Hello, data leaders, data practitioners, data stewards, data managers, data people, data aficionados, data wanderers, you name it. Hello all. I am Malcolm Hawker. I am the host of the CDO Matters podcast, and we’re here today to talk about data, specifically master data.

We’re gonna roll up our sleeves, and we are going to embark on an MDM journey. I’m gonna tell you how to get it done.

If you are a CDO, VP of data and analytics, director of data and analytics, maybe you are in charge of a governance program or an MDM program, or maybe you have been given an MDM mandate as a part of a governance mandate.

If you’re interested in MDM, maybe you are trying to get an MDM program off the ground. Maybe you’ve tried in the past and failed.

I did. First two times trying to make MDM work were dismal failures.

So maybe you’re resurrecting an MDM program. Maybe you’re trying it for the first time.

Either way or maybe you’re just interested in learning more about MDM, master data management. Keys to success, things to avoid, things to do.

I have learned through this school of hard knocks.

This is real gray hair, And, because I’ve I’ve done it.

I’ve implemented MDM. I’ve implemented government’s programs. I have consulted to companies launching MDM programs. I have been a consultant on successful MDM programs. I have been a vendor helping people get MDM programs off the ground, both from the software side through my role with Profisee, and on the data side, through my role previously with Dun and Bradstreet. Not to mention my role as a Gartner analyst where I talk to people all day every day about how to succeed at MDM.

So I’m what they call an expert, and I’m here to guide you through the process.

Ah, before we go into the details, let’s cover a few logistical items.

One, thank you for being here and listening to the podcast, downloading the podcast, sharing the podcast with your friends, liking, subscribing, all of that stuff that that that we do.

Your your ongoing support and patronage is awesome.

I I go to events, like I did at at at Gartner, recently in Orlando, and people would walk up and say, hey. I listen to the podcast. I really get value out of that, or I really appreciate the stuff you post on LinkedIn.

Guys, I can’t tell you that has just that is so awesome because that’s what I’m here to do.

Not post content per se. I mean, that what I’m here to do is to help you succeed as a data practitioner, as a data leader, as a CDO, or somebody who wants to be a CDO. This is my job. I’m here to help.

I’m here to promote the value of better data management, of better data governance, of master data management, of all the things that we know are important to succeed as a day leader. That’s what I’m here to do, is to promote those things and to support those things and to support you in your journey. So when you come and you tell me, hey. This is really helpful, or this is really useful, or I found that content useful.

That just means a lot to me. It’s it’s wonderful. Part of the reason why I’m having this conversation today is is because, I was on the road a lot last year, and I’m on about to be on the road again. As, you’re watching this is probably mid to late April.

By this time, I would have been on the road a couple of weeks already. Most of April, most of May, and most of June, I will be traveling.

But what I would hear when I was on the road was, hey. You know, this stuff is great about strategy and high level things and things to think about, like data culture and data strategy, and and those things are awesome. We need to do that as data leaders. But I also need to focus on fundamentals, basics, things like governance and MDM. There’s important lessons to be learned there.

It’s not just high level stuff, and it’s not just all, you know, get it done stuff. It’s both.

So today, we’re gonna talk about getting it done as it were. Before we dive into the details, logistics.

I’m coming to a city near you, maybe.

Maybe.

Sometime very soon.

Like I said, you’re probably listening to this or watching this on the YouTubes, in mid April, mid to late April.

And I am I probably would have already come and gone from Data Universe. That’s for me next week in New York City.

The week after that, I believe I am in Dallas, soon to be, in Boston, Chicago, New York again, Denver, Atlanta, a number of other cities, soon around the corner where I’m we we, Profisee, are doing a roadshow with our partners from Nudesic, and Microsoft, where we are going to be talking about the power of MDM within an AI enabled architecture.

So we’re gonna talk about things like the data fabric. We’re gonna talk about how MDM plugs into a data fabric. We’re gonna talk about how to enable AI at scale.

We’re gonna talk about how to make sure the the custom chat bots and copilots that you’re building, maybe in Azure OpenAI Studio or in any other tool. That’s okay. But how do you make sure that those copilots are providing accurate and consistent reliable information?

These are some of the things we’re gonna talk about at these roadshows.

So if you’re in one of the cities that I’ve mentioned, we’d love to see you.

We do a half day event at a Microsoft Technology Center.

So if you’ve not been to a Microsoft Technology Center, if you live in a in a major city like Chicago or Atlanta or Detroit, and you haven’t been to your local Microsoft technology center, they’re kinda cool.

They’re kinda cool. They’re they’re where Microsoft holds a lot of meetings, local community events where they will wine and dine a lot of their bigger clients.

Through our partnership with Microsoft, we have access to these amazing venues.

So come out. See us.

Somewhere down below, I’ll I’ll make sure that I post a link for registering to those events in one of those cities.

But if you’re interested in spending an afternoon with your fellow data practitioners from some of the companies in your area, if you’re interested in meeting me, if you’re interested in hearing from Microsoft architects, if you’re interested in hearing from other data leaders about really building out a next gen let’s call it a next gen data and analytics infrastructure, particularly when it comes to, leveraging AI at scale.

Come on down.

We’d love to see you. We’d love to chat. Would be great.

Beyond that, there’s gonna be a number of conferences as well coming up.

It’s spring, so it’s kind of conference season. I already mentioned Data Universe in in New York.

Soon to be, speaking. I’ll be speaking in early June. It the DGIQ, data governance and information quality event in the amazing city of San Diego, California, the first week of June. So I’m actually speaking on Wednesday, whatever, June fifth or sixth, I think.

The days are all melting together.

But the DGI Q events are fantastic, particularly if you are looking to kind of grow, maybe even if you’re looking to get certified as a CDMP, certified data management professional.

Great place to do it through Dataversity, which is the kind of the the coordinator of the DJI Q events. Great, great event.

Followed very soon in July by, CDOIQ.

The CDO event in, it was on the MIT campus in Boston. It moved just slightly off campus to a a great hotel, adjacent to, the MIT campus and adjacent to the Harvard campus and right across the river from from Boston University. Boston’s an amazing place. Anyway, we’ll be there in the second or third week of July as well. So got a lot coming up.

Come on down. We’d love to see you if you’re in one of the cities that I mentioned.

Without ado, let’s talk about MDM.

If you are listening to this podcast, if you have downloaded it, I’m gonna share a few slides while I’m talking because I think that’d be useful.

We do, publish these podcasts on YouTube. So if you want to see the slides, go to YouTube or, ping me, DM me on on LinkedIn.

And if we’re not connected on LinkedIn, folks, please connect with me on LinkedIn.

If you’ve got questions about anything I say on any of these podcasts, just send me a message through LinkedIn.

I will absolutely respond.

But if you’ve got questions or if you want actually a copy of the deck that I’m talking about, this deck, I’ll I’ll send it you through LinkedIn, or I’ll send it to your email, whatever. I’ll send it to your carrier pigeon. I’m happy to distribute the deck. So if you’re seeing this on YouTube, you’d want actually want these slides. Maybe you wanna use these slides to help pitch MDM within your company.

Right on. Happy to share. That’s cool.

The slides that I’ll be showing on the YouTube version that you may not be seeing as you’re jogging or driving to work right now, but but you can get. I happily provide you. These slides were an adaptation of a presentation. I still I would I did in my last few weeks while at Gartner.

So, I was doing a presentation on how to launch an an MDM in ninety days, that I expected to give just before I left Gartner at at at a Gartner event, in twenty twenty two, but I didn’t get a chance to just because I made a career shift, and that’s all good.

So what you’re seeing is an adaptation of that, and I and I think the the information here is extremely useful.

So let’s get into it. Let’s get into it. Do we need to talk about what MDM is?

Maybe? I I I hope not.

I’ve got to the point where I honestly believe in my conversations that MDM is now what I said when I was a Gartner analyst in some of my last research is I said that MDM has transitioned from nice to have to must have, and I firmly believe that.

At a high level at a high level, the way I describe MDM actually is a little bit different than a lot of people describe it. The way I describe MDM is is literally the connective tissue in your organization.

It exists in the gaps. MDM exists in the gaps and is the glue literally in your company between functional silos.

So you’ve got a marketing organization as a functional silo. You’ve got a finance organization as a functional silo. You’ve got a procurement organization as a functional silo.

Master data is the data that literally binds them all together. It is the thing that enables cross functional collaboration in an organization.

If you got a a business process like quote to cash, quote happening in a CRM, cash happening in an ERP, there’s a jump between those two systems.

At some point, your quote will become a contract, and at some point, that contract will need revenue recognized in a GL, general ledger.

And maybe there’s a customer. Of course. Not maybe. Of course. There’s a customer name on that contract. There’s an account name on that contract. Maybe even before it was a contract, it was an opportunity, and there was a customer or prospect associated to that opportunity.

But the concept of a customer, a buyer here, it it traverses quote to cash. But, usually, what we see is in a CRM is defined one way and in an ERP is defined another way.

Main maybe in the CRM, it is the ship to.

Right?

Who’s buying it? Who’s receiving it? Who’s using the goodies?

And maybe in that ERP, it’s a bill to. Who’s paying for it? Very common to have somebody pay for something and and somebody maybe a division of a company actually use it, but the headquarters is paying for it. Bill to, ship to. Two different kind of views of that customer where the definition may even be different. Maybe the data is even different across those worlds. And how do you enable cross functional efficiencies when the data doesn’t look the same across those two silos?

Answer, MDM.

It’s what MDM does. It’s the connective tissue that holds those processes together. It is the data that is shared widely across the organization that needs consistent standards, consistent structure, consistent hierarchies, consistent governance to enable cross functional coordination across the organization. That’s what MDM is, and that’s why it’s important.

And will AI change this? Do we see a future where we just have this kind of magic AI happening in the background where all these things go away?

Maybe in ten or fifteen years, maybe.

But as long as as long as companies are functionally aligned, and as long as we as CEOs and COOs and CFOs want an operating model that allows those functions a certain degree of autonomy, Autonomy is good.

Domain centricity is good.

As long as we, as business leaders, want those functions to operate a little bit autonomously, maybe have a little bit control over their own data.

This is the whole idea behind the data mesh folks.

Autonomy, domain centricity, domain control over their data.

We generally recognize these things as good things. I think that’s a good thing.

The problem is the problem, the challenge, the opportunity for data people just like you and me is that when marketing calls a customer one thing and finance calls it another, how do you make them coordinate? What do how do you create that common language across those processes?

It’s MDM, and MDM isn’t going away.

As long as those functions operate differently and as long as we see value in them operating differently, we need a way to tie them together, and we need to do it at scale.

There was a time when I led an IT organization where we created basically a really complex spiderweb of point to point integrations, where we hardwired the CRM to the ERP, where we were hardwiring the ERP to the procurement system, where we hardwired the HR system into the ERP system and on and on. At a certain point, that is no longer scalable.

Again, there’s where MDM comes into play, where you have a single place to manage business rules, single place to support, stewardship functions and governance functions, a single place to where you can configure and manage all of those integrations instead of having fifteen different integrations.

This is why I always kinda chuckle when I hear data engineers say that MDM is a solved problem because you can create something in an ETL. You can create a what my dear friend, Juan Cicada, would call a fancy integration, where you could take data over here and move it over here. And in that process, you could do some complex mappings. You can do some transformations. You could do some some normalizations. Maybe you could do some cool stuff like some matching to see if the first three letters of the customer name are the same here as they are over here.

You could do that point to point. You could do it in a pipeline. You could do it in a data engineering workflow if you wanted to.

Not scalable. Not scalable. Not repeatable.

Years ago years ago, we used to do these things in, like, stored procedures, right, right, where we actually wrote code to do all of this stuff, to take data out of this database and drop it into this database.

And then the person who wrote the stored proc leaves, and nobody knows why they’re seeing what they’re seeing, right, in a consuming system. What happened? I don’t know. The guy left.

Again, MDM. So I’ve opined here for about what MDM is and the value that it brings. It is all about scalability and flexibility, folks. And you need it. And you need it. Not just because the company who pays my mortgage thinks you need it. I’m telling you you need it for all the reasons that I just, I talked about.

Proof positive, guys. I can tell you.

When I was at Gartner, during the beginning of the pandemic and through the entire pandemic when I was at Gartner, what we saw when the pandemic hit, I was like, oh, man.

This sucks.

Right? Things were being deprioritized. You know, there was a recession hitting, and everybody’s getting freaked out.

I was like, oh, man. This sucks. You know? Nobody’s gonna wanna talk about MDM anymore. I’m gonna be I’m I’m out of a job. Right? Like, who who’s gonna wanna talk about MDM?

It’s gonna be p two, p three, p four. Priority x.

Right?

Oh, that’s what I thought, and I was totally wrong. Thankfully, I was wrong. What had happened and what we learned what we learned with the pandemic, and we continue to see to this day, is that in a in an era, in a world of, like, constant disruption and change where your business needs to adapt to changing customer behaviors.

Right? At the pandemic, it’s like, okay. Nobody’s shopping anymore. Right? In in in, like, in real life.

Nobody’s going to the mall anymore. They want it like, everything’s online. Everything had already been transitioning online, but it, like, it went all in through the pandemic. Buyer behaviors completely, totally changed overnight.

Supply chains changed overnight. Everything changed overnight. We remember this.

When that happened, what happened was organization said, oh, okay. We need to pivot.

We need to respond to the changes. Makes sense. We need to respond.

Metaphorically speaking, what happened was companies that were like, okay. Let’s hit the gas pedal, and let’s turn left.

They hit the gas. They turn left.

Nothing happened.

Why? Because they didn’t know who their customers were. They didn’t know who their suppliers were. It took them three weeks to pull an aggregated report to show how much business they were doing with Acme Incorporated or how dependent they were on Beta Incorporated for supply chain.

They didn’t know the complexities of their supply chain. They didn’t have a complete view of their hierarchical relationships that existed with the supply chain. So when a ship gets plugged in the Suez Canal and all of a sudden, you no longer you no longer have the materials that you need in order to make the stuff you sell. So what we saw in the pandemic is that in order to respond quickly, to change quickly, to pivot, to digitally transform, you need a foundation of master data.

It just is what it is. Because you can’t afford to take two weeks to pull a report to see how much the inform how much business you’re doing with Acme Incorporated.

You just can’t.

Right? That’s just the so so what we saw is that instead of companies running from MDM during the pandemic, they ran to it, and they continue to run to it to this day, thankfully. Gives me a little bit of job security. Let’s cut to the chase.

How how how how do you get MDM off the ground quickly?

Let me see if I can actually share my screen. That would be Fabuloso.

Fabuloso.

By the way, that stuff just just I’m horribly allergic to Fabuloso.

The cleaning fluid. You know what I mean? The Fabuloso stuff. I’m horribly allergic to it.

As you can see, I presented this actually at a conference in Philly last year. So for anybody who’s listening while they are driving or jogging or vacuuming your house or doing whatever, again, some I I will be speaking to slides, but I will I will share verbally share what what I’m sharing through the slides. And, again, if you wanna see them, check out the YouTube version. Or hit me on LinkedIn, and I will happily send you a copy of this deck.

I already talked about kind of MDM becoming what I would argue is a must have as companies evolve and need to pivot. For those of you able to see this photo on YouTube, yeah, I took that photo.

This is a photo of a, SpaceX Falcon nine rocket taking off here from the Space Coast of Florida where I happen to live. Isn’t that cool?

Alright. There are four things you need to get right if you are serious about getting m MDM program off the ground. Now my focus here is how to do it quickly because I would argue you always need to do it quickly. The idea that the business can wait nine, twelve, eighteen months for MDM, those that ship is sail, folks. We can’t wait. Business does not want to wait.

So the deck that I’m sharing, the information that I’m sharing is how to do this quickly, and I would argue this is absolutely positively a best practice. What you need to do is to implement MDM quickly, show value to the business, put some wind in your sails, and then expand the scope slowly over time.

Best practice, get it up and running quickly. You need to deliver value quickly. We don’t have time to wait eighteen months for an MDM program to launch.

So four things you need to get right. One, you need to have a maniacal focus on managing for scope, and you need to choose the right approach. We’ll talk about that. A maniacal focus on scope, and and and I mean that.

You you need to be really rigid in your scope. We’ll talk about how you do that. Number two, you need to find the right business partner. What do I mean by that?

You need to find somebody who wants to work with you. Go where you are wanted.

If your CFO doesn’t see the value of MDM or doesn’t get it or doesn’t understand, well, then maybe you shouldn’t be partnering with her. Maybe you need to be partnering with somebody who really, really has an acute problem that MDM can help with. Maybe your chief revenue officer is struggling with cross sell or upsell or identification of white space to go sell products to.

If that person has an acute pain, you wanna work with them. Go where you are wanted. Find the right partner. Number three, be agile. Be iterative.

And and and and this is not just, like, conceptually agile. I I mean, really, be agile and be iterative. Take an MVP approach.

Carve this into the smallest portion that you can and still deliver compelling business value.

Lastly, avoid some common pitfalls. We’ll talk about that.

So what does it mean to have a maniacal approach, a maniacal approach to managing for scope?

Well, what that means is is that, really, folks, there’s two styles of MDM. Yes. There’s four implementation patterns as identified by Gartner, and my name is on those documents. I was responsible for saying there’s four implementation styles.

But honestly, when push comes to shove, there’s only two, operational, analytical. Operational is where you’re creating a a a version of the truth, maybe a single version of the truth or maybe even multiple versions of the truth, and you’re pushing that into operational systems where you create a gold master record in a MDM hub, and then you push that into ERP, CRM, whatever.

Where the ERP now defers to the MDM hub as the system of record for customers, accounts, suppliers, you name it.

That’s operational MDM.

That’s where the most value will be, but that’s the most disruptive form of MDM because you will inevitably ask your business partners to do things differently.

You will onboard your suppliers differently. You’ll put some controls in your CRM about who can create a new account in Salesforce and what some of the data quality standards are that you expect to see when you create that account. The minute you start asking your business to do things differently is the minute things really slow down with MDM.

So, yes, operational MDM is incredibly valuable. It’s incredibly valuable. It can be transformational, but it will take time. If you wanna be up and running in the next three months, you can’t focus on an operational MDM. You must focus on an analytical MDM. This is the idea of creating a three sixty of something.

If you’ve got four versions of Acme Incorporated in your CRM or in your ERP, how do you get to a place where you can at least see one version of Acme Incorporated and all of the transactions associated to Acme?

Whether those transactions are sales or maybe open sales opportunities or maybe past recognized revenue, the list is long. What we’re trying to get to in an analytical style is the ability for you to have confidence in saying, this is how much business we’re doing with Acme, or this is all of the HR records that we have for a given person if if you’re trying to manage a person. I could keep going, but the point here is analytical styles of MDM can be launched far quicker.

The reason is because all you need to do is find a way to correctly, consistently, and accurately link records together.

That’s an analytical MDM.

You can be up and running with that in a few weeks, not a few months.

You absolutely can.

So that’s what I mean by our approach.

Operational MDMs can be incredibly valuable.

It’s where most of us want to get to in time, but where we need to start with is those analytical styles of MDM.

One thing here that you need to keep an eye on is that you don’t need to break every single data silo. So when we say manage, take a maniacal approach to scope, find two or three silos to break. Don’t break every silo. A great example. If you haven’t figured out how to have a consistent view of customer across CRM and ERP, great place to start. Just focus maybe on those. Maybe focus on one application.

Maybe just one. Maybe if you if you just focus on your CRM, I guarantee you, you’ve got data duplication issues within your CRM.

If you’re a reasonably large company, I can pretty much guarantee you’ve got multiple versions of the same thing already existing.

So maybe just even focus in within one silo. But my point here, folks, is take a very maniacal focus on managing for that scope.

Another thing you need to do here is to not limit your scope through the lens of a domain.

Meaning, focus on a business process.

This another key here to your success is will is building some idea of a business case.

You absolutely, positively need some high level understanding of a business case.

Business cases are are are keyed off of KPIs. Things like, I’ve mentioned them earlier, cross sell, upsell, identification of white space.

Could you implement an analytical MDM with the goal? Well, this is a hypothetical. It’s not could you, should you.

What I’m saying is is that you need to have an idea of a business outcome.

Like, just one, maybe two. But an example would be, I’m going to increase cross sells by five percent.

What I need to cross sell, we’re selling product a to customer a, but customer b is part of the same corporate hierarchy.

Maybe it’s a division or a subsidiary or an operating group. We’re selling to one division of of company a, but we’re not selling to another division of company a. That’s an example of a cross sell. Could you use MDM to support that use case? Answer, yes. You absolutely positively can use MDM to support that use case.

But you have a goal here of creating some idea of a business case where you are measuring business impacts.

For example, I want to enable a five percent increase in cross sales.

That is an example of a business outcome that your stakeholder will care about and that you could absolutely positively measure.

Notice it’s not about a domain. I mean, it is about a domain. It’s about customer data. Right? It’s about customer data, but your measurement of success is an improvement of the business metric of a KPI, not a domain.

So I would used to ask this at Gartner all the time. How are you measure success in in your MDM program? And people would say, well, I’ll I’ll have better data, or I’ll reduce duplicates or I’ll increase my matches. I’ll reduce the number of none null fields we have. That’s not a business metric. Those are data metrics, and what we need to be focused on is business metrics and focusing on one or two KPIs.

One or two.

Another example may be, I’ve been talking about supplier related data.

Can you improve the speed at which you onboard new suppliers?

Can you reduce supplier risk?

Can you reduce the amount of inventory you have sitting around in a warehouse at any one point in time? I could keep going here, folks, but the idea key the key message here is is that when it comes to limiting scope, don’t limit your scope on a domain because it’s a false limit.

When you say, oh, I’m gonna limit my scope. I’m just gonna look at customer data.

That’s a false limitation on scope because customer data is everywhere.

Now if you said on the other hand, I am simply what I’m gonna do is I’m gonna increase our cross sell rates by five percent.

That has a galvanizing effect on where you need to focus. You immediately know if you use that as your success lens.

Is you if you use that five percent, in this case, five percent cross sell, as a way to put a fence around your scope for MDM, you know exactly where to look.

You know exactly you you you limit the amount of data that you have. You you limit the number of systems you need to connect to. You limit the amount of governance policies you have to apply, and on and on and on. So focus things on a business outcome. Focus on one or two key outcomes. Use KPIs to measure your success. Don’t limit your scope based off of a domain because it’s a false limitation.

So I already touched on the idea of number two here, partnering with the right folks. I’m not gonna spend a lot of time here, but what you need to do is to find people who have a lot of MDM related pain or pain that an MDM solution could help you resolve.

I’ve been giving example after example. Right? But find people in your organization who have, you know, a lot of the problems with MDM are almost always articulated in slow analytics, or it’s time to build a report or a lack of trust around report. If you have people in your organization, in your company who are saying, man, I just don’t trust that report, or it takes forever to build that report or the data that I see in the report is duplicated or I don’t get it, it doesn’t make any sense, or it doesn’t align in my business process.

Find those people. If they’re being verbal about it and and if they’re if they’re being outgoing about complaining about it, find those people, sit down with them, and start to understand, hey. Could an analytical style of MDM where I find a way to reduce my duplicates link like entities together, would this help solve their problem? Chances are it will.

Those are the folks that you need to be working with.

Those are the folks you need to be working with.

Don’t try to force engagement. I used to see this all the time.

I I would talk to people and and, I I’d say, okay. Well, who are you working with in in your MDM? What well, we’re working with department a or department b, but they’re kinda reluctant. They’re not showing up to any of our meetings, and they’re they don’t wanna participate in in our governance committees, and they don’t see the value. They’re kinda pushing back. They’re reluctant. It’s like, all else being equal, if you can find somebody else in your organization who wants to work with you, find them.

Find them.

Number three, being agile.

Being agile.

What does this actually mean?

Well, it means start small.

It means think big, but start small.

And what I would see so often when I was at Gartner is so many data leaders saying things like, well, I need to do MDM.

Gotta do this. Alright. Fantastic.

But I gotta go do governance first.

Oh, and before that I do that, I need to define my data strategy. Oh, but before that, I need to work on fixing our data culture.

Right? I need to I need to fix our culture because we’re not data driven.

So I’m gonna need to embark on some sort of data literacy program maybe, and I’m gonna have to try to fix our culture, and I’m gonna have to spend the next eight months defining a data strategy.

Yeah.

And then I’m gonna need to work on governance, so we’ll go hire some consultants.

And then all of a sudden, this little MDM program has gone into just mushroomed into this gigantic thing that’s gonna take twelve months of dependent task work before I can even start an MDM.

Don’t do that.

If you went to the Gartner data and analytics summit in Orlando about a month ago, in the keynote, you would have heard and this this actually touched me.

I was like, fantastic.

Rock on. We need more of this.

What you would have heard is Gartner analysts saying, be ninety seven percent execution and three percent strategy. Doesn’t mean we don’t need a strategy. Those things are true. Doesn’t mean we don’t need to focus on governance.

Also true. Doesn’t mean we don’t need to improve our data culture. Yes. We do.

But if you are on YouTube and you’re watching this, I’m I’m creating a I’m showing a high level framework and a high level operating model as it were.

We’re basically what I’m saying is that there are certain things that are way out of your control in the short term.

In the short term. As a data leader, do you need to try to become more data driven? Absolutely. Do you need to work on it? Absolutely. Are you gonna move that needle anytime in the next three weeks?

Nope. You’re not. You’re not gonna change your governance maturity in three weeks. You’re not gonna become data driven in three weeks. You’re not gonna change your operating model in three weeks. You’re not gonna become data literate in three weeks.

None of those things are gonna happen in three weeks. So what you need to focus on are the things in the middle of this model.

Again, I’m showing something on the screen now. Where you do focus on things like, okay, what technology do I need? Right? How do I define success? What does my team look like? What sort of governance do I need in order to support this MVP approach to MDM?

This is what I mean by being agile. Start small.

Start small. What is the minimum amount I need to do for all of these things across what I would argue is an MDM operating model. And, again, I’m sharing something on the screen that you can’t see if you’re just listening. But the the the these these five things are the core of an NPM operating model.

Technology certainly is one.

Some idea of a business case, some idea of outcomes, and some idea of a road map is another.

Having some idea of a governance framework is is a fourth component here. So we got technology. We got a business case. We got a road map, one, two, three. Governance model, number four, and your team, number five.

Right? You need to figure out what do I need across those five things in order to drive the business outcomes that I expect in this first iteration of MBM.

The example I’ve been giving over and over again is five percent increase in cross sales.

What is the minimum amount of governance I need in order to enable this MDM use case? That is a totally different perspective than I need to go do governance.

Yeah.

You need to do governance, but what you need to be focused on is the minimum amount of governance to enable that outcome.

What’s the team I need to deliver on that outcome? What’s the tech that I need in order to deliver on that outcome?

On and on. What’s what does my road map look like? So when you apply this lens of an outcome driven approach, things get smaller immediately.

Instead of needing to spend a year on a governance program, you could spend a few weeks on defining some policies related to data quality, matching, maybe survivorship. Well, you don’t even need that when you’re talking analytical MDM. You just need to link things together. But these poll the set of policies you’ll need to define is drastically smaller than it would be if you’re just saying, oh, well, I’m gonna go focus on customer. Okay. So customer is my scope.

What are the governance I need for customer?

Boom.

All of a sudden, you’re in you’re in a year long governance dependent activity for MDM.

When you say your scope is customer and you acknowledge that you need to define some governance policies in support of MDM, If you say your scope is customer, then all of a sudden, that governance initiative has become huge because you need to look at all aspects of governance related to customer data. You need to look at access and security. You need to look at ethics. You need to look at all every aspect of our beloved data governance framework, which is I’m referring to the the DAMA wheel, d a m a wheel.

If you if you’re you’re not familiar with it, I would search DAMA and wheel or the d m b o k, d m b o k. It’s a very it’s a functional it’s it’s a it’s a practical way of looking at data governance, but what it will tell you what it will tell you is that there’s a lot of work there, folks.

And if your scope is better customer data, then you can be guaranteed that you’re gonna spend a long time trying to figure out all the governance policies needed to enable better customer data.

Better way to do it is to say, my outcome is what I’m focused on is five percent improvement across sales. What’s the governance that I need to do that?

Totally different perspective. That’s what I mean by being agile.

Speaking of agile, when you focus on one or two KPIs, you launch that, get it done, you show value, you you make your chief revenue officer ecstatic because you delivered.

Find another KPI and do it again, and then do it again, and then do it again.

In, let’s say, a year and a half, if you keep following this pattern of agility and MVP driven iterations, in a year and a half, you’ll you’ll you’ll have crossed off probably most of your high priority KPIs, probably most of your high priority use cases.

It’s such a better way to approach this problem instead of top down. What we’re talking about here is a bottom up approach to MDM instead of a top down approach where you focus on all customer data, all product data. Start from the bottom. Work your way up. Focus on KPIs. Focus on business benefit. And if you do it iteratively over time, you’ll get to the you’ll get to where you need to get to far faster with less risk, and you reduce the the the chances that you’ll have this giant monolithic MDM program that blows up halfway through, which has historically happened to a lot of MDM programs.

So lastly, in our last few minutes together, today, my friends, what are some of the bigger MDM pitfalls?

Well, consultants would be one. Ninety percent of MBM solutions and MBM programs are supported by consultants.

So we love our consultant friends, and I know a lot of consultants listen to this podcast. The key thing that you as a data leader need to understand and manage for when it comes to consultants is you, data leader, define the scope of this initiative.

You need to put a box around MDM.

You need to be very specific about what your SOW looks like.

We often we we all aspire to having fixed price SOWs.

Right? But we fall into these traps of saying, okay. And it’s it’s it’s factually true.

The only way that I could get to that fixed price bid is if I spend the next six months figuring out what the requirements are.

It seems like a logical approach, and it is.

But the problem is is that if you have no defined scope at the beginning and all you’re doing is spending time and materials, spending money to bring in a consultant to help you understand what your scope should be, again, if you start from a a world, a starting point of all potential use cases, all potential KPIs, all potential domains of data, it’s gonna be months before you sort it all out.

Instead, what you should be doing, again, start from the bottom up. Identify one or two or maybe even three, but I would say one or two key business processes that you’re trying to make more efficient and support through this analytical style of MDM.

Use those business outcomes to put a fence around your your consultants. Just like you’re using it to put a fence around your scope, you can use it to put a fence around your consultant. Meaning, hey, consultant, you guys are awesome. I need your help.

You know MDM software vendors. You know this world. You know governance. You know these things.

And I really need your help, and I need your help to improve our cross sell rate by five percent.

That’s the SOW.

Build me an MDM solution and deploy an MDM solution that is purpose built to deliver five percent increase in cross sales. Again, that’s that’s the example metric I’ve been giving as as our kind of desired outcome.

But when you take that approach, it’s a drastically different approach than saying, okay.

Well, let’s work together to figure out what things could be you and and and our total universe of possible outcomes is is all. Sales love that stuff, by the way.

They love it.

Like, that’s that’s the seven month engagement where you’re only halfway finished, halfway through cataloging your data and glossarying your data and inventorying your data and doing all the lineage for your data. When you start at the top and sky’s the limit, boy, money can fly out the door quickly.

Instead, you need to do the groundwork to get to the point of having some idea of your desired outcomes, having some idea of what your scope is going to be, having some idea of what your approach is going to be having some idea of what the governance model is needed to support that, having some idea of the people required to support that. So this is gonna be work on your on on your end here, folks. You’re gonna need to do some of that work. But when you go to that table to have that conversation with a consultant and say, these are the things I want.

Here are my constraints. Here are the people that I’m adding to the mix. Here are my desired outcomes. Here’s my success metrics.

Here’s my approach to MDM. Here’s maybe even some two or three of my preferred providers. That’s totally different con conversation than it is.

Hey. Help me figure it all out. Totally different conversation.

So put a box around your consultants.

Consultants are our friend. We need them to support MDM. We need them to succeed, but you define the scope, not them.

Don’t number two here from a pitfalls to avoid perspective.

Try not to put MDM in a back seat. Meaning, all those being equal, MDM should be able to fund itself.

MDM as a program should be able to stand on its own.

I see so often over and over and over again where MDM programs are tied to large scale transformation initiatives like an ERP consolidation is is one, or maybe, tied to some broader ill defined notion of a digital transformation, whatever that means.

All else being equal, you need to justify MDM on its own and not tie it to some giant thing that’s gonna take six years to deliver value, which I I mean, I I I talked to people at Gartner who gave me timelines of three to five years regularly for their ERP migrations.

Well, we’re moving to SAP HANA.

That’s for HANA. That’s what we’re doing. Yeah. Three to five years.

Business does not have three to five years to wait for the insight and the value of MDM. They’ve got three to five months, maybe.

And if you but if you’re tied to that ERP consolidation, less likely you’re gonna be able to drive any meaningful value from MDM in the short term. So have it have it kinda stand on its own is is the point here.

Lastly, this is gonna sound a little heretical.

And if you’re if you’re watching the the, the YouTube version, you may be, like, scratching your head saying, what?

Because I just put up a slide that said to avoid data cleanups.

And and let me tell you what I mean.

An analytical style of MDM where what you’re trying to do is to create a single view, right, where you create one Acme with a master ID where every other version of Acme and every other ID associated to those multiple versions of of Acme are are tied to that master version of Acme.

Basically, a lookup table is what we’re creating here. It just doesn’t sound very sexy, but, basically, that’s what we’re doing.

When you take that approach, I would argue you can still deliver meaningful value in the creation of an analytical MDM where you don’t need to embark on any costly data cleanups to do it.

There will be some situations where the data maybe is is so incomplete or so poorly structured or so problematic that you actually cannot match, where it falls into when you run a match and you and and match algorithm is like, I don’t even know what this is.

That’s gonna happen. That’s gonna happen. And at that point, you could ask yourself, okay. Am I gonna try to clean that thing up, or am I just gonna let it percolate for now? I would let it percolate.

What do I mean by that? You could embark on large scale data cleanups.

But if you don’t even know what that record is or how to match it, how do you know what’s even worthwhile to the organization?

Probably a lower value record. May maybe it is, but probably it isn’t.

The key message here is don’t get tied up into multi month long data cleanups that aren’t probably gonna meaningfully move the needle in the short term. The Pareto principle will apply here. Eighty percent of the value of what you are trying to drive will be buried in twenty percent of the data.

And if you’re out there trying to clean a hundred percent of the data, that means eighty percent of your time is probably gonna be wasted on that data cleanup. Not to mention, chances are also pretty good that you can still get what you need done and deliver meaningfully value.

You can still probably deliver that five percent improvement in cross sales even if some of your legacy data remains problematic.

What I the way that I would use to describe this to my clients when I was a Gartner is I could say, you can price in you can price in the fact that some of the data will continue to be problematic.

Meaning, can you build a business case that justifies an investment to improve cross sells by five percent?

Again, this is just an example.

If you build a business case that says, I can still deliver the five percent improvement here. I can do this even knowing that some of my my source data will remain problematic that I can’t even match, that that will remain in my CRM and it’s still problematic.

If you can do that and your business stakeholders are aligned to it and they’re behind that and they’re excited about it, that’s all you need.

Go for it.

That’s what I mean by don’t get too mired up in cleaning up data that probably isn’t gonna be meaningful to help you drive value in the short term. When it comes to the point where we do get towards more of an operational style of MDM, where you’ve done enough of these iterations and you’ve driven enough value within the organization where you do need to have a single trusted curated gold master record that is being consumed in source systems, Yeah.

We may need to talk about some data cleanups then then. But now I wouldn’t do it. I wouldn’t do it. And I’ve learned this the hard way.

I’ve I’ve spent a lot of money on offshore resources to try to clean up data.

It took me longer it’s the right way to say this.

I spent a lot of time, literally months, creating rules that people that offshore resources could use to do the data cleanup.

And in the time that it took me to define all of those rules, I probably could have stood up in India. Not to mention the time that it took to clean up the data itself.

By my estimates back of the napkin, a single human working at a pretty good clip can clean up about a thousand records in a month.

If you’ve got two hundred thousand records that are problematic, you do the math. You’re talking about a small army.

Now, again, doesn’t mean we don’t need to try to figure out a way to resolve some of these things in the future.

But for now, if you wanna launch quickly, do not get mired in a prolonged data cleanup. Focus on the data that you can match.

Focus on the data that will help you support that business case that I was talking about.

So with that, I will stop sharing.

I hope this has been beneficial.

Instead of stare staring at me in this, my video quality is is still bad.

Sorry. I’m blur I’m blurry at least on my monitor. You’re pro I’m probably blurry, and I keep I’m shoving my hand here because I’m thinking maybe it’s an autofocus issue and it’s not.

Anyway, I hope you found that beneficial.

Apologies for me ranting about my video quality again here. You’re probably listening and thinking, so what’s wrong?

My video quality is low.

MDM is a critical enabler of business flexibility, a business transformation, the ability to have consistent and clear insights that you can make decisions on, MDM is critical on that.

It’s not going anywhere. As long as our businesses operate as functional silos, we will need MDM. And if you don’t have a focus on it, you really need one, and I’m here to help.

If you found this content beneficial, please like. Please subscribe to the podcast.

If you’ve got questions coming out of this, if you want a copy of the deck that I shared, please connect with me on LinkedIn.

Please say hello. Maybe even drop a comment below on YouTube.

I hope you found this beneficial.

Thanks for tuning in to another episode of the CDO Matters podcast. Words coming up on episode fifty. Just crazy.

Thank you for checking us out. I sincerely hope you’re getting benefit from my content, whether that’s written, whether that’s video, whether that’s speaking content, you name it. We’re creating a community here, folks, and it’s about your success.

I will see you again in another episode sometime very soon. Talk to you soon. Bye for now.