and financial reporting. Infrastructure, metrics, and processes have evolved around accumulating, balancing, and reporting financial data. Sounds like an awful lot of information management, data governance, and, yes, data quality applied to that specific domain. That’s one. What about the rest? Is it any less important to get that data right?
If the data is really important to get right, then they’ll do it.
Information security is another example. The role emerged in the 1990s as cyber threats became more sophisticated through the emergence of networked systems and the internet. Those risks required a dedicated executive role, the Chief Information Security Officer, to oversee strategy, governance, and risk management. More information management and data governance.
Perhaps those assumptions about the correctness of financial data and the protection through Information Security that are taken for granted in the executive suite are extended to other domains. In most cases, though, those same controls don’t exist. And nobody bothers to mention it. Or notice it. Or do anything about it. For a long time, leadership was just happy to get anything resembling data. Just happy to get the reports. I had one executive once tell me that as long as the report was “directionally correct” that was good enough.
CEOs care about direction, alignment, and risk. Not so much day-to-day operations. And there’s not much more day-to-day operations than data. But with unmanaged data comes incredible risk, and that IS a CEO’s concern. The CEO, and probably the CIO, need to be educated, but not with fifty slide PowerPoint presentations and a DMBOK wheel poster. Heavens, no.
Instead, executives need to be shown the direct connection between data and business drivers, and to be made aware of the risks. They need to know the questions to ask their subordinates. And they need to know how to evaluate the answers and how to follow up if necessary. They need to have confidence that quality is being measured, and actions are being taken when quality falls short. You want them to understand enough to be supportive. To be aware of what you’re doing and the importance of it being done. CEOs need to get ahead of this. Yes, data needs to be added to the CEOs plate.
It’ll take exercising muscles that haven’t been used before. Getting the data correct takes effort. Committing to and providing the ongoing support required for getting the data correct also takes effort. We can’t replicate the entire financial data management edifice at once, but we can start building it piece-by-piece. It starts with leveraging the AI initiatives to ensure that the data is well understood and data quality has been established.
Even then, though, CEOs will still not care about the data. They’ll care about what happens with it. They’ll understand the consequences and potential impact of quality metrics moving in the wrong direction.
