 Welcome, my name is Shannon Kemp and I'm the Chief Digital Officer for Data Diversity. We would like to thank you for joining today's Data Diversity webinar, Data Management Best Practices, sponsored today by RELTO. It is the latest installment in a monthly series called Data Ed Online with Dr. Peter Akin. 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 would be collecting them by the Q&A section. If you'd like to tweet, we encourage you to share highlights via LinkedIn or other social using data, hashtag data Ed. And if you'd like to chat with us or with each other, we certainly encourage you to do so. To open the access either the Q&A or the chat panels, you may find those icons in the bottom middle of your screen for those features. And just to note, the Zoom chat defaults to send to just the panelists, but you may absolutely change that to network with everyone. To answer the most commonly asked questions, as always, we will send a follow-up email to all registrants within two business days containing links to the slides. And yes, we are recording and we'll likewise send a link of the recording of this session as well as any additional information requested throughout the webinar. Now let me turn it over to Chris for a brief word from our sponsor, RELTO. Chris, hello and welcome. Thank you, Shannon. Happy to be here. Really excited about this today's best practices. You know here at RELTO, I'm going to share a little bit here for you, but here at RELTO, you know, we turn data into action and we really believe in a great component of course of data management. It's our love and our passion here as it is for I'm sure many of you if you're taking time out to listen to Peter share some great details with us today. We really do believe that data is a key to success in today's business environment, right, and it means managing data that is most important activity to an organization. So we all can come in to play here. Now at RELTO, you know, we see like you all see we see what I call data spaghettification still right. Average organization has around 450 applications and that's only growing right as people move to the cloud we're actually you know integrating with more niche applications to support very targeted business needs within our organizations. So we're writing, you know, more data more silo data within those applications, and then a framework around it that just says data kind of going everywhere like you see on this chart here. Again something I call a data spaghettification because it kind of just looks like a plate of spaghetti. Right. So, obviously data management plays a key role in how you make that this spaghetti here and the actual something consumable something we can use and bring value within our organizations. These are really kind of modern solutions right so we're seeing this shift towards data management as a practice and those practices and the standards haven't changed all that much but the way that we apply tool sets and the way that we automate things has changed. So this is really around kind of these modern software solutions that scale to that explosion of data that we're still seeing exponentially across all, all, excuse me all verticals. So, as those data supply chains really shift to support, you know, both analytics and operational use. So data management tools, needing to shift and support data that's dual with dualistic in nature, meaning, you know, we don't need to replicate data necessarily to support analytical and operational use cases anymore. We're using the same data to support both of those and distributing those in different ways. We're seeing a lot of dependency to operate in more real time. We integrate with everything with all the cloud applications in the proliferation of cloud there's real need there to integrate and move that data and be a focal point to move data around in organizations. We see the need for enabling both you know that federated and centralized data governance approach and management processes and architecture so a lot of business flexibility right and we need our solutions in our practices that we're building around these solutions for data management to actually be flexible to provide those needs for for either way that our, our businesses try to tackle data. And then lastly they need to deliver value quickly right that as, as we've seen in this, this cloud age, right. Value is key value is king and you need to prove that we're providing value very very quickly, which can be complicated in a data management space that is historically owned by it and kind of is a and a lights out type of operation for the business. Dieter is going to talk about some great things today all across data management. He's going to help us understand several specific practices and how all of this really needs to work together right around the wheel so I'm really looking forward to it. I'm really happy to be a sponsor for this event and hopefully future ones that well, look forward to learning a lot myself, and I'll be around for questions after to participate and I look forward to to listen into all your great comments and and learn in myself today. So thank you. Shannon back to you. Thank you so much. And as Chris mentioned, he'll be joining us at the end in the Q&A portion of the webinars if you have questions for Chris or about relative feels free to submit them in the Q&A portion of your screen. And thanks to relative for sponsoring today's webinar and helping to make these webinars happen. Thank you to our speaker for the webinar series Dr. Peter Akin. Peter is an acknowledged data management authority and associate professor at Virginia Commonwealth University, president of Dama International and associate director of the MIT International Society of Chief Data Officers. For more than 35 years Peter has learned from working with hundreds of data management practices in 30 countries, including some of the world's most important. His books are many first starting before Google before data was big and before data science Peter had founded several organizations that have helped more than 200 organizations leverage data specific savings have been measured at more than 1.5 billion US dollars for through his efforts. His latest current company is anything awesome check it out. And with that, let me turn everything over to Peter to get his presentation started hello and welcome. Thank you Shannon and Chris. Thank you for a great intro on this good afternoon morning wherever everybody happens to be top of the day greetings. As we do this the topic for today is data management best practices, and I always kind of rewrite the headlines underneath it because what we're really talking about is practicing data management better, because you will be doing data management from now until your HR goes away. And I like to bring people into that kind of a mindset early on. And because it's a relatively new profession compared to accounting or even HR. It's actually quite good to say that we are practicing this process that there is things there are things that we need to do in order to make it do better but we certainly don't have all of the answers right at the moment we're getting better to have instead of an easy button I like to have a button that says easier because I don't think any of this stuff actually does become easy and the reason for that is because at the moment data volume is still increasing faster than we have the ability to process it. That means we're always out running our ability to do the analysis work and this data interchange and overhead and other practices are measurably taking resources away from our organizations. There's a new source of productivity in addition to some of the other ones that I've mentioned in some of these previous events, and that is the knowledge workers and how much time they spend doing this and it's quite obvious by looking around by the popularity of solutions such as relative and others that relying on existing technology approaches, as the answer will not address the gap. Again, I'm a university professor. And I can tell you that we have failed miserably in the university community, in terms of serving the public, the demand that you all have that you think you're going out there and acquiring data scientists for are really people that who should probably be more about what we're going to talk about today, rather than the data science people of it and we can go to that in the questions and answers section on this. Finally, it is an uphill battle. There is an industry type who sole purpose is to extract data from citizens and use it to make money, or let's be very clear about it. And this is to sell them advertising advertising influences behavior influence behavior means actually population control. And if you haven't had a chance to look in some of those topics or I've never Googled the phrase surveillance capitalism I would certainly challenge you all to go off and take a look at what we have out there. Doing data better means that you really understand the vastness and the quality that data plays in everybody's life and that you're motivated to increase your individual as well as your organizational skills because poor data skills cost you for data skills steal increasing amounts of your time, they deliver less and they present greater risk to the organization. And if we recognize this critical role of data management in modern life, you should ask the question legitimately, why didn't they teach us this in school and I don't have an answer to that but I do know that at least now that we know that we have to encourage all of our various academic partners to do more with data. One of the things that you'll see is that other groups like TV w I, and EDMC and others are now increasingly moving into data management because they understand the critical importance of this. It's also critical to build our skills out around these topics in a way that can help all of our knowledge workers as well as our organizations, understand the difference between good and bad quality data on this and let's start out with a basic assumption is of unknown quality and that's a very problematic activity. In addition to that you want to assign values to some of these things in here and take advantage of the increases in workload so that we can more effectively manage data in our professional as well as our personal lives. I want to show you a quick chart here this is a standard Gartner hype cycle chart for data management here and I just want to point out a couple of things on the chart that are kind of important first of all, if you're in the data mesh space Gartner has said it's obsolete before it's going to plateau so it's just, nothing's going to happen in data mesh, according to Gartner now we may have some other people that are there data fabric, same thing a very popular term recently. But over the top of the hype cycle and it's now on its way to the trial of disillusionment as well doesn't mean it won't be there going forward but it is nevertheless a problem. Finally, just to circle a couple things here master data management data lakes are now at the start of their climb up the slope of enlightenment so these are not things that most people think about when I'm buying a master data management in years previous to this it was probably not a good idea to see it falling to the bottom of the trial of disillusionment that occurs in there and that's going to be a challenge for us all the way around so lots and lots of things that we can talk about what we're going to do today is look at some motivation and understand that we're unsatisfied with the current state and quite frankly among us chickens on this webinar we are not making progress in substantive means. The other one is how did we get here and it's by a kind of an unusual process, I happen to be part of it but lots of other people want to make sure I highlight some of my colleagues that have done this as well, but there is a push in industry for best practices around and there are two real important parts, one of them is the DMM the data maturity model. The other one is something that we call the dem Bach, that would be the data management body of knowledge will talk briefly about those and then look at how they work well together. The key is that the entire architecture is held together by a weak link in the chain architecture means your weakest link is the area that's going to cause you the most problems in this field, going forward, looking at it will do just a tiny bit on strategy and talk about the need to look at all of these solutions as holistic solutions. I call it a three legged stool on this, for some very obvious reasons, and we'll look at the process of practicing and of course this is it for those of you that are musicians. How does one get to Carnegie Hall well it's not walk up Fifth Avenue and take a left it's practice practice practice. So then we'll talk a little bit about where we're going further on this but last part is we'll get Chris back up here and start to talk about what your specific questions are around this in our q amp a segment. So let's jump in and get started. There are tremendous amounts of un productivity on this knowledge workers stress at having to work with data. 33% of knowledge worker time is spent reworking and recreating knowledge that already exists. 10% of their time is spent creating new knowledge and content. Oh my goodness, I think you just demotivated me for the rest of the week here Peter you know this is not a very good thing. 53% of knowledge workers would rather do household tours than work with data and 52% would rather pay bills they use your content management system and repository solutions that organizations are working with people. 44% of knowledge workers feel unhelp, unhappy or overwhelmed when working with data and one third of them spend at least one hour a week procrastinating on data related subjects. Now, that last point right there I'm going to tell you we used to justify a repository a glossary whatever you want to call it, the thing that you're going to put all your common definitions in there to manage your control vocabulary. We had an organization that I said if we could just save one hour per year for these would it be worthwhile. And they clearly came back and said oh yes we have a number of knowledge workers and if we just looked at one hour per year of savings that would more than pay for itself but certainly at one hour a week. There is absolutely no question that these are cost justified measures. Moving a little bit further with the measurements. When I draw your attention to something in the bottom right hand corner here which is the data literacy project link. They've done a great job this is a initiative between click and Accenture. And they have a series of surveys they've just refreshed the website and I haven't had a chance to get through it yet. These are some of the last refresh which was pre pandemic but nevertheless still very interesting 14% of everybody has a good understanding how to use business data that's self reported one and five of the new employees that we're going to be bringing on board though feel this way so they're slightly above average but nevertheless we can still conclude that our future employees are under prepared for these data driven workplaces and that our business decision makers have similar types of activities that they are stuck with 24% of them feel I'm so sorry I'm going to go back 24% of them feel like they can actually make decisions on this. Hang on I hit the wrong button here and I'm going to go backwards one because I'm going to make sure I cover this point here very very well so again won't read you the stats again but let's try it. Decision makers, again in this process, looking at one in four of them feel very comfortable making data decisions. Some third are able to create measurable value from their data, 27% less than 30% say that my analytics projects produce actionable insights, but most importantly when I'm sitting around a table with a group of executives as I often do. I can say to them and look at them across the table and say four out of five of you would pay me under the table to make you smarter about data. Let's just bring that up to the top and realize that our data literacy efforts need to be focused on the executive decision makers and board members as well as the knowledge workers that we're looking at in this again just a couple more of the stats on this but these are a little bit more frightening. Many organizations feel that they just have to put better quality data in the hands of their employees. The problem with that is that even when they get better quality data, 48% of employees make a gut decision instead of incorporating data into this. And it's even worse when we look at sea level Swedes two thirds of them make their gut decisions so I guess stop thinking is a criteria for being promoted. At this point that's a little bit mean but it is nevertheless, torn out by the data for sure. Similarly, when we look at data skills in the workplace, and you present somebody with a data task, half of them will take it on. 36% of them, however, will find an alternate method to complete the task without using data. I don't even know what that means but nevertheless it is a valid saving and team percent almost will avoid the task entirely. Now, this leads to very easy answers to the question. Why weren't my data problem solved when we moved everything to the cloud as many of you probably are in the process of or planning to, or have recently accomplished that. And I'm going to talk specifically about that because just because you have your data in a cloud or in a warehouse or anything else. There are three attributes that you should be concentrating on from a data management perspective, in order to make sure that these are different than data, not in the cloud or not in the warehouse. The first one, I hope you all agree with me, is that data inside the cloud should be cleaner than data outside the cloud. If it's not, what is the point. If it was less quality, that would really significantly impact our ability to achieve a return on that. Similarly, the volume of data in the cloud should be smaller than the volume of data outside the cloud. And that immediately stops people and say, well, wait a minute, what are we talking about here? Well, the other aspect of it, it should be more shareable or, I guess, shareable is probably the right way to say it in there as far as what we're doing. But if you architect correctly and take the cloud or warehouse or digitization transformation, whatever it is that you're looking at with your initiative and make that an inflection point, you will find that data can be designed to be more shareable. Therefore, you don't need to have as much of it. Let me be even more specific about this. If we're talking about warehousing in this case, or clouds or whatever, we tend to take this big pile of data here and we do what's called forklifting it in here. And the problems with forklifting it in there. Notice, first of all, the cloud got bigger when I forklifted into it because we just said, let's just make it big enough and put everything into it. But there's no basis for making the decisions that we have about that. There's no inclusion of architecture or engineering concepts in this. And there's no idea that these concepts are even missing from the process. Excuse me. So that's a real type of an issue and it gets even more important when you come back to the absolutely true statement that minimally 80% of the data in your organization is redundant, obsolete, or trivial. If that was the case, why would you spend any amount of time putting redundant data in the same data store, putting obsolete data in the data store or putting trivial data in your data store in order to do this because this is simply not the way we should be investing our money. So a more proper way to do this is actually to look at this inflection point as an opportunity to transform, clean up, make shinier, whatever it is we're going to do, so that the data in after this initiative in the cloud in the warehouse, whatever it is that you're doing is less and volume is cleaner in overall quality and is more shareable by definition than others that are out there and again this is the same whether you're doing warehousing or clouds or any other of these topics. The idea is that we really need to try and find out the opportunity of what can happen with data branding and what I say to organizations. I remember the very first ones I did was when Alcatel bought Lucent on this. We had a very, very big challenge around this and the strategy that was adopted was that we're going to use data that is of known quality. So this is an opportunity to brand your data to say it's being put into these new ways of doing it again data lakes anything else that you're doing here in terms of repositories in a way that allows you to do this. So of course, what will happen though and we'll see this happen. Again, I've got a sheet going here with now got 14 companies on it that have installed Salesforce.com and then decided to clean the data afterwards well, fixing the data after you put it someplace is kind of like wearing gloveboxes you know it's just not exactly what you want to do you certainly want to do this to protect yourself from bad things that are happening out there but at the same time, it's not the easiest way to do and if you could do those changes before you put the thing into the glovebox, it would certainly be able to much more effective way of working with this. I mentioned a three legged stool and the three legs of course are people process and technology in order to do this. If I just have a one legged stool I don't have a very comfortable ride if I have only a two legged stool. And yet, when we look at what's actually happening out there. This is by the way a wonderful survey that my colleagues ready to be in a time to have a board have been doing for years. The URL is right here you can go look at the data yourself but they've asked people and said you know are you driving innovation with data less than half are you competing on data and analytics again, much less than half managing data as an asset. And still not good numbers creating a data driven culture, or creating a data driven organization again all of these things are really problematic but the most important findings that they found in this set of surveys is that if we go back to your 19. In 2018, it's a 8020 rule as to whether your problems are largely caused by technology, or by people and processes and of course you can see there's 2019 results. There is our 2020 results and there is our 2021 results they are virtually identical, which, again, minimally 80% of our data challenges are people and process base just the same as 80% of our data out there is right. So leadership is the only resource that we have that can come back and address these challenges, because of course, there's this problem with perceiving these things that again, I can't ask you all directly but I know that there's a bunch of you out there that have gone to your management and tried to have real good differentiation sessions to talk about the difference between data management data governance and all this sort of thing and you know what those of you that are younger that's an old teacher talking in the old Charlie Brown Christmas Carol story that shows up on a regular basis they just don't get it they don't care. What you really need to do is focus their attention on a data program so that they understand that this is all part of what we have to do for our data program in order for us to do more with our data. So try and over explain this to folks instead keep them concentrating on the ultimate objective which is making data better as we go through this. I love this particular quote from Alice in Wonderland you know the Cheshire cat asks Alice where do you want to go and Alice is I don't care. Sorry I don't know says Alice and the cat said then it doesn't matter right. So I get this question a lot we want to move our data management program to the next level. Where are you currently at. If you don't know where you're at. You can't measure it. You can't manage it effectively and you don't know where to put time and energy into this process. Now this wonderful problem was presented to me when I first joined the US Department of Defense in the late 1980s. Given the title US Department of Defense reverse engineering program manager. I'm not going to say I got that for any good reason. Mainly because they knew it was a function needed to be done and I was hired to fill that particular function in fact, the actual task that I was first presented with was the fact that the Defense Department had 37 payroll systems. Of course one, but the fight to get down from 37 to one was going to be a political fight, as I've just shown you, not a technical fight. And consequently, we had to invent a technique that was something that the rest of the world could agree with. I say the rest of the world the rest of the Defense Department world that was in there. So the task of many of the things that we were doing at the Defense Department was we sponsored some research at Carnegie Mellon software engineering Institute, and we were asking the questions how can you measure the performance of the Defense Department and our presenters that are doing this, the partners that are working with us. So the questions of, if you're being asked in do to do things can we say that some things are being done better than other things. And that is really where the origin of what I'm going to talk about on this happened on this in response to research that we sponsored out there. So with that, I was also told to go and check out to see what the Navy was doing I was DoD corporate so I was working out of the Pentagon, and I went to go see what was going on in the Navy and found out it was a joint seminar by Clive Finkelstein, and and that was sort of my introduction to the space in particular. Back to the SEI story though software engineering Institute responded with an integrated process and data improvement approach. It was very, very nice. And, interestingly enough, DoD came back and said your name of the software engineering Institute that's what the S here stands for is software so you have to take out the data portion that you did before you can turn it back over to us because basically as a software engineering Institute unqualified to talk about data. Now, we'll just let that sink in for a minute that's the kind of knuckleheaded thinking that's gotten us so so many lumps in this particular business. What happened was I was up at the SEI at one point and Dennis Smith who was the PI out there said hey Peter, DoD didn't want something that we produced for them would you like it. And I said, Well, absolutely. I'd certainly take a look at it and it grew into this CMMI dembach type of work, etc, etc. I want to introduce you to our departed colleague, Bert Parker, who really took this on from me he went over to the MITRE Corporation in October of 94 and started to work on this this is the paper we wrote out of it if you have any interest in it at all I guarantee it'll put you to sleep, but the real key for it is that it gave a very generalized approach to processing improvement. And that's what we're going to talk about in here because you will need to look at your own areas and find out at what level are they, and then how can you improve them in order to move next so this gives us a normative model for data management. It's required, and you need to understand the scope in order to do this so that we can organize these key practices and report on what actually happens here so again I'll include this paper in with the slides that come out of this particular event. So that you can read this if you want. I've never had anybody complain they didn't go to sleep so it'll probably be quite useful for you. Quick bit on the CMMI at Carnegie Mellon. Again, they came out of this as a federally funded research and development center and have now a 20 year history of looking at this process where they've taken the same improvement process and applied it to software development acquisition in the area, services, people management of course to data management as well. There are training and certification programs that they have but recently they sold all of this stuff to ISACA. So ISACA is now the group that owns this portion of what goes on here. I want to also introduce you to my friend and colleague Melanie Mecca, who sat down with a wonderful opportunity from the CMMI and produced for us. This wonderful thing that we talk about the DMM. It's the idea that we're now going to be able to say, hey, here is a good way to do data practices better. 1.0 was released in 2014. We're looking for some funding to do a second version. There's been a little bit of change, but the thing is actually held up reasonably well over the years. There were a lot of contributing authors. And most importantly, it provides very specific practice statements that will help you understand a bit more about where they go in terms of measuring this. So it gives you the idea that you can take a look. Here again is the basic structure. There's a core category, a process area. There's a purpose, some introductory notes, goals. Then there's a one to five set of scales that we put in place. And again, give example work products in order to do this. Each area follows the same process. And it focuses on what you actually do. The model emphasizes behavior as all good research in this area should do. They should take a look at the changes that you're doing so that you can carry out repeatable processes and then take those processes and extend them across the entire organization. These produce work products that allow the organization to standardize what it is they're trying to come up with and what they're trying to gain by looking at their data better. Key here is of course reuse and extend these find an area that works and leverage the heck out of it if you're going to do this. It reflects real world organizations and real world processes that allow us to take everybody and align them in the same ways. And most important of all of this, the exception of CMM and I tell these are the only ones that actually show improvements in some key measures. So for example, when we look at projects on budget CMM I produces the biggest improvement in these areas in order to do this and actually things are going to get worse if you go to the regional process or Kobe or PMIs area around the seas are just really, really solid statistics that are problematic same thing for on time delivery. Once again the CMM I process allows us to make the greatest improvement here, whereas you can see the others sometimes make things go in the other direction rep isn't quite so bad here but certainly Kobe and PMI are very problematic. So the first thing is that when somebody comes at you and says I'm going to do an assessment for you and if you're going to pay them. I would suggest that it would get better results. If you look specifically to the CMM I process, because I've gone out so many times and I will see company X told me they were doing the Dama process will Dama doesn't have one right, but CMM I does. And people say oh it's the CMM I process right now look at it and say no it's not I'm sorry it is not they are lying to you or misinformed but it is not in fact the process so many organizations try to make up their own on this, and they just make things more complicated. If you look at the fact that we have 20 years of good quality research in this area. Why on earth would you start something new unless it offered something significantly better than the results that we've been gaining so far and again remember, these are the only ones that produce on time and within budget types of results. The other part of this is something we call the DIMBOK now before the DIMBOK, we just have these little handbooks here here's one that I've got on my desk give a handbook on data management and information systems. Right well, they're academic they're not helpful. Instead, what we did at Dama was put together a multi year effort this is the second edition of the DIMB data management body of knowledge, published in 2017 but still current today. We started to take a look at this and what you can see is that while it's a better step it's easier to understand this and to say that these are activities that need to occur if you're going to successfully manage your data across the organization. Obviously there are 10 high wedges, and the thing in the center is called data governance so we keeping that very much at the center. Unfortunately, this icon this meme I guess we'd call it nowadays is is not as perfect as we'd like it to be because it doesn't show the functionality and dependency to things that are very important for data management. One of the things that happens is that people will look at this and say okay Peter. You're the president of Dama Dama published the the DIMBOK here and it says I must do data integration and interoperability or I must do document and content management. So you can see how that would be understood by putting this out here what we really should have said was document and content management something that you may do as part of your data management activities but should only take on if it's going to produce a positive return on your investment. So really we did a little bit with the governance piece and say that you know dependency wise data governance is really kind of the heart of all of these things but I don't think it quite articulates it in the way that we'd really like to and perhaps some of you all will have the ability to come along and talk to us more about this as well. Let's talk and spend some time here talking about how to apply them together. Now, this is where I'm calling you from today. This is my home in rural Virginia. I am at the end of a twisted copper pair, a wire set, and I'm not showing you that for that reason but instead showing you pictures of a foundation that is in my front yard. Now, the foundation eventually became a barn. If I had a live cam I could show you the barn is out there and populated with horses, which is a wonderful to do but the reason I'm showing you this is because I borrowed money from the bank in order to build the barn. The bank gave me exactly this much money to do it and then they said stop. Now that you have put the foundation in place document it and show us that you've passed a foundation inspection for the county that you live in in this case Western Hanover County in rural Virginia. The banks pretty good at this and they said you know you build a good foundation and then you can build a good one on top of it but if you put a four foundation in place. Peter because we know that you're what's called a horse husband, then I will pay the components in terms of any bills that would come up from a bad. Barn that I might build on top of this. So it makes good business sense from the bank's perspective to say we're going to give you this much money when you prove to us that you spent that money well on a good foundation. You will be able to borrow the next tranche of money that comes out in order to do this and the point here is there is no it equivalent. I'm working with one company now that that has, you know, 100 connections to a hub and they want to move it to another hub, and they want to know if they can be done by Friday. Well, it's Tuesday, and probably the answer is no if you have 100 connections. We need to understand those connections but they've already spent the money they've put it in place there they're getting started on the process, and it's a day late and many dollars short in order to do this. I like to show this because it's very much related to a concept that I learned in high school as many of you all did as well Maslow's hierarchy of needs. Maslow was very astute in understanding that if we have food, clothing and shelter needs that are unmet, we will not be able to be safe. The second layer is dependent on the first layer in order to do this. If we are never safe, we will never be part of something that is bigger than ourselves. And if we are never part of something that's bigger than ourselves and we will never know ourselves in respect to other people I was talking to two young women that were thinking about becoming PhD students, and it was a great conversation that they had, and they said, you know, I didn't realize that the demand was so huge out there well yes, absolutely it is. And it needs then you to have self esteem before you can get to where you'd like to be which is what Maslow called self actualization. I'm showing you this for a particular reason because lots of data represents itself in exactly the same way. Unfortunately, the TED folks came along a couple of years ago and hijacked the self actualization piece and now they call it flow. Okay, well whatever you're going to call it is the idea that you're working at your best capability you're happy at work you're producing things that are useful, and you're doing it in a way that doesn't inhibit you. It doesn't cause you frustration that doesn't cause you to go out and work on household chores instead of actually the data in your organization, and data is an awful lot like this that I've already mentioned, the technologies that people are using again data mining right IML data ops mesh crypto all these things that we've looked at here. These are just technologies, and they really represent the tip of the spending iceberg that we have around this in order to truly do investments and do them well. You need these foundational data management practices to be put in place just the same way as my barn needed to have a good foundation out there in order to get it right and these things are not technologies they are capabilities. The organization has to learn how to do hint hint hint practice practice practice right now. I'll show this a lot people will look at these lines and diagrams and things and then they'll say, well that's great Peter but can you do it faster. And the answer is yes, I can do it faster if I do it faster it will take longer. If I do it faster it will cost more I do it faster it will deliver less if I do it faster will present greater risk to the organization. And the reason for that is because people not understanding this will go out and hire a bunch of data scientists in the top part of this thing and say get started doing whatever it is data science does. And they'll turn around and say great where's your data dictionary where's your capabilities, where's your good clean quality data because I don't want to work on the bad quality data, and they get blank stairs back as a result it takes most data three years to become useful and productive in an organizational capacity. And if you know how much we spend on data scientists, it's not a very good investment. In fact, the better way to do it is in fact to hire data engineers at first data analysts, and have them put things together and then, when you're ready for it bring the data scientists on. The organizations don't don't have that kind of patients. Now we're back here I know I'm mixing things back and forth but I'm back here now to the DMM, and they have some things that are similar to the pie wedges here but what I want to show you is how they go about the process of looking at them so they look at it as a data management strategy which says we have to manage the data coherently. We need to manage them professionally we have now after 20 years a cadre of people who are data management, excuse me data governance professionals, who we can have talent to do the right kinds of things that are out there that there is a data operations, the life cycle that we need to pay attention to that works for our organization whatever it happens to be that data quality is a search for fit for purpose data so that we can again go with the branding things and I was talking about before. So of course we need to have the operation stack in order to do this. All of this of course needs supporting practices. Now, the problem with this as I mentioned before is that if you have a weak link in the chain that your entire data foundation can only be as strong as the weakest link. Now, turns out there's five levels that we spend our time talking about in the data management area here. And again just to give a little brief thing you get one point for having a pulse that's a pretty low standard if you have any sort of repeatable processes you get two points, even if the process is give it to Peter because he'll know what to do with it. All right, if you have any of that process written down at all any documentation at all you get a three points for being defined. If you measure any of the things that happen within your defined processes, you have four points, and you get five points, you actually optimize something based on the measurements that you've collected from your defined processes again very very brief these things are all recorded you can go back and rework it again but I'm going to give you an example of why that's so important. And that is that in this organization hypothetically that we're talking about. They are at a three for their data governance their data quality their data operations and their data platform but they are a one with respect to their data strategy. Because it lowers the entire organization down to a one, because if you don't know where you're going any road will take you there. So we've taken these two components here on the left hand side. Those are the five areas but they can just as easily be the pie wedges that we're talking about and we perform them at a one, two, three, four, or five level. We can go through and there's a method that we can go through and produce charts and things that are out of this. Again, the idea is if you know you're not very good don't spend a lot of money getting somebody to tell you that process. This type of detail is less useful at that point in time but as I mentioned before, we can do this over time and look and see how things are going into place with the various measurements here. So these are ranges of where the practices are and ways you can help to dial in the overall process. As I said, not a lot of time investing to tell you that you're at the beginning of your journey, but do look at this assessment process to uncover points of excellence that you do have within your organization. I've never failed to walk into an organization as Shannon mentioned. I'm up to over a thousand organizations that I've individually looked at specifically in this and there's always wonderful pockets of excellence that are in there where the people don't know how good the data management practice is actually being. And then let's see what we can do to take those pockets of excellence and expand them to other parts of the organization as our first step in the process. Now when we get these results, we can produce something like this. This is a chart that I put together several years ago when the insurance industry asked me how they were doing. You can see the answer here is they weren't doing very well as a group. These results are dated. The insurance industry has come much since then and in fact the insurance data management industry took these results and particularly focused in efforts in order to do that. Same thing here. This is an airline that I did this with and if you can imagine the airline looking at this and going, so I see some ones and twos up there. Executive management at the airline really doesn't care until I say, by the way, here's where your competition is and they go, oh, the competition's ahead of us. Now even if you only get that across to them with this type of a chart, it's actually a really good way of taking a look at it. We can also look at this compared to all respondents and see what the airline industry is better or worse. Most importantly though it shows the challenges. Remember our model is if I don't take these ones and make them into a two, there's no way the organization can ever get to a three in order to look at this one last example on this. I'll mess this one. Notice I have the World Bank logo on here because the World Bank actually published these results. This was kind of a fun one. We went in and looked at their treasury operation and their information systems a group, and both of them were kind of low. Okay, ones, but when we went to their business component and said how are you doing from a governance perspective they were doing world class. And that was something that they hadn't realized until you compare yourself to everybody else. You don't really know so in this case, the matter of learning more about data, how to do data better was a matter of walking down the hall go from Treasury and the information systems group to the International Finance Corporation, which was doing excellent, excellent work around all of this piece of bad news for all of us in general between 07 and 09. They've been measuring these maturity levels and they just haven't changed, particularly when you consider how much more data has come into place is another measure we can put on this as well this is something called the national assessment of adult literacy. It's recently been transitioned to another group, but nevertheless they still been measuring literacy, numeracy, and digital problem solving for years and years. And of course, once again just as the chart before we're not getting better about this and data is of course, crazy out of shape going nuts, in terms of what we're trying to do. I like to use the definition of strategy that comes out from the military, as opposed to the one that most management consultants use strategy is a pattern in a stream of decisions. For example, if I am looking at a cycle of a problem. I like to apply what's called the theory of constraints to it. You start out by identifying your problem exploiting your constraint subordinating everything else to it if the quick fix doesn't work, and making it an actual program, if we don't get it right. Again, these are things that we can do but if it looks a little complicated here this is really just plan do check act standard Deming quality improvement cycle around this. You need strategy though because you can't have something that says, well what do we have we pull up the manual that says the good guys are here and the bad guys are there. Or if we find the good guys are here, and the bad guys are there. Or the bad guys are here, and the good guys are there. All of these are problems from a strategy perspective and if somebody tells you that strategy is really good. Just ask them to go back and read what their strategy was in 2019. I think that most everybody's strategy was completely disrupted, but that those organizations that had a more flexible and adaptable strategy that was focused on a pattern in a stream of decisions did better throughout the pandemic than other organizations that had a very rigid strategy. This entire discussion is summed up very well by a quote from Dwight Eisenhower, who in addition to being our great general was also president and he said in preparing for battle. We found that plans are useless but that planning is indispensable in and of itself. Now, we need to also talk about the idea that the leverage point in this for data has got to be well performing high performance automation, and that that's difficult to do if you're relative literacy of the organization supply of the data, and the standards that are used in the organization are of uneven quality. So a focus on data management all the way around is to try and address these to harmonize them to make sure that they will work correctly together, so that when we put them in place, and put them in place that starts to achieve our organizational objectives, we understand the value that engineering and architecture have to do. In fact, I had to travel all the way to India to see a Deming quote on a cash register at this very T farm in India that said quality engineering and architecture products do not happen accidentally. And if we add the word data into that of course it is still even more true so let's take a look at how this would start to work. So you may have an idea that we're trying to do something with data, and we think that we need our three legged stool well the three legged stool also says you probably need to have at least three components. In order to do this so here's our first attempt at this that we're going to do one exit data government we're going to do one unit of data quality management we're going to do one unit of data warehousing. This is the three level stool that you really need to have an event. Most organizations do not have success, trying to operate within a single pie wedge. However, as your project evolves same project you may go in and do some more data warehousing, and some more governance because those two were going to need, but now we've shifted our focus from data quality over to metadata management and put together our project in this fashion. Our third version of this. Again, depending on what happens, we may discover that the metadata management was good but we also need to include something called reference and master management. Notice what's happened here in the process of going through this exercise. The organization has had three practices at data governance three at data warehousing and one each at reference metadata and data quality. In order to do this, the idea of finding also where you start your project is going to be something that I like to call a lighthouse type of an activity here. The idea that there are certainly things in your organization that if you had better data support for it, it would further your organizational strategy similarly you have in your organization data that's being used by the business but can be improved and the intersection of those two will be better than either with the resources by itself. Finally, you may have the opportunity to practice some data skills and only when you have that really good set of all of three of these things coming together is the sweet spot really set up to well set you for what needs to happen going forward on these particular projects. As I mentioned before, I like to do this in a Bruce Springsteen kind of way I won't play the whole clip here for you but he's at the moment getting ready to play a song that came out in high school when I was there. The key of course for this is that his band is a wonderful band, but more importantly, they've taken some very good music and not just improve the data that goes into it, but in fact, practice and practice this band has been together 40 years. So they are absolutely phenomenal at making music. And the only ticket that's hotter out there at the moment is the Taylor Swift tickets instead of the Bruce Springsteen tickets around all of this. So let's talk a little bit about where we need to go next in order to do this. And that is the idea that there is a different set of skills that we need to have in here called change management and leadership. And if we can get people to do things the new way, instead of the old way, we've actually achieved change. It's one of the hardest things that we have to do in organizations. I can recall when I was first brought into the Defense Department and I had my swearing in day you have a whole day throw ceremonies and things and you swear to uphold the Constitution of the United States and all the rest of the things that we do in that. And I had somebody come out from the TQM office, who came out and said, Oh, and by the way at DoD here we also do TQM it is in fuzz in everything that we do. And of course their use of the wrong word there infused is the word they should have been doing. They should have actually made a difference in there but they were not. It was just lip service to it. And it was unfortunate because TQM is actually quite a good way of improving quality. There is data TQM that's in there but actually everything that I've shown you here today will help you get started around all this. The only else that we're going to have to explain to our kids as well is what is a physical lock, right all of our locks are digital today so we don't have this but notice the way this is showing you, it's showing you how the tumblers are raised to the right level and only when you have all things lining up. Can you in fact achieve organizational change. It turns out from a data management perspective you have exactly the same type of equation here so I've walked into organizations where I see there's a good vision, great skills, wonderful incentive and action plan, and lots of frustration. What are they missing of course, whether missing resources. Similarly, I've walked in and seen vision incentive resources and action plan and anxiety and when I see that anxiety I know that they're missing skills. I wish I could tell you I invented this chart I did not. This came from Mary lipid, who did a great job of talking to us about organizational change and the most important part of this of course, is that the only way this works is when you have vision skills incentive resources and an action plan in order to change and get what happens with most organizations is that they go out and buy some technology, and when you buy technology that's not going to help your organization as much as it could. If you instead looked at the people and process dimensions of your data management challenge. All of these are very very important in order to do this and this is not a new field change management exists. Many of your organizations have groups that are already competent in this area and able to assist with all of this because culture is in fact the biggest implement, excuse me impediment to the shifting in our organizational thinking about data. These are all very significant challenges in order to do this. I have a case study that you can simply go online and download with no registration that talks more about this it's about a 70 page piece so one of those other things that will put you to sleep at night for sure around this. So let's go back to our beginning, where we started out about 50 minutes ago data volume is increasing faster than we're able to process it it means we need to have some good things in place and we have not educated our folks to do this we don't tell our knowledge or anything about data we tell our data people barely about this particularly data scientists are not well equipped to handle this that data interchange overhead and other costs are sapping the organizational procedures in order to do this that technology based approaches can only solve part of the problems and are not materially addressing any of these and finally we have to of course keep our mind. The surveillance capitalism industry is a great place to go for models on how to do it, but probably not great models on what to do. Process is much more important than results at first and that failure itself is a lesson I finished last night with my classes and my executive sponsored last night and one of the things that the students really felt good about was that they knew some things that definitely weren't true on the exercise they had gone into. Well, if it's definitely not true that's still good data points that's a good thing to have out there and that the people in process aspects are simply not receiving enough attention in order to do this, but that best practices do exist and the idea is simply work your way up through those pie wedges start off at a one, then start to find a repeatable process once we found the repeatable process. Now let's go a little bit further and document that repeatable process now that we've documented we find we're following the documented process let's start to measure how long different aspects of that process take. Once we have sufficient measurement, we can now look at reengineering in a way that actually does create some very significant improvements in the area overall. So to practice data management better it is best to think of it as something that you are in fact going to practice. We're going to start, we're going to do it, not perfect the first time because we're humans. And because there's no wonderful thing like generally accepted data practices out there that we can follow in order to do this but we are unsatisfied with the current state and we are committed to making some progress in there we're going to get there by following a group of things that have been put together through a combination of the Defense Department as well as data international these are the two icons that I'm showing on the right here. The data management on the left is the one from CMMI and the one on the right is from DEMA, and you can put them together and make one and one equal 11 with these it's a very, very powerful technique again. Each of those have to be applied in a way that focuses on the thing that you are worse at at the moment. And you want to try to get a bit better about that. Of course there's lots of things we can get better at so we need to have a focus on strategy to say which things are more important than others, and make sure that we put together is structurally sound that it has a three legged stool instead of a two legged stool, and that we're not going to get anywhere fast, but we are going to start practicing in order to get to Carnegie Hall around all of those bits and pieces. So the idea here is, let's not take a look at this as one time in order to come up with this but instead say, this is a skill that we're going to start and get better at and I'm going to tell you a quick story around here of a related field. The related field is human resources I mentioned it once already HR used to be just like data, it was diffuse it was done it was figured that the person who was going to be the best person judge of the person to hire would be the people that they're working with so hiring practices were decentralized around that. Anyway, long story short about 80 years ago corporation started saying you know, there's a lot of risk associated with the old model we should do it centralized and nobody's ever gone back to them and said, we should get rid of HR. Data management is exactly the same practice. You've got to start somewhere and make it work your way towards that. And we are now at the very top of the hour and it's time to invite Chris back on and see what sort of questions you all have for us so thank you guys for listening and Christian Shannon back to you guys. Peter, thank you so much for another great presentation. If you have questions for Peter or for Chris feel free to submit them in the Q&A portion of your screen. And just to answer the most commonly asked questions just reminder I will send a follow up email by end of day Thursday for this webinar with links to the slides and links to the recording along with anything else requested throughout here. So diving in recommendations for how to collect information about the data management practices at the industry level. Good question. Chris if you know of anything please jump in what what I'm looking at is, we've got about 20 years worth of data in this process of collecting it the way I described it here so here for example. The results that we pulled out for the World Bank that we did the purple and the teal lines that are showing the first two in each of the groups are World Bank as well as the long line the IFC the sort of blonde colored line there but we do have industry benchmarks and I can go back through some hooks and crooks and have some data that we can say hey, you know, for example. You can tell this one to the airline industry about just tell you that the the airline when I showed the competition, they told me well they weren't going to do theirs until we had done a competitor of theirs and they wanted to use this process to see what the results of their competition were. There are things out there Gartner has some sort of measurements they won't really tell us what their method is. So we are starting now to see that groups are looking at these types of practices within industry, and it's probably no surprise to you the financial industry being the most heavily regulated is also the industry that does best in these practices, mainly because there's no reason to comply with Basel three or any other things that go on finger Cohen. It's the one that came after and Ron Dodd's Dodd Frank Bill, I think it was acquired them to pay attention, they, they are better as you move further away from that manufacturing is going to be your next most mature industry because manufacturing quality very very well. And that is a really good place to apply in this type of a context here as well. After that though with sort of catches can, and I would also say one other piece of movement measurement here is that one in 10 organizations formally attempts to try to do this so if your organization is making some attempt, you're already in that top 10% of organizations out there because remember people who don't understand or care about data aren't going to be on a webinar they aren't going to be looking at their data management practices so the fact that you have somebody in your organization besides yourself interested in this is a really wonderful thing if you've got specific questions about industry feel free to reach out to us. We'll be happy to dive into more detail there but that's a, I think a good answer to that Chris anything on industry types of performance in there you've seen in your travels and things. No, just, it kind of is a negative space type of thing. Right. I don't see it, right, I mean you just don't see it you don't see a lot of organizations that actually understand the level of their own data maturity, and it's kind of a shame, but we don't see a lot invested into that level of discovery and organizations. I think it's something that we should invest more into. Again, the, the other analogy that I like to say is, if you've got a brand new Tesla, that you've just spent a fair amount of money for and you've got a 16 year old, and it just snowed outside in the Washington DC area, something I'm intimately familiar with. You probably aren't going to hand the fob to your Tesla to the 16 year old say good luck go out and try and drive in the Washington snow traffic because nobody in Washington DC can drive in the snow people who know how to drive stay home. People who don't know how to drive in the snow get in the four wheel drive things or their SUVs and go crash is basically the way it works. So don't don't look for a happy outcome based on that and do the same kind of a rational process. Okay, so we're doing something. We know it costs us money again go back to my data sandwich analogy that I had earlier. And we're having a lot of one organization I worked with actually had to stop selling things to customers when they ran into 10,000 errors that occurred in a day, because they knew they would never be able to fix the errors before they next started and so they would always be at the end of an ever growing pile of things. Well, at least they could do math. In order to do that but it was about a 10 year journey for them to completely get transformed around, which is one of the reasons we have to make sure that management understands that there is a journey in place that we are improving, but it's not going to happen on their watch it's probably going to happen on their successors successors watch, but you still need to put the effort into it because if you don't, you'll never have the chance to get better. And one more thing I'm sorry just keep going on but I had a brief chat this morning with somebody I hadn't talked to for literally 15 years and they were at a company that got acquired by another company I said gosh you guys used to be really good at this stuff. What happened when you were acquired and they said they threw it all out. They didn't see any value in it and guess what we're struggling in the marketplace now, not surprising information. Thanks, Jen a great question. Just to note that the questioner was is from higher ed. Anything about the industry that you want to add on because I know you are in the industry as well. Yeah, well first of all we do have an interesting thing happening in higher ed if you all haven't heard we have about a 15% drop in students coming up in about three years the reason for that is quite obvious we were all around here and oh seven when the oopsie occurred and people decided to stop having lots and lots of school age kids so we have about a 15% drop in enrollment I don't know about any of your businesses that can turn around and say, oh great in three years we're going to have 15% less customers in order to do this so clearly we have initiatives that are out there like how to sell more goods and services to students that are coming to class and it's like no no no that's not the right answer we need to actually readjust our capacity downwards that's going to be interesting on that. And of course researchers in the university community are particularly interesting animal. Yes, there are some things that have been done in the university community and be happy to take them offline, but that's probably not for general for everybody else. But absolutely there can be some things happening in those areas. Okay, so any recommendations for evaluating data management practices I've already pushed purchase your book monetizing data management and how to measure anything by Hubbard. I'm wondering if there's a resource specific to expressing the value of data management practices as a tool for justifying a C suite investment. I think that the, the, the measuring date, excuse me, the monetizing data management book is the closest thing that's out there you've already mentioned Doug Hubbard's book, I'll put a little clue for it there. Doug was inspirational to me he's out of Chicago gives great seminars well worth your time and effort. And as the questioner said there's a wonderful book out there called how to measure anything it's a great book that'll give you lots of inspiration on how things end up costing. So it's it's I think the subtitle is how to find the, the business value and the intangibles that are out there that's probably not exactly right and I apologize to Doug for that. The other book I would mention although it's probably not going to give you the guidance that you want it's certainly worth reading is Doug Laney's infonomics, wonderful text that he put together about the same time. It's one third PhD dissertations it's one third really good advice it's one third Gartner Gartner marketing stuff, but it's still well worth reading in there. That said, the process is pretty straightforward what you want to find out is where are your knowledge workers spending most of their time in those unproductive areas know the things we can do to help them. It's very very clear that if you're in the process of creating new knowledge as a knowledge worker, your inputs are going to be important and if you spend 30% of your time looking for inputs that should be right in front of you. You should be able to improve the flow into that process of data into that process and help those individuals become more productive. So I like I like doing process modeling process architecture and things like that to help out. Again Chris I'll go back over to you when you walk into an organization and see they're just getting started how do you tell them to get what sort of resources do you point them to what sort of things do you have them think about. Follow the data, right is is really the key concept that we tend to provide to them that that works right that that is an action item that they can they can take on and on and execute, which is essentially what really they're looking for. And what I mean by that is it's really about the process engineering right that you're talking about there Peter, but it's getting to which process and oftentimes when it teams are the owners and managers and operators of, you know the data platforms and the data supply chain with an organization. They're, you know, less than directly associated to the business, and they talk in a completely different language. So the way for you to get on terms we're able to have conversations and get to what the business value really is is follow the data get it to where it's actually used what business processes does your data impact, how are you changing that data really the data availability, these type of things and understand how it impacts that business process. That's how you get to value true business value and to what the folks on the business side within organizations will see as value and what they care about. Absolutely couldn't agree more it's it's the case and really this is following the Tom Cruise advice from Maverick on the original movie right follow the money. You can follow the data, you can really look at this I this is another concept that leads us into and I have a big issue with this and because I'm old and cranky I can actually really pontificate on this. Don't let anybody own your data, the minute somebody owns the data they start to do the same thing that everybody else does they start mining it. So I can control the requirements never know the only thing you can let somebody own in this is the requirements for the data at a particular point in time, relative to the process that you're looking at. So Chris just exactly as you said, what are the inputs, what are the processes what are the outputs if I can follow them through. I can look and say, why would it take somebody this is an example from IBM credit, two years to decide whether excuse me, two weeks to decide whether a university that has always paid its bills in the back and the past and is backed by the full faith and credit of the Commonwealth of Virginia which we have a triple a rating in Virginia so that's actually a good thing. Why would it take them two weeks to decide whether to give us another loan when we just rolled over the last one on this. Well it was a bad business process and it turned out there was only about 15 minutes of work that occurred during those two weeks, which meant that IBM could actually sell a lot more credit in their process in order to do this so exactly as Chris was saying, follow the money, so follow the data which is money, and you will find good areas to start working on on all of these practices with. Perfect, so let me know in here if a tiny team performs an initial data management maturity, like assessment on a very large organization, where should the team concentrate in order to get the best information. The fun part about the way I get to do things is I get to walk into a room and some executive announces we're doing more with data and I look around the room. And I want to see the people who are rolling their eyes in the back of the room. We've been here before you know we you guys have tried this before but it's never worked the skeptics, if you will. Again, particularly when I was do the the response usually was I've seen your kind come and go Peter and I'll still be here after you're gone so you can go ahead and tell me whatever you want but it ain't going to change nothing right. Well, those skeptics know what the real problems are in the organization was preventing the organization from making that graduated leap to go from an initial place to a managed place from a managed to a defined. By the way, something else really quick on all of this to it is not necessary for all of your data practices to be a level five. Your organizations are making money and are for the large part profitable at this point in time. The question is if you improve some of your practices to an optimized level with that resulting even more money to the bottom line for your organization or more service if you're a service based organization. We did one in Virginia here where we took childcare interventions where somebody's actually making a decision about whether a child should be ripped away from parents or not, and found out that 40% of the interview questions that they were being asked had no material value at all. And so they could stop asking the questions and most importantly deliver over a million dollars from administrative overhead into service delivery. And that's a huge amount of money, particularly for a state agency. And that's the kind of thing that you're looking for by doing data management, better around that process. So the question of where you look for find the people who know what's going what are going on in there. We don't have AI that can actually do anything other than write English essays perfectly well, which should scare the heck out of anybody in the liberal arts community at this point in time. If you haven't had a chance to look at the the open sources that are out there. It's where we need those people and the people know how to do things. They are the ones that are going to talk about the primary blocks and things that are most important for the organization to be able to do, and they'll know for example but if you could sell, we could sell a billion more eggs if we had a different kind of container to hold them and because, well actually I can make a very specific example Amtrak would be made a more profitable organization if they learned how to manage food better. And that sounds really weird but Amtrak's expertise is running trains, not actually managing food and they have huge food spoilage problems that are there and of course if you've ever been on Amtrak I'm not telling you anything that's surprising to you. But Amtrak clearly needs to learn how to manage its food better and the data around its food so that they don't have as much spoilage in the warehouses which is the thing that's really been hurting Amtrak from a profitability perspective. And that just seems crazy doesn't it, right? Amtrak runs trains, why would they be, why would we expect them to be good with food? Well, food's part of the process and it's clearly something that Amtrak is looking at in order to get better and become more profitable. Again I babbled on a bit, Chris when you walk into organizations, where do you tell them to start? What sorts of things do you say? What are the first things that they should be looking at when applying something like your technology into the organization? I look for the problems, right? And you're kind of stating that as well, right? The people that will share the most and most poignant information and the most information with you are often the people that are having the most trouble. Right, so in understanding what some of their problems are whether it's, you know, a business process doesn't function optimally because of perceived problems with an application or things like that. So understand that application's owner's frustration, see if there's, you know, how data may impact that. And these are ways that you can do two things by understanding that. One, in general in my experience, you just, you do, you get more information, you get somebody's attention, people like to talk about problems more than they like to talk about resolving problems or resolved problems, right? And that's the human nature thing. When they do that they share a lot more information with you and a lot of that information can be flipped around so that once you've attached how data is part of the problem or maybe part of the solution as well, you're understanding how data is contributing to the value of that business process as well. It's literally just taking an issue or a problem and understanding the impact of that and flipping it around inversely to the value that that data is providing for that specific process. And once kind of that connectivity is made, it does go back to the previous question around value. That's where you get people willing to invest time that it takes to concentrate on these assessments because at the end of the day, people aren't going to spend on time on something if they don't perceive there's any value in it. And when you attach it to how it could solve a problem for them, they're much more apt to volunteer and spend that time, make that investment. Interesting. And the way you were describing that Chris, I have another talk that I do where I talk about a 12 step program for improving organizational literacy and, you know, you take it as a 12 step program because it's a problem and the organization needs to admit to the problem and people kind of go wow that's really good but actually what you described there was closer to cognitive behavioral therapy, which is really what we need to have for most of these organizations and say, do you like bumping your head against the wall. Okay, we can get you to stop bumping your head against the wall and you'll actually feel better and things will get done faster. One more quick example in there as well and this is a story that was relayed to me and I think it's imperative to get you all to do this. Xerox had a expert system that they had put together at one point in time that was supposed to help Xerox repair and repair the Xerox machines now this was back in the days when Xerox machines were big. You know they would occupy corners of the room and things like that. And nobody in their organization would use it. And they finally went to the Xerox repair people and said well how do you diagnose the problem. And they said well I walk into the room where the Xerox machine is and I pick up all the paper that's in the trash can, because those are the copies that didn't work and they'll show me what the flaws were that were in there and that's a whole lot more insightful to look at those errors that they had coming out. Then it was to try and you know go through several steps of diagnosing the machine and getting its general health and all the rest of that sort of thing out there. So yeah absolutely there's some definitely good ways of doing this and there's some ways that are less helpful which is just sort of generally interviewing everybody. Right so as Chris said if you can find the people that are motivated to do this, they do want to talk to you and they will be very happy if they figure that you actually have a solution that doesn't involve sort of silver bullet and that's of course what everybody wants but doesn't exist. Shannon great question let's do another one here. Indeed so what do you do if one side of the company data governance is on the slow journey to crawl walk run and the other side of the company wants to move things to cloud quickly despite all the data chaos. So how to get the organizational alignment to ensure a data lake instead of a data swamp, there's no point moving the junks to the cloud. Yet it happens over and over and over again. I can't tell you how rich Amazon is getting by having just rubbish stored out there it's just Amazon Azure has the same thing Google has the same same kind of cloud. Keep with our cognitive behavioral therapy theme and say what we really want to ask is what is the organization's strategy long term. If the strategy long term is to get to know their customers better than by all means let's make sure the data that goes into Salesforce is in fact quite good. Let's put the extra time and work that we need to do and they're not just to put data in there but to also clean the data before we put it in the process. In order to do this. There are several dependencies Chris mentioned a couple of them in there. Again there's some technologies that can help. This is another thing that we don't do well with our data scientists. They understand Python very very well but will probably not understand that there are some really good utilities out there that will help introduce you to a data set in a much more efficient fashion than just sort of poking around or dropping it into a spreadsheet and then doing various bits and pieces on on all that. Chris you want to take a crack at that one too that's sort of a real good one. Yeah, I mean it's not uncommon right and I would say first data governance kind of clashing with it wanting to take large leaps isn't necessarily a bad thing. Right, so you do want a little bit of that that pressure from data governance on there now. So it would really depend on what are the things that are the blockers right most specifically to this question to answer that to get, you know, or expedite actually moving data to the cloud. But understanding which of your data assets are kind of cloud worthy. I would suspect there's ways that if you've got mature data governance hopefully you've got mature viewpoint and visibility into your overall data quality that's a great place to start to understand where do you have data quality where you know where are you managing data in a way that it's not junk and start to instead of maybe one fell swoop of moving all the data and all of everything over to the cloud. You start with your assets you think are you know cloud ready and if there's not an agreement on what that is, then then start with a definition there that you can align with the data governance team on what makes data cloud ready for us in our organization, and a little more from a program standpoint of how do you iterate through that then to make sure that you're moving your, your quality assets to the cloud for usability, and then building a program where you can actually now fund some of the data quality initiatives that are going to take place in those projects to clean the data and get it called ready now that you define what clock ready actually means for your organization. Super good guidance again phased approaches, having conversations, making sure that the strategy isn't conflict because one part of the organization can't be optimized for strategy that says faster. If another part of the organization says no better. Right. And so, very, very important to make sure that happens in there and they are conversations. Again, there's nothing out there in the technology world that's going to magically do this but there are lots of good things that can help you, and we don't teach the data scientists what they are. They actually are going to have to find somebody who understands a bit more about what's in there I think Shannon one of our webinars next next year's could be focused on data management technologies if I'm not mistaken, where we can dive into that topic a little bit although it's kind of hard to do in an hour isn't it Chris it's just a lot of a lot of things going on. As always, Shannon will tell you over and over again Chris and I are not hard to find and you guys are welcome to reach out with further questions and suggestions and sometimes they turn into these webinars so we love the interaction. I think we have time for at least one more question here we've got about seven minutes left so maybe I can fit in a couple but Peter, can you discuss and Chris, any, any advancements that you may be aware of using AI to cleanse data at the edge before persisting it in the data ecosystem that is using AI to prevent garbage in. One of those wonderful terms that everybody's using now at the edge so what that is implying is that before data becomes part of the general data practice data processing that goes on in the organization. I'm full circle on that now I originally started out I was a data processor. And now we're going back in and realizing yes you do need to process the data in order to do that. And there's lots of things mainly they have to do with design, which has become a skill that business has not paid attention to. So my university mandated curriculum here they get a course on analysis and design in the same semester. And oh by the way let's throw in project management as well which is just a real disservice to the students. And so the idea is we would understand these requirements but then we would design them in such a way so let's for example, I like telling this particular story but it does represent. I remember a hospital called me one day into a meeting, I was there, he wanted to announce this, this very new initiative that they were going to do. And so I was on the podium with the hospital director who announced that we were going to now become the knee surgery capital of the upper Midwest. And we had done so much knee surgery at this particular hospital they were going to put a new wing on the hospital there was grant money that we were going to use to buy new equipment and we were going to train everybody up because we were doing just an amazing job with our knee surgery stuff that we were doing at this hospital. Now from the podium, I'm observing those people in the back that Chris referenced earlier, who we definitely want to talk to and of course I went right to them and said, you know, okay why you guys laughing what's going on here. And they said well, the hospital director doesn't realize that the default hospital admission code is knee surgery. And you just go, Oh, no, right. This is not really yes sure enough this and very smart individual had grabbed data and tried to make decisions with bad data, and just had no ability to do a good job with this because they forgot this one little fact that if you don't have anything else going on, they get checked out as knee surgery and it turned out to be way overemphasizing the amount of knee surgery that was actually occurring in the hospital. They didn't bother to cancel the grants and went ahead and invested in it but they clearly knew that it wasn't all about knee surgery. I don't know Chris you got something similar to that or another bit of guidance for everybody but Oh, yeah, it's, it's a, it's a chuckle that my son and I have once in a while he is a chemistry grad not chemical engineering just straight on chemistry so definitely. He used the word lovingly he's definitely a chemistry nerd loves it. So we always laugh because we see the words GMO put on everything it's kind of a trademark buzzword now right and to be honest with you I feel like that's where we are with AI and machine learning still today, whether it's actually applicable to the specific technology or not you're seeing it stamped on anything and everything so it's a little difficult to see and truly understand I'm in the software side of things so obviously we don't in a lot of editors also don't really publish how they're using technology or what it really does, you know that that's all stuff that comes in a sales conversation. So it's tough to assess how far along we've got what's really out there how people are integrating at the edge. We know that, you know everybody's still focused on from a data management perspective in the data supply chain. There's no verification validation cleanse and enrich activities. There's investment there definitely but again it's within the data supply chain within specific aspects and capabilities around data management practices in their entirety and and I'd say any of the software solutions that and to focus on either all of the data management capabilities in the wheel or even just individual ones. So they are looking to, you know, machine learning and AI, but again primarily from the data quality aspects of cleansing validation standardization and enrichment at this point in time. There's enough going on in these areas that we've managed now to, as the questioner said, sort of start nibbling around the edges. And so it is possible to put, again, I don't want to say AI but things that can check things, right, which is really more than AI based but nevertheless passes for a lot of what AI passes for in these days in order to look at that so when you get a number in one of my favorite things that I've seen over time is that a sensor will go out and record temperatures of zero in certain parts of the country because we didn't have electricity during a storm Well, you wouldn't want to make sure those zeros go in as the average temperature because they are really nulls instead of zeros and we need to make sure that these things are recognized but take a look at your various devices that you have and go into your photos app on them and ask it to find a picture of a horse or a zebra or a snake. This is what machine learning is capable of if your phone is sitting around doing nothing. Some piece of software crawls through all your photos and reads every piece of text that you've taken a picture of that's kind of interesting for starters and can look at various animals and identify them. Well, that's nice, and it's helpful for us but we didn't have a business problem that we were asking these folks to fix and that's really what it comes down to how well are we able to understand these things and say this is not an isolated case of this thing just sends us the wrong data in this case but we can put some intelligence out there to try and catch these things early on and that just follows the general Deming quality cycle in order to do that. So we're getting closer. But at the same time, we're not to the point where we can recognize bad data necessarily when it comes in, but we can do some things to make it better. Again, not perfect. Not easy but easier. Right. I guess that's a good piece to conclude on isn't Shannon. That is because that is all the time that we have for today. So many great questions. But thank you to all of our attendees for being so engaged in everything we do and thank you Peter and Chris for joining us today. And of course, thanks to Royal Tail for sponsoring today's webinar to help make these webinars happen just again a reminder I will send a follow up at the end of the Thursday for this webinar with links to the slides and the recording from today's sessions I'll get you the links to the books as well. Thanks everybody. Hope you all have a great day. Thanks Chris. Thanks Peter. Thank you. Thank you Chris. Thanks everybody.