 Hello and welcome. My name is Shannon Kemp and I'm the Chief Digital Manager of DataVercity. We'd like to thank you for joining today's webinar, Top 7 Capabilities for Next Gen Master Data Management, sponsored today by RELTO. Just a couple of points to get us started. Due to the large number of people that attend these sessions, he will be muted during the webinar. For questions, we will be collecting them via the Q&A in the bottom right-hand corner of your screen. Or if you'd like to tweet, we encourage you to share highlights or questions via Twitter using hashtag DataVercity. As always, we will send a follow-up email within two business days containing links to the slides, the recording of this session, and additional information requested throughout the webinar. Now let me introduce to you our speakers for today, Ajay Khanna and Erin Zorns. Ajay is the Vice President of Marketing at RELTO. His technology expertise stems from various leadership roles at large public enterprise software companies, including Viva Systems, Oracle, Khanna, Progress, and Andox. He holds an MBA in marketing and finance from Santa Clara University. Erin is the founder and chief research officer of the MDM Institute. Prior to founding the MDM Institute, he was founder and executive VP of Meta Group's largest research advisory practice for 15 years. Erin received his MBA in management information systems from the University of Arizona. And with that, I will give a floor to Erin to get today's webinar started. Hello and welcome. Thank you so much, Shannon. Now that I'm off mute, that was fun. Okay, so today we have a great topic. It's What's Next? That's the new year. It's time to think about resolutions or revisiting resolutions. It's time to think about what do we want in our MDM package over the next several years, our MDM solution, our master data management, and governance-related software. So that's the topic of our one-hour session today. I'll do about 30 minutes as an industry analyst to bring together some of the high-level concepts and trends, technology trends and business trends. And then AJ will give us some use-case examples from his real-world experience. And then lastly, we've saved 10 or 15 minutes at the end for your questions and answers. You're encouraged to, again, pop them into the Q&A window. And Shannon will moderate those for us at the end. So let's carry on. So the MDM Institute, some of you may not have heard of us, did around 15 years basically. We work with very large companies to help them work on their requirements for master data management, reference data management, and data governance. We help them do their pricing, their negotiations, their reference checking, et cetera, et cetera. Again, we have 150 large companies that are our counsel that get free unlimited consulting to that effect. And as industry analysts similar to Gartner and Forrester, by the way, my company Meta Group was acquired by Forrester 15 years ago, which is what I spun off to do, nothing but master data management, was acquired by Gartner, I should say, my company Meta Group. So as an industry analyst, what I do is I also take positions, just like a stock analyst. When I say something's going to happen in the next one or two years, here's what you need to do. Something's going to happen in the next two to three years and three to five years. So here's what you need to know and what to do about it. So next up, on the right-hand side, you'll see some of those companies, international companies, some pro bono work we do with Doctors Without Borders companies, but mostly very large corporations. Westpac, the very bottom one, is the seventh largest bank in the world out of Sydney, Australia, and so on and so on. And I do work in both San Francisco, I have two homes, San Francisco and Munich, Germany, and so I work out in both. So the top seven capabilities that we're going to discuss today, we're going to look at cloud, we're going to look at microservices, we're going to look at graph, big data, machine learning, artificial intelligence, action enablement, and then information or data governance. And in particular, our thought here is that all enterprises need to focus on what's next, because our current crop of widely installed MDM platforms, the IBMs, Oracles, Informaticas, etc., some might even call them legacy systems, they were not designed with digital transformation, the digital enterprise, etc., in mind. And so we've had to sort of patch things together the last five or so years to really get there. And in particular, the notion of system of record as a master, your golden sample, your customer 360, to add also something called system of reference, and likewise to get us into system of engagement, where we're keeping track of the actual transactions and their complex relationships among the transactions and among the consumers and citizens, etc. So we're going to be looking at relationship driven analytics more and more. You hear that word a lot, cognitive applications, everything in context. And that's what we really need to do is to have MDM built into our systems or available as microservices to our systems, so that we can take that dream, that target that we're a future state looking for, the digital enterprise, that basically everybody, the markets, governments, etc., all need to do this. Now cloud, of course, has been around a while. And again, I apologize for the analysts speak, but the text you see on this chart is basically an analyst strategic planning assumption from the MDM Institute. And what we're saying is that the small to medium businesses have definitely picked this up and also the large enterprises to do some of their MDM. Now smaller enterprises tend to do this because of cost, because of labor, the ease, etc., whereas the larger enterprises are looking at scalability and power among other things. So partly why this has been very attractive, we all know it's because you pay as you go as opposed to a big lump sum of several million dollars. You can get away with several hundred thousand a year, etc., etc. Likewise, there's challenges, though, when we look at how do we do federated architectures or blended architectures, you know, the co-terminus type where we have both on-premise and cloud. And then we've got private cloud and public cloud. How do we bring this all together in terms of an organizational architecture structure, etc.? And then we've got other areas where native cloud, as opposed to things that were, you know, as we say, shift and replace, simply moving an app into the cloud, you know, doesn't do it or a system piece of software, doesn't necessarily cut it. Now being designed with cloud in mind, not just cloud pricing, but, you know, overall the mega vendors, IBM, Oracle, Informatica, SAP are still struggling with cloud because in particular the pricing model takes a huge cut to their earnings. Whereas they're looking at a much smaller upfront payment, even on an annual revenue recognition model, it's the challenge for them to get there. And so it's put a lot of stress on the large vendors in particular. And we're also having challenges with, you know, I won't say run a month, but sort of the Wild West cloud approach where organizations functional and enterprises functional organizations such as sales or marketing or research or engineering, customer service will go off and buy or rent their own cloud apps. And then we have that challenge where a sales force, SAP business by design, work day, success factors, you know, trying to bring all that together is a challenge. And at the same time, you know, privacy, not just because of California Privacy Act and not just because of global data protection, right? But also because the fact that it's a huge experience and loss of market capabilities, you know, and financial funds, et cetera, where we have these privacy intrusions or hacked systems. And so we really have to look closer at how well, MDM being the keeper of that golden data, okay, how it addresses privacy and security. And that's an ongoing challenge for most organizations as well as the vendors. And then reference data is perhaps an exception of those static tables of codes, ISO codes, ICD, international classification disease, the various industry standard codes and private codes within your company, those tables of codes, those are something that do definitely fit into a cloud model. That's something that definitely works and is low risk in terms of putting out there on the cloud for sharing. So our bottom line thought is that cloud economics are very compelling with their challenging both for customer, employee, and citizen data, but they are a key sales enabler. They do allow all of us to move into MDM easier in terms of a test bed, in terms of an evaluation model, and in terms of delivering. But at the same time, they'll remain technical challenges. And likewise, even business challenges getting business out of some of those sales force installations, managing multiple hierarchical orgs across multiple discreet sales force implementations in your large company, for example. Now microservices is another thought of the seven trends that we'd like to talk about. And in particular, microservices is basically deconstructing. When you think of a deconstructed taco or whatever food, this is the same thing here. We're deconstructing MDM down into its granular components. And of course, it's various degrees of granularity. There could be components such as high level components such as order to cash or onboard new customer or resolve telephone conflict, et cetera. And then there could be the very micro crud type read update, just simply insert or read or update or delete. And so what we're seeing is that MDM and data governance, which are often blended together in a software vendors package or solution or platform, MDM and data governance are increasingly deconstructed into these components and to these microservices that are executable as web services and other capabilities. So when we see that, then that starts getting into the space of some of the mega package vendors like SAP, Oracle, et cetera, and four, where we see that the basic functionality that MDM is providing in data governance has historically perhaps been the domain or bailiwick of an application package. So we've got a little bit of that going online, not to mention business process management and its new hype cycle type name, robotic process automation, RPA. So those are also something that has to be thought about when you look at the core functionality of your next generation MDM. Now we're seeing that graph, just to insert graph here a little bit, also provides a lot of the capability to crosswalk across these micro components or microservices as well as across domains. We're going to talk about graph more later. But basically the thought is that data driven or MDM and AIDs, that is something that has MDM built into it rather than as an afterthought, the market size of that is going to exceed that for the raw MDM platform software. And so when we look at microservices and terms of the seven, top seven evaluation criteria for next generation MDM, that's the de facto architecture we want to see in our MDM solutions. Now graph technology has been around a while. Franco's, you know, Neo4j, et cetera, et cetera. There's a bunch of different graph technologies out there. And just like there were a bunch of SQL or SQL databases originally, there is now some standard efforts to unify a common query language. There are some other efforts to congeal or create a more or less industry robust set of capabilities that comprise a graph database in particular. Okay, so a graph database, as you probably know, allows us to manage very complex and hidden relationships and or hidden relationships. And in particular, it's very good at doing analytics across the relationships. The intersection of these domains. And so it's very good for modeling the real world. In fact, it's a simplistic way to model the real world compared to the foreign and primary key relational database data models we had historically. So when we look at graph technology, it does provide an ease of use in terms of working with the stakeholders and business users. It does provide a lot of power and flexibility. There have been some concerns about the scalability when we get into large scale systems in terms of inserts and deletes, et cetera. So we have seen a lot of, if you like, analytical MDM or batch style MDM capabilities within the initial graph capabilities. But again, this is a key technology. It's a query that has to be there. It's an interface, a UI, a UX that has to be there. So because the users expected Google like or LinkedIn or Facebook type interface. And when we look at Google search system or Facebook or LinkedIn, we see what a graph database functions as, how it can manage those relationships and follow and traverse the intersections of the planes and so on. So graph database, graph technology itself, the UI and the underlying database, which sometimes can be separate, sometimes together, the UI and the underlying database. This is providing the missing link between domains so we can crosswalk among customer and product, for example. And it also provides big data analytics and also gives us some fork, if you like, foray or entree into Internet of Things and Internet of Everything and also some people might call it. Now, big data, we all know as data management professionals that big data is simply the ongoing slog that we've had all of our professional lives where things are getting deeper, more complex, and lots more of it. So big data for however you feel about it evangelical or in terms of dogma, it's here, it's coming, and it's going to continue to impact our systems as well as our customer and our stakeholder experiences. The problem is that most organizations are getting about a 30 or 60 or 90 degree view of their customer, product, supplier, citizen, etc. So the challenge is to get rid of those blind spots and to be able to bring in data out of the big data, sometimes which is not, in a tagged, metadata, tagged type format. And we have to be able to either, after the fact, go in and tag that big data or we have to tag it as it streams into the data lakes and so on. And so there's a challenge. How do we reincorporate big data which is sort of strayed from the structured view of IT? It's vastly unstructured for most organizations. How do we reconcile that, repatriate that back so that that data provides and rounds out and augments the so-called single view of X where this customer, product, supplier, citizen, good guy, bad guy, whatever. So we find that big data is key. We find that MDM provides the capability to do identity resolution, that is to match the data over in the big data lakes or data markets, etc., with the data in your master data systems, along with your transactional and analytical subsystems. So it's very important that big data be there to allow us, I'm sorry, that MDM comes along to crosswalk, if you like, across the domains again, the length of domains, and also to even clean up and tag the data going into the data lakes. Now AI and machine learning, likewise something that's been around a long time, I don't know about you, but 30 years ago I did my degrees and back then we were doing neural nets and all sorts of AI fun things. The challenge was the compute power and the databases needed to train these systems. It was pretty funky back then in my old age, early age experience. So where we are now is we have, you know, AI being used for data profiling, to understand the relationships as a sophisticated software crawls across your data landscape to identify the relationships and the sources and targets and so on. We also have AI, Artificial Intelligence coming along, to assist as sort of an expert system in the AI language or parlance to have an expert system to help data governance stewards in their day-to-day work to eliminate a lot of that repetitive work that they have to do and to also capture it so that it becomes rules in an expert system. So clearly the AI machine learning is needed to help us get there when we have these more complex and larger, much larger data landscapes and much more complex data landscapes. And so not just the data scalability, but the human side to be able to deal with this as architects, as users, as data stewards or subject matter experts and so on. So it's not going to replace, it's basically going to augment and make us do what we do better, faster, easier. So when we have a bottom line thought about machine learning, it's really all about scalability, complexity, and agility, which is some of the problems being solved by machine learning when it comes to data management, especially governance and master data management. So action enablement. It's one thing to have analytics tell us something, it's another to turn those analytics into action. And in order to do action, you need some sort of workflow integration with your operational systems, with the UI, the UX on mobile platforms, set tops, car dashes, et cetera. And so action enablement says we need to somehow bring back together business process management, robotic process automation, et cetera to bring that stuff back together because the process management tools need data and the data management tools need process management. And here before they've sort of gone their separate ways architecturally and religious origins and so on as how they've split and diverted. So how do we bring it back together? Well, you'll see that many of the modern MDM platforms have sophisticated workflow built in and workflow that's able to integrate with the enterprise workflow as well, not just to kick off an email or a semaphore or something, but also to be able to integrate at multiple levels, RSS, Java message, you name it, it has to be there. So action enablement is a critical capability that we see in the more modern MDM platforms but also as a retrofit, you know, when IBM buys various BPM vendors and Informatica and Oracle, now they're basically making all that work as well. Meanwhile, we've got people like TIPCO and other BPM, software AG, other BPM vendors that have also started blending in MDM into their overall platform for BPM. So the thought on action enablement, our bottom line thought is that from the enterprise perspective, if we're going to have a real modern MDM solution that we want rules and we want reference data and we want metadata, we want it all to be applied across domains and that requires some good amount of workflow or process automation. On our next chart, last chart for me, it looks like, in terms of the seven, I could have numbered these and that would have been better for all of us, but in terms of information governance compliance. Okay, so we all know that data governance is people, process technology is all that happy stuff that allows us to administer, you know, functionally the management of our data assets, you know, capture them, grow them, retire them, manage them, et cetera. And so data governance itself needs that end-to-end lifecycle capability. You know, we need to be able to onboard data and then roll it off, et cetera, at the end of a seven-year or other legal life. And likewise, we need workflows that will support, you know, bimodal, both IT-specific and business-user-specific capabilities. And we're going to see that the data governance winners that were out there here before, including Calibra, who just got, what I guess, another 100 million yesterday or so, and are now valued at one billion, they're a unicorn, Calibra. But look at them, they don't have the MDM integration. Okay, and so what we do need is MDM and data governance integration. And that's what you're going to see in a modern data management platform. You're going to see sufficient data governance to also drive the MDM side and likewise sufficient MDM to drive the data governance. I mean, what's the use of doing data governance if the end result is another PDF, Excel, Visio, whatever? You know, you need action, you need integration, proactive integrated data governance with MDM. And so that's what we're talking about here, is that the mega-vendors themselves, you know, they're a solution as well as a lot of other data governance platforms, as well as, you know, CSV, Excel type, input output type integration, as opposed to round-trip push-a-button or round-trip hyperintegration through RSS, JMS, or whatever it might be, to give us that integrated data governance capability for our data assets. So a possible action plan for the next two years. We all know that we should be promoting MDM as a business strategy, not as an IT refresh, okay? So, you know, we find the low-hanging fruit, we go out there, you know, cross-channel marketing, omnichannel marketing, data governance, you know, compliance, those sort of things, so fine. It's a business strategy. We also want it to be more than just reporting and communication. You know, it's got to be operational at some point. You know, it's okay to start out with reporting and analytics and compliance, but we need to get into operational transaction as well. We're probably going to start either with party or thing, either with customer-citizen or with product supplier pricing, et cetera. And we should, while we're evaluating our MDM capabilities, we need to keep our eyes out on the radar in terms of what's the next two years, fixing metaphors there a little bit. We need to look at, you know, not what we need now, but also what's going to be hitting us in the next two years in terms of this increase, move into, you know, digital everything. And so it was not just from the mobile experience, the handheld experience, but likewise the integration of IoT for certain industries and likewise the integration of just the complex extended supply chains that comprise the modern corporation. I mean, those systems, like your Apple, HP, IBM, Lenovo, whoever, you've got a lot of business partners in your supply chain that you have to integrate with. And that's a huge challenge. I mean, you're not just a solo entity. These days, you know, you're hyper, you're being increasingly integrated with the rest of the world. You know, we're not just talking about, you know, Amazon or Yahoo or eBay doing your website of your company and you running the brick and mortar operational systems. And it's not that way. We're talking about your whole supply chain being increasingly call it fragmented, call it deconstructed. Now where we get the best price, where we get the best capabilities, and where we get the needed capabilities to address the consumer or the business user's requirements. So when you're looking at MDM these days, you may need a certain MDM just to do reference data or just to do compliance. But if you're looking at an enterprise level solution, you're going to be looking at multiple domains. You're going to be looking at linking that together. You're going to be looking at making that a modern system as opposed to a legacy architecture where you don't have data governance integrated, where you don't have workflow integrated, where you don't have graph. Modern data management, here we've given you seven evaluation criteria or seven capabilities that you can use to have conversations to structure your confirmation, your conversation with the various people that you interact with, whether it's your implementation partner, whether it's your software supplier. This is a good outline to structure a conversation. Seven good things that we need to know about. And so, again, to revisit the concept, in my 30 or so minutes here, I'm at the 25 minute mark, so I'm doing pretty good. All enterprises need to focus on next-gen requirements because we're moving from the older system of record, golden record, customer 360, et cetera, to system of reference. And increasingly, we're ultimately moving into system of engagement. All those transactions, all those interactions that we have to analyze that we have to manage, and increasingly the dirty detail that we'll get into IoT. Now, relationship-driven analytics implies graph because the older relational database systems and hierarchy management systems, et cetera, don't do the relationship-driven that we need, the analytics and the UI, the user interface. And so, this implies graph is something that we need in order to have those cognitive apps and in order to help ourself move into the next generation of digital enterprise, what that's going to look like. Where can you get more? Well, my organization for 15 years has been doing summits around the world. We do like 3,000, 5,000 people a year. Singapore, Tokyo, et cetera, Sydney. And there you go, that's us. And again, here's where you can find my organization, myself on Twitter, on LinkedIn, phone, email, et cetera. So now, I'd like to head it off to... I was surprised to hear the job title, VP of Marketing because I think of Ajay Kanna as a very technical, savvy guy who's been in the market for a long time. But I'll just call him a RELTO's representative to discuss use cases relative to the seven evaluation criteria for modern data management. Okay, Ajay? Thank you, Aaron. Thank you so much. Yeah, this was really, really exciting discussions that you just kind of went through. And it's just a really exciting time to be in data management, right? I mean, data is no longer just an exhaust coming from various applications and system now. We all kind of consider it as a enterprise asset. So, yeah, I'm pretty excited to be in this space at this point in time. So thank you for the seven core capabilities for next gen. So the attempt from my side here is to kind of discuss some of the use cases and how these technologies are coming together in a modern master data management platform. So let's go and get started. And I see a lot of messages from people which are from really cold areas. So, friends, please stay warm from Chicago, from Boston, and from St. Petersburg. So let's go and get started here. So when we speak of modern data management or modern data management platform, it goes beyond core MDM capabilities of matching and merging and creating that single source of truth. We also talk about including graph, as Aaron mentioned, to find the relationships, just like LinkedIn, finding the relationships between customers and the products and the stores and the parts and the suppliers, or some of you guys who are coming from life sciences and pharma space into this session, it will be finding the relationships between a physician and the hospital system, patient and the plans and the payers and the prescriptions, so on and so forth. So just like LinkedIn or Facebook, it provides a new context across various data domains. Next is workflow and collaboration capabilities, very BPM-like. You can design your processes about data change request, data deletion request, data access request, especially when they are mandated by regulations like CCP or GDPR, we must adhere to those. Analytics and machine learning, not just to get next best actions like you would see in Netflix or Amazon, people who like this would also like to buy this, but also improving the data governance part, also improving the data quality, also improving the data matching within your system. Data as a service where you may want to consume third-party data, how easily you can consume third-party data from vendors like Denon Bradstreet or in life sciences vendors like IQV or MedPro. So those are the capabilities that are required today as part of the model platform and then deliver that information to business users for action enablements. That's where you deliver all that information to the business user in form of these data-driven applications where they can actually view the profiles, where they can actually work on various tasks generated by workflow, where they can do the collaboration and get insights into the information about certain profiles as well as certain relationships. And as I mentioned that we need to deliver this as a big data scale, bringing in on each other interactions and transactions and incorporating some of the newer technologies like microservices, architectures, containers using Kubernetes so that we can leverage this multi-cloud architecture. So I will get into a little bit more details into each of these. So first let's see how these things kind of come together and work. So I'm starting with point number one from the left-hand side. The first step is to organize your data. You connect to various data sources, your internal, external third-party data sources like Dunham Bradstree or IQVR or from your CRM systems or even like public data, right? Whether it is data like NPI data or any kind of like data from postal services and then you merge that data together and refine and reconcile to create those consolidated profiles and then find the relationships, relate those entity peoples to people, people to products, products to locations, to stores, et cetera. And also correlate that with omnichannel transactions and interactions. So if it is a consumer, what have they purchased? What have been their web visits? What kind of emails they have opened? So all that information is also correlated with the master profiles to create the true views and then make it available to the underlining analytical system where the machine learning, whether it is Spark-based or whether it is in GBQ kind of environment is utilized to gain further insights about that information, create further insights about that data and bring those insights back, bring those insights back to as part of the master profiles and then as a step number three, visualize that information in the contextual view whether you want to deliver that view to your marketing people or account reps or field service agents or call center agents, deliver that view about that product or the customer or a consumer or a supplier in procurement in the context of their business rules. So everybody get the relevant view which is as per their business goals and objectives. And then also making it more operational. So all this information is then made available to your downstream, any kind of analytical or operational systems. So when we talk about storage approach, so there's a lot of different type of data which is relational and which is kind of RDVMS type and then we want to maintain relationships which is more suited for graph databases then we have a lot of interaction, transactional information then we have a data on big data scale for which you want to use maybe something like NoSQL Cassandra. So an ideal approach will be this multi-model data approach where you store different type of data in the storage which is right for that particular data type and whether we are storing relationships in graphs and then reference data in something relational like RDVMS, we have interaction, omnichannel transactional files in some storage like S3 and related to those records and then bringing all this together in a contextual way where you can provide relevant information and create those reliable data profiles of any kind of data entity whether it is a customer or an B2B organization or a supplier or a product. So specifically talking about the graph so what are the use cases? So here are some of the use cases that we see out there. First is bringing in for example data from Dunn and Bradstree to create legal hierarchies. The hierarchy across this organization what are we just business units within this company and then creating custom views of those hierarchies. I want a product penetration view of my customer or I want to see what are the green spaces and white spaces in my customer organization where I can do more upsell and cross sell or I want to create a hierarchy where I can manage my sales alignment my zip to territory to product kind of alignments or I want to do risk roll-ups or I want to do value roll-ups across the hierarchy. So when you have this graph technology and have these relationship across various entities now you can do very sophisticated hierarchy management within your data profiles. And then some ad hoc kind of for use cases where you want to create a profile of your customer, B2B customer and you want to see what is my account team like or what is the procurement team or approval team like in my customer organization or in live sciences. So how does the formulary team looks like and so all those kind of relationships can be used to create these teams and committees which then become part of your master data profiles. And then some of the other use cases around affiliations and relationships are to identity resolution or to find an influencer within your network like who is the most influential or key opinion leader in this disease area or who is my most important influencer in my B2B customer who should I talk to or who can introduce me to this particular person what kind of new product should be recommending to this particular person. And so these are the kind of use cases even householding for example in retail where they want to understand that okay I'm marketing to this particular person but what does the family look like I mean is this person married, do they have kids and then based on those insights you can market much more personalized offer much more contextual offers to your customers so householding is another important use case for using graphs. Getting into machine learning so some of the things that we see out there are like the lot of companies are investing a lot into data science and bringing on data scientists and the challenge has been that the data scientists are spending 70 to 80% of their time doing data cleanups or doing ETLs so the whole idea here is first of all making that reliable data available for these data scientists so that they can run their algorithms and in an agile way in kind of like shared object wave if you add more data attributes to your MDM profile those are easily available to the analytics data models and the data scientists can just focus on their algorithms and bringing that business value and when they get those insights back in a close loop those insights or those recommendations are pushed back to those master profiles or those actionable applications for business users. For the use cases side we see these three categories of use cases one is around business use cases which is in the center getting the next best action that what is the right channel to engage with this particular customer whether it's a consumer or whether it is a physician in life sciences or what should be the next product to recommend to this particular person through which channel to engage in business use cases where you need these reliable data and integrated machine learning algorithms to make those recommendations and then finding those relationships suggesting new relationship and then third is around data quality so this is like really important that how to use machine learning to improve the data quality itself so the first use case of machine learning is to improve the data where you can use it to score your profiles rank your profiles what I mean by that is for example you can score your profile for data quality that if I have a consumer profile I mean have we done the address verification do we have the all the data that we require have we done email verification or phone verification do we have the right licensing information NPI numbers for this particular profile so all that things can be collated and given a score where you can see that what is the completeness or richness or of this particular record from the data quality perspective or you can rank the profile from business value perspective right so is this is the right target segment for me what is the business value for this particular customer and then you can juxtapose these values business values and data quality ranks to see where should your data stewards focus on on to the people who have high value but low scores and then how your marketing should be using this data for better segmentation to run the next campaign by using these scores and rank and identifying other data issues suggesting new matching rules so that's how we see some of these use cases where machine learning is being used to improve the data quality itself so now just kind of going a little bit back here so before we move forward so here is like now we are creating these profiles whether it is for a person or an organization or a product we need to connect to multiple data sources right internal external third party data sources but as part of the governance as part of the whole audit history transparency requirements today we need to make sure that we have the right data lineage that which data is coming from which particular data source and if we are using particular attribute value why did we use that where did we use it where did we get it from and it is important from data quality perspective and it is also important from compliance reporting perspective so we are bringing all this data together from internal external and third party data sources and then correlating with omnichannel interactions and transactions so that we can get the complete view of profile for action enablement now your profile has information regarding the demographics if it is a person regarding particular insights what is their channel preference what is their product preference what kind of products they have purchased in the past using relationship what kind of households do they belong to using relationships you can see what kind of devices they use to visit my websites or what are the various interests that they have and again providing various scores like what is the data quality score what is the business value of this particular profile what should we recommend them next so all those things are becoming coming together so that you can now provide this information to your field service who is making maybe going to do a house visit for fixing something or it is this information is delivered to your point of sale system so that your store person can print out a relevant coupon or it is able to authenticate so this information becomes much more actionable so this is kind of bringing together the analytical aspect as well as the operational aspect of these profile together to provide that action ability and then last piece I would hit on is the whole idea of governance and compliance including BPM kind of capabilities where you can design BPM and 2.0 based process models so that you can kick off these workflows whenever the data change requests come from your field users from your business users and even be able to push these requests to your third party data providers right so these kind of processes which kick off certain processes which kick off these workflows for review of data changes and then once the data change request is reviewed and approved it goes to the data store they make those changes and then persist those changes so that your data is continuously data quality is continuously maintained your data is continuously improving and you are addressing all the requests as per mandated regulations such as GDPR general data production regulation where you want to manage the customer data in a proper way you want to understand that where did the consent come from where did the data value come from again this is where the graph technology comes into play this is where the correlation with interactions and transactions come into play as well as information governance come into play where you want to address some of the customer change requests like access to data requests like data changes request like data deletion requests right to be forgotten so all those things can be possible once we are kind of bringing these things together in a modern mass data management way where you are leveraging correlated transactions where you are leveraging graph technology where you are leveraging workflow and BPM kind of capabilities in a one in in kind of like cohesive way so at a higher level I have like a couple more minutes left so at a higher level if you see some of the industry use cases in the life sciences you will be thinking of these technologies to maintain reliable healthcare provider profiles or hospital data healthcare organization HCO profiles you want to understand the affiliations across plans and pairs and HCPs and HCOs and patients and prescriptions and even compliant product 360 with IDMP compliance or even later on managing patient experience which also leads to healthcare use cases where you want to maintain right patient experience right member experience as well as maintaining the reliable provider directories as per CMS center of Medicaid and Medicaid compliance purposes and then a lot of use cases in retail are around digital transformation are around delivering the connected customer experience across the buyer's journey are around addressing the household issues as well as adhering to use cases or regulations such as GDPR and very similar use cases you will see in financial and insurance industries as well and some in like high to high tech kind of industry where the selling is mostly B2B it is understanding the hierarchy of their B2B customers doing hierarchy management maintaining the sales alignment calculating the sales or distribution or distributor compensation sales effectiveness and even supplier 360 kind of scenarios where you want to create complete profiles of suppliers and to see who is your top most supplier or which supplier or is the supplier maintaining the right compliance requirements so all those kind of use cases are can be addressed by bringing these technologies together so these are just a sample of use cases and definitely we can discuss more when we deep dive into a single industry so just kind of a recap the idea is to make a data heart of every decision by organizing the data whether it is internal external third party data or interactions and transactions bringing it together understanding the relationships and then delivering that information to business users for action ability in personalized views infusing analytics using machine learning or AI as part of those business processes and then using that information to continuously improve your data continuously improve the customer experience continuously improve the business outcomes and the idea is for your organizations to gain competitive advantage better provide better customer experience improve the operational efficiencies within the organization and then reduce compliance risk across the organization so with that I with and my presentation and if you want to learn more definitely you can contact us at any time or visit relative.com we will also be at conference called modern data in San Francisco so registrations are still open if you are in the area or if you are interested definitely go to the website and see what all this is about and we will be discussing some of these use cases with some of the top pharmaceutical companies and top retailer and consumer and high tech industries in that conference as well so that's it from my side Shen over to you Yes, thank you both and Aaron for this great presentation just a reminder I will be sending a follow up email to all registrants within two business days so by end of day Monday with links to the slides and recording of this presentation by end of day Monday to everybody and if you have questions feel free to submit them in the bottom right hand corner of your screen we did have a question come in earlier so one of the major complexities in MDM is growing is growing list of suspect match pools how do we tackle it in new gen MDM So I can take that one and then Aaron definitely please provide your thoughts as well so I mean there are multiple ways for suspect matches or what we also call like potential matches so there are kind of discrete rules then there are fuzzy logic and smart matching rules and you can also use machine learning to improve that matching as well so for example there are you added a new data source and there is a match happening and system says that this is maybe like a 60% match in that particular case but it is not a high confidence level these kind of matches will go to a data store they can eyeball it and say that okay we should merge these records and then system in the background is learning and identifying those patterns and then saying that okay you have been merging such kind of records so this is the new rule that we will suggest we want to persist this rule so that is one part of improving your matching using machine learning and then second aspect is graph as well right so for example if we do not have these kind of fuzzy matching and we do not have like right keys but we can use a relationship that this person like John Smith and Jonathan Smith and Jay Smith has the same relationship with the same person in the same kind of zip code so even though the data is different maybe this is the same person and we want to train our machine learning to identify these relationships as well to improve those suspect matches historically Aaron here historically we have had both database driven or data driven and and probability driven matching okay and so that has been the historical means and sometimes the various MDM platforms could use one or the other the trend now is to use both okay that is you know to match against historical databases for accuracy synonyms whatever and likewise to use thresholds probabilities etc to fine tune a system which over time will change in terms of what is a suspect perhaps or what is a duplicate or what is a false positive or whatever so we've got a number of ways that we could do this now the good news is that you know machine learning AI and all that which is not the end all everything but it is being applied to help train these to create these databases if you like to curate the databases for the the deductive matching versus the probabilistic matching the probabilistic matching those rules are often are increasingly I should say are being determined or found by our expert systems by the machine learning processes themselves so these are areas again where where AI and machine learning can help us do better what we've done in the past do it faster do it more accurate and make it take those mundane tasks out of our hands but at the same time kick it out as an exceptional or appropriate so that you know a steward or analyst can review the suspect to make sure that it does either modify the rule or does go in the right database or not go in the database for onboarding etc next question yeah everyone's very quiet today so we have questions in the bottom right hand corner but we've got some other questions here do you really need modern MGM for GDPR do you want to take that well there are various degrees of compliance with GDPR yeah and like a lot of American companies are sort of sticking their head in the sand even though they're doing business in Europe especially smaller companies because of the cost of just bringing in consultants and determining how much GDPR capability we need to continue our business in Europe or whether we should drop Europe because it's the cost of servicing Europe versus the cost of GDPR compliance for Europe anyway back to the big question you can do GDPR with any number of tools I mean we've had all these you know we've had anti-money laundering databases good guy bad guy watch lists all this sort of stuff in the past MDM itself is you know fundamental to GDPR you know where do we have data on what who's allowed to access it when was the last access you know the audit trails the protections of the data etc the ability to tell a customer you know that this is where we're keeping data on you and why we're keeping it the ability to purge a customer when you get a forget request okay from a European consumer etc so these are all classical capabilities that are simply magnified when you get into a GDPR scenario and I'll pass it over to AJ at this point but again any fundamental MDM can do it the good news is that with graph we're able to suss out a lot of relationships of other data sources about that individual or entity that we weren't necessarily explicitly aware of the hidden relationships and the hidden data files that are out there when we get audited we get in trouble so I'll pass it over to AJ just to carry on that thought thanks and I agree I mean you can go ahead and do the basics but the whole idea is that understanding the data lineages as well how easily we can understand that okay if a customer asks that okay where did you get this consent from right amen which is permissible in GDPR we need to be able to tell them that okay you provided me this consent from this particular website or this particular digital asset or so that kind of information you need to be able to provide pretty quickly or if you are interacting with a minor for example like 12 year or 15 year old kid you need to say that okay which parent gave consent for that particular interaction whether it is for healthcare purposes or whether it is for other commercial things and so that's where the whole idea of graph technology also comes into play to manage that particular consent and be able to trace back and see that we did that particular consent came from or we did that particular attribute value came from and as we discussed earlier that having little bit sophisticated business process management kind of capabilities workflow and governance capabilities are also required as part of these systems and then propagating that information back to downstream system so that we can either eliminate that data or anonymize that data as per the requirements is is required as per GDPR so making the system more efficient to meet GDPR requirement I think that is the key how well we can adhere to the requirement and how easy we can make it for our customers as well as for our data stewards and business users and marketing users making easier for them to be compliant and be confident that when they are running their campaigns they are pretty they are following the regulations and they are compliant so I think that is the essence as well. So one of the major complexities in MDM how do we tackle it in the new gen MDM? Sorry he is just repeating his question there so we kind of tackled that already I am just re-reading questions so let me move on then what are some of the early adopter industries for modern MDM? The ones that we are allowed to talk about are the retailers especially online retailers we of course we can't talk about certain government functions but you can imagine that there is a lot going on to track down bad guys and the relationships among bad guys across many different operational systems but in terms of the early adopters in classic MDM the early adopters were TELCO finance, financial services banking insurance high tech and pharma I think AJ has seen a lot of pharma stuff with his company so I will pass it over to him but those were the classical MDM early adopters and from my experience watching uptake of products like RELTO I have seen a lot of pharma in particular a lot of biotech You are right Aaron we have seen a lot of quick adoption on life sciences whether it is pharma, biotech, medical devices and again the relationship played a big role there because they want to have the right data of course because of their marketing purposes or even for compliance and reporting like Sunshine Act and all but they also want to understand the relationships between physicians and hospital systems and peer organizations and plans and prescriptions and patients so that has been there the one is as you pointed out is retail where they want to have that connected customer experience they want to provide better personalization they want to provide better contextual offers as well as making sure that they are meeting all the compliance needs and understanding the households within their consumer base so that has been a big adopter and then similar use cases we see are around healthcare as well as PNC insurance where we are our customers have been saying that we want to be more customer centric than policy centric organization to understand the customer to understand their needs and then delivering more customized products for those particular customers so that is another big one that has been in the forefront of adopting these modern technologies and we just have a couple minutes left I think we have time for one more quick question how easy is the path to upgrade to the new gen MDM from the existing traditional I will take that for a high level thanks AJ real quick we are talking about graph often being a layer across legacy systems including legacy MDM to pull everything back together to do crosswalks across domains to find hidden relationships explicit implicit whatever to go from product to customer to supplier to pricing which are often in discreet MDM systems or commonly in discreet systems so graph often provides the glue to bring things back together as you migrate to the next level or next generation platform back to AJ exactly right I think I was going to say the same thing so it is not necessarily kind of a rip and replace kind of a solution like any other good solution should do so it is more like bringing your information from your legacy systems together and then providing that kind of a step approach towards moving on to these modern technologies so you can kind of bring the data from your old systems whether it is MDM or your old data warehouses or even like sometimes the data lakes we have seen out there and then fulfilling a business purpose and then as you move along then you can kind of first say okay you want to kind of move the data from your legacy system into the new system as well but it is more of like a phase approach where you can still continue but pull the data from your older systems into the new one and then in a phase manner remove the older systems alright well that brings us to the top of the hour Aaron and AJ thank you so much for this great presentation and thanks to Railto for sponsoring today's webinar just again reminder I will send a follow up email by end of day Monday to all of you this morning and we hopefully you all have a great day thank you so much thank you so much, thank you everyone bye