Data governance is the unsung hero of maximizing the ROI of data. Afterall, without established processes that ensure data quality and security across the organization in place, your data science projects will have zero credibility. Zero.
To share his perspective on building a data governance framework that drives business value, we invited Mark Wilson, Head of Data Governance, Handelsbanken to present at the upcoming AI for BFSI virtual conference this March 1-2, 2022. To give you a sneak peek of what you can expect to learn, we recently sat down with Mark to discuss some of the common mistakes data governance leaders make.
If you like what you read here, REGISTER NOW to attend his session on “Data Governance From A Business Perspective” taking place at 9AM SGT on February 2, 2022.
Seth Adler, Editor-In-Chief, ADA:
I’ve heard you're not thrilled with the phrase "being a data-driven company." Why not? Could you unpack that for us here?
Mark Wilson:
Well, we've got shareholders. I'm guessing they probably want us to run a successful organization and build shareholder return. Whether I'm an insurance company, a bank, large retail outlet, internet provider, if I was to start selling that to my shareholders, "Relax about our internet business, because we're going to become a data-driven company." Well, but what about the day job? Where's the money coming from, because isn't that what a data center does?
I think we need to be careful. I think it's too easy to find the cliched phrase that suits us. They're great at symposiums. The big companies that sell us these phrases like big data and blah-de-blah. But we need to be mindful when we reach out into a business world, they're not going to get some of these phases and whatever we say to them has got to resonate.
Fluffy one-liners are great, but they imply the wrong thing. Yes, we want data to be at the center of our business decisions to help drive our growth, but there's lots of cliches I think people find too easy to throw in that really don't hold up.
Seth Adler
What about making sure that we've got that data democratization going on, that everybody knows what their responsibilities are around the data. Furthermore, if that's the construct, how does somebody like you get out in front of any issues that would stop us from collectively doing our jobs?
Mark Wilson:
When it comes to how to handle remediation activities, how to structure an organization to, I think every organization is different.
In some industries we find very heavy regulation that's growing and growing. Others, less so. We do not want to reach out to the customer on a regular basis getting or asking for more data. It looks we don't know what we're doing if we've got different parts of the organization going out and saying, "Oh, we need to confirm your address." Then another part of your organization is going out a month later saying, "Oh, we want to confirm your email."
The internal structure of how we handle remediation is really important.
You have a customer services team who are there to work with the customer. You start giving them loads of stuff to clean, so how do you centralize that role so you can actually get the right people fixing the right data?
There are things I know. US states, it's a finite list. I know what they're called. I'm sure there's 200 different ways of spelling them all. We have a lot of problems in the UK. Henley-on-Thames, Henley upon Thames, Henley in Thames. Now, I've seen some instances where we've gone in to cleanse a bit of customer data while they've gone into a website and changed it back to what they know it to be.
The central team's gone in the next day and said, "Oh, that change didn't take, so I'll change it again." Now the customer knows, "I'm not living at home now. I'm staying away for three months while I'm having my house renovated." Centrally in a company we don't know that, so they keep trying to put in the official address.
I think the remediation activity is a core thing to really think about before you start building flashy dashboards or get too big, because once you start telling everybody, "Here's our data quality problems. Look how bad they are," if you can't fix them, you'll lose all credibility. You're not doing anything.