 Hello and welcome. My name is Shannon Kemp and I'm the executive editor of Data Diversity. We'd like to thank you for joining today's Data Diversity webinar, Best Practices with a Data Management Maturity Model. The latest installment in a monthly series called Data at Online with Dr. Peter Akin brought to you in partnership with Data Blueprint. Just a couple of points to get us started. Due to a large number of people that attend these sessions, he will be muted during the webinar. If you'd like to chat with us or with each other, we certainly encourage you to do so. We also have an icon in the upper right corner for that feature. For questions, you'll be collecting them by the Q&A in the bottom right corner of the screen. Or if you'd like to tweet, we encourage you to share highlights or questions via Twitter using hashtag data ad, data ad, excuse me. To answer the most commonly asked questions, as always, we will send a follow-up email to all registrants within two business days containing links to the slides and the recording of the session and as well as any additional information requested throughout the webinar. Now let me introduce two of our speakers for today. Peter Akin is an internationally recognized data management thought leader. Many of you already know him or have seen him at conferences worldwide, certainly at many of the Data Diversity conferences. He has more than 30 years of experience and has received many awards for his outstanding contributions to the profession. Peter is also the founding director of Data Blueprint. He has written dozens of articles and eight books. The most recent is Monetizing Data Management. Peter has experience with more than 500 data management practices in 20 countries and consistently named as a top data management expert. Some of the most important and largest organizations in the world have sought out his and Data Blueprint's expertise. Peter has spent multi-year immersions with groups as diverse as the U.S. Department of Defense, Deutsche Bank, Nokia, Wells Fargo, the Commonwealth of Virginia, and Walmart. He often appears at conferences and is constantly traveling. And joining Peter today in the discussion is Melanie Mecca. Melanie is the CMMI Institute's program manager for the Data Management Charity Model. A primary author of the DMM says January 2011. She's secured corporate sponsor funding and created methodology for DMM-based benchmarking of an organization's data management capability and has led DMM assessments for the Securities and Exchange Commission, the Office of Financial Research, Microsoft Corporation, Fannie Mae, and the Federal Reserve System Statistics Function, the Ontario's Teacher's Pensions Plan, and Freddie Mac. She has architected and implemented data strategies and organization-wide data management improvement programs and managed programs and projects to design enterprise data architectures, master data hubs, data warehouse components, and shared data services. She has a master's degree in philosophy. Now let me get us forward to Peter to start today's webinar. Hello and welcome. Thank you, Shannon. It's a pleasure to be here and a pleasure, Melanie, to join you again. This seems to be an annual thing with us that we get together once a year and sort of talk about the state of data management maturity. And this is a subject that's near and dear to both of our hearts. So just for those of you that are joining us, some of this material we have had on before. And if you know this, if you haven't been familiar with the DMM, much of what we're going to have here is reference material. So we're going to flip through some slides. And instead, what we've asked Melanie to do today is to really bring us up to date. Because we're going to be assuming here now. We did this the first time in August of 2014, which was the year this came out. And what we want to do now is not just present to you what we've presented in the past, but really talk about how it is evolving. And it's a very interesting story. Again, as Shannon mentioned, Melanie has got very much experience in this. And it's one of those funny things that we both worked in and around the Washington DC area for years and never ran into each other. But I'm very grateful that she sought me out. And I count her on a very good friendship basis at this point. So Melanie, let me just turn it over to you. And let's talk about what is the DMM structure just so that everybody gets it for the starting place. So the DMM was essentially designed around the topic, the broad topic of enterprise data management programs. As we all know for the past number of decades, our biggest problem through the industry with data management has been the piecemeal approach to implementing various practices in an unintegrated fashion. So the DMM attempts to, in five broad categories and 25 different disciplines, or we call them process areas, attempts to express all of the fundamental practices that any organization needs to do to do a great job with its data assets and to make them available and trusted and shared where necessary. So we have essentially five big topics. How does the organization develop strategy for what they're going to do with data management, whether they have a program and they want to improve it, or whether they have a new program that they're launching. We have a complete approach to data quality from top down to the very bottom detail of cleansing strategies and impact analysis. Data operations is doing a great job on requirements, deciding what your providers will be, and maintaining and building data lineage across the entire lifecycle as needed. Platform and architecture is the DMM's advocacy for business engagement in every aspect of architecture, going from the approach to the standards to the platform, et cetera. And data governance, of course, is critical. About 25% of the model is focused on data governance, business glossary, and metadata. And we have five levels. Peter, you want to talk about the levels? Sure. So the real key with the levels is that they're simple enough for everybody to understand. And most importantly, your boss has already understand them as well, because the folks at the CMMI Institute have done a great job of pushing this in a number of areas not just related to this. Melanie will get into that a little bit later on. But basically, I hate to say it this way, but if you have a pulse, you get one as you're starting place. And going all the way up to the top, that means you have the ability to improve your existing processes. And we'll talk about the intervening levels on a couple of other slides. So what is the DMM? It is a big, sick document that is primarily a reference model framework that, as I said, has best practices. One of its main uses is as a measurement instrument, so that organizations can know exactly what they have and where they are right now in their capabilities. That allows you to quickly identify gaps. And then there is a built-in outline of a path to improvement for capabilities and extension of capabilities across the organization. So this was developed with four corporate sponsors and over 50 expert authors over a period of 18 months for the content model and then another year and a half plus of formalizing and putting in the best principles of the reference model architecture and doing a big peer review. And Peter was one of our top reviewers and he's at the top of the list because he has the first letter of the alphabet. Anyway, what should we do next is the purpose of this. And it lights the pathway so that anyone can see where they're headed. Emily, you actually posed for this shot, right? You actually put the lamps on the pathway. I did myself, myself. So a couple of key things about the DMM. It is architecture and technology neutral. So we have been talking to firms and visiting firms with the DMM when they have Enterprise SAS and Hadoop and a very, very, very contemporary architecture. As well as firms that primarily still have main frame architectures, some firms with very large enterprise data warehouse layers and everything in between. So the fundamental practices of managing the data are really not different and they're not technology dependent. We did not want the DMM to be prescriptive about technology. It's also industry independent. The original group of experts was primarily from the financial organizations who, of course, with their regulatory needs, had a huge incentive to participate in this. But we had experts from all over many industries finalizing the model. And now you can use it anywhere. Advertising data, scientific data, medical data, financial data, retail data, does not matter. It is abstracted to the point where you can apply it anywhere. And it focuses on the current state. Because you cannot really plan well for the future or do a good strategy for how you're going to build your capabilities unless you have a very precise view of what do I have here, and particularly, what have I done around the organization that I can leverage elsewhere? That's one of the main values of using the DMM. You surface things you didn't know about that actually are good and can be used. And the purpose is to launch this capability improvement or if you have a capability improvement plan in place to give it new life and energy and vitality. And you will have to sustain this forever because the data assets are forever. If you manage data, the DMM will benefit you. We like to do this in context, too. And to help place it in context, most everybody remembers Maslow's hierarchy of needs from high school. It's the idea that if your food clothing and shelter needs are unmet, you will not likely self-actualize. Self-actualization, you'll notice on the chart there, it's very small, but it says it reaches your full potential. And in order to reach your full potential in data practices, many companies come at you with the idea that you should be working in this golden triangle, but the golden triangle only operates at the tip of the iceberg and that there's an awful lot more. And these foundational practices that Melanie just mentioned to you on the previous slide all occur in that foundation. If your foundation is of poor quality, then the practices become much more difficult. Most organizations look around this and the things in the golden triangle are technology-focused. The foundational practices really talk about organizational capabilities. So when, not if, the technologies change, the organization is more resilient as a result. We get asked all the time collectively, can we do this without doing the DMM? And the answer is yes, you can, but it will take longer, cost more, deliver less, and present greater risk to the organization. And if it's dead, you provide a great foundation for those business results. And if you take a look at the picture on the right, that represents the Buddhist path to nirvana. So you can see in nirvana at the top, which the DMM describes for data management in a sense, people are riding on these beautiful white elephants, the angels are there, and everything is harmonious. And if you look at the bottom, you can see that starting out when you don't know what you're doing with your data and you can't get your arms around and you're spending too much money, you can see that little black figure at the bottom being chased by the evil elephant and there's demons down there. So we're trying to help the organization, wherever you fall in this path, we're trying to help you get to the nirvana state of managing your data beautifully so that your business strategy can be carried out. And what does that get you? Trusted data, it improves your risk posture with the data, data errors, you're not gonna make extremely expensive multi-million dollar errors because your data quality was bad or the data wasn't timely. It definitely improves analytics and Peter knows this well and so do all of you, but when you talk to people who are data scientists or they're working with predictive analytics, they end up spending 70 to 80 percent of their time hunting sources, integrating on the fly, cleaning up bad data, going back to the business to resolve anomalies that they find. And maybe 20, 25 percent doing the creative art for which they were trained. So definitely improves the analytics of readiness of the organization. By doing effective processes and extending them across the organization, you also attain gains in cost reduction and operational efficiency. So cost reduction, everyone wants, if I don't have to develop a data quality profiling process 15 different times over five years in 15 different projects, I'm gonna save money. And operational efficiency is not only good for business but it also assists your employees in being happier and more creative and more calm because they don't have to reinvent the wheel for everything. Finally, good quality data that the DMM will help you manage and build is important for regulatory compliance. So this is really the motivation in order to do this. The next question is how did we get to where we are? And we're not gonna go through the entire history of it but the basics were very similar to Alice working with the Chester Cat. She asked the Chester Cat, where did she go? And the Chester Cat says, well, where are you trying to go? And he says, if you don't really know where you're going it doesn't matter which road you take. So we funded a little bit of this research out of DOD, did some things in that area. There's an original article published out of MITRE that talks about this at a base level but it was a nascent effort and really until Melanie was able to step in and get some corporate sponsorship on this we weren't able to really bring this to fruition. So the CMMI Institute just for those of you who haven't been familiar with it was a spinoff from Carnegie Mellon's Software Engineering Institute which has been around for about 30 years which is a federally funded research and development center. And the CMMI Institute supports all of the CMMI reference models, offerings and services. So we have models for software development and engineering, acquisitions, services, staffing and people and now data management. We also have a new profile coming out which is the Agile profile for software engineering and that's going to be launched shortly. It's doing a pilot right now. And the Institute also has extensive training and certification program in all of these areas including data management and a thriving partner program. So we have organizations who want to use our products, they train, they certify and they work with clients like Data Blueprint and other valued partners. Thank you Melanie. And last, we're now owned by ISACA which is the home of COVID and the CISO certification. They are 140,000 member trade association and they do IT governance and have a big presence in cybersecurity. So that just happened at the end of March and we are very happy to be owned by ISACA and working with them to harmonize and plan joint product offerings. I changed your slide a little too early there but it is wonderful news that ISACA has picked up on this because they're a very appropriate place where this should be housed. So would you like to speak to this one? Yeah, as I mentioned before, your managers know this already and this is the really wonderful thing because people like Melanie have been working in a really dedicated fashion for a long time to promote the CMMI statistics. They have literally 10,000 organizations plus. I'm sure these are older statistics in here. 94 countries, 12 national governments, 10 different languages, 500 different partners. As you mentioned Data Blueprints, one of those lots and lots of folks out there that have gone through the appraisal process and becoming partners certified and you're probably over 2,000 appraisals at this point in order to come up with that which really does give you a good set of benchmarks. So this is just some of the organizations that have worked on this. There is some science behind here as well and I really like this slide because it shows in a decent study the CMMI process outperformed on budget and on time both ITEL, RUP, COVID and PMI. Now you can imagine because the ITEL organization kind of gets outperformed. It's good to have competition with these things. I'm sure COVID is going to go back and try to harmonize a lot with this too. On this, as you mentioned, Melinda, there's a bunch of different components. Again, the services component. So can you deliver services on a repeatable process? Can you develop software? You mentioned the agile process is coming out. Acquisitions process and workforce development. These are all flavors that management understands. And when you talk to them about CMM, CMMI, they go, oh yeah, I know what it is you're talking about. And that's important because a lot of people will come to you and say, why should I do this? What is the point in looking at this? And again, this Melinda gets to your and my personal motivation of being coming involved in this. I think you're still on mute, Melinda. If you want to put in place a good solid program to better manage the data assets, get the most out of them and engage everyone in the organization from a data entry person all the way to the CEO in appropriate responsibilities and involvement. You need a strategic effort. So it is not a project and as Peter's been saying for years, data management is not even a program. It's more of a lifestyle. It's more of a change in culture. It's a change, proactive change in behavior. So the reference model was designed to help organizations build that. And we have written it such that it does help unify the understanding and the priorities of business staff, IT staff, and the data management organization. And it is a foundation for collaborative capability building over time. So the timeline on the right, you already know this, we started it early and from January 2011 until August 2014, we were in development and now we are taking it forward and with our partners, we're launching it across all industries in the US and abroad. So again, it's been out for about two years and you've gotten a lot of good experience which is gonna now help to build the groundswell of support that we really need to have. So the question of who wrote this and who got into it, this is something else that's very critical to the component. So there were a lot of grayheads. I'm gonna use that. I'm a grayhead so I can say that. Lots and lots of grayheads who were passionate about data management involved in this and giving really unstintingly of their time. We estimate that if you added up the time of everyone who worked on this, the model costs would have been approximately four and a half million dollars to develop. In order to get anything into the model, any author would have to prove that it works. So when we were having these discussions, 90 minutes, two or three times a week, working extensively, the test was, what would this do to implementation? How does this practice that you want to include in the model? How does it affect business? How does it affect IT? Is it implementable and manageable by the data management organization? So everyone was drawing on all of their years of experience, well over 1500 years of experience in the primary author team. So why did we even do this though? We have been passionate about data management our entire careers, the author team. We needed it badly. We wish we had something like this 25 years ago. Looking back, we know we could have done a great job for the organizations that we worked in or for. So we wrote for ourselves to have this kind of capability to let everybody come to consciousness very quickly on data awareness and their capabilities. And we wrote it for you to use to benefit your organization. So Melanie, one of the hats that I wear is that I'm the past president of Data International. We always get this question, how does the effort of what you're doing tie in with what Dama is working on? So first of all, all of the authors are familiar with Dama and we're big fans of the DIMBOK and we always used it in one form or another in our previous practice. And we still refer clients to it if they wanna say, well, just tell me what are the levels of governance that I should be implementing? We say start with Dama. And then we refer them to 50 other books. But the tools are designed differently. So the DIMBOK is a body of knowledge which contains distilled practices for implementing various data management disciplines. And the DIMBOK also contains things that are more technical than within the scope of the DMN like content management, data design. So a lot of those areas of data security in the new DIMBOK too is coming out. A lot of those areas are for the data management professional as a whole with all of the disciplines that the various data management professional roles may encompass. Whereas the DMN is based on the enterprise program and leans very heavily towards maximum business engagement and measurement of capabilities. So that's the difference. We're developing, it should be out the next month or two, a certification, a DMN associate based on our three day or 10 hour web intro course. And this will be recognized by Dama International as one of the optional modules for the CDMP. So we have very good relationship with Dama. And the real key there is of course that no one organization has a lock on what's actually happening out there in this space. And that we at Dama recognize the fact that the TMM is a tremendous addition to this as well. And we're looking forward to these further collaborations as we go forward on this. Yeah, just another powerful tool in your toolbox. Exactly. And again, we're certainly doing a whole lot better by working together than we were trying to put competing standards out. That's kind of like having two watches, right? So we've talked about motivation and how we got here. Let's dive in and talk specifically about the DMM. So while it was released two years ago and represents, as you said, 1,500 years of collective focus that's in there, I don't know that we really need to talk about the practice areas and things. I'm assuming people have heard that. But let's talk about how it is really the focus of you are what you do, Melanie, because this to me is one of the more powerful aspects of the model in there. There are objective ways of measuring how this stuff all comes about. A lot of organizations, let me take a step back to answer that question, a lot of organizations feel bad about the state of their data management because the executives and the key practitioners on both the business, data management, and IT sides of the house, they feel like they've been whistling in the wind for a long time. They can't get traction. And this is because people are doing the same thing the same way. They're not thinking about extension and they haven't all come together to realize everything that's required and that many hands make light work. So what the DMM actually does is it does help the organizational culture insofar as the data assets are concerned. It emphasizes behaviors and that is repeatable, sensible, sound, clear processes to do certain things like populate a business glossary or do a data quality assessment for fitness for purpose that you can then leverage and extend. So it creates calm, rational approaches where a lot of it was scattered around in projects. It also brings people together to make decisions because there's governance activities everywhere throughout the model. When you do these processes and you're performing them after you have established them, then it results in work products. So when we're evaluating an organization against the DMM, we're looking for do you do these things and do you have a sample of work products like the following ones, which are typically evidence that you are doing the practices. Now that said, the process areas, each one of them are standalone. So we worked very hard. We will, in the text of the process area, we will talk about where it has, like the business glossary has touch points into quality and to governance and to data requirements. Of course it does. Everything is indeed connected in the end because you have one data layer in your organization usually. However, you can look at these practices according to their own definition and see how they're going separately. It also is very flexible because certain organizations only concentrate on certain areas of the model. Let's say you have an organizational mandate to improve data quality. If you want to, you can just use those four data quality processes which are self-contained together to focus only on data quality. And we do indicate relationships among process areas because everything operationally in the end has some kind of connection or dependency. And two years ago when this was introduced, it was theoretically going to work. But what's happened in the intervening two years is you've seen a tremendous adoption of this process and a super validation of this. I've not encountered anybody in the past two years who said it's wrong. Everyone tends to like it because it's very encouraging. It has as a premise, just like in self-help, you know, the self-help books that say wherever you are, there you are. So there's no blame. Everyone knows that organizations have spent lots of money in various places doing this, that and the other about data management and that the efforts over the past decades have not been well coordinated. And there's no judgment or blame about that. It says here's where you are now and here's where you can go from this point on. If you get Melanie and I alone away from the actual formal presentations and apply us with an adult beverage or two, we'll tell you some stories about it. But what it really comes down to, Melanie, is that data management is currently done in virtually all organizations well at the workgroup level. But the workgroup level is not the organization-wide processes and what we say is imagine the power of having all those workgroups actually trying to do this and being able to integrate their information from workgroup to workgroup. What they spend their time doing now is flipping data back and forth between workgroups and that's really death by a thousand cuts for many organizations. So our process of looking at this, we've mentioned it before, one for performed, two if it's managed, which is the idea that there's an ability to say, this is what should happen. Once you have it managed and you have the opportunity to define it, you get three points for defining it. Defining it usually implies some kind of documentation implying that we can now standardize things out there. If we get to the process of being standard and consistent, then we can start to measure things and it makes a difference. And once we are measuring things then we can move to the next level, which is optimizing them. And this is the basis for TQM or ISO 9000 and all the rest of these process improvement frameworks. Again, one, two, three, four, five and you can see on the right-hand side of this diagram the idea that as you go up in levels, your risk decreases and your quality increases. That the ability to do ad hoc querying, which is what most of this looks like, changes to the ability to reuse data, which is of course its highest and best use in there. And that the stress level in the organization decreases as the organization gets there. You may have heard of a fire call or an alarm bell or something like this. Melanie, we had a client that we were working with recently that will just pretend they sold Harry Potter books around the world. And the problem was their systems weren't optimized well. So they could tell how many Harry Potter books they sold in any specific country. But if you ask the question how many Harry Potter books are being sold all over the place, it was a two-week fire drill. And oftentimes they came back and provided the right answer to the wrong question, causing lots and lots of stress from the organization. So again, these are all things that are gonna help the organization as you increase the competency and the capability in this area. So I just wanna take about 30 seconds because this topic comes up. What's capability, what's data management maturity? So we're gonna disambiguate that quickly. So capability is the ability to carry out something, to achieve something. So we do that. So the data management maturity model is a capability evaluation model in that it measures are you doing these practices and have you documented the processes that you're doing? In other words, do you have some work products? Do you have a documented process? Do you have guidelines, templates, standards, policies? All of those things would be typical to create if your behavior has changed. So it's chicken and egg, work products behavior, work products behavior. So that's capability. Maturity in the CMI world in the architecture has a different dimension. That is about stability, strength and repeatability of the processes themselves. So under stress, can you still execute them efficiently? And those are very sensible supports for the process like policies, resourcing, planning, training, quality assurance. So it's two levels, capability and maturity that the model can evaluate. And Melody, where would the word attest come into here? Because this is some of our financial products like that word, right? Work products and maturity attestation would be the next level down where you would be looking to verify that these processes are actually being performed the way that you set them out. Those of you that are in the financial community know the importance of that concept. It is something that people are in fact being held accountable for. So this slide talks now about how the actual large, as you said, document, don't let the word large scare you. It's not an immense document. It just has a lot of material in it. Once you understand the structure of it, it's pretty easy to get around in. And the Denbock II is currently sitting at over 750 pages prior to editing. So we're nothing compared to the Denbock II. So this diagram just shows you that what is evaluated is whether or not those functional practices are being performed. There are 414 of them in the entire model all the way up through level five. So are those practices being performed and the work products you see are not scored in and of themselves. They're just used to provide evidence that the processes are being performed. Now if you have an enterprise-wide governance implemented and you don't have a charter and you don't have any action items and you don't have any meetings, but you say you're doing it, we would question our point of contact and say, you must have documented something. But it is an indirect relationship. And then for maturity, what we call infrastructure support practices, like the examples I gave, policy, training, resourcing, et cetera, also for maturity evaluations have to be satisfied. Right now, most people are using the model as a capability building accelerator. What have I got? What should I do next? And that's what they get out of using it. Next year we will be introducing a more rigorous audit method, possibly based on our current one with CMMI. And also it might have elements of COVID. We're still working out exactly the approach we'll take, but it will be completely auditable. Yeah, what that gives you then is the ability to look in an organization. And what we're gonna do on this slide is just take you sort of gently through this. I always put the fireworks on here just because we in the community were so grateful that the work was actually done here for this. So this is showing the next level of decomposition. Melia mentioned there were 418 individual components, but this will help you to get a sense on it. You know, if you're not managing your data coherently, you're managing it incoherently. And of course, nobody wants to manage anything incoherently. Similarly, if your data assets aren't being managed professionally, then they are managed unprofessionally. And of course, we would prefer to have our sole non-depletable, non-degrading, durable strategic assets managed, excuse me, managed professionally. Melanie, you mentioned a calculation for fit for purpose. Again, it's easy to say that. You've actually got some techniques in there to show how an organization can figure out what fit for purpose is for specific data uses in there and whether the architecture and lifecycle implementations are going to be corrected whether you have the appropriate organizational support in here. So the next couple of slides, Melanie's going to briefly talk about each of the five areas up there. Will the supporting processes come in at the very end? So it's actually six total. Here's the first one. So the data management strategy is about the organization-wide program. Do you have a strategy is the first one? We think also that communications is extremely important in data management because as we said, you want to manage data management like you would any other critical asset. Human resources, finance, facilities, you're going to manage them forever. So you really need to have a lot of people engaged in one role or another and you need to communicate with them in such a manner that you can sustain all this good work forever. So we talk about the communication strategy and its importance. Governance to IT, IT to governance, data management to the program, to everyone, et cetera. The data management function is going to separate the data management role proper from governance. So governance is collective decision-making, particularly about shared data or shared concerns. Whereas data management are the backbone of persistent data products, such as the enterprise data model, the business glossary, the metadata repository, quality processes. And we talk about how the data management function can be implemented effectively. Business case should align with the strategy and data management funding should ensure that there is continuous and nondiscretionary funding for the program. We've run into organizations where there have been wonderful work products and wonderful achievements done and the funding died at the end of the project so it gets dropped and all the work is wasted. Maybe you're not done yet, Melanie. That's really the question both of us run into on a regular basis. Right, absolutely. So governance is very key and governance management is our name for governance structure, how you set up the governance groups to maximize engagement and minimize time and increase the value of collective decision-making as well as what you do to help foster informed decision-making on the part of participants like training, for example, governance training and also metrics and development. Business glossary is the business metadata that is a part of metadata but it's pulled out because business meaning around shared data is critical. We were at an insurance company recently and they said that there was a differentiation that was not made the same way between product and program and that was bubbling all the way up to their financial reports and to the CEO reports on performance. So it was a critical issue and there are many examples of very, very major disagreements on the use of business terms across business areas. So that is by itself so that the business can pay a lot of attention to it. Metadata is everything else about metadata, categorization, agreements, level of depth, technical metadata, process metadata. So everything that you need to know to have the basic fundamentals for putting into place sound metadata management is in that process area. Data quality, and I have to refer to Larry English here who is my first guru as well as the guru of many of you about data quality, so he total information quality with his term and we have replicated that through many, many years of iteration and program experience into these four process areas which we think are a total 360 degree approach to data quality. Have a strategy. Data quality, right after the overall program strategy, the data quality strategy is probably one of the most important things the organization can accomplish because data quality is always vital, always of concern and needs attention in and of itself. Otherwise you have improvements here that don't flow down to the data warehouse. You have improvements there in one hub and then the same data is in another hub and it doesn't get the same treatment. So it can be chaotic and expensive. We wanna help the organization bring it into harmony. That's what the strategy is for. Profiling, we know what that is, discovery of what is going on in the actual physical data stores and does it match the metadata or descriptions that we have of it? And if not, what defects do we find? Data quality assessment is a business-driven evaluation of fitness for purpose, and we also emphasize the setting of thresholds, meaning how bad can this data quality be and I can still do my business process and targets which are aspirational targets. So in healthcare, for example, you would really like to have 0% duplicate patient records so that you never make an error in care for a patient because you can't differentiate that patient from another one. Data cleansing is not just cleansing but also business process improvement that prevents data errors. So it's doing the cleansing work and doing that in a sensible and strategic manner so that you're saving as much money as possible and cleansing the most important data stores and areas and data sets first and also how you go about changing your business processes to have better data. Platform and architecture is fairly intricate. We have five different process areas about it. Don't build an architecture for the target data architecture until you have involved the key stakeholders in the approach to the architecture. So we've run into organizations where let's say enterprise architecture has come up with a target data architecture and these, you know, are many, many year projects that, you know, in the transformation of data stores over multiple years. And if the business is not thoroughly engaged in this from the beginning, it can be rejected. You know, or someday in year two, you might surprise me and say, this is the year you have to redesign your entire database. Well, it wasn't in my budget. You know, you can just imagine the organizational issues you get into. So the model tries to help you avoid all of that by having a sensible approach that is organization-wide. Same with data standards. The data standards for many, many years have primarily been set by IT. We want to put more business engagement in that and get them to care a little more about it. So that essentially is data standards. The data management platform, we were recently at an organization as an example of this. And the IT group was doing very good job coming up with features and doing a selection for a reporting tool and supporting BI and visualization tools. But they hadn't involved the business and it happened that these business individuals were highly experienced in reporting and SaaS and many other technologies. And they needed a voice because they had certain key requirements that weren't really being met by the direction IT was going. So let's prevent all that by following the processes and practices here. Data integration is essentially to ensure that the business is engaged and informed on what is happening with data integration. And probably the best example of that, and you all have familiarized yourself with this, is a master data hub or any kind of data store where you're bringing together multiple sources and multiple business lines supplying and consuming the data. You really have to get it together on how you're going to integrate the data and what the decisions will be made. Historical data archiving retention is important. How many times have you tried to pull up something from a backup and it's not there? So we want to make sure that there's business continuity that records retention is done according to internal or external legal and regulatory requirements. And the next area is supporting data operations which is essentially doing a great job at specifying data requirements. We do a very good job across our entire industry of doing good functional requirements specifications, whether it's waterfall, agile, or any other approach. And we don't do a good job correspondingly of specifying data requirements. So this talks about best practices for that and how to make sure that you are doing a parallel good job with the data requirements. Data lifecycle is mapping of data to business processes. You don't have to do that for every business process or for every data set. But for your critical processes, especially when there's handoffs from business line to business line, it's an extremely good thing to do and we talk about how to do it. As well as tracking data lineage which can be very critical for regulatory purposes or for accurate data integration. And provider management, we've run into many, many firms that have problems in that they have so much external data. The contracts are not standard. The SLAs are not standard. They haven't been able to get their data quality and performance requirements into those contracts and SLAs. And it costs a large firm, probably many millions of dollars of the lack of efficiency in this. So the practices to do that better are in provider management. You're gonna need to drink a water after this last one. That's right. Now these are adapted from CMMI development because they've applied well in life cycle and they're not applied very well with respect to data but they're equally as valid. So metrics and measures, everywhere we go, we find that people sometimes have qualitative measures but they don't usually have quantitative measures. And not doing that, you're shooting yourself in the foot because you can't track progress among your peer groups. You can't report progress to your management and you can't bubble up the achievements and progress and improvements to the funding executives. So this process area has tremendous information in it so that you can start getting a set of metrics program-wide and by discipline, measure them and monitor them and get your money. Okay, so process management. The policies, processes, work products, templates, reusable report, reporting standards, et cetera. How do you manage those and how do you make them available to those who need them? Do you do quality assurance for key processes? Like how well did you do the data requirement specification for this data warehouse evaluation and audit capabilities of the processes themselves? How do you mitigate schedule and budget risk for data management projects and programs? And finally, for all the work products and assets that you're creating, have you configured them properly? An example that we all can relate to is many times in releases of an operational data store, the developers and DBAs will change the data store to reflect the new requirements in the release but you never have what the database looks like now today because they don't bother updating the physical data model. That's an example. And Melanie, correct me if I'm wrong, but these supporting processes are the same across all of the CMM product line, which means if your organization already has another CMM component in place, they're doing these for them and the question should be, why aren't you doing these for our area as well? Yes, because as we know the systems development lifecycle as the engine of new functionality for the organization has always had more attention and more attention has been paid to it because any failures in delivery or budget are bubble right up to the CIO. So that has a long history of attention and now we'd like to give the same kind of attention gradually and as it makes sense to the data layer and the data management. And if you'll remember from our previous slides, this actually does result in better, faster improvements than other frameworks that put it in place. So let's talk about how organizations can use this and you've got one more long slide here before we let you take a break on it. Okay, so obviously this first bullet, if you're going to establish an organization-wide program or you're going to greatly accelerate or move your program along more rapidly, you really want to use the DMM prior to doing that because it will very quickly tell you exactly where you stand and many of the gaps and difficulties will fall out like ripe fruit just from looking through these practices and finding out whether you're doing them or not. And then just keep showing bullets. Well, we'll make this very quick. So yes, if you're going to redesign your whole architecture, the third bullet deserves a little attention, establishing data governance. When we're at governance conferences, people come up to us and they'll often say very quietly on the side, we've tried governance a couple of times that it's just not working out. We don't know what really to do with the governance groups and we're having trouble. So the DMM's approach to governance, we are not the consultants or the model that's going to tell you exactly how many groups you should have or that it should be the same as organization be. We're going to tell you how you engage the people and the kinds of decisions they need to make and how you need to nurture that along your path. So governance improves when people have a stake in it, when you have accomplishments for the governance groups, when you have projects for them, like let's say this year, a retail firm is going to use their governance stewards group to standardize customer information for a master data store. That would be a very proactive, engaging and interesting use of governance groups. So we say accomplish things and use the DMM to help you and they will be energized. And many other things you can see. And as we said, for data quality program, bullet five, you can just use those four process areas if you want to concentrate on one particular set of disciplines. So you get a lot of different types of uses that can be done. And what this gives us then as practitioners in the field is on the left-hand side are the five areas that she had, plus of course the supporting practices. And each of those gets evaluated on that one-to-five scale based on the actual provable pieces that are able to demonstrate. Again, the word attestation comes in a lot. And this gives us now the ability to put together a set of things. So here are a couple of ways we've used these. This is one we did for the insurance industry a couple of years back. We were looking at them and trying to see exactly how they were looking. And the group that we had done here, this was a group of about 30, I think, in here, showed that there was room for improvement around there. So here's the entire industry can get behind this and say, this would behoove us as an industry to do a little bit better in that type of a process. Notice we're not labeling any one specific organization in here. Here's a very definitive airline that I did. Excuse me. You can imagine the executives in the room going one, one, two, two's into one. What does that mean? And I say, well, here's the competition. And they all of a sudden realize they're the ones and the competition is a two and it probably isn't a good place to be behind the competition. Melanie, we're talking here today, the day after Delta airline went down. Of course, we're all curious to see what happened there. Nobody really knows the answer to that, but gosh, if it was a data management practice area, that'd be a really interesting story, wouldn't it? And of course then, as Melanie mentioned earlier, the way forward, what's next? And in this area, because we've already mentioned it a couple times, the foundation is only as strong as its weakest link. So rather than making the data stewardship and data development pieces higher, instead what should happen is that the other areas, all the ones should become twos before you try to make any of the twos into threes on this. I mentioned, we hadn't mentioned any specifics of which airline or anything like that, but the World Bank told us we could use these results. They had actually asked us to do their treasury group, their information systems group, and the business component of it, and the business here actually produced world-class results. So for many years, these were the highest results that we had gotten out of any particular group. And again, we can come back in and throw in the industry competition and what are the benchmarks that look relative to this. I also use this in a paper. Most of you may or may not know that we're in the era of post-big data. So the big data era is over. And the reason big data era is over is because it worked about as well as most IT projects, which was to say not so good. But the idea was in 07, we had a certain level of maturity, and in 2012 we had another level. And these areas, while they look relatively close and are actually showing slight improvements, they are statistically unchanged. And that's important because of course, we all know the volume of data increased an incredible amount in that same amount of time. So this really is all about measurements, Melanie. And the point is measurement gives you what? That's your point to say confidence there. We didn't rehearse this one. Measurement gives you confidence. There you go. Tried to set you up. Yes, we've done a good job. We know where we're going. We've improved our data quality in certain key data elements by 25% from our profiling and assessment processes. So it definitely does help fuel enthusiasm and energy on the part of everyone involved. And it also, as I said, in the end it helps you keep your funding because you can demonstrate tangible improvements with business benefits. So the assessment method that we developed along with this model was designed to maximize participation from everyone involved in contributing to or managing the data assets. And to get them to think together, to collaborate together, to reach consensus about where they were with respect to these practices. And to dive a little deeper with executives or sometimes EA or sometimes business architecture or sometimes internal audit to anyone with a big stake in the data management program and look at work products. So all of this happens in a very short time. The way we've structured it is a very high intensity one week at the organization who is prepared by getting people together, choosing their participants and gathering work products. And it allows you within a scope of a couple of weeks to develop a very substantive report with very specific scores, specific gaps, specific recommendations for improvements and always at least eight to 12 initiatives that you write up as a mini business case that the organization would benefit from acting on in the near term. So let's switch the slide. And as soon as you're done with the evaluation, you get among other things this chart which shows you if level three, which is our target for maximum efficiency, it's like the 85-15 target for all organizations in all process areas, that would show you the blue line where this organization scored in these areas. And it would be substantiated by the scoring spreadsheet and all of the notes and all of the work product evidence. So it's one page, how am I doing? It makes it very easy to come up with road maps because it becomes obvious not just to you whoever is leading the assessment and your team, but to all of the participants together. They'll say, for example, in one area, they'll say, you know, we really haven't done a good job at metrics and they'll look at each other, they'll all nod. So there's a tremendous amount of agreement that is easily attained through this process. Literally getting everybody on the same sheet of paper. Yep. I wanted to spend a minute here. This is a moving average. So those score ranges are the scores that all of the organizations we know about. And there have been organizations over time who have come up to us at conferences and said, we just bought a bunch of copies of the model used at ourselves internally and built our whole program from it. Some large organizations that you would recognize if I said them, and we're very happy about that. But for those who either are partners or selves have worked with directly in evaluating against the DMM, this is what we have so far. So there are some superstars. You can see those really high bars in a couple of areas. Some organizations have really pulled out all the stops. They've done a great job. And you'll see how low it goes, right? So if you look at data lifecycle management, and part of that is data lineage, you'll see that there's some organizations that have barely scratched the surface. And the triangles show our moving industry average, how the industry is doing currently. And certain areas are better than others. Note that governance, one, two, three, four, five, I think it's the sixth line from the left. Governance is pretty good. Most people have instantiated governance and they're moving it forward. And other areas like measurement is low. Data lifecycle management is low. And business glossary is low because they usually have it in chunks from business line to business line. And they haven't undertaken the discipline of standardizing it for the shared data. So those are some observations. Peter, do you have observations on this credit chart? Yeah, it's just wonderful that we actually have facts and figures now, which is relatively something that we didn't have prior to this. So we're gonna just got a few minutes left, Melinda, I'll let you talk a little bit about how DMM helps the organization overall in this general way of thinking about things. Common language, same terminology. That's often a problem in data management projects that people are using the same words to describe a discipline or a process. So at the end of the evaluation, everyone will have a common language that they'll then understand each other better. The work products, there's 596 work products in various places through the model at each level in each process area. So it gives you a really good specific idea of the kind of things that other organizations produce when they're building and implementing these capabilities. Then the model itself with the five levels has a path built in for improvement. And the practice statements are easily understood by all parties. So it helps you accelerate what you're doing. So we're really at the top of the hour, Melanie, and I think it's important for everybody to understand too. This is just a portion of how CMM can help in these organizations. There's really an entire eco structure that you all have built out that includes training classes, certification programs, et cetera, et cetera, that can help out. Obviously more information on your website in terms of getting people started on the program. And I'm sure we're gonna get a couple of questions that we have on this. You yourself are committed into this as well to continue to produce the results, the best practices, the case study, the types of things that are coming out of here so that we as a industry can continue to move forward and make progress, bringing more professionalism, more discipline, and frankly more rigor to what it is that we do here in data management. Absolutely, and we, of course, we love case studies. In fact, that's our favorite presentation at conferences, is one case study after another. So we did one in Enterprise Data World with a small advanced startup, NaoA Business Solutions from Brazil, one of our partners conducted his first assessment there, and they did a wonderful case study on governance and agile development and how the DMM helped them improve their data quality, during the middle of high-impact agile development. So that was very good. We have one coming up at Financial Data Governance. This is Fannie Mae, where we visited twice once two and a half years ago and once recently. And their case study is going to be on governance and the integration of enterprise data management, best practices with agile delivery, and they're having tremendous success. So we get very excited about all of the firms that use the DMM. And be able to pull them in and actually make them able to be used. So while we move to our Q&A session, I'll just mention the next event we're going to have is on September 13th with Dr. Stephanie Byrd, my good friend who's going to get a great title, How to Get Stuff Out of Your Crystal Ball and the way of predictive analytics. And Melanie, we've got your contact information on the slides here. You are absolutely fine with people giving you a shout over the e-mail and things like that. That is the way we want to get the word out if you'd like to find out more about this. And now we have a real good Q&A session, hopefully, and we can answer some questions around what we either didn't present clearly or what you'd like to dive in deeper on the model. So Melanie, thanks as always for doing this with us. Thank you. And thank you both for this great presentation. And of course, we do have a lot of questions coming in already. And if you do have a question, feel free to submit it in the Q&A section in the bottom right-hand corner of your screen. And to just answer the most popular question that comes up, just a reminder, I will be sending a follow-up e-mail within two business days. So for this webinar, by end of day Thursday, with links to the slides, the recording, and anything else requested throughout. So let's dive right into it. Can we do the DMM without additional financial outlay for training and consultants? Yes, just as you can do data management without any guidance whatsoever. However, to get the best results. Go ahead, Melody. Yes, well, the model, just like the dimbox, the model is sold for a modest fee. So an individual copy for your use alone is $100. And we have multi-user discounts. So you can get to DMM, you can use it for yourself. And if you use it, let's say you're a consultant. If you want to use it all on your own with an organization, the only thing that's needed is that that organization needs to buy copies of the model for those individuals who will be using it or referring to it. So yes, you can do it on your own. We like to think that it is easy to understand for those who have been in the field for a while. You have to have a fairly embracing view, though. You can't have just done one thing, because if you have, if your career has only been a few years long, and let's say you've only been engaged in data governance, you will have a big learning curve in terms of using this for your organization or another one. Because one lifetime is way too short to learn everything that we need to know for data management. Yeah, and another example might be somebody that has had a lot of experience with DBA and data operations, might not be as versed in the data strategy area as well. So while it is a good set of reading materials, your recommendation is that they get some guidance in order to help do this properly. Yes, and we have a web-based intro course that will take you very thoroughly through all of the data management concepts in the DMM, and that's easily available. And by web, it's basically an eight to 10-hour self-paced course that gets you started. That's on that website, right? Yes. There we go. And, you know, I think you just really answered that, and maybe we can get a direct link from you, Melanie, that I've included in the follow-up. There's somebody interested here, also in the three-day, ten-hour, you know, the DMM certification. Excellent, okay. That's on that to you. Perfect, and I'll get that out to everyone. Next question coming in. Peter mentioned that operating data governance with working groups is inefficient. What is your opinion on starting data governance with work groups and then expanding as the organization gets accustomed to the idea? Boom, boy, I'll have to go back and listen to the recording so they're really said that or not. I said that in many cases, data governance can be perceived as inefficient if management doesn't see a value that's coming out of it. It is on us as data management participants, professionals to demonstrate the value of what we do overall to organizations. And that we've seen a criticism of data governance be that it is often seen as not really focused on achieving business values, and that can be a problem. So, Melanie, let's go back into the model here. How does the model help us keep governance focused in on something that will be perceived as useful and hopefully necessary in some cases? Well, we recommend that when you, let's say the organization does not have a well-established multi-layer governance structure in place. We do recommend working groups as the best way to start data governance. In other words, if you're, as many people ask us to do and other consultants to do, prove the value of data governance. So the best way to prove the value of collective decision-making and collaborative effort is by doing something. So I've used this example, I think, already in the presentation, but let's start putting together a business glossary for our customer data. So if you have a working group with business representatives from all of the key supplier and consumer areas of the company, and you can come up with a scheme for a business glossary and show those executives that this is happening and that the disagreements from misuse of terminology or different uses of terminology are being solved, that is a great way to energize governance, give people a collaborative work experience. Those people can then easily make the transition into becoming permanent data stewards. So that's an example. I'll give you another one, Melanie, too. We're working for a small savings and loan at one point, and we're actually in a business meeting where the senior leadership was trying to present the executive committee with information that showed that sales were going both up and down. And of course, we know that's impossible, and it turned out it was a definitional term. And so that's one of the reasons they said, and that's why we're starting this governance effort, because I cannot have you guys presenting this information to the board. It does make it look like we don't know what we're doing. Absolutely. Love it. So the next question is, you know, can you talk a little bit more about the difference between the DMM and DEMA? Peter, I'll let you take this one. You're talking about the differences. Well, first of all, this effort was set up initially funded by the Defense Department that became part of Carnegie Mellon University, and then the CMMI Institute was looked to commercialize because, as we all know, universities haven't had as much practice commercializing commercial property as businesses have. So DEMA has really, as a volunteer organization that I've been affiliated with for, I think, as long as Melanie's been in the business here as well, a volunteer organization, we put a lot of our efforts into the DEMBOK, which was the data management body of knowledge in here. And in parallel, Melanie, you were working on this in a way, and what we said was, let's both agree to get our products individually out to the marketplace. So the purpose of the data management body of knowledge if you haven't seen it, it's the circle that most people are familiar with that gives people what it means to do data management. And we think it's a tremendous contribution to the field because 20 years ago, if you'd asked 10 different people what data management was, you would have gotten 10 different answers. Now we've put something out there that we can rightly criticize and improve, but at the same time, it's a great starting point. Similarly, Melanie, you were doing this, and we actually had a conversation at one point about, you know, should we try and integrate these things before we send them out? And the answer was no. Each of these is useful in a standalone component. So the DEMBOK tells you what are the things that data management does, and this talks about how data management practices are improved in there. Do you want to add anything to that? No, I think that was an explanation. So, you know, and no one should be afraid of a new framework that would be my true advice. You are a professional and you're a practitioner, everyone on the call, and whatever you can find that's useful, you should, you know, check it out, and if it seems valuable to you, make it your own and use it in your work. Just like the mile-and-a-half high stack of literature that's been accumulated on best data warehousing practices over the last 30 years, so data management is the same. Put it in your toolbox. Should I say we should fear this less than we fear the new release of Microsoft, whatever it is they're going to put out us? I would say so. That's not, but we love Microsoft, don't get us wrong, we actually have a Microsoft case study we didn't get a chance to get to today, but yeah. So again, we're collaborating, and we believe that we'll be able to harmonize these things without an extreme amount of difficulty here. So this questioner says, I may have missed it. Does the DMM have equivalent of ITIL service catalog? No. That's a quick answer. Now, we have had organizations ask us to comment on data management services that some of the more robust and mature organizations would like to offer internally, and we have helped them to determine what those data services should be based on their strengths and where they were in the model. So you can use it as a baseline for services, but that was not its primary purpose. Sure. And it looks like that's all for the questions. Oh, here we go. So we'll back to the difference between the DMBalkan and DMM. Are there companies using both? Do you guys have... Peter, I think especially you might know. Yeah. No, absolutely. Melanie, you can counter the DMBalk when you go in with places and I encounter the DMM when I go in places as well. Yeah, absolutely. Yes, absolutely. Recently for us, a large federal agency had been using DMBalk for some time, and they wanted to apply the DMM for evaluation and as a check on the performance of specific practices. They liked that aspect of it, the measurement aspect and the evaluation aspect. So they're in total harmony. We shouldn't be saying anything. I think in DMBalk 1, there's nothing in the DMBalk 1 that would contradict anything in the DMM. Yeah, I certainly agree with that. And you talk about harmony, we'll go back to our Nirvana slide, right? So really the goal with all of this is to help organizations get from the elephant chasing them to them riding the elephants. You think about the elephant as a perfect metaphor for what your data is. Most organizations are probably lower on your diagram here, and this is a process to say, we can give you a roadmap to get you up to the top and how to tame the beast. I love that comparison. Next question, I just joined our company to help with our data. I come from a data governance department in a company on one ERP, but this company uses multiple systems. Is it possible to get our data cleaned across multiple systems? What pitfalls should I watch out for? Is it possible to get your data cleaned? That's an interesting question. So how can the DMM help with that? Yeah, we could apply the DMM to that. Essentially what you would need is to get the key business representatives and the data store owners in agreement that they're willing to do some work across these different systems. You can get those business data stewards to help you work on the quality rules that they want to apply to it, and you can decide which of those application data stores constitute the point of origin or the multiple points of origin across the life cycle, and that is where you would be gaining the agreement to work the data more intensely as it is considered to be a source. Then you can apply the quality rules and the data quality assessment process. You can profile the data in the key data stores of interest, and then you can cleanse and improve the data so that the DMM would take you through those four areas of concentration. And let's add one other piece to that too, Melanie, which is a real good description on it. Let's look at this as well. If you approach each of those four ERPs at level one, it means you will have people doing data cleansing operations, but the risk will be higher. The ad hoc-edness of their approach will be greater, and the stress levels that they'll be dealing with will be later. But by applying the DMM model to this, they can concentrate on reducing the risk, increasing the reuse of the methods in cleansing across the various systems, and to help, again, go back to your Nirvana comment earlier, and find the best way of cleaning data across their systems. And again, if you have to do something, if you have to clean data, and most of us do at some point, it'd be better to have a standard process for doing that rather than simply starting each, oh my God, there's a bad data element out there. How am I going to clean it? You know, that makes no sense whatsoever, and get us what most organizations are faced with. So it's exactly the type of problem that this framework is designed to address. And it looks like that. Maybe all the questions we have for the day, anything else that you want to add, give people a second or two if they want to add anything else? No, anything? No, thank you so much for having me. And I wish you luck in improving your data management capabilities, and you might find that the DMM is a very useful tool for you to help do that. We do, and we thank you for your contribution in there as well. So we'll see everybody next month then to talk about crystal balls and how to get stuff out of the crystal balls, right? Don't smash it, Peter. That's all I have. Don't smash it. There you go. Thank you again, Melody. Thank you, Shannon. Thanks, everyone. See you both again for this presentation. And just a reminder, again, I will send a follow-up email within two business days with links to the recording and the slides of the presentation and the additional information requested throughout the webinar. And we hope to see you next month. Thanks, all. Bye.