Master Data Management: De-Risking Data Modernization, AI & Fabric Initiatives

×

0:12
Hello everyone, and welcome today’s Senturus webinar on Master Data Management DE Risking Data Modernization, AI, and Fabric Initiatives.

0:22
Thanks for joining us today.

0:23
It’s always great to see you here on our webinars.

0:26
Before we launch into the presentation today, just want to talk over the agenda real quickly.

0:31
We’re going to do a quick introduction of today’s presenter.

0:34
Then we’ll go over some fundamentals of MDM, what it is, why it matters, talk about the value drivers for implementing MDM in your organization.

0:45
We’ll talk a little bit about what it actually takes to do the implementation and how you can get started.

0:51
Last but not least, we’ll cover some additional resources that Senturus has to offer.

0:55
And as I said before, we’ll do some live Q&A.

0:58
To wrap things up, our presenter today is Greg Frasca.

1:04
Greg is a solutions architect here at Senturus.

1:06
He brings nearly 25 years of experience in the data and analytics space to the table.

1:13
Greg is passionate about data governance and he loves helping companies focus on the right strategy at the right time to maximize the impact of becoming data-driven.

1:23
So welcome, Greg.

1:24
Thank you for being here today.

1:28
And before I turn the mic over to the Greg for the presentation, we want to do a quick poll.

1:33
Let me just start this up here.

1:34
And the question is, what is the current state of master data management at your organization?

1:38
Are you just learning about it?

1:39
You know, you have awareness, but don’t know a lot about MDM yet.

1:44
Is it initially defined?

1:45
Are you kind of reacting to a known specific need for MDM or are you proactively pursuing MDM and so you’re kind of managing and optimizing that process today?

1:57
So we’ve got quite a few answers coming in already.

2:01
It looks like by and large, most of you are kind of in that initial learning phase, but we do have some folks out there that have done some initial definitions and also some who are who are managing and optimizing proactively on the MDM side and just kind of share the results out so everybody can see this.

2:19
So again, it looks like the lion’s share of you are aware of MDM, but just kind of just in the beginning of your journey learning about what MDM is, how it can make a difference in your organization.

2:32
So you’re in the right place for that.

2:34
And again, a handful of you are working at some little on an MDM initiative already.

2:42
With that, I’m going to go ahead and stop sharing and I am going to turn the mic over to Greg to walk us through MDM.

2:54
All right, thank you, Steve.

2:56
Appreciate it.

2:56
Good morning and afternoon.

2:58
Thank you to everyone who’s joined today.

3:01
Appreciate you taking the time.

3:03
So first thing is a little bit of just setting the table and providing some context.

3:08
You know, I noticed in the poll results, as Steve mentioned, the majority of you are in the awareness phase and learning.

3:14
So I wanted to quickly cover, you know, what is MDM, why is it important and where does it fit in your overall technical data state?

3:23
So we’ll cover some of that first with MDM fundamentals.

3:27
So I won’t be labor this too much, but really what is MDM?

3:31
These are just some terms you’re going to hear me speak about over the next 40 minutes or so.

3:37
Number one is domains or data domains, which are basically logical groupings of data relevant to stakeholders within a business.

3:45
So these things could be people.

3:49
You’re going to hear me refer to it as customer quite a bit because customer is, is your typical use case when starting a master data project.

3:58
But it could extend to employees or it could extend to things like students if you’re in the education space, voters, if you’re in the political space, but really just anything, anything people that you want to sort of come up with a golden record of who a person is relative to your organization.

4:16
Additionally, products are important as well, locations, vendors and assets.

4:22
And I sort of put these in the order of what I’ve seen in my time around the, the level of engagement we’re seeing based on these domains.

4:32
So that’s, that’s a domain.

4:34
If you’re in the dimensional or data warehousing world, you think about those.

4:39
Next is master data, right?

4:41
So master data is essentially just the elements or attributes that make up those domains.

4:46
So if you think about it from a product perspective, maybe a product color, a product size, product availability, all those are attributes are around a product, right?

4:55
So that’s the master data and the master data management is essentially bringing those things together.

5:00
So it’s conforming those data across an organization’s shared data assets.

5:05
So if you think about your company or firm, you might have ACRM and ERP, you know, OMSWMS, depending on your organization, point of sale system, whatever those systems may be, might share commonalities when it comes to those data domains.

5:21
So around people’s product location, vendors or, or other, right?

5:25
So, but those are the primary ones.

5:28
So again, you’re going to hear me refer to those throughout.

5:30
So hopefully that puts a little bit of context around the, the terminology you’re going to hear later.

5:36
So next, why it’s important, why are we here today, right?

5:38
So if you look at why any project is important to an organization, it starts with ROI, right?

5:44
You’re, you’re typically not going to get funding or budgeting for a project unless there’s ROI attached to it.

5:50
And there’s this concept of hard ROI for soft ROI.

5:54
And what I’m focused on here today is the hard ROI and that more directly affects the bottom line of the organization, right?

6:01
And what’s typically going to be approved to move forward with on a project.

6:06
So the three main pillars of hard ROI that I tend to focus on are an increase in revenue, a decrease in costs and risk mitigation.

6:15
Typically those are the three things that if you do an ROI analysis, you can put dollars and percentages to those things, right?

6:24
So luckily MDM covers all three of these buckets.

6:28
You know #1 increase revenue through a strong master data, master data management program will give you more cross sell upsell opportunities, is going to give you better decision making ability.

6:40
It’s going to increase customer satisfaction, loyalty and retention, which in turn is going to increase your overall customer spend.

6:46
If you’re A B to C organization, decrease cost.

6:50
This really where operational efficiencies come into play.

6:53
You’re going to improve IT operating costs, perhaps not upfront when you’re implementing something like MDM, but in the long term, not having to reinvent the wheel multiple times and getting, you know, sort of that proverbial hamster wheel type approach of fixing things as they occur.

7:08
Setting things up front and building out that single version of the truth is going to ultimately decrease your operating costs.

7:16
And thirdly is risk mitigation, which is very big in the data and analytics world.

7:22
Depending on your industry, it might be larger than smaller, but you know, right now there’s a lot of talk if we look at customers around PII, which is personally identifiable information, personal credit information, personal health information.

7:36
Additionally, you know, you might be familiar with like CCPA, which is the California Consumer Protection Act or GDPR, which is the European version of that.

7:43
And more states are coming up with their own versions of those laws.

7:48
And unfortunately they’re all slightly ever so different.

7:51
So you need to account for all those things.

7:55
You’re also going to get a better holistic view for regulatory auditing.

7:57
If you’re subject to audits, having your data in place at the right time in the right place is going to make your auditors more happy and make the audit process go a lot more smoothly rather than be reacting to what they need in a real time sort of situation.

8:10
And lastly, the one thing I want to bring up with risk mitigation, why and you’re going to see this as a theme throughout of why MDM is so important is that it can really remediate against costly errors and disruption and, and things like supply chain, which I used in the slide as an example.

8:27
Imagine having the wrong sort of master data in place and you’re targeting 10s of thousands, hundreds of thousands of customers with the wrong next best offer campaign.

8:37
Or you’re choosing the wrong supplier vendor and, and costing your organization a lot of money.

8:43
Not just money, but time and effort too in in mitigating that, remediating, you know what that risk that occurred, you know what will happen.

8:53
So the compliance and regulatory piece, as well as mitigating the chance of error within your organization is super important.

9:03
So where does this all fit within your company’s overall data and analytics program?

9:10
Number one, if you’re at the step of just thinking about implementing master data management, the first thing you need to think about is that it must align with the current goals of the organization.

9:20
It’s going to be really hard to obtain funding, buy in backing from the rest of the organization if you don’t currently align it to the organization’s current goals or vision, right?

9:31
And that doesn’t necessarily need to be the corporate goal.

9:33
It could be exclusive to a line of business, but really aligning it up to the strategic goals of the business is, is very important.

9:43
Next is data architecture.

9:44
Working clockwise around the circles.

9:47
Data architecture.

9:48
If you don’t have the foundation in place and you want to embark on an MDM engagement, it’s really important to first make sure you’re doing an assessment of what your master data goals are and do you have the foundational components in place to execute on that.

10:03
Business sponsorship extremely important to this.

10:07
It goes back to strategic alignment.

10:08
And the next couple of slides are going to cover this in more detail.

10:12
But really, you know, what we’re going to see in the next slide is that IT is funding and is the backbone of a lot of these MDM efforts.

10:19
But without business sponsorship, it’s not a, if you build it, they will come type scenario.

10:26
You really need the business to, to be on board and involved from the initial steps, you know, week zero, as I call it, of the project and, and getting involved to understand what the benefit is of the business and so that they are conveying and communicating to their stakeholders what the benefit of this project is, right?

10:45
So really important for business sponsorship, data quality, you know, really MDM is a function of data quality.

10:55
However, it is dependent on basic quality routines being in place when you’re considering an MDM implementation.

11:02
And a basic example of that is looking at address standardization.

11:06
So you know you want to stay.

11:08
If you’re trying to master a customer record and your addresses aren’t standardized across all of your sources or in some sort of hub or repository, it’s going to be really hard matching on an address to determine who the correct great Frasca is, right?

11:23
So to establish that golden record, you really need data quality in place.

11:27
And to have a good data quality program, you really need a good MDM program.

11:30
So there’s a recursive relationship there between the two.

11:33
And lastly, data governance.

11:35
Data governance should be at the head of the table really driving what specific domains you’re going after, what those priorities are and how you’re going to implement them throughout the organization.

11:48
So depending where you are in your data governance journey at your organization, whether or not you’ve established, you know, data governance community and you have named resources within that is really important that data governance is driving this journey and helping you with the prioritization of which domains to tackle, when to tackle them, and you know, how to advance or enrich them over time.

12:13
So really MDM is at the center of your entire data state.

12:17
So going back to the business sponsorship, I want to get into a little bit about the value drivers of master data management.

12:25
And the first piece I want to bring up, which is in my experience, something I’ve really had to encounter multiple times, is that IT is typically the one responsible for the spend and the initiation of a master data project.

12:43
However, the presumed outcomes of this project are really business oriented.

12:49
So there’s this great McKenzie survey or study out there.

12:54
I encourage you all to look it up.

12:55
It’s the master data management survey.

12:57
It came out just last year from McKenzie, which makes a great case for master data management.

13:02
But a couple of the stats that really hit home for me were first of all, like what are the objectives that an organization wants to achieve with more mature master data management?

13:13
And if you look at those top 4I won’t read them verbatim, but the top four are really business oriented.

13:18
They’re all business oriented outcomes, right.

13:21
So for example, just revenue growth by better cross selling and up selling opportunities.

13:25
I brought that up in the previous slide.

13:27
Why so?

13:28
So that’s a big reason that’s a big catalyst for a master data program.

13:33
However that is solely on the business, that’s solely on your sales and marketing team and perhaps operations in some regard of, of delivering those, those revenue growth opportunities through cost selling and upselling.

13:45
The bottom two reduced IT operating costs, which I brought up as a decreased cost ROI and lower risk of non compliance is sort of a shared, shared resource, a very unwelcome shared resource, but necessary across the organization.

14:01
So the reason I bring this up, you see that the vast majority of respondents are business oriented, but how is the NDM program funded within your organization?

14:09
And that same set of respondents, you see the vast majority of MDM programs are being funded by IT, right?

14:17
So there’s clearly a disconnect there of the business understanding the power of what good master data management will do for the organization.

14:28
So it’s very important upfront to build that Better Business case.

14:32
If, if you’re an IT practitioner understanding what do I need to do to get the business’s attention and obtain the funding for this?

14:40
And if you’re, conversely, if you’re on the business side and you’re a business stakeholder, how do I engage with IT and, and get this to the top of their priority list in terms of implementing master data management?

14:53
So, so in order to build a Better Business case, there’s a few things you need to think about just upfront.

14:59
Many of these bullets you see are really bullets that could apply to any sort of obtaining of funding for projects within an organization.

15:10
And MDM is really no different here.

15:11
So first, first and foremost, if you’re in IT, it’s important to find a business partner for your effort.

15:20
Many MDM projects are unfortunately a reaction to something that has just gotten unruly and unmanageable over time or it’s a specific thing that is sort of quote, UN quote blown up for the organization and they need to be very reactive.

15:35
So I’m not necessarily suggesting that hopefully that’s not happening within your organization, but it is important to find the business stakeholder that where that is truly an opportunity where you can mitigate that before it happens, right?

15:50
So it’s important to look at that.

15:51
It’s important to look at the sort of the lowest hanging fruit and the highest economic value for your company.

15:57
You may have a sponsor that’s looking for mastering of a product domain, but you’re really seeing that customer is the more important piece.

16:07
So you really need to get all of these business stakeholders at a table and work out what is the prioritization of that effort.

16:15
So to going through that exercises, it’s really important and prioritizing what the business domain is and then calculating the cost of not taking action.

16:26
You need to do the math.

16:27
You need to figure out even if it’s a swag, right, what the cost of not taking action is.

16:32
Again, it’s very much more of a reactive type of approach.

16:37
Sometimes MDM, unfortunately what I’m seeing, but the companies that are proactive, that’s kind of the Three Little Pigs, the third little pig is proactive in, in building that house to prevent it getting blown down, right?

16:48
So calculating that cost upfront, defining what those KPIs are clearly articulating the benefits to your business stakeholders, and ultimately your financier of the project.

17:01
Figure out how you’re going to demonstrate value early and often.

17:04
And that goes along with any agile principles.

17:06
And then assess and present the value to the team early and often so you don’t lose the momentum you’ve probably built.

17:13
And why this important, if you look at the numbers on the right, is that 62% of consumers are lost due to a poor personalization experience.

17:22
I know I can vouch for that in, in my personal life with, with my brands and things that I choose to buy from.

17:28
83% of company leaders are saying operational efficiency is critical for an organization.

17:33
Having good MDM in place reduces your operational inefficiencies.

17:39
And then 80% of company leaders say trusted data plays a critical role in responding to market disruption.

17:45
So nobody knows when the next Uber is coming.

17:47
Nobody knows when the next big disruptor is coming within the within your organization or, or industry.

17:54
So it’s really important to get ahead of that and make sure that when that does happen, you’re able to react to it in a in a really quick and nimble manner.

18:01
So those are all advisements of building a Better Business case.

18:05
We’ll cover that a little bit more when we get to implementation.

18:10
Next piece is everyone wants some piece of API or AI right now, right?

18:17
I was dealing with an API issue earlier today.

18:19
So a lot of people on this call are probably piloting, implementing, or at the very least ideating about how to implement some function of AI at your organization.

18:30
And a lot of those same organizations in their haste, are trialing AI and putting it out to actual external and sometimes internal customers without considering data quality and master data management.

18:43
And it’s really imperative not to lose sight the importance of that in your rush to innovate, right?

18:48
Everyone wants to fly before they crawl, walk, run, but it’s really important, especially when it’s AI because you are entrusting AI applications to represent your organization much like a human would or a manual process would, right?

19:02
And if you look at the stats on the right to reinforce that, first of all, 35% of global consumers trust how AI is being implemented.

19:12
And, and rightfully so, they should, because in many of the same cases, people accountable for that implementation of AI are not even aware of how those models are being built, right?

19:25
Not only is it, it’s going into a black box for someone on the business side, but it’s going into a Gray box for some of the IT practitioners that are, are doing the data engineering architecting piece of this.

19:35
It’s really the data scientists and the training of the model to make sure that you have the right pieces in place to do that, right?

19:42
And those same consumers that are saying they do not trust how AI is being implemented are also saying they think organizations must be held accountable for their misuse of AI.

19:54
And I’m actually surprised that number is as low as it is.

19:57
I consider me part of that 77% when you’re thinking about implementing an AI program, again, it’s like they are representing your company, your brand.

20:08
They’re responsible for your loyalty as much as a marketer is or as much as someone in operations would be responsible for defending your brand.

20:19
So really think about that as you’re as you’re implementing AI, how important it is.

20:24
And like I mentioned earlier, data quality is a subset of MDM.

20:29
So having that right data quality in place is going to lead you to better decision making, right?

20:35
So, so those are the things around ported to quality and customer attrition.

20:39
So a couple other things I want to talk about was because of these numbers and because we know that AI is representing our organizations, back around 2016 or in 2016, Microsoft, Facebook, Google and Amazon, well, Facebook now Meta all got together and they formed a partnership to promote the responsible use of artificial intelligence.

21:07
And they came up with the responsible AI principles.

21:11
And those principles at a very high level are fairness, reliability and security, inclusiveness, transparency, and accountability, right?

21:20
And every organization should be striving for those things.

21:24
They’re all very important to your organization’s success and your success of rolling out AI and faltering in any of those places could have a huge effect.

21:33
You know, I mentioned brand loyalty.

21:34
That’s a big one.

21:36
And one large mistake that maybe hits the press might be detrimental, entirely detrimental to your organization.

21:44
But another piece is I’m sure many of you in this call are aware of ESG scores.

21:50
So ESG scores are, are something that has really been coming on in recent years.

21:54
And those are your environmental, social, environment, social and governance scores, right?

22:00
And they look at everything a company is doing for the greater good and betterment of society, right?

22:06
Very loaded definition.

22:08
But those scores can be severely impaired or impacted by poor AI practices.

22:15
So when you think about just, you know, trying to innovate and getting the cart ahead of the horse when implementing AI, it’s really important to think about.

22:23
Am I sure that AI is consuming the right data to make the right decisions to represent my brand?

22:29
And master data sits at the top of that pyramid, I guess, right?

22:34
It’s looking at all those things.

22:37
You don’t want AI to reach out to the wrong customer.

22:39
You don’t want to implement a chat bot that thinks it’s talking to AI.

22:44
Know there’s a Greg Frasca up in New Jersey, there’s a Greg Frasca here in Florida.

22:49
It could be, you know, you want to make sure that they’re talking to the right Greg, right, Greg Frasca at the right time, A, to keep their loyalty, but B, there could be privacy infractions there.

22:58
When you think about sort of the implications of that, if they’re talking to me and releasing some of the personal demographics or information about the Greg Frasca in New Jersey, when you’re talking to Greg Frasca in Florida, that that could be absolutely detrimental and financially cost you from that risk mitigation standpoint.

23:19
The other piece of AI, this conversely to what I’ve been talking about, but is that AI is really, it’s a really mutual, mutually beneficial or, or symbiotic relationship to, to MDM in that the proper AI implemented under the water line now is able to better infer matches on data.

23:39
Once it’s matured, it’s better to able classify data, especially this comes with unstructured data.

23:45
As, as you know, I won’t give you the, the stat that everyone hears about, you know, the exponential size of data increasing annually, right?

23:53
So in that there’s certain data elements that need to be identified and classified.

23:58
That’s well beyond the, the scale of what humans can do And, and you need machines to do that and, and the ability of AI to classify those things, bring them to the forefront and then maybe have some human interaction to classify them.

24:15
For an MDM program is, is very important.

24:18
And then scalability, again, the volume of data allows you to add.

24:22
Once you have those classifications in place, you know what to look for.

24:26
You know, the, the percentage of data in the wild, so to speak, is going to lessen more and more over time and you’re going to be able to better classify that data on the fly as it’s coming into your organization.

24:37
So again, before thinking about AI, while thinking about AI and if you’re trying to pilot it, thinking about what your master data approaches and, and where you are from an overall data quality.

24:48
This goes beyond MDM, just an overall data quality perspective is, is very helpful.

24:54
So the next piece that everyone’s talking about from, you know, not a future state, but a very current state is the unified data platform.

25:03
And today I’m going to wrap everything around Microsoft Fabric as the unified data platform.

25:08
So for those of you unaware of the concept of the unified data platform is something that sort of brings together all of your resources across the data estates and leaves them together within a specific fabric.

25:23
And it’s gaining a lot of traction right now.

25:26
And one of the key benefits, if we talk about it from a Microsoft Fabric, and I’m, I’m leaning on fabric today because it’s very much to the forefront of this.

25:34
It’s gaining a lot of traction.

25:37
You know, if you’re in the industry and, and you stay on top of, of all the news and everything, you’ll see that fabric is 100% top of mind for a lot of organizations, right?

25:47
And the other cloud providers are catching up in some regard, whether it’s from a marketing perspective or an actual implementation of resources perspective, they are catching up.

25:56
So the concept in Microsoft Fabric is 1 lake and one Lake is a logical data lake that brings together all of the sources within your organization that have uncommon definitions for the purposes of this call, right?

26:14
The One Lake can be defined as many different things, but for the purposes of, of master data management, it’s really providing that logical lake across all of your sources, right?

26:25
And the reason that this going to be really beneficial in the long run to things like MDM is that the concept of 1 lake allows for you to virtualize MDM through pointers or mirroring, mirroring across all of those disparate data sources within your organization.

26:41
So if you think about if you have a point of sale system, if you’re AB to C or nearly every company has ACRM or an ERP, etcetera.

26:50
You are now able to land that data in the data lake.

26:52
And you can have one virtualized piece of your solution over that data lake.

26:59
That is implementing the business rules, monitoring the data as it’s coming in and applying those business rules and surfacing them up for your development teams and for your business users to understand the nature of the data coming in and apply that to your master data principles.

27:15
And one thing that’s great that we talk about the hard ROI, something that’s going to be more of the decreased cost or even soft ROI from employee morale and, and automating them out of sort of manual processes they’re doing now is they will greatly benefit from the lack of propagating these things over multiple sources, right?

27:36
So if you look at if you look at the reference architecture on the right, you see the data engineering piece, the data warehousing, data science integration, there’s Synapse, real time analytics and surfacing it through Power BI.

27:49
Having one place to store those common business data rules across all of those things is going to be an actual game changer for development teams right now in terms of the fact that not everyone has to worry about it.

28:05
There are people that will have to worry about it, specific defined people, whether they exist in your data governance organization physically or logically, or they exist in your data architecture, data engineering side or on the front end or analysts consuming the data.

28:22
Those are the ones that will have to understand the business rules.

28:26
And if applied properly, it’s going to cost a lot, going to benefit by causing a lot less development costs for your overall teams when implementing any new data integration process, right?

28:40
Rather than reinventing the wheel every single time or repeating the process every time of implementing the same business rule across multiple environments.

28:48
Now you have it in one place through one lake, right?

28:52
And it’s very much in a burgeoning phase right now.

28:55
Again, if you follow the trades or if you’re a Microsoft person, you’ll understand that Microsoft is pouring everything into Microsoft Fabric right now.

29:02
It’s, it’s really exciting right now.

29:04
I’m really excited to see how it matures.

29:07
Microsoft does not have a specific master data management tool.

29:10
They are very much partnered with a tool called Prophecy, which we’ll talk about momentarily.

29:16
But having just the ability to apply those business rules for a master data piece is, is really important.

29:26
Couple things I’ll cover quickly around real world applications of MDM Couple of war stories for me honestly, but these are things I’ve seen in place and implemented that have been really successful.

29:40
One was an automotive retailer that got to a point they were in a heavy acquisition mode.

29:46
They’re one of the largest automotive retailers in their space in the country.

29:51
And they were in very heavy acquisition mode.

29:53
And as you know, people in heavy acquisition mode really aren’t prioritizing how the data looks when they are acquiring, you know, firms, you know, nationwide, right?

30:05
So what happened was, is A cause of that or an effect of that was that you’re getting unmanageable products and SKUs and their SKU and product management just completely went to a point where it was impossible for them to understand what their current inventory was, what their inventory in stock was, you know, and also what was out of stock, right?

30:30
Their stock outs and they were having a really hard time moving inventory around.

30:34
They had inventory sitting in certain regions or at certain dealerships that could have easily been transferred to other dealerships and then utilized.

30:41
So if you’re in the supply chain or production space, you might understand that that case, right?

30:47
So what this automotive retailer did was implement an MDM solution to better understand their inventory on hand, consume all of that inventory, take it into a master data management hub.

30:59
And they built an enterprise hierarchy that was able to not only take what they’ve already purchased, but also able to bear the burden of bringing on new dealerships as they bought them.

31:11
What they did was build an enterprise hierarchy and product management solution where they were easily able to determine what a product in SKU was and apply it to that specific hierarchy, right.

31:22
So they may have a part that could be from a different manufacturer is represented by a different SKU in four different places.

31:33
They were able now to get a fifth different place and implement that into their product hierarchy.

31:39
Or if they bought a dealership that was using one of those four examples that they had already implemented, it’s plug and play for them.

31:45
The SKU management is taken care of.

31:48
So it really led to a ton of reduced costs and something that’s not here.

31:52
That really also led to a lot of new opportunities for them in the service.

31:57
And service became really one of the things they started marketing after implementing this because they knew they could handle the burden of additional services.

32:06
The next one I wanted to bring up was a clothing retailer, slightly different and another B to C company, but the clothing retailer had been leveraging this clothing retailer decided that they were to use a lot of different third party vendors to satisfy all of their digital marketing needs.

32:23
So they chose that for a reason.

32:26
They thought that it was, you know, less expensive to do that.

32:29
In a lot of ways it was they didn’t want to invest in, in building out that internal marketing team, that martech team that was responsible for doing that.

32:37
They wanted their marketers to be marketers and not technicians.

32:41
So they really facilitated this decentralized marketing solution that they wanted.

32:45
But it started getting a little out of hand for one reason is that they wanted their customer 360, which is important to everybody, really important to someone like AB to C clothing retailer.

32:56
They wanted to build that customer 360 and they used another vendor that was mastering their records, taking all their customer data from all those disparate sources that they had from online sales, in store sales, third party sales.

33:10
They were selling to different companies that were selling their products.

33:14
And they were consolidating all that information and providing it to a specific third party vendor, which was then mastering their data, creating that golden record, enriching that data with demographics around the customer.

33:29
Think about, you know, approximate age, income, marital status, all those things that you see and then sending it back to them.

33:38
And what they really found was a big problem was they were dependent on this single firm to do all their match emerge.

33:43
And then they are also responsible for them distributing that out to all of their other marketing firms for things like e-mail blast or any sort of digital marketing campaign they wanted to do.

33:56
So there is this baton hand off between all these firms, which was very costly to them from a timing perspective.

34:04
If they if they decided they wanted to do some kind of promotion the fly because of some extenuating circumstance, they had to wait on vendors and there were a lot of dependencies that were preventing them from doing that.

34:17
So what we did was come in and ask them to build a customer hub where they’re doing all the mastering of the customer through an MDM solution.

34:26
Most MDM software tools they use have data enrichment involved with them to do things like address standardization and all the demographics.

34:35
So they were able to build a golden record of the customer themselves within the hub and then gave them control to more to better provide those third party vendors that were handling things like e-mail blasts and IT.

34:48
It really led to a lot more timelier, more accurate and consistent marketing efforts across the board, which led to, you know, further acquisition, customer retention, customer loyalty, overall spend, kind of those typical metrics you hear from a customer 360 perspective, they were able to see true results by implementing an internal customer hub.

35:11
We’re getting to a minute around build versus buy and, and the importance of those things.

35:16
And it’s not the same solution for everyone.

35:19
There’s a lot of different permutations you can make around this, but for them building that internal customer hub and by the way, since then they started tackling their product in Skew hub as their next domain of choice because the customer one works so well.

35:34
So just a couple of war stories from me that where I’ve seen MDM really, really work in a great capacity.

35:42
So moving on to implementation, I’ll be very brief on this.

35:50
This more just academic and educational, but there are four MDM implementation styles.

35:55
If you were to take MDM one and 1 in your class, you would see consolidation, coexistence, centralized and registry.

36:03
And then what we’re focused on today is those first two is consolidation and coexistence and give you very quick high level difference between them is the 1^st two work on consolidating data from all those multiple resources.

36:17
If you remember when I was talking about the one Lake, bringing in data to a single source and consolidating that, it’s doing that and it’s, excuse me, conforming that single golden record over all of those sources.

36:32
The difference between consolidation and coexistence is that consolidation is more of a place where you go to get the golden record.

36:42
You know, if you need the golden record, if you’re doing analysis or analytics or reporting on this, you have to go ask for it.

36:49
And it’s in the hub.

36:50
It’s available to you, easily available if you’ve implemented it, right.

36:53
But the data lives in that MDM hub in coexistence.

36:56
The difference is that it’s fed back into the individual source systems.

36:59
Now there’s extra complexity there, but there’s extra value in most cases, right?

37:05
In some cases, that could maybe be a nightmare because you want to see how the data was entered in the 1st place.

37:10
And if you’re implementing some strategy that’s overwriting that data, you’re going to run into a lot of trouble because you want to see how the data was it was entered, right?

37:20
Which leads me to the centralized and registry approach, which is more of a, I’m sure some of you have heard a lot of the shift left nomenclature that’s been in a lot of data and AI communications lately.

37:33
So centralized and registry is more of that shifting left where it’s everything is maintained at the source.

37:40
So centralized is, is very controlled and it’s storing in that data within the centralized environment within the source and registry is doing the same.

37:53
However, it’s keeping an exact replica of the data that you can look side by side with what your master data is and very beneficial for some companies, but it’s also very expensive and is the most challenging of all these implementation styles.

38:10
So it’s really up to what your industry is, what your line of business do your homework, talk to us around what is the best implementation style for me.

38:19
But typically consolidation and coexistence in most industries is going to get you where you need to be.

38:26
However, centralized and registry, depending on what your requirements are, could be a good option for you as well.

38:34
Thing I want to talk about is building versus buying an MDM solution.

38:38
So a lot of companies when they’re first starting are afraid to dip their tail in the water of, of buying an MDM tool, which absolutely is, is could be a great choice for them in terms of building, you know, I always kind of think the 80 20 rule of, you know, if you build something, you’re going to get 80% of the way there.

39:01
And if that’s what you want to do up front, that’s great.

39:03
If that’s what you want to do to demonstrate and evangelize the benefits of MDM to the organization, building that thing up front is great.

39:13
A it’s tailor made solution.

39:15
So you’re building it for your needs.

39:19
And then the second bullet that I have under advantages is I should have it as the first bullet, because it’s by far the most important is you’re getting all those data quality routines I mentioned earlier, which are a dependency of master data management out of the way.

39:30
When you build, you’ll it’ll be much more front and center and sort of in your face.

39:34
If I just can’t throw bad data into a machine and it’s going to come out with good data, right?

39:39
There’s no garbage in garbage out solution yet.

39:42
Hopefully we’re on our way.

39:44
But it will remediate the quality issues by, you know, putting eyeballs on it and seeing what those data quality issues are.

39:52
And if eventually you want to buy a tool or when, when it’s time or it’s a necessity to buy a tool, you have some of those things in place already.

40:00
And it’s going to really lessen the burden and pain of implementing a new tool as any tool.

40:05
Causes burden and pain.

40:07
And then also there’s no licensing and software costs.

40:09
But the disadvantage is there are development costs said it’s not.

40:13
I’ll throw away the development costs because a lot of that is those data quality routines that are necessary if you buy a tool regardless.

40:20
And then you do need to look out for a data and cleansing and enrichment vendor.

40:26
So there’s lots of things like I’ll use LexisNexis as an example or Axiom or a lot of tools out there that you send them their data, they enrich it, they standardize addresses, they will build it, they will give you the demographics that they understand around a customer and send it back to you.

40:41
That ingress egress piece is a bit of a disadvantage when as you’re waiting, but you do need to look, look for a vendor as well.

40:48
And there’s obviously costs associated with that, but the costs are also built into the buying piece, right?

40:53
So the advantages of buying something because you’re going to go faster to market.

40:58
Again, I’ll bring up the 80 20 rule.

41:01
You’re going to go faster to market.

41:02
However, you’re more prone to error the faster you go to market.

41:06
If you don’t have good data quality in place, you’re going to pay the not only the upfront licensing fees, but support and maintenance over time.

41:13
There’s going to be regular updates that your development team needs to worry about and but the typical typically cleansing and enrichment are built into many of these sorts.

41:23
Not all of them, but almost all of them.

41:25
Any of the best of breed ones will have some connection and usually it is a hook into a LexisNexis or an Axiom or whatever tool they have a partnership with to do that cleansing enrichment.

41:34
But it’s built right into the tool natively.

41:36
So it saves you the extra steps and time and effort to go out and find a vendor, send the data to the vendor, build the processes out to share data with them, etcetera.

41:47
So the disadvantages, the additional software cost, you know, you are obviously building software costs, there’s still some development costs necessary.

41:58
There’s a lack of customizations when you buy, typically these firms will work with you to customize as with any organization, there’s a lot of customization features you can find, but there are customizations that you may lose from a build buy perspective.

42:12
And then you’re also dependent on a vendor to make sure they’re up running reliable and getting you the data at the right time as you need it.

42:20
So in my experience, just blanket recommendation without knowing any customizations, it’s billed first and then buy the tool and fit what you’ve built into the vendor you’ve purchased.

42:38
As for tools on the market, if you’re eyeballing tools to buy, some of the top ones there are Informatica, there’s Prophecy, IBM and a few others.

42:51
And really choosing these tools very much you should go through a proper vendor selection process of see what’s the best of breed for your industry, see where your internal skill sets are or what you’re looking to hire, where, where they lie.

43:03
If you’re a long term customer, Informatica, Informatica is a bit expensive.

43:08
However, if you’re a long time customer at Informatica, it might make more economic sense to choose them because a lot of what their MDM capabilities are builds right into their existing technology tools, right?

43:19
So you have that opportunity.

43:21
Prophecy is where most of my experience lies in using Prophecy or the D365 which I’ll get to.

43:28
Prophecy is made for one thing, and that’s master data management.

43:32
Some of these other companies and tools have different areas of concentration.

43:37
Prophecy is strictly a master data management tool.

43:40
It’s all they’ve ever been for as long as I can remember.

43:42
It’s all they ever planned on being.

43:46
If you’re a wholly Microsoft shop, might not make sense to go to IBM, right?

43:49
So, but if you’re an IBM shop, it may make might make all the sense in the world.

43:53
And D365, I mentioned earlier that Microsoft does not have a native master data management tool.

44:00
I misspoke.

44:01
There are certain elements or modules to D365 which are customer, customer insights and product insights which I have some experience with.

44:11
If you are a D365 customer, it might very well be worth looking into those because that is natively built into the tool.

44:21
So other tools on the market as well.

44:22
Happy to talk about those offline or whatnot.

44:26
I just chose these as these are sort of top of the forefront right out of what people are looking at.

44:33
So lastly, to sum up everything, what an MDM domain implementation might look like, we covered the definition and prioritization of this, right.

44:42
You’re going to define what that business problem is, identify what the domain is, scope the effort, scope it as intelligently and as you can.

44:54
Don’t sort of, you know, under scope what it might be because it, it does take effort and there is, you know, long term work in implementing MDM, attaining the sponsorship from the business and then building your team, whether or not that’s internal resources that can be committed to this or hiring from the outside, etcetera.

45:13
Is building the correct team to implement this building.

45:16
These are all very similar to any of your existing projects right now.

45:22
It’s profile, the data greatly create that MDM repository and hub depending on your technology stack, that’s going to be depending on where it lives, curating all that source data and getting it ready to be consumed by an MDM solution, building out those data engineering routines and then integrating it with other solutions.

45:41
And the last piece before I hand it back off to Steve is promoting and applying it.

45:46
And I want to say that this not necessarily just promoting it to prod and walking away, right?

45:50
There’s a huge change management and user training and knowledge transfer component to master data management and getting people to understand why it’s important.

45:59
That goes to the very surface level of your organization of people that are frontline facing customers, your analysts that are making decisions, your report developers, All those people need to understand what’s been implemented, why it’s been implemented and how it’s going to benefit them.

46:14
And that is an ongoing process.

46:16
In my opinion, it starts at UAT and then it goes through as long as you’re using and enriching the, the solution, it’s really important.

46:24
And then creating that feedback loop with users to again keep refining, refining, refining on what your solution is and making sure your customers are on board.

46:35
So typically that’s AMDM domain implementation.

46:39
You’re obviously going to get economies of scale after implementing an initial domain, you’ll know some of the pitfalls and some of the advantages and, and building out your second, third, 4th domains are, are going to be presumably a lot easier.

46:52
So with that, I want to hand it back over to SRP and I want to make sure we have a few minutes for questions.

47:01
Thanks, everybody.

47:04
All right, thank you, Greg, Next up, how to get started, right, Everybody wants to know what is the best way to start tackling MDM and in your organization.

47:14
And so in terms of next steps for getting started with MDM, here at Senturus, we offer 1/2 day planning and strategy workshop to help you kind of frame out what your MDM implementation might look like.

47:26
Our planning and strategy workshop really involves 4 pieces.

47:29
It’s identifying and prioritizing the domains that you want to tackle for your MDM initiative, estimating the overall cost and timeline, helping you with vendor selection and, and really building out the business case and the ROI for implementing the MDM in your company.

47:47
So if you’re interested in that workshop and, or if you’d like to just talk more broadly with us about MDM options for your organization, our events team has posted a link in the chat.

47:58
You can grab some time on our calendar and you can get a meeting scheduled with us to talk through workshops, to talk through your MDM needs.

48:06
We would love to hear from you.

48:08
Beyond that, just want to cover a little bit of additional content we have available on our website.

48:16
We do have at Senturus.com/resources a plethora of information, everything from data governance, which is really our focus here today, broadly across the entire analytic space.

48:28
So if you’re interested, check out Senturus.com/resources.

48:31
You’ll find on demand webinars, blogs, and details about our data governance offerings.

48:39
And beyond that, we also have additional resources on our site including technical tips, product demos, reviews of tools.

48:48
So it’s a really valuable set of resources.

48:52
Check out Senturus.com/resources.

48:57
And before we go into the Q and A, just want to talk a little bit about what we do as a company.

49:02
So Greg, if you just jump on to the next slide there, actually the next slide after that.

49:07
Thank you.

49:08
So what we do as a company, we are a data and analytics company over 20 years of experience.

49:16
We provide a full spectrum of analytic services and enablement in addition to proprietary software accelerators.

49:25
We, we’ve been in this business for a long time, 20 plus years, 1400 plus clients, 3000 plus projects.

49:34
So we’ve kind of been there, seen that done a lot from small projects to large projects, everything in between.

49:43
We’re here to support you and your analytics enablement journey.

49:48
So let’s just take the last few minutes here, do a little bit of Q&A.

49:52
We just have a few questions that came in during the session today.

49:58
So the first question is what does it take to do an MDM project?

50:03
Like what would just kind of effort wise, Greg, can you give us an idea of you know, how, how does a company tackle something like this and what should they expect in terms of the effort to launch an MDM initiative?

50:18
Sure.

50:18
Yeah.

50:19
So as I mentioned earlier, I think, you know, the first domain is going to be the biggest chunk to bite off, right?

50:26
Because you’re building your, you know, MDM capability muscle, so to speak.

50:30
So really, again, that’s, there’s, there’s a big planning and strategy piece to determining what you’re going to go after.

50:39
I’ll just use the example of customer cause like I said, it’s the vast majority of what I’ve seen as people’s initial domain that they want to tackle through the requirements for master data management.

50:51
And in order to do that, it’s understanding A, where all of your customers are coming from, B.

50:57
What, how are those, what’s the conformity rules around those?

51:03
And, and you know, what is your first source that you’re going to use as the system of record to apply all the other sources to and conform those sources to that initial source.

51:13
The third piece is where I’m going to give you the big it depends line in the sense that depending on how mature your data estate is and how much you’ve implemented data quality routines and processes around customers and integrated those customers for the purposes of say a customer dimension or surfacing those up.

51:34
For even if you’re doing list creation or the example I use of the clothing retailer sending lists of customers out to other, other vendors, all those things are dependent on the lift it’s going to take to conform those customers to a point where you’re ready to determine what the golden record is of a customer.

51:53
So there’s that aspect.

51:55
And then there’s the secondary aspect of I mentioned implementation is not just promoting to prod.

52:01
It’s also going through the process of educating your users on how to consume master data management or master data, what the golden record is, incorporating that into your sources.

52:14
I’ll give you an example of a company that was their, their call Senturus were heavily using Salesforce CRM and we had to implement back into CRM our conformed Golden record single version of the truth.

52:29
Customers back into Salesforce and then train their inbound call staff to understand what they were looking at because it was somewhat of a shift in their processes.

52:40
Their rules required them to change their scripts and where to go to a little bit.

52:44
So there’s some training and effort around that as well.

52:49
So it’s hard to quantify a number.

52:51
I give you a lot of it depends right there.

52:53
I’ve always kind of baseline it is maybe the same, the same level of effort it would be to create a new say like a sales data monitor something like that, a new analytics reporting type of tools.

53:06
It’s typically very similar to that.

53:08
All right, thanks Greg.

53:13
Another question here question is do you have examples of metrics that have been defined to measure increased revenue or decreased cost related to MDM and the asker says I know the common metric is to show improvements and improved data quality, but that’s difficult to translate to reduced costs or increased revenue.

53:37
Do you have any examples or thoughts on how one might measure the impact on revenue or cost from an MDM implementation?

53:48
Sure.

53:48
Yeah, I don’t have any you know, right here, you know prepared to demonstrate.

53:55
However, if we take again the customer example, all of those, those typical customer KP is that marketers are looking at right now from acquisition, retention, loyalty, spend, customer lapse, all those things.

54:14
The lift after implementing MDM are typically the types of KPIs we’ll look at.

54:20
So six months after an implementation of MDM, you know, are we retaining more customers than we were previously or is there a customer lapse or attrition in that?

54:36
And then also, you know, if depending on what your use case is, you know, the acquisition of customers, are we acquiring more customers because we’re making better targeted marketing campaigns?

54:45
So the devil’s in the details there.

54:47
It all lives in existing KB is that probably exists that that people are looking at.

54:53
And again, product, I’ll use the example of that automotive retailer.

54:56
They knew within that that was one of my favorite stories.

55:01
They knew within three to four months that their services industry was being more efficient.

55:06
They could take on more customers, parts were more available, so the time for repairs was lessened.

55:13
And they knew within, like I said, probably 1/4 that this was making real dividends.

55:22
Is there an example, you know, the first week you implement?

55:24
Probably not.

55:25
It takes a little time to incubate, but yeah, typically those KPIs and taking a snapshot of the way it was and then taking a snapshot of the way it is after implementing it is the key way to see the lift overall from MDM, right?

55:42
Yeah, that makes a lot of sense.

55:43
Yeah, I can see where customer satisfaction, churn rate, that kind of thing is really positively impacted by an MDM implementation.

55:52
I think internally, you know, you talked earlier about operational efficiency and even though that’s harder to kind of measure engage internally, you know, you’ve got things like employee morale and just reduced effort spent on, you know, kind of error correction internally.

56:07
All right, one quick last question here before we wrap up.

56:12
I know we’re getting near to the end of the hour.

56:15
Erica asks what aspects of data governance do you think are most important to have in place before working towards MDM.

56:24
And she mentions that her team is being pressured to provide data for analysis, you know, now, which is very common, of course.

56:31
And governance, you know, kind of seems like an afterthought.

56:34
So what aspects of governance do you think matter upfront before you tackle MDM?

56:42
Yeah, that’s a great question.

56:43
And governance is very near and dear to my heart in recent years.

56:48
You know, I think it’s important to not, you know, governance is, is such a balancing act in terms of there’s a book called non invasive data governance that I’ve subscribed to a lot of theories in that, but it’s really implementing data governance on as needed basis.

57:04
So if you see a specific problem, again, I’m going to pick on customer.

57:09
But if you see a specific problem around customer establishing those named resources, OK, you are the data owner of our CRM, you are the data owner of our point of sale system.

57:21
You are the data owner of any other, you know, customer facing applications that are gathering customer data, identifying who the stewards are.

57:29
Typically those are the people who quote, UN quote, know where the bodies are buried and understand how the data is moving throughout the organization, naming those owners and stewards, giving them a mandate to move forward right through a formal data governance program.

57:44
And you might not want to communicate it as a formal data governance program and you want to sort of finesse that into the organization.

57:51
If data governance is a bad word, like it is at a lot of places, but giving them the mandate to, to own and steward that data and identifying those roles and just having that, that baseline and then joining that with your business stakeholder and sponsor that I mentioned earlier.

58:09
Doesn’t have to be one person could be stakeholder sponsors along with the IT team.

58:14
And, and those, those groups of people coming together from a governance perspective and determining what the overall effort is, is really important.

58:25
And then the one thing that I would say is the most crucial thing and important thing about data governance is data government should be looking at the rest of the organization and determining the priorities at all times.

58:36
I understand this the priority right now and other people need to understand this the priority.

58:41
And if you want to get on the list of things that data governance is going to look at and that could change in the blink of an eye.

58:49
Like I said, if you have some sort of disruption or, or something that happens within the business, it could priorities could change overnight, but understanding what the priority is and just putting those, those base roles into place from an org chart perspective, again, it could be a logical virtual org chart.

59:07
It doesn’t need to be named for HR or anything.

59:11
I think is really crucial to that.

59:15
We’d love to talk about that further if you want to.

59:17
All right, thanks, Greg.

59:21
So we’re at the end of the hour here.

59:23
So let’s go ahead and wrap up.

59:26
Thank you, Greg, for being here today and sharing MDM details with us and answering the audience’s questions.

59:34
We do appreciate all of you joining us for today’s webinar.

59:38
You can reach us by multiple, multiple routes.

59:41
Like Greg, if you go ahead and jump to the next slide, we are reachable by telephone, by e-mail, by website.

59:49
So reach out to us.

59:51
We’d love to help you with your own MDM initiatives.

59:54
Again, thank you for joining us today, and we look forward to seeing all of you on a future Senturus webinar.

1:00:02
Thanks and have a great day, everybody.

1:00:05
Thanks, everybody.

1:00:06
Thank you.

Read more