Data Governance is entering a new stage. It’s no longer the obvious-to-a-few, confusing-to-many discipline that a few of us started evangelizing several years back. Now it’s a given that if your organization has any size and any technical/data complexity, you need to formalize governance as well as management of your information.
So now thousands of us are working within our organizations to establish governance. We’re doing the right things, although maybe not at the pace we’d like. We’re agreeing on common activities and responsibilities. Frameworks are being used to help people communicate and stay in sync. Things are good.
So why don’t I feel great?
It’s because of the world’s economic crisis and what we all might have done while it was all brewing. Yes, I understand there is much that none of us could have changed. We’re not in a position to stop fraud and crazy Ponzi schemes. We’re not in a position to stop leaders from taking crazy risks. We’re not even in a position to stop corporate leaders from taking calculated risks that might just turn out bad.
But we are – or should be – in a position to help leaders and other users of information to understand how well they should trust the information they’re using to make critical decisions.
The truth is that many of us know – and have known for a long time – about problems with the data we work with. We know when the data is polluted, or incomplete, or just wrong. We know when data quality hasn’t been tested (meaning that all we have are assumptions about how good it is). We know when users of information have all validated data definitions – and when they haven’t and instead are assuming they know what a particular report field means. We know where streams of information from various sources converge into a big repository, and we know when those data streams have been carefully sorted into trusted sets of information and when, instead, we really aren’t sure what we have to work with.
We know so much. And yet we haven’t made sure that the people making world-changing decisions understand whether they’re basing decisions on trusted information or on assumptions.
And that’s just not right.
Yes, I understand corporate politics. (Believe me, I have the scars.) I understand how hard it is to push information up the executive ladder. I understand what happens to people who break protocol.
But many of us do have access to executive users of information. (Or we have the ear of those who have access.) I propose that 2009 be the year of Proactive Data Governance. Let’s use our voices to make sure that leaders understand key facts about the information they rely on when they’re making critical decisions.
I suggest that we keep our messages simple. Let’s all tell our top-of-the-foodchain data users the same thing until they get it.
Let’s tell them that they need to know the answers to 3 questions concerning their data, and that we as Data Governance leaders are ready to help them get answers. Let’s tell them that they should know whether the information they’re relying on
1) Comes from trusted and controlled sources
(or the alternatives: we know that the source is suspect, or that nobody knows because lineage hasn’t been documented and analyzed)
2) Whether it means what EVERYONE thinks it means
(or the alternatives: we know there’s confusion, or that the data definition validation work just hasn’t been done)
3) Whether the data quality is “good enough”
(or the alternatives: we know it’s not, or that it hasn’t been tested, or that rules haven’t been established to say what “good enough” means).
Maybe we can’t do anything about the situation that got us all here. But we can be part of the solution. We can make sure that decision-makers know the right questions to ask, and we can help them get answers. We can help them understand whether they are basing their crisis-recovery decisions on trusted data or merely assumptions.
Now let me stop a few of you who are thinking that my three questions are too simple – you could probably add another half dozen to the list. And you could add a bunch of words to the simple questions:
1- Trusted source?
2- Agreed-upon definition?
3- “Good enough” quality?
Yes, you could cloak these with techno-speak, use words like “profiling” and “lineage” and “metadata repositories” and “taxonomies.” And when you’re talking to techies, you might want to do that.
But if you make this too complex, you won’t get executive attention.
So here’s my suggestion:
Let’s all tell our leaders the same simple message, and let’s use simple, non-technical language to get their attention, and then (if they respond with interest) let’s help them make the decision to proactively do/fund what it takes to get the quality of information they need.
Let’s all send the same message:
If you have answers to three questions about the information you need to make critical decisions
(1-Trusted source? 2- Agreed-upon definition? 3 – “Good enough” quality?),
then you’ll know whether you’re basing your decision on solid data,
and you can be more informed about the risks you’re taking.
Right now, around the globe, leaders of corporations, governments, not-for-profits, educational institutes, and other organizations are trying to decide how to react to the current economic crisis. They’re trying to decide how to be proactive, to avoid making future mistakes. They’re pulling report after report after report, looking at their data while they ponder critical decisions.
What do you know about that data? Even more important, what should THEY know about that data? What questions should they ask?
Let’s do what we can, troops…
