Chief Content Officer,
Techstrong Group


In this Digital CxO Leadership Insights series video, Mike Vizard talks to Brett Hanson, chief growth officer for Semarchy, about why organizations are now focusing more on master data management (MDM) in the age of artificial intelligence (AI).


Mike Vizard: Hello and welcome to the latest edition of the Digital CxO Leadership Insight series. I’m your host, Mike Vizard. Today we’re talking with Brett Hansen, who is the chief growth officer for Semarchy. And we’re talking about master data management and how to get the C-suite on board for this because, well, we’ve been at this for a while, but it turns out this AI thing is coming along and it’s becoming a much more pressing issue. Brett, welcome to show.

Brett Hansen: Thank you, Mike. Thanks for having me.

Mike Vizard: What is your sense of… Where are we with MDM in terms of our overall level of maturity? I mean, I think the first time I talked about this was three decades ago. So why hasn’t the problem been solved yet?

Brett Hansen: Well, obviously a lot is happening in the data sphere. And so in the three decades, there’s a wee bit more data out there. It’s coming from work and sources. The importance has changed. The number of organizations have changed. The expectations of being able to access by different users. So it’s a different world, obviously, to three decades ago.
Master data management itself is going through a fairly significant evolution from those initial heavy-handed siloed environments of the IBMs and TIBCOs and the Oracles to a cloud-based, nimble, data fabric approach that you’re seeing today. And so it still has a lot of opportunity for adoption, but the good news is that we’re seeing more and more companies recognize how important MDM is to their data strategy.

Mike Vizard: To that end, I feel like a lot of organizations are not very good at managing data, to be honest. A few of them would get what I would call a good housekeeping seal of approval for the way that they have been managing their data over the years. There’s a lot of conflicts between the data. The quality is suspect. What is the root cause of that problem?

Brett Hansen: I wouldn’t say there’s a single root cause. I think there’s a lot of different elements that create this challenge. And I talked about a few of them a few seconds ago. You have more data. You’re getting more data collected. There’s also this expectation of more ownership of data by functional groups. So if you think of the evolution of data in large organizations, it was at one point kind of highly centralized. And now, as the expectation of, “Hey, I’m a line of business executive and X, Y, Z function, and I want to be able to utilize my data to go and make better decisions around my product, or my ingredients, or my marketing to customers,” that push is realigning how we handle data.
So for us, most of the MDM projects we work on don’t originate with a chief data officer. They’re not originating with that central organization. They’re actually originating with a line of business organization who’s saying, “I need to accomplish X.” And X isn’t data. It’s a business problem. “I need to get better insights on my customers. I need to be able to handle my ingredients more effectively. I need to be able to address compliance challenges.” That’s where this is starting from, and it’s a whole new way of thinking about, “How do I utilize my data effectively?”

Mike Vizard: Do you think there’s a certain amount of distrust of data in IT these days? Because the C-level execs know full well how that data was created. And so when IT shows up with a report, they’re a little circumspect because they know that a lot of the data used to create the report is somewhat flawed. Has this become a vicious cycle that we need to put an end to?

Brett Hansen: Well, there’s definitely a data quality issue. Right? And I think that’s why master data management is so important, which is, how do we get to the root truth? Right? What is master data management all about? Creating that golden record. Right? That record that is indisputable.
And so when we work with organizations on their business challenges, we start with, “Okay, what is it you’re trying to absolutely understand?” I need to know who Mike Vizard is. I need to understand where he lives. I need to understand his core product behavior elements. Right? How do we bring together different data sources, improve its quality so that we can create that golden record, that true source of truth, that then can lead to better insights?
You mentioned AI. AI starts with good data. And I know we kind of want to jump over that hop and go right to, “Oh, AI solving all problems.” Well, it can’t do that until you start with good data to go train that AI model, otherwise you’re likely to have a very bad outcome with your AI efforts.

Mike Vizard: There’s a lot of talk about how to operationalize AI. Is that forcing this conversation about data management principles? Because, to your point, if the AI model hallucinates, it has a lot to do with the fact that the data is flawed.

Brett Hansen: We are finding more companies are having that conversation which is, “We knew we had opportunities to improve the quality of our data, but there’s these broader macro objectives in the company around adopting AI and being more effective utilizing AI, and we can’t take that step.” We can. We can’t take that step intelligently until we are confident that the data that we have has a higher quality.
So there’s still a myriad and reasons why people are adopting MDM. And again, as I said, most of it is sitting in the line of business function. I would say AI is a additional motivating factor. It’s not the cause, but it’s sort of saying, “All right, well, we want to get here. AI is a really compelling proof point. All right, I got to bite the bullet and go down this MDM path because I have to have high quality data, or ultimately we’re going to have a hallucinating AI.”

Mike Vizard: Is MDM getting easier? I think part of the issue has always been, historically, it was a heavy lift and I had some big giant thing that I had to build and maintain in an on-premise environment. And now we move to the cloud here, as you noted. But is this getting to the point where mere mortals can actually do this?

Brett Hansen: Yes, it is getting easier. I would not say it is easy. And anyone who tells you it’s easy is probably hiding some facts. It’s a still challenging project because it is bringing together multiple different data sources from multiple different locations. What is important though is, A, having a AI solution that can be built for your needs. So if I look at the legacy AI solutions and then even the next sort of phase of AI, it was highly templatized, right? Hey, I have a template for you to use to do Customer 360. The challenge is every company’s a snowflake. Every project’s a snowflake.
So if you start with the template, you’re going to find it doesn’t really fit your needs. Then you got to break the template and now we’re doing custom coding and it’s a bit of a disaster, right? And this goes to that hardness that you talk about and projects that would be estimated nine months ago, 18, 24 months.
So the next evolution of MDM solutions are no-code data application builders. So you can define what is it that my business needs? I need to achieve these objectives. And then through automation and no-code, the MDM solution can provide exactly what you need. And then as you iterate, which we know you will, it’s just again, inputting business requirements and then the code generation creating an MDM for you. So that’s a really important evolution from, “Hey, I got a template which I want you to force you down into,” which is going to break to a no-code automation approach, which is far more effective at allowing you to achieve your goals.

Mike Vizard: Ultimately, are we going to need to apply AI to the way we manage data so that we can build more AI models as there?

Brett Hansen: Yes, absolutely. So I mean, we already have customers who are using AI as part of MDM. Right? So it is not… It’s a pragmatic approach because obviously you don’t want to be black boxing and trusting AI to help you create your golden record and then find out that your golden record is not golden but burnished and damaged. But there are opportunities to streamline that MDM manual process through utilizing AI to do things like matching.
And what we have found working closely with customers is it’s not eliminating those data stewards and eliminating that manual process. It’s streamlining it, making recommendations, and then allowing a line of business or a data steward owner to say, “Yes, this is a good recommendation,” and then train that AI slowly to be more thoughtful. It’s using AI to help the business owners be able to be more effective at helping their MDM be successful by allowing them to ask query questions in natural language and being able to provide responses.
So there is certainly a big role for AI to play in MDM. What I would caution customers about is, again, just like everything else, be pragmatic, start small, know what you’re doing, and then you can start to expand in a gradual manner. Don’t just jump into the black box and, “Hey, X, Y, Z vendor has said they have an AI model. I’ll trust that.” That’s scary to me until you really have done the training to make sure that it meets your requirements.

Mike Vizard: So what do you see organizations doing that makes you shake your head and go, “Folks, we’re going to be smarter than this”? It seems like we’re struggling in some shape, manner or form. So what are you seeing?

Brett Hansen: The three things I see most frequently when organizations come to us and say, “It broke. We tried MDM, it’s not working, what’s happening,” the number one thing is there wasn’t good alignment between the line of business and the data professionals. So 90% of the organization or the individuals who we engage with are data professionals. They’re data architects, data steward, data analysts, data something. If they don’t have a strong working relationship with their line of business and have a well-defined set of business requirements, it likely leads to challenges. If you’re just doing MDM for the sake of the data, it’s not going to have the outcome that you want. You need to be thinking about, again, what is the business outcome I’m trying to solve? So that’s challenge number one.
Challenge number two is line of business executives often want to go and do everything right now. And so they set these unrealistic objectives that then creates an MDM project that is not going to be successful because by the time they actually build it, their requirements have changed, and now we have to start all over again. This goes back to the approach I was talking about, instead of using a template, being able to utilize a automation engine. So you build something, you get it out the door, you get it in the hands of the line of business, they can test it, they can feel it, they can taste it, and then we iterate again and we do more and more and more.
A number of our most successful companies who have adopted MDM, they started really small. Chipotle started off with just doing MDM for location management, and now they do ingredients and compliance and menus and all kinds of things. But they started small and they were successful, and then they went to the next one.
The third thing we see is they don’t have a really well-thought-out execution plan. It’s like, we buy the vendor, we’re going to deploy it, and we go. We actually have rapid delivery blueprints that we sit down with all of our customers on and say, “Okay, let’s chart out your journey.” Because as you called out earlier, MDM is not easy. It’s gotten easier, but it’s not easier. And unless you have that well articulated plan, it’s more likely than not that you’re going to find issues along the way and that’s going to slow you down and be less successful. So make sure you have a good line of business relationship. Make sure you have a reasonable plan. Start small, grow quickly. And then make sure you have a plan for implementation and you’ve thought through all those nuances that come with trying to bring together data from different sources.

Mike Vizard: Is this something of an awkward conversation for IT people? And I bring this up because for years they’ve been telling the business that, “We’ve been managing data on your behalf and we’ve been doing a great job,” and then suddenly you have to show up and say, “Well, actually, we need this new MDM thing to fix all this stuff.” So how do I approach that topic in a way that doesn’t quite make me feel like I’m telling somebody that the emperor has no clothes?

Brett Hansen: Well, honestly, it’s a good starting point. Things have changed, right? So at one point, yes, you had all your data in one location. I think you and I both can remember the days when it was relational databases. That was where all my data was stored, right? I had Oracle Edge AI or Microsoft SQL or IBM Db2, and that was it. Right? Now, think of how many different places we’re putting our data. Think of how many different places we’re collecting data. So it’s a pragmatic conversation with your line of business. Things have changed. Things are more complicated.
Now, that being said, think of what more we can do. I’m amazed at some of the exciting changes that we see with our customers who have implement MDM. Brown-Forman, a company you might not have heard of, but I’m certain you have drank. They do Jack Daniels, they do tequila, they do a number of different well known brands. They went and they reconciled all of their different brands.
And so just as example, Jack Daniels is called a dozen different things around the world. And so they were having a difficult time like understanding, okay, so I’m trying to do market shares. I’m trying to understand how successfully we’ve been around this brand. I need to be able to go and pull information from different sources with different nomenclature naming to get a common view. And then once they’re able to do that, their line of business was amazed at how much more thoughtfully they can plan, how much better they can target their investments, how much more they can do around being successful by focusing on certain areas.
So again, it’s not easy, but the impact can be truly profound on the business. And I think that’s the conversation that has to be had. Let us work together in partnership to deliver this outcome that you’re seeking.

Mike Vizard: All right, folks. Well, you just turned it here from the horse’s mouth. If data management is driving you to drink, you clearly need to have a different approach. Hey, Brett, thanks for being on the show.

Brett Hansen: Thank you, Michael. Appreciate the time.

Mike Vizard: All right. Thank you all for watching the latest episode of the Digital CxO Leadership Insight series. You can find this episode and others on our website. We invite you to check them all out. Until then, we’ll see you next time.