multiple-multi-color-doors

Data Talks Episode 21: Assessing MDM Software Choices

Selecting the Right Technology Involves Careful Research and Consideration

Host: George L'Heureux, Principal Consultant, Data Strategy
Guest: Howard Poppel, Solution Design Consultant

There are many benefits that enterprise software can bring to organizations pursuing a Master Data Management (MDM) strategy, but as the space has expanded, so too have the options that companies need to sift through in order to make an intelligent choice that meets their needs. It is easy to get overwhelmed. 

Dun & Bradstreet Solution Design Consultant Howard Poppel has worked through and consulted on many MDM journeys with companies around the world. In this episode of Data Talks, he shares his expertise on which factors have the greatest influence on companies’ choices of MDM software packages, which options to consider, and how to prioritize and staff properly to maximize the value of such an investment.

 

 

Read full transcript

George L’Heureux:
Hello everyone. This is Data Talks, presented by Dun & Bradstreet. I'm your host George L'Heureux. I'm a Principal Consultant for Data Strategy in the Advisory Services group here at Dun & Bradstreet. In Advisory Services, our team is dedicated to helping our clients to maximize the value of their relationship with Dun & Bradstreet, through expert advice and consultation. And on Data Talks, I chat every episode with one of the expert advisors at Dun & Bradstreet, about a topic that can help consumers of our data and services to get more value. Today's guest expert is Howard Poppel, a Solution Design Consultant at Dun & Bradstreet. Howard, how long have you been with the company?

Howard Poppel:
I've been here 21 years and 13 of them in Technical Advisory Services.

George L’Heureux:
And can you tell me a little bit about what it is that you do in your role as a Solution Design Consultant?

Howard Poppel:
Well, the solution design consulting team, we're technical advisors, and we do a post-sale technical consulting, mostly specializing in integration and automation of D&B tools and data within our customers’ ecosystems.

George L’Heureux:
Thanks, Howard. I think our topic today is really relevant for a lot of the data practitioners that are out there and that's how to choose the right MDM application. Let's face it, right? There's a ton of different software packages out there. How is one supposed to figure out what's right for them?

Howard Poppel:
Well, I think they first have to figure out what they're trying to accomplish. What is their point of arrival? Because every package is different, and once they understand what that point of arrival is, then they need to start then breaking that down into the different components of what's important to them. So whether it's alternative hierarchies, operational hierarchies, multilingual functionality, global access, in-product identity identification, they need to understand what, what really are the needs. And then that'll help them to determine what's the right package for them.

George L’Heureux:
So just then, you mentioned a few of them like alternative hierarchies, multilingual functionality. Are those the most common functional needs that we hear clients saying that they need? Are there others?

Howard Poppel:
Yeah, there's a few others. But the most important, and one of the main reasons why we see our customers purchasing MDM packages, are for those operational hierarchies. So when we're talking about that, we mean like sales hierarchies, geography hierarchies, territory hierarchies. D&B can provide the standard levels of hierarchy. So we can do standard corporate hierarchy, alternative linkage, minority linkage. But a lot of these customers need operational views. And that's really where MDM comes into play, because they can then separate and slice and dice the data and provide it into those operational views, taking our data.

Also, a lot of customers are looking to be able to rationalize data from disparate sources. So whether you've got two, a hundred different sources, what they're trying to do is find out what data is across the organization and where is it duplicative. And then MDM helps to solve that.

George L’Heureux:
So when we're looking at how to choose, I imagine that these types of functional factors, the ones that you've mentioned just now, are important. But also its ability to really handle the volume. If you've got 200 sources and they're all small, that might be one thing. But if you've got five sources and they're all massive, it might be a completely different calculation. Is volume something that comes into play when you're looking to evaluate MDM offerings?

Howard Poppel:
Absolutely. The two things in that aspect are volume and throughput, because it's not just how many records or how much data you want to push through the system, but what kind of speed do you need? With some customers, just the ability to rationalize data in days, is sufficient. In others, they want to do subsecond response. They want to know immediately, and they need to clean their data for that.

So that's your throughput, is how much data can we consume in what period of time, but then the amount of data and also the number of different source systems. Because keep in mind, no two data sources are ever alike. So when you're looking at disparate data sources across an organization, the more you have, obviously the more technology you're going to need, and the more complex the process is going to be, because it has to account for all of those different nuances from each of those systems.

George L’Heureux:
So do the distinctions between sort of on-premises software and cloud based software, do those play into it as well or mobile accessible, and like you were talking about before, multilingual capabilities. How should people be assessing each of those in terms of their importance to the overall decision?

Howard Poppel:
Well, we're starting to see going it across the spectrum that on-premises is minimizing. It's a lot more cloud based solutions. A lot of our customers are really ... Everything's moving to the cloud, but there is that need for on-premise. And we see that a lot more in the financial services industry where security really becomes a priority. They need to keep everything behind their firewall. So we're seeing them lean towards more of those on-prem.

But for the most part, a lot of our customers are really going cloud based and they really need the security of working with cloud-based systems, but also they need to know that they have connectivity. Whether again, it's mobile, and from any region across the globe. Being an international or global corporation like Dun & Bradstreet, and we're delivering data across the globe, our clients could be anywhere. And I mean that not just from their position as a company, but as their position as their employees, I've done work personally in multiple countries around the world, reaching out to the D&B database to get that information, so that has to be accessible.

George L’Heureux:
Talking about Dun & Bradstreet, here we often talk about the importance of really understanding your data before you try and put together a technology stack. Now, are there particular considerations around data that you feel you need to consider when you're assessing your software options? Is data the first thing that you need to think about, or is there something even before that?

Howard Poppel:
Well, right before you decide what data you need, it goes back to what I had said previously. You really have to focus on what is the use case. A lot of times, our customers come to us with an MDM problem or opportunity, and it's really based in the sales and marketing space. But keep in mind today, master data management really goes across all spectrums of data. Whether you're doing finance analytics, whether you're looking at your supply chain across the globe, or whether you're using it for sales and marketing purposes. The first thing you need to do is what is that use case. Then once you've got the use case, now you start looking at the data, because what data do I need?

What data do I have and what is the integrity of that information? Because again, as I mentioned previously, when you've got disparate data sources and different levels of completeness or accuracy of information, your MDM software has to be able to manage that. Then the last piece and again, I was saying just previously, was when you start looking at data enrichment, do you need finance data? Do you need diversity information or compliance information? Do you need sales and marketing? Do you need URLs, IP addresses?

All of that different information all correlates back to what was that original use case, and then quality of the input and then quality of the enrichment.

George L’Heureux:
And you mentioned that there too, kind of that integrating the different sources and that's what a lot of the MDM software is really trying to get you to, is this idea of a golden record. Are there different ways that these MDM providers are building these golden records? Or allowing you to sort of integrate toward a golden record, that you need to take into consideration that might be different? It might get you different results in the end.

Howard Poppel:
Well, a lot of them are using the same type of processes to get you to that point of arrival. A golden record should be the same across every organization, but it's not, because again, it depends on the input data, and then what data are you matching it up against. A lot of these, keep in mind, MDM is really a tool to get you to a point of arrival. But these companies don't offer the in information that D&B is offering.

So when you're trying to get to that golden record, you have to look at first party data with which is the customer's data. You have to look at second party data, and then you've got to look at that third party data and again us, when you bring them all together, then you can get yourself to that point of arrival, which is finding out what is that best record. Now, a lot of different MDMs do it slightly different. But at the end of the day, they're trying to get you to the same point of arrival.

George L’Heureux:
Now, a lot of the providers have packages that come pre-loaded with, like, pre-defined logical models. They cover a variety of different domains. In what you've seen, in your experience, do these end up being helpful? Is that something that can help differentiate between different products, different packages?

Howard Poppel:
They're helpful as a template. You have to start somewhere and it's much easier to start with that template, and then bring that use case in and then make those configuration changes. Doing this for 20 years, I've never seen two customers’ data look exactly the same. So there shouldn't be just one process for everybody, and that makes sense, and a lot of these vendors know that.

So what they do is they'll provide you with the template, get you started, but give you a toolkit or a toolbox so that you can do configurations based on your needs. Again, you brought it up just a minute ago, multilingual. Are we dealing with diacritic characters? Are we dealing with double byte characters? How is the data provided? Is it name first, address second? All of that needs to come into play. Are you using ISO standards? Are you using FIPS standards? So when you look at all of the information that's available, you've got to look at what customizations are available, but those templates will get us the starting ground. But they should never be looked at as the end all be all to get you to your point of arrival.

George L’Heureux:
Well, I'll tell you, Howard, as a guy who has an apostrophe in his last name, you're talking about diacritic characters, I'm just thinking, can they handle apostrophes right? Another thing that you need to think about though, anytime you're bringing on a new software package is resources. And are these packages that these MDM providers offer, are they going to require specialized resources generally?

Howard Poppel:
A lot of the MDM providers have a package that comes with it, where they'll at least get you implemented. But yes, absolutely. Every MDM’s process that I've ever worked on, you need to have resources really in three different areas. Number one, you've got to have it at the customer base. They need to be able to provide a project manager. They need to be able to provide DBAs. They need to be able to provide reporting people.

In the old days, it was easy. We would deliver a flat file of data. You'd have a DBA massage it in Excel or some other simple program and they'd spit out, "Here's the answers." But today things have so much more complex and it's so much better because we can really dig into the data. So at the customer level, you need all of those different resources to really get you to your point of arrival.

Then second, you need the software provider to have their provision. They need to be able to provide the training and technologies and support at the application level. And then last but not least is where D&B comes into play at the data level. We're going to work with the hardware and software providers, but we also need to be able to explain how the data fits into those structures. What should the canonical models be? Where does the data update structures? All of that gets really important, and that's why it's important to have resources on all three levels.

George L’Heureux:
So we've been talking for a while now, and we've gone over a lot. Now there's probably some people out there watching this or listening to this and they're asking the question, "Can you do this without enterprise data management software?"

Howard Poppel:
That's a great question, George. And to be honest, it depends on the two things. Number one is what's your point of arrival? So what are you trying to accomplish? And in what I found in my experience and working with my team, is it also depends on the size of the organization and the size of the engagement. I've seen very successful engagements without using master data management software, but bringing it into a data lake, a data warehouse, or even a CRM. But keep in mind, those are kind of more outliers. A lot of the big companies, they have to have this software because it really does the job for them.

But when you're a smaller organization or you have a smaller footprint and you really just need to be able to, again, dedupe across two different data sources, and then provide external data onto it, that's not overly technically complex. So that's where you might want to look at again, a data warehouse or a data lake to manage that kind of process. So the short answer is, yeah it does, it's necessary to have it. But the longer answer is that it really depends on the use case and the size of the organization to make that decision.

George L’Heureux:
So, Howard, as we start to wrap up, what's one thing that you'd want someone who's watching this or listening to this, to walk away with having learned?

Howard Poppel:
They have to understand one thing. And I think if you're going to get anything out of this video and out of this conversation that we're having, is that enterprise data and master data management itself, it's not a project. It's a process. It's a journey. You've got to have that starting ground. You're going to have then the implementation. But what a lot of companies fail on, and that's why we as consultants work with them closely, is also understanding the most important part, and that's the maintenance or monitoring strategy. Once you get the data in there, if you don't do anything with it, then all you have is an expensive Rolodex.

But if you have that maintenance strategy in there, then you've got the technology in place that will get you guys well through, keep the data clean and keep it actionable, because that's also important. And I know we haven't touched on that, but I want to finish with that. And that is at the end of the day, master data management is really, nobody does master data management just to master their data. There has to be an end game and that's make the data actionable and useful, regardless of what area of the organization is asking for it.

George L’Heureux:
Well, thank you Howard so much for joining me and sharing some of your expertise on this.

Howard Poppel:
Thanks, George. I appreciate being here today.

George L’Heureux:
Our guest expert today has been Howard Popper, a Solution Design Consultant at Dun & Bradstreet, and this is Data Talks. I hope you've enjoyed today's discussion, and if you haven't, I encourage you to share it with a friend or a colleague. If you would like more information about what we discussed on today's episode, visit www.dnb.com or talk to your company’s Dun & Bradstreet specialist today. I'm George L'Heureux. Thanks for joining us. Until next time.