 Hello, and welcome. My name is Shannon Campion. I'm the Chief Digital Officer for Data Diversity. We want to thank you for joining the latest in the monthly webinar series, Data Architecture Strategies with Donna Burbank. Today, Donna will discuss emerging trends in data architecture. What's the next big thing? Sponsored today by StarDog. Just a couple of points to get us started. Due to the large number of people that attend these sessions, you will be muted during the webinar. For questions, we will be collecting them by the Q&A panel, and we encourage you to share highlights via your favorite social media platform using hashtag DA Strategies. And if you'd like to chat with us and with each other, we certainly encourage you to do so. And just to note, the chat defaults to send to just the panelists, but you may absolutely change that to network with everyone. To open the chat and the Q&A panels, you will find those icons in the bottom of your screen to enable those features. And as always, we will send a follow-up email with and two business days containing links to the slides and recording of this session and any additional information requested throughout the webinar. Now, let me turn it over to Navin for a brief word from our sponsor StarDog. Navin, hello and welcome. Thank you, Shannon. Good afternoon and good morning, depending on where you are. My name is Navin Sharma. I am the VP of product here at StarDog and looking forward to a great webinar with Donna here to walk through the emerging trends in all things data architecture. Just real quick, from a StarDog perspective, I want to make sure you had at least a perspective we're bringing to the marketplace. It's important and probably most very relevant to the content Donna is going to share here as well. One of the things we talk about emerging data trends and data architectures is we often think more along the lines as we've been trained to do as data practitioners from more of a bottom up perspective, right? From a data production, data factory perspective, setting up the data factory, setting up the data pipelines, bringing data down to some sort of a centralized data lake. And what gets lost and often frankly where the biggest need that gets unfulfilled in a large enterprise is ultimately the goal of why companies invest in data architectures and data plumbing is to really democratize data and information to as many users in the enterprise as possible, right? You don't have to have the necessary data skills in order to make data driven decisions. And so the challenge we find that will continue to remain in a lot of these organizations is that last mile, that last mile, the closing, the gap in order to achieve that data democratization and why is that? We look at it. It really boils down to three main themes or thematic challenges, right? Number one, data still lacks business context. So yes, investment has been made to modernize the underlying data infrastructure platform, things are in the cloud, things in a warehouse. Perhaps you've invested in a lake house. But at the end of the day, the representation of a lot of that data is still siloed in ways, is still not well understood. Yes, it's all co-located, but it's taken in the organization a lot of time and energy and resources to get the data to a reasonable state. But at the end of the day from a researcher or a scientist who needs to make important decisions around drug discoveries for life sciences or for a factory floor analyst that needs to make decisions around, you know, the resiliency around the supply chain and decisions around the impact of a particular product recall or looking at it from an operational risk perspective and a financial services, your compliance officers, your compliance teams. It's not easy, right? They're not skilled at a lot of these areas of the data plumbing. And so not having that context and not being able to look at all the data that perhaps may live beyond just one data lake or one data warehouse and not having the right tools that are designed specifically for those users in mind is certainly a big problem. And we see a lot of this happening all over the place. So when we break it down, it really breaks down to, you know, let's start looking things or pivoting away from the data producer perspective to more of the data consumer perspective. And so you have to start to organize information in ways these users understand things, not strings on the left, right? So, yes, traditionally we've created data models that are tightly coupled with the data infrastructure and supported, you know, good query performance. But at the end of the day, it's really driven by how a consumer wants to visualize information, understand information in context of their use case and how they describe those concepts that are more meaningful to them, independent of where the data is located or how the data is structured. And that's where representing that in ways that, you know, conform to open standards, decoupled from the underlying data infrastructure, being able to specify and create business logic rules against that being able to define data quality constraints against it. You start to kind of come at it more from a business information model perspective, akin to a conceptual model and a relational database paradigm, right? And then having the ability to then use that, those concepts or those business ontologies as a way that you can share and reuse across both your within your organizational boundaries, but also across a larger ecosystem that you're working within, that may operate outside of your organizational boundaries. And so that's the benefit of sort of coming at it more from that mindset than from a data producer perspective. We also know that data exploration and discovery for citizen data users isn't encumbered because there's a dependency on these highly skilled data engineers or data scientists in order to produce results or queries that are important for them to understand and make decisions. And so enabling the ability for them to search, explore query information independent of that specialized skill is still a very much a high demand need, right? Self service for citizen data users in ways that they can explore and visualize and search for information without requiring these hardcore IT skills or dependencies on these skilled individuals. And last but not least, you think about data access, right? We still follow this very archaic approach. Everything has to be sent pattern across a centralized pattern, right? Everything's got to land in one place. And that certainly makes sense. You know, for most use cases and a lot of times you want to build that centralized data pipeline architecture, but what happens when you have new sources of information coming in or you want to look at external data sources, or sources that are just unstructured. How do you support the ability to query those without having to wait on, you know, creating additional complex data pipelines and more importantly limiting data sprawl because you don't want to wait and take a snapshot view of that information. But as that information gets updated, you want a real-time ability to access and present that information to the user to support, you know, whatever their ad hoc query needs are. So federated data access becomes just as important. And ultimately the goal of closing this last mile is you start to think about integration more from the perspective of a data model that harmonizes data based on business meaning rather than where data is stored and how it's structured. You are able to connect to data in a federated manner that allows you to link and virtualize access to enterprise sources of data to enable more dynamic data orchestration and ultimately enable data exploration and discovery or the ability to uncover new patterns and answer questions across different domains in some sort of a self-serve capacity and that becomes just as important for the citizen data users that you're trying to enable, right? Why are you putting the effort and energy in modernizing your data platforms when at the end of it, and the end result is you're not in a position to truly democratize information and data to as many users in the enterprise as possible. You know, it's no surprise with the gardener of the world, gardeners of the world come out and talk about this notion that, you know, more organizations are going to find it harder to exploit their data assets efficiently and effectively. We know that, you know, in terms of the recommendations there starting to make is this notion of adopting a semantic approach to their enterprise data. And that is also reflected in the idea of a data fabric or a data mesh architecture where a dynamic composable and highly emergent knowledge graph essentially is at the core underpinning of a lot of these are, you know, modern, you know, modern ways to think about data architectures, which starts to reflect everything that happens to your data, right? This core concept in the data fabric really enables these capabilities allow for more dynamic data integration and data use case orchestration. And, you know, Stardog's been in this business for a long time where for us, it's all about providing companies and organizations and data teams the ability to connect data based on business meaning into a flexible reusable semantic data layer for better insights faster. And we see this across industries from manufacturing supply chain to financial services to healthcare life sciences and pharma that are working with data supply chain across these silos and are trying to create data and make data more fair. By saying fair data, you're talking about making data more findable accessible, making data interoperable and more reusable. And that's the ultimate goal for a lot of these industries as well. And for those that are interested in learning more about Stardog and how we enable a data fabric or data mesh architecture, there's an opportunity for you to do that. Go ahead and sign up for free at Stardog.com. Get started. For those that are Databricks users, we certainly have the ability for you to connect through their partner connect capability, which allows you to look and search for the semantic layer category and look for Stardog there as well. So I think I'll stop here and I'll pass this back to Shannon. I will introduce Donna. Look forward to your questions. I mean, thank you so much for kicking off this webinar. And if you have questions for Naveen, you may submit your questions in the Q&A as he'll be joining us in the Q&A at the end of the webinar today. And now I'll introduce the speaker of the monthly series Donna Burbank. Donna is a recognized industry expert in information management with over 20 years of experience helping organizations enrich their business opportunities through data and information. She currently is the managing director of Global Data Strategy Limited where she assists organizations around the globe in driving value from their data. And with that, let me give the floor to Donna to begin her presentation. Hello and welcome. Hello, Shannon and hello world. Nice to see a lot of folks joining us in the beginning of the year. And if those of you who are nice to see some familiar faces. I can always say because it's just nice to see some names that kind of show up every month. But some new names as well. And if you are new to the series, it is a series and we have this every month. One of the common questions that I'll just answer is, you know, will these be saved and recorded and can we get the slides. Yes to all of those. And there's an archive for years, I think that diversity keeps so not only can you get this after the fact, but if you're interested in any of the other topics. And so the ones that we just talked about things like data mass and data fabric, MDM data governance, lots of great topics coming up this year if you'd like to join us again. But what are we talking about today and this has been a bit of a tradition that we'll talk about each January kind of, you know, let's look back at the past year and then kind of look ahead for what some of the new trends are going to be. And as always, what we try to offer in this webinar is, you know, practical real world practitioner value for there's always new hype and new words and we always have to have some new buzzword and data and how much is just the same old thing with a new name. And how much really is transformed informational and and also what are other people doing, you know, is this stuff folks are just talking about, or how a real companies like myself or organizations like myself doing this and this is what we try to offer. And this type of webinar. So the research this is based on real data because we're data folks. I'm just talking to Shannon before the webinar I think this is I think our fifth or sixth year doing this, which gives some really nice data trending so diversity and my company global data strategy we partner and do a research survey each year and I think that's coming up shortly in March or so. If you want to participate in next years. So you can be part of the trends and then what we do is collect all that and put it into a survey. This year you can see some of the metadata about the survey how many participants across, you know, not just North America is 35 countries 32 industries. So a lot of different perspectives. So at the paper, we kind of are doing a little subset here and in some of the 50, you know, facts that come from it, but you can download the paper either on diversity or on our global data strategy web page and read it all yourself if you cannot sleep one night. So, let's jump right jump right into the meat of it so if you've heard me speak, you've probably heard me talk about the data driven business and how data is really a business initiative. It has been for years, but I think even more so now everybody and you look at these, these quotes are from the Wall Street Journal there from Forbes and many organizations not just for profit. I really looking to get that value out of data partly because there's a lot of cool technology out there but at its core, it's a business asset and more and more organizations are realizing that. Another thing I always like to kind of bring up is there's different ways of looking at a business asset or why we're doing something from the organization. And I know there's a lot of talk about this in terms of things like data governance. Are we doing offense? Or are we looking at this? It's more of a defense perspective and neither one is right nor wrong. And it really depends on your industry, right? If I'm, you know, pharmaceutical or I'm, you know, a hospital or, you know, I'm handling all your money. I hope I'm looking at regulation and risk reduction and I'm a bit cautious with some of the data. I think there might be a brand new startup with this cool product and we just really, really all about growth and we're all about benefit and we're all about opportunity in a normal typical organization. There's probably bits of each of these, right? Even if I'm in a highly regulated industry and I have patients and I want to reduce risk. There's also a lot of opportunity. How can I better help my patients or things like that, right? So there's always aspects of both. And my kind of two cents that I often kind of rant about it would always frustrate me in years ago when I would go to kind of some of the diversity or Dama conferences. It seems to be a lot of negativity on business doesn't care or why are we doing this or so. I always wonder, you know, be careful we ask for the business is definitely paying attention. So I'm seeing less and less of that and more excitement, which is what's keeping me in data because I do think there's a lot of business centric opportunity around data. So back to the survey, I find this interesting we literally ask that question, you know, why are you doing data management, you know, people like me and Naveen might just do it for the fun of it. We like this stuff, but there's generally a business reason for it right year after year. Number one is always reporting and analytics is getting the reporting analytics is getting the business insight. How can I find new patterns new insights into how I can be better my organization. You know, part of that is the next bullet that saving costs increasing efficiency. The more I know, you know what they say you can't manage what you can't measure. Right. If I don't know how much I'm spending in an area, or how much time people are spending or etc, etc. You can't optimize that. So that's often a very classic way to have dashboards and reporting. I don't want to say it's new, but it's definitely gone up on the rankings. This idea of improving customer satisfaction. Probably again cus companies been doing that for decades but using data to really get that customer view and that 360 customer view is even even hotter. So digital transformation is another one and we'll talk about this in a few slides so I won't kill it too much here but you know what does that even mean and is it a trend is it something that's just becoming kind of like calm just becoming business as usual. And of course everything's digital now. Kind of an interesting topic and you could see some of the other reducing risk regulations. I don't know what you like or sort of at the bottom and maybe they'll be popping up as well. Product quality. I find that super interesting. Again, something we've probably been doing for centuries and decades is to look at data about how to improve our, if we're a product company. But more and more a lot of products are internet of enable their internet of things or web based and you can literally see the click through traffic of what people are doing so there's literal. We're think of medical device, you know, things that people can even actually see how people are walking or moving, you know, you literally are getting real time feedback from your product which is, you know, definitely a data driven initiative that's kind of new and upcoming. The last one I like too is that improving outcomes because we so and I'm as guilty as everyone else. So often when we talk about data what it was the first thing you hear single view customer product and profitability and of course that's a part of it. So, improving health of your patients. We're working with a number of both higher ed and kind of children educational and a lot of it's doing longitudinal analysis of how are we improving educational outcomes right and that's all data driven. Again, it's not always the, you know, financial dashboard people are looking at this dashboard for other things and analytics and predictive analytics and a lot of things what's the best thing we could do for our children to make sure they're going to have the best education when, when they're adults right so. So interesting data and what I think is fun about my job because we get to look at a lot of these different types of organizations and go, oh, how do we apply data to this use case. So, kind of going back to that idea of carrot and stick and maybe this is a goofy slide but heck, never stopped me in the past. So what I find promising is that I mentioned before in the old days a lot of why people did data was reducing risk regulate the regular regulators going to come get me if I don't or I'm going to get sued, or you know maybe increasing the saving costs and things like that but more and more, I think organizations are really seeing that business value and definitely much more the carrot so I did my carrot versus stick meter. But when you look at these top bills business drivers top ones really are the carrots and there's more carrots and six and I think that's a nicer way to look at it of yes of course, you know, if we have to do things because the regulators are going to get us you might you might do it but is that so exciting is it going to stay is it going to grow within the organization and are we really helping things, or we're just pushing paper out because we have to right so even more because we've worked with that might have to do do some, you know better management for customer data for regulation can also then use that data that you've managed and use it for, you know, for better analytic run your customer and supporting your customer journey or customer satisfaction right so this isn't either or stand right but but definitely I think thinking of that what what's our business opportunity here through data is definitely growing trend and what I'm really really happy to see. I know the digital transformation I mean I'm going to tease buzzwords left and right here because we use them and what is digital transformation that I always like our nurse definitions they pretty practical ones. And it can mean a lot of things. Or is it becoming business as usual, I don't really love their comment at the bottom of widely used for public sector organizations for modest initiatives like putting services online. I don't know that doesn't seem modest to me I'm happy I can renew my driver's license online. I would agree with them and that that's sort of expected now. And again will will digital transformation cease to be a thing. Not because it's not a thing but because it is that doesn't make much sense but it's because it's become so business as usual, we don't call it out as a separate thing that just the default is digital right, kind of like the default is ecommerce now everyone, you know kind of you have a website on your company. So I think digital transformation is heading in that direction that is expected it's not something necessarily novel anymore. So on that note what I found interesting on the trends year over year because you know on purpose, we tend to ask the same questions right so you can see those trends, obviously adding new ones as new technologies come along, but you're able to see these trends and this thing that was interesting is that digital transformation, in terms of its year over year how many people chose it is down by about 7%. And again, I don't think it's because less people are digital. I think it's because more people are this just not a new initiative anymore. But what's gone up is an idea of improving customer satisfaction and my two cents on this is that yes I can renew my driver's license online but is it easy can I do it with my cell phone can I do it with one click. Can I you know people are being much more savvy about not enough just to have a web page but how does it work. And from a business point of view do I really understand my customer experience both the data I can capture from it, within, within guidelines, or you know how I can then use that to improve my customer and we say customer could be customer slash students less patients less citizens slash, you know, the human you were serving. I think that's interesting because again, to me that kind of speaks to the evolution of digital and again, you know we're really doing it for a purpose and not just for the sake of being digital. I think that's kind of showing the maturity across a lot of things so you know this is kind of going to be a medley of a lot of different graphs from from the survey, but one of the other things is what are your top priorities in terms of and when you look in 2022 a lot of it is no surprise business intelligence data warehousing self service reporting analytics so it's a lot about that good old fashioned but will never go away. Reporting and be I because that that's why you're often looking at data how can I optimize my business but when you look at what people are looking at in the future. The other question was what are your priorities for the next coming two years. I find interesting it's governance quality metadata master data and a bit of strategy. To me, my, my color commentary that you know we've got the report but do we trust it. And I think that's a normal evolution of, you know, companies that maybe. Yes, we as an industry have been doing behind analytics for, you know, decades but is every organization. And that's what I find fun about my job is, you know, we've done data strategies and implementations for museums and nonprofits and you know a lot of different things that maybe, you know, 20 years ago it was your financial services and your government and the big players. Now a lot of people are kind of jumping in new to different areas that maybe are new to them and looking at, you know, now we have the nice shiny report, but the data behind it and you have to do the hard work to get that data right. And I quote from a customer a few weeks ago that was a very quotable person had a great sense of humor is like I don't need to keep drilling down into more layers of wrongness my dashboard. You've got the drill down and you've got the slice and dice but I'm slicing and dicing garbage. And that's what people are finding so kind of going back to the drawing board so, you know, we don't want to do self service reporting analytics, I think they're realizing that once we look at the data is not a high quality it may not be understood trusted but it's just stewarded all of all of that. Which brings me to my favorite topic which is metadata. We all need more metadata in life and and I was again on the same sort of flavor is when you look at what the trends were year over year what fell into the top five. The self service bi over the years has been in the top five year of year and it's gone down and it's no longer in the top five, and it's actually decreased, whereas metadata jumped into the top five and that's increased by 22%. It's not exactly a, you know, one to one match but to me that said again I don't think it's that people don't want to be able to easily create their own visualizations and reports went on demand. They're doing that and they're disappointed by the quality of the data or the, you know, it could be a whole webinar but data quality has many dimensions right. And it could be, you know, the data is right but is that calculation understood by me or is the context right or what sources are coming from all of that hard stuff that we in the data architecture and data management community live and breathe is becoming more and more important in the spotlight because people are realizing they need to understand the metadata not just the data. And so if you are new to metadata and wonder why I get so excited about the word, it really is what makes data click and I won't really spend a lot of time this isn't a metadata webinar it's a trends webinar. But if you are new, we've gotten some good feedback on this slide, you know really just kind of simplifying metadata it's really your who what where why when of data or your context right what's the definition, where did it come from who created it who's the steward when is this data is a new data. How is it formatted and stored and all of that good stuff that really makes data management thing. One thing I meant to mention the beginning. Yes, this is the data architecture webinar and the survey was on data management, but I think in a way that's fitting and to be sort of touched on that in the beginning to have, you know, data without business context or just without management around it with governance and things like that really isn't as effective so we kind of jump in architecture is definitely a piece of this, but architecture is much more broad in this context that we're kind of also putting things like, you know, governance and things like that. So, you know, this is a formatted data and I don't necessarily need to go and read each one of these layers, but but again you're starting to see some themes here right we have governance we have quality, and of course, data warehouse and bi we kind of loop those together here. But it's still an important folks want to do it and they want to be more efficient and agile and do data science and analytics and some of those other things you see in the list. Until you get the governance right until you get the quality right until you get master data management which is that single view of truth for your again patient customer student product location, you know all the stuff that makes your, your organization saying the rest of it isn't going to come easily so you know all of that is super important. So again, we asked, you know, in the survey you know what are the biggest challenges you see the top one, no surprise, whether you're doing, you know, the good foundational things like a data warehousing or even a data lake, the number of data silos and how you get a solution that as well graph is one solution to that governance is another solution to that warehousing and a lot of different ways to slice that pie there, but definitely a challenge of how we manage that doesn't mean they always go away, but how do we manage them in an effective way. Often that's kind of lack of governance quality again you're seeing those over and over that way some themes there, but the bottom two I found was interesting because it didn't necessarily pop up in the others but in terms of the challenge. Do we have data literacy or just skills, are there enough people to do this hard stuff, I don't find it was hard but kind of fun right a lot of us on this call get data management. But how much does the business need to know about data and in terms of literacy and how what kind of skills we need to create those pipelines and that ecosystem and the architecture kind of behind a lot of what we're looking to do. So a lot of frustration there. Again, I'm going to go back to Gartner for the definition of what do we even mean by data literacy and again, each one of these slides can and probably is a full webinar. But several aspects of data literacy one is from the business point of view do I even understand the data sources, how to communicate or read data in context. How do I understand, I know it may sound silly but even how to interpret or read a report in it I understand to ask the hard questions about how it was sourced. Do I understand the business rules behind it do I know what kind of analytical method was used on this data right this can get super super complex or just also very very simple. We do a lot of data governance in our practice and almost the sign of when you've kind of gotten it right is when the execs in a meeting kind of ask where that data came from right to come from the trusted sources is this from the corporate dashboard or the data on spreadsheet before the meeting to make the numbers look like how you wanted them to. Right. And I think part of that feed with data literacy, but there's also, you know the data skills literacy right do I know how to build a pipeline. Do I know how to build a warehouse and why a warehouse is different than a lake and all those good foundational things that you know data management association and the dama dm block kind of help us with. And so I mean part of some of the issues around this idea of data literacy and the need for it and I think, and again this is heartening. We often say in this quite questionnaire, who is driving data management and this is a multiple choice because I think that's correct because it should never be just one person supporting data management or data architecture. It really needs to be holistic. This top one, the data governance lead is really jumped to the top and past years and I think more and more organizations do and should have a full time data governance lead. That's that's supporting data management organization management or separate things, but they are related. The data officer is definitely one that I think didn't even show up in the past years. And now is number two. And then for anyone on the call who is or wants to be one of those roles. I often see that that data governance lead is often a really nice stepping stone to the chief data officer because what is was hard about the data governance lead role and where someone is successful is that you need to know the tech you need to leave these cases around data. And you need to just one of my clients once said, it's someone that tells you what to do, but you like them at the same time, right that you do have to put rules and regulations in place, but you're also getting buy in and championing data and then. And the need for quality data right. And so once you've really got that, it's really stepping stones that executive manager can take something like a chief data officer. Chief information officer. And I have many Donna Rance is that generally that's not an officer is generally, you know, it is that really knowledge is often the servers or the platforms and things around it, but certainly they should be involved in data management. And then some of the others analytics data architect great business stakeholders is another one that I'm very, very pleased to see. Chief executive officer absolutely kind of wish that were higher than little less than 10%. But I do think that's growing and what we didn't show here is that there is a right in option for this and generally all of the right ends were C level positions right. Chief operating officer, my chief marketing officer, my, etc, etc, because again, think of even those two examples, marketing, absolutely a data driven right was our click through right how did our campaigns go. So a lot of these executives are hungering for good data and good analytics so that are now starting to realize probably through things like governance that they need to have a voice and I say, So, hopefully that one is interesting to you. It's always interesting to me and this one this is one means some of them. Again, I'm not doubting doing the survey every year but there are some definitely sometimes wonder why I keep asking this question behind analytics has been number one. Each year we've run it doesn't mean that's going to be the same next year right and and this is one that I have seen has changed probably most drastically every year and the things that are changing is more and more data governance. So C level, which I think is kind of heartening when you look across. I know this, a lot of this presentation has been kind of on the softer skills that the less techie, which I think says a lot and maybe talk about that even in his intro right is that we all love tech and we're probably on this webinar because we love tech, but unless you get the rest of it right. So, but this is these this this in the next slide or sort of my favorites, because it really gets to all these cool hot technologies out there what are actually that the platforms people are using. And this is one that absolutely has not changed year over year relational database despite being maligned and called old fashioned and called not relevant anymore continues to be absolutely number one. Yes, has cloud sort of moved up and on prem, perhaps the lexical it's not a little less is that it's almost everybody's still on prem. But there's also a lot more year over year to the cloud, because the relational databases still work really well for what they're defined for. Does that mean that's the only tool and you'll get absolutely not I mean if you're only using relational, I would question that. I wouldn't throw away relational and go to something else if you're running your P system or you're trying to have referential integrity and high data quality is really good for that. The one that I do lament every year and keeps me up at night is the spreadsheets spreadsheets are fine I use them a lot and they're excellent but they're not an enterprise data management tool. The number of unnamed very high revenue, well known companies we go into. When we ask where your, I just learned not to be surprised where's your customer master product master location master, you know analytics, and it's in a spreadsheet, somebody has on their desk practically. It's scary for a lot of reasons for, you know, efficiency for regulation for optimization is just really not a great idea. So kind of moving spreadsheets into something else is highly recommended but you didn't need to come to this webinar for that did you. But again, there's a lot of other things but I'm also interested that each year we still see a lot of quote legacy systems yes there's still mainframes out there, and they're still running so again we love to kind of knock old fast and things like mainframe but the ones that are running around are still running and often have a lot of great business rules that were implemented correctly back in the day. But again, there's a lot more choices than relational database, which is why I always like to show the next slide. Again, still relational but more on the cloud. And I do see that as a trend. You know, I don't think on premise on premise. There's still needs to be. It's still a good choice, but definitely cloud is a good option. In Naveen mentioned kind of this data lake house one of those funny words we always like to have in our buzzword dictionary, but this idea of using cloud object storage either for backup or you know, things you want to keep or even in a way it's sort of your, your landing area before you get to a warehouse right and you can really store a lot higher volume you can store unstructured data you know this is the place for both. Yeah, I don't think stop at the lake and magic happens I think we've passed that kind of evolution there was a day where we had to keep convincing folks that you can't just dump things in a lake and and that's the only thing you need to do. Obviously, there's a lot of manipulation of that data to make it usable and trusted, thus, all the other things we've heard about in terms of the rise of data governance and data quality. But what I do also like about this slide is that when you see the previous one is definitely sort of peaks. And it really hasn't over time straightened out as much still just a whole lot of relational and yes some ERP systems but I can argue that's a relational database behind it. You know some folks doing data exchange through XML and things like that, but not a whole lot of experimentation or new stuff going on here when you look at the future. There's a bit more, you know graph, of course we just heard from star dog really, really cool solution, especially in that semantic layer and you know is it necessarily the best for cleaning up data quality or doing governance, you know, there's different so tools for each toolkit. But definitely, especially in some of the areas that Naveen mentioned, you know pharmaceutical and when you're trying to do exploratory analysis and really see those relationships. That's great. Things like real time data I mean the technology that's out there to do a lot of these things that we couldn't in the past is really valuable doesn't match every use case now. But looking at some of these other systems non relational key value pair. A lot of other good things that are out there that we can play and have in our toolbox. The other one that I do find interesting and I love the honesty, I don't know, right and I don't think you're dumb. If you don't know, because I don't know, and I've been doing this for almost 30 years and it's hard to not know because there's just a lot of technology out there right and hopefully diversity helps some of these webinars help. There's a lot of tools in the toolkit and understanding that fit for purpose that you probably do need, you know, a relational database something like a graph some non relational. Maybe some real time streaming. I almost see this new architectural world as I often just call it as zones right there's my relational zone for those use cases. Maybe I might have a streaming zone I might have a graph on the front end to really do some analysis and I would my kind of recommendation as you think about this and if you're one of those in that 30% of folks, probably higher because a lot of folks don't like to say you don't know is that that's a very healthy place to be because there's a lot of research to be done. Yeah, obviously start start with some of the basics like make sure your relational database and some of your core data asset. After data right, but this might be 2023 is your year for some experimentation right to really kind of look at some of these new technology and kind of a use case. I do find it interesting interesting with the honesty that some folks do say they will continue to use their spreadsheets in the future, probably even higher than that. But it didn't come up and often wasn't the top 10 that surprised me was legacy systems that folks that yeah that's still going to be around, you know, maybe this kind of final stages of some of this legacy migration is sort of wrapping up and worry able to kind of move to some of these new environments. So, I hope you find this one as fun as I do because I always like to hear there's just so much technology out there but what's really realistic in terms of you know the big companies on the planet. What other people are doing and am I that FOMO, am I missing out and what other folks doing or am I maybe a little ahead of the curve. So anyway, hopefully that was helpful. I always like and I'm asked to kind of leave time for questions because usually folks aren't shy. So we can get into that but just to summarize. I mean, the idea that we have business insights analytics as a main driver, not a surprise will continue that's often why in a normal healthy way, we use data. But that idea of being a little more on the offense of how do I, how do I drive use data to drive business innovation and particularly this idea of customer satisfaction is a really nice way I think to be able to think of data. But because of that, this idea of trusted data, whether it's governance data quality metadata, etc, etc, is growing and you're probably not alone if you're looking to do that, which drives that data literacy and skills right so data literacy. Everyone has their own definition right I would say data literacy is more in the business side do I understand enough about data to be literate. And then skills I see is more on that architect type side. And if you have children looking to go to university tell them to work on data. Really great space to be in. And there's a lot of demand for what we do. And part of that reason is that last bullet this just is a lot of great tools and technologies and I guess I guess part of my frustration about this previous slide is that there is a lot of cool technology out there and are we not using it because folks don't know. There's a lot of skills and I think that's maybe some frustration for management of you know all the things we can do we're still kind of just cleaning up our messes. But yeah, I mean that was too negative I do think there's a lot of great things going on. So, before I open up to questions a little a little plug for the rest of the series if you can join us on in February we'll be talking about data strategy. And how you put all this architecture together and to that point of making it a business initiative. How do you do that this really clear steps for that. I'm blatant plug for what we do for a living global data strategy if you want help with any of this. That's what we do. And I will now pass it over to Shannon to open it up to Q&A. So thank you so much for another great presentation. If you have questions for Donna or for Naveen feel free to submit them in the Q&A portion of your screen there. And to answer the most commonly asked questions just a reminder I will send a follow up email by Monday for this webinar with links to the slides and links to the recording. So, a couple questions that have come in. So on slide 16, what was the source for the data for that graphic was it the research paper, the survey. Probably, but let me go back to 16. Yes, this all of the graphs and I didn't put one on each one but I would assume all the graphs here are from that research. From the survey. I love it. And where do you see data mesh in 2324. I will firstly do a plug from ours because we'll have a whole webinar on that. I sort of didn't mention that on purpose. I'm skeptical and I will come out and save them a bit skeptical of data mesh and data fabric I think there's a place for this idea of distributed processing. I think things like data virtualization can be really strong. I think there's a lot of talk about it. If you notice one of the questions for some of these is data mesh and it's not in use yet. I think there's definitely some I have some customers that are using and I know data virtualization isn't the same necessarily as data fabric and data mesh and go to the webinar. There's a whole, you know, what are the slices of that. I think part of it is a technology is that there is things are things like data virtualization or semantic layer, a graph that can kind of look across different data sources. Absolutely interesting. You, I think what's sometimes missed with the fabric and that also ties into the governance is. Yes, there's one of the big challenges with silos. And maybe sometimes with this fabric, it's a bit too much to distributed. You don't, you don't solve the silos by creating more silos and saying well then marketing needs to do their own thing and sales needs to do their own thing. That's sort of implicit. But when you're trying to do things like master data or common KPIs, I think that's where data fabric breaks down. So I think it's an and I think with any trend like data lake, right. So data lakes work. Yes, but not in isolation. So I guess that's my answer to David fabric which is, it's a tool and the toolkit pieces of that, but some things do need to be enterprise wide right you're your product data that you're selling. There should be a central view. Yeah, you might you have a different product code in Europe than the US, but that should be an orchestrated thing. So, but I'd be curious and maybe I'm sure this is right up your alley to have some thoughts on that. So if you look in the context of data mesh, you think about it. Basically, the idea was, you retreat looking at building data as a product. Organized by domains, the enterprise, and the organization by domains basically allows you to have some sort of federated control or independent control over governance policies associated with the data product that you're building and creating within within that domain itself, rather than everything being centralized and managed through a central governance, sort of paradigm. So in that, in that same context, when you describe, you know, a data product, you know, essentially what a data product is essentially is a business consumer driven representation of the data that's relevant to the, you know, relevant to the context of the use case right so within that kind of context of the use case, you may have different business logic, you may have a very specific data model in mind you have a set of business logic associated with it. And the collection of the end data access policies associated with the collection that is this data as a product that can be shared across the organization. Sometimes I've even heard people saying, well, is it a data fabric or a data mesh that we should be thinking about. I mean, architecturally, I think it's sort of the hybrid it's coming together the two approaches right at some point. Yes, you want that independence of building domain centric data products, but then there's a need that you need to fulfill in the enterprise that brings all those domains together. And almost like a connected knowledge graph and so it's the combination of the two approaches frankly that would make sense in any given context. Yeah, and where it's like your warehouse and like, and marks right. Finance can have their own mark but there's also kind of enterprise. Correct. Yeah, no interesting approach. Shannon what's next. I think there's a lot of questions around data mesh here. So, you know, the idea behind data mesh and data fabric have been around for a long time is this just repackaging. I didn't want it like what data medicine data fabric in this presentation. I think it's, it's a buzzword it's repackaging and is it is it oversimplifying a problem and maybe I just summed it up but you know the idea of that different data samples and when is that describing the problem that we have data silos, and when is that describing a normal way of some, some data is local to your group. And some data, which is maybe forgotten in the message is does need to be centralized and managed and you can't skip the hard work you know I did master data is hard because you do need to have that central view. Common KPIs reporting to the street need to be hard. So there is new technology that wasn't as easy to do. I mean, we've had some folks do do really great things with data virtualization that doesn't work unless the source data is good right so that's the kind of the both and graph databases I would put kind of in that category as well that you can look across different data sources but again those data sources have to be good so I do think there are some tech things even the cloud right and a lot of the volume of things we can store. I think when it comes to the governance and way of working in use cases and business rules. I don't see too much new there right it's just kind of what we've been working with for years but maybe I'm old and crusty and Jaden. I don't think so I think you're spot on with that right look ultimately it's, it's companies you think about working with data. What's the problem they're trying to solve and that's why it's important I can I set up set it up in the beginning or you got to look at it more from the context of the consumer data consumer and if you start there. Then you kind of work your way down and say okay which which data sources are relevant to answering this question that's been framed in this now semantically defined model by the consumer right so data modeling to me has been this place where the problem with data has always been hard data models tightly coupled to the underlying data storage infrastructure, make it very hard to serve the needs of many different users and many different use cases, because they've become very rigid in it and they're sort of shaping of the data. So if you can shape data to your needs through the way you define and describe your somatic model for example a data model or conceptual model and using techniques like virtualization, you federate access to where the data lives, because you've described things in terms of concepts and you've mapped them to where the data lives, and then at query time you're accessing, you know you're sort of providing federated access to the sources. And then you've you've sort of, you know moved away from these hard rigid data models I think there's a question here on, do you still need an enterprise data model I think that's just that's insanity at best if someone and try to build a world that will support the need of the entire organization with a single enterprise data model. Yeah, that's just, that's a man. I was worthy until you started knocking data models. No, I would agree with you. Sorry, we're going to go in a whole full out. I love it. I would agree with you when you started with the isn't business driven yes and I think that that to me when as soon as you put these use cases into what we're trying to solve the tech becomes easy and I think you're right that that's often was missing. Do we need a data model because those hard rules absolutely it's a different use case then what you were describing at the end of a graph right but yes, can a customer have more than one account at the bank that either is or isn't right that isn't a fluffy thing what's my, I mean I still got company we worked with two years ago their entire enterprise website went down because someone changed the length of a product code. I don't know why, but that's an absolute there are places and to me that's the, the was missing with that to flexible of an approach there is a place where you have to have those rule our product code is 10 characters and this is how we use it kind of customer more than one account can someone have more than one address right that's those are those rules is that separate from can I get a quick view at a semantic layer from those data sources but if someone did the hard work. I think it's a both and right so I hate to knock data moms because I use them all the time. And even the enterprise, I think at a conceptual level absolutely at the physical level do we need one centralized the, you know, data model that has everything. Or even the logical, maybe the logical but absolutely at the conceptual level do we even understand the different flavors of our customers and products and I don't know we in our practice have solved so many hard problems, just that that conceptual model doesn't have to take years. What are we even talking about here whatever core semantic rules that would be okay. I think that's exactly a point that the conceptual level that conceptual level has to be in place. That that is your semantic understanding across the enterprise. I think one of my point was more on the physical data model. Okay, but still wonderful. Yeah, we're there where they're needed like the master data and things like that off to hurt you then we're okay. Professional disagreements are so good. It's so good to have differing opinions and sorted out and I love it. It's great. And it means we're thinking about it. Um, so, uh, so, so the, the question came in though, while during the con that the discussion is enterprise data modeling still relevant and I think you've answered that already but any additional things. Oh, I will go on. Oh, if I answer data mess I get to answer data modeling. Absolutely. I have so many success stories around data modeling at both levels from the bottom up. Often that's where you find the gnarly business rule that was hidden in an application we're not sure why things aren't broken are broken. Absolutely at the top down from the conceptual or logical. So for me is setting the scope of, you know, the number of things and I have sort of fun ranting at websites or customer experiences that when they don't work it comes down to the data model like kind of customer more than one account can, can I have more than one address the number of websites that still can't figure out that you have a mailing shipping billing. All of those and get those right. And the number of customers we work for that, you know, and not to belittle that because it's complicated how many I'm talking about my customers do I need to roll that up and do a global entity or their hierarchies there which my product, you know, all of those hard, hard business rules are around that core data. And that's where I think, then this maybe to sum up that discussion of the high level I think we all agree there's a some sort of semantic business driven conceptual model, but then down at that logical level there is a subset of your data that you absolutely don't want to burden the hard rules for your data quality. And then that that is that's driven by your relational model that's why they're so popular. Do I later want to do something like a graph or like a virtualization to get the value from that guest but if unless you do those hard rules. That's why when you see the in the survey things like governance like quality like metadata are so high. That's the hard stuff. You can't skip the hard stuff to get the juicy stuff at the end. Yeah, I don't look I'm getting aligned on the, so a lot of conceptual models are great but at the end of the day the implementation of the exceptional down to a relational model is where things fall apart right so we are trying to model a universe that has many to many relationships. You know complex hierarchies in the mix. That's where the relational paradigm just falls apart from our experience and that's where you have to think about it. How do you implement that conceptual model ways that has to your point you know the logical model with explicit rules that can be created constraints that can define around data quality. I think that still is something you can enable through the power of, you know, using a somatic layer that sits on top of your silos right your data lakes your, your enterprise applications. Yeah, for the, I mean, I think I got agreed with you when we had the, there's a conceptual model and then from there you pick your implementation use case that they're just not, there still is a place for, I have an operational system and I want some hard data quality rules that is a good relational or a good kind of traditional data model which is different than kind of that graph exploratory so yes I think but you need to start with what are the rules and then kind of to that last picture. There's a lot of rules and the toolkit solve that, but I think the data model helps me as it is doesn't mean the relational goes away if this could be added it could be after could be a graph or the answer could be a HTML hierarchy. I think you don't know that until you do that conceptual so I think it's a hand. We have just less than four minutes here but I think we have time for one more question so back to that. So, with data mesh architecture and domain ownership, who then owns MDM. Hello, exactly. No, that's my problem that's my problem with my I mean to me, certain things that are, I mean almost think of your there's certain things in our enterprise, and there's certain things that are local. I think a conceptual model helps with that data governance helps with that. And I think me one of my problem often when we build the data governance framework, you know one of the questions is how do you how do you define your stewardship is it by business process is it by org is it by capability, or is it by data domain and so many people out of the box it was so much easier to have. There's an owner for customer and there's an owner for product and there's an owner. Who owns customer or who's patient is the person that checks me into the hospital and gets my insurance card the same as the person that enters my diagnosis I hope not. So there's we often do all of the above is who in that process owns the piece of the data customer right there's from marketing, there's sales there's support there's, you know, so much around customer or student from when you are the, are you the instructor of that student are you the registrar at the university are you the, you know, emotional counseling rep I mean, there's so many different people that touch those. That's where that idea of having just one owner for those that said you know maybe I can say that HR owned my salary or is that finance right but there are certain things that certain departments own but that actually is where data model comes in really well. We often build a data model and then we get down to those slices. Maybe this area is absolutely owned by HR, but these are the common attributes around an employee that we can all share. And that's the best use case that kind of solves that discussion we've been having you do that architecture the guy of through a model, and then depending on which areas, some of it can be centralized needs to be centralized, and someone that can be localized by domain but you can't skip those hard questions and that that's why governance and quality kept coming up high, but I'll give me a chance to chime in there too. I mean, look, you can't, you can't walk away from the hard work that needs to happen right so. Ultimately, it's, what is it that's important to the business, I still kind of repeat what I've said before, you've got to have the business context use case in mind, not all data is created equal not all data is important so you know, try not to boil the ocean right. All right, well that's perfect timing bring us right to the top of the hour here. Thank you so much to both of you for such great presentations and conversation. Thanks for attendees are being so engaged in everything we do and thank you to start off for sponsoring today's webinar and helping to make these webinars happen. And just again a reminder to everybody I will send a follow up email by end of day Monday for this webinar with links to the slides and links to the recording. Thank you all so much. Thanks Donna thanks. Thank you.