 Hello and welcome. My name is Shannon Campan on the Chief Digital Manager of Data Diversity. We'd like to thank you for joining this month's Data Diversity Webinar, Data Governance Strategy, sponsored by Xgerian. It is the latest installment in the monthly series called Data Ed Online with Dr. Peter Akin brought to you in partnership with Data Blueprint. 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'll be collecting them by the Q&A presentation in the right-hand corner of your screen. And if you'd like to chat, we encourage you to do so. Just click the chat icon at the top right-hand corner for that feature. Or we encourage you to share the questions via Twitter using hashtag data ed. 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 will likely send a link to the recording of this session as well as any additional information throughout the webinar. Now, let me turn it over to Jenna from Xgerian for her word sponsor today. Jenna, hello and welcome. Jenna, you might be muted. Are you there? Are you muted on your end? Can you hear me? Y'all, can you hear me out there? I don't hear Jenna. Let me know what you're hearing on your end out there. All right. Thanks, Glenn. You guys, we did have sound. Jenna, you were there. Let's see if we can get Jenna going again. We did have her on. Technology is so great when it works. We are going to have to come back to you at the end here. We can't get anything going. Peter, I'm going to take it over. I'm going to transfer to you. We'll see if we can get back to Xgerian and their presentation at the end. Let me turn it over to Peter here. Peter is an internationally recognized data management thought leader. Many of you already know him or have seen him worldwide. He has more than 30 years of experience and has to receive many awards for his outstanding contributions to the profession. Peter is also the founding director of data blueprint. He has written dozens of books or dozens of articles and 11 books. The most recent is your data strategy. Peter has experience with more than 500 data management practices in 20 countries and is consistently named as a top data management expert. Some of the most important and largest organizations in the world have sought out his and data blueprint expertise. Peter has spent multi-year immersions with groups such as diverse as the U.S. Department of Defense, Deutsche Bank, Nokia, Wells Fargo, the Commonwealth of Virginia and Walmart. And for one of the few times in history, Peter and I are in the same room together at the Data Governance and Information Quality Conference instead of at opposite sides of the country. Peter, hello and welcome. Hi, Shannon. This is also even fewer one of the times that we're actually together. We don't have a bottle of wine between the two. We'll save that for later on. But it's a pleasure to be here at Data Governance and Information Quality Conference. We have over 600 people out here. It is a phenomenal group and we have been enjoying the conversation all the way around the conference. It's just getting started and the real question is why aren't the rest of you guys out here enjoying the beautiful San Diego Sun? It is fabulous. But we're here to talk about Data Governance and I'm going to start out by putting out perhaps something that's obvious. But there is confusion and there has been confusion for many years. And if you didn't get that from all the intro that Shannon gives me on that, it's a wonderful piece. But I'm an academic. I also am a consultant and also an entrepreneur in that process. And all of these three groups, academics, entrepreneurs and business owners think that IT thinks the data is a business problem. And the typical response is if they can connect to the server, then my job is time. IT thinks it's a business problem. And this looks around and says there's an individual in our organization that has the title chief information officer who else would be taking care of this. And so what happens is that data falls into this enormous gap that we have between business and IT. And our goal collectively as a profession here is to help repair that and reestablish those partnerships. Now I'll tell you two sort of conflicting things about data just to get started. The first one is I hope you agree that better organized data increases in value. If you could find something that will save your knowledge workers time and reduce the amount of money that it takes to do certain business practices. So therefore the corollary also has to be true. Poor data management practices are costing organizations a lot of time and money and effort. And then this is kind of amazing but minimally 80% of the data in your organization is rot. Now you guys must say this guy is an insane consultant that's coming here and yell at people and tell them that their data is rot. Rot is an acronym and it stands for data that is redundant, obsolete, or trivial. Actually my wife corrected me on that said it's actually redundant and complete, obsolete or trivial but I've already called your data rot if I call it a riot. I'm not going to make any progress with you at all. So the question comes up which data do you eliminate? The answer is well, it's kind of hard to tell. It's one of the chief drivers of data governance because on the other side data being largely rot, that data that is left over, data is the most powerful underutilized poorly managed organizational data asset. It's the only asset that you have in your organizations that doesn't deplete. It doesn't degrade over time. It is durable in nature. And at the strategic level data assets really win when you compare them to these other assets that organizations manage. I was over in Saudi Arabia recently and they love the phrase data is the new oil but I really don't like that description. I'll tell you the reason specifically is because we don't think about what happens to gasoline after we put it into our cars or long millers or chainsaws whatever it is we're filling with gasoline. And data is not a production function strictly. Data I like to say just change this a little bit. By the way, take this stuff. Use it in your organization. That's the whole purpose of this conference is for all of us collectively to help move the profession forward. So next time somebody says data is the new oil, I challenge you to just wander up there and say let's think about it a little differently. Let's put the letter F in front of oil and make it soil. There's two things about gardening here that make for good metaphors. The first one is it takes a while to prepare your soil. You don't just randomly sprinkle seeds around the yard and hope good things are going to happen. And the second thing is you don't let stuff on Monday and expect it harvested by Friday. It takes time for these things to mature and write that says your data program. On the other hand, if you have to call it bacon to get it sold, that's okay too. We're really glad people are working in that. So data by itself as an asset deserves its own strategy, it deserves attention on par with similar organizational assets and it does require professional administration to make up for past neglect. So a couple of questions here if I'm going to get a little story. Most of you are familiar with the target data breach and of course the Ashley Madison data breach, probably less of you are familiar with that. And I'm going to throw in a third piece here which is something called the office of personnel management data breach. Let's look real quickly at how some of these numbers pop in. Federal employees, there were 44 users that signed up for Ashley Madison and if you don't know what Ashley Madison is, it's a dating site for married people. Okay. And 44 people who had the email address whitehouse.gov signed up for that service. In addition, thousands of military and government email addresses. Lots of Canadian citizens, don't worry I'm not trying to provoke a war with Kennedy here but it is interesting to notice at least one-fifth of the population of Quebec, Canada, their version of Washington, DC were also registered on that. So we asked the question, are we managing data with guidance here? So this is a title of a report I'll flip up here just a second. How the government jeopardized our national security for more than a generation because you see there were 25 million Americans in the office of personnel data breach. Most of them have security clearances. The Ashley Madison data breach was 37 million and the target data breach was 70 million and all we need is the intersection of people who were top secret security clearances who signed up for Ashley Madison using their government email address, not the brightest thing to do in the world. And why does Target play a role in this? Well, Target actually has data on our habits. So they may know that somebody named Peter stopped off at the Target store at short pump Virginia on his way home every Friday. And when we have those kind of habits you can then be targeted by bad actors. And those bad actors present the biggest threat to the national security of the United States that we have ever, ever faced. Some data should be locked up in a safe and not put anywhere. And that was the conclusion of that report. The people who made the decision to put the OPM data online were unqualified to make that decision. Ouch. Pretty bad. Not really good. Now let's go back to Target as well. Target had some issues with it. The CIO was forced out. The CEO was forced out. And interestingly enough they came after the board of directors. So we're starting to get attention for this at the board level. And that is really a good thing because people are starting to realize now that decisions about data have consequences and those consequences can involve jail time. Data governance is critically important because organizations spend millions a year in lost productivity, redundant siloed efforts, poorly thought out hardware and software purchases, delayed decision making, being reactive instead of pro-acting. And finally, I tell you you can save 20 to 40% of your IT spending through data governance. I get a lot of people that say I'd like to see how that saving would come about. Many of you are familiar with the DAMA guide to the data management body of knowledge. The wonderful piece of work we're actually at version two but you'll notice that data governance is central to that particular process. It connects all the other pie wedges, if you will, the data management functions that we've articulated in there. And the Denmark in the first version gave us a wonderful, nice one-page piece of this. This is essentially the plan that we're going to cover today. It's a fair amount of information here. I'm not going to read it all to you, but please do use this information describing inputs, activities and deliveries of data governance and see what works for you because that's a big take away from this. While data governance can be prescriptive in nature, and those of you that are in regulatory environments obviously have a primary direction on this, most of the rest of the organizations need a little bit of guidance around it. This is where the word strategy comes from. So strategy will talk about recent usage, give a little context of organizational IT and data strategy, and talk about some fairly difficult choices that are around it. We'll then get into data governance, talk about what is it, why is it important and what are the requirements for effective data governance. Then we'll look specifically at governance components, frameworks, building blocks, checklists, maybe some worst practices in here. The most effective way we've seen data governance work for most organizations is in action. The action usually involves storytelling. So we'll finish with a couple of stories and then get to some Q&A. I'll look for the last part of this where we get the experience back online and hopefully they'll participate in the discussion. So let's start off by looking at strategy and governance. I'm willing to bet that hundreds of you that are out there online have experienced data governance but you're sitting in a meeting and you're going, why are we here? I'm sorry, I've lost the purpose of this. Well, data strategy is what gives your data governance initiative a purpose. What are the data assets supposed to do to support organizational strategy? Similarly, governance feedback into the data strategy says how well is that data strategy working for your organization? Let's add on to that how an organizational context, the data is only there to support the organizational strategy. And if you listen to Peter, Peter says that IT projects should be subordinate and the data governance should control those aspects of IT projects. Maybe Shannon and I can talk about a webinar to tease that subject out a little bit further in the future at some point. Add in some feedback loops here and we've got the whole picture. It's kind of messy but the key for this is data strategy is what helps the data governance group what is important, what is key. And I would never show that other picture to most managers. I like to simplify it by bringing it down to this level here. And say that data strategy has to be expressed as specific business goals that are addressed by the data governance group. Similarly, the language of data governance has to be metadata. If you're talking about something that you cannot connect directly to some aspect of the business then you are too far disconnected from the problem and you need to reconnect with it. So let's talk for a minute about strategy here. If you haven't had a chance to see this particular TED talk, I've given you the link in the upper right hand corner. Simon Sinek has made an entire industry on this for himself. But his original 10 minute TED talk out there, I'll give it to you in a minute here, talks about organizations and people are not okay about describing what they are. We're less good at describing how we do things and we generally are pretty poor at describing motivations. For example, again he says it's not what you do but it's why you do it. It's people excited about it. I just want you to imagine Reverend Martin Luther King instead of saying I have a dream. He says I have a plan. Well, it does not have the same gravitas as that. Back then he said I had a dream and of course I can't even imitate him but you know the way he said it. Everybody has heard that particular phrase. It talks to you about the why. It does not talk about a plan. So hugely important. Strategy is something that we haven't used a lot in business until recently. You can see the usage of that term has gone up more since about 1950. My favorite definition of strategy is that strategy is a pattern in a stream of decisions. Let me give you three brief examples. The first one, you know it sounds military. It's actually kind of important. The situation here is that militarily the French in blue in this little diagram here are facing two armies, the British and the Prussians. The British are in red. The Prussians are in black on this diagram. Now the question is Napoleon is outgunned, outmanned. There is a larger army out there. How do you defeat them? You guys know this answer. Divide and conquer. This was his strategy. Now his strategy had to be implemented in two parts. The first one, Napoleon noticed that there were lines of supply. There are always lines of supply for an army. An army travels on the stomach. When you get hit in the face or attacked, you tend to run towards your food instead of away from your food. Napoleon in blue here said if I could hit these two armies sitting right there at exactly the right point, the British army, the red, will retreat towards Ostend, which is on the coast of France. And the Prussians will retreat towards their supply line which was out of the Belgian. Now the attack process was hit them exactly at that right part and the two armies will separate. I'll show you that one more time. Hit them and separate them. That was only one part of the strategy. He still had to win. He's now got them apart and they're running in direction of their supply lines. The next piece of this was of course then conquer. So he's divided them. And then he says to his troops, everybody turn around after we separate them and attack the Prussians. After we defeat the Prussians, then we'll defeat the British. I want you to imagine yourself as a soldier in Napoleon's army. And you're going to die if you don't get this right. So you have two things to remember. Hit them hard and hit them here in that exact position. And then first go after the Prussians and then the British. If you guess holders that are confused, do I go after the British first or the Prussians first? They lose. By the way, Napoleon lost. It's still presented as a great example of strategy. Second example of strategy, Wayne Gretzky. The ice hockey great. Had a strategy that was very simple. He doesn't skate. After the puck, he skates to where the puck will be. Where he thinks the puck will be. Because if you're chasing a hockey puck around the field, you're not doing much good for anybody because the puck is going to be much faster than you. Finally, third example here, Walmart, a company that comes to work with him in the past and I'm not telling anything out of school here because you all know this. Walmart's business strategy is simple. Every day, low price, that is inculcated in every associate at Walmart. Every decision that they make is guided by this particular philosophy. Strategy must be simple or it will be difficult for people to follow. Let's look at a couple of definitions of corporate governance. Corporate governance, of course we've seen, has been increasingly important around the world. I'm not going to define it here, but people are saying, yes, it's important. Then we get down to IT governance. We have aspects of IT governance. Again, very, very important stuff. Unfortunately, most people think that corporate governance and IT governance should then be followed by the data governance. But my colleague, John Ladly, did a great talk a couple of years ago, told me I could use this slide of his, which just says something that he's talked about a lot. There's a little challenge when you try to map what's happening in the organizational strategy and what's going on with their IT portfolio. We call it the alignment gap. It exists in most organizations and when we've seen this happen it just shows that people are doing things but they're not really getting the right aspects of it. Again, we have organizational strategy, IT strategy, and data strategy, but this is wrong. This is wrong. The reason this is wrong is because IT strategy and data strategy should actually be parallel with each other. Most importantly, your data strategy can help you to reduce risk, increase delivery in your IT area in very, very many major ways. There's of course a little bit going back, but IT strategy should not be superior to data strategy. They should be co-equal around there. This makes for some difficult choices for some organizations. Let's now talk about data governance. I have a little billboard here that I love. What we want to be perceived as is probably a good negative role model. The committee. This is the file naming convention committee. The committee decided that the file naming convention will start with the date in the order of month, day, and year. Yeah, we know that's wrong for starters. Then a space and the temperature at the airport and the hat size of the nearest squirrel, to be perfectly honest, it's a long meeting. We probably didn't do our best to work towards the end. Well, we don't want to be accused of that. Here are seven data governance definitions all done by very fine people, some of whom are here at this conference speaking on this, and they are good definitions. Even our DIMBAC definition here is kind of okay. I actually like to do a simpler definition here. The first question that you want to ask is what does data governance need to mean for my organization? I like to say it's managing data with guidance. So it's getting some people whose opinions matter to form a body that needs a formal purpose or charter, and who will advocate, evangelize for increasing the scope and rigor of data-centric thinking in your organization. Let me give you a specific example. This is a healthcare company that we did some work for. And one of the things that they all said was, you know, getting access to data around here is like that Kathleen Dada Jones scene where she's having to get through all those lasers. So I'll show that to you again just to get it. It's kind of a cute little movie. I think it was her and John Connery in the movie. Can you imagine if this is what it takes to get people to get data in your organization? I mean no wonder we're happy with small advances for this organization they use this motif around their data governance. It actually made a huge amount of difference. One of the questions we get often is what's the difference then between data governance and data management. Data governance is policy level. It's a general guidelines and direction. For example, you might say that all information not more public should be considered confidential. Data management on the other hand though is a function for planning, controlling and delivering information assets. So making sure that we have the data that exists there is a planning and a doing difference between the two of these. And things that should be included in your data governance program eventually include security, quality, life cycle management risk, standards, content valuation tool kit, and also case studies as well. But I don't want you to jump in there all at once. Instead take a general approach. So if we look at goals and principles, what we're looking at is, okay, how do we approve things? How do we track compliance? How do we oversee the management of data delivery projects? How do we resolve data issues? And how do we get people to understand that less data is generally better, but that we also have to increase the quality at the same time around that. Leading to deliveries such as policy, standards, issues, get lists and lists. All right, roles and responses. I'm not going to read these all for you. In fact, I'm going to take and put them all together in one piece of serious practices and techniques. And I take those four slides and put them up here and go, oh, my goodness, where do I start? Well, the challenge around this is that most organizations tend to overthink this process. They want to know it up. Three-year plan. I have one organization that wanted a 10-year plan on this. Back off, right? Take a breath. Come back and say, look, what can we do that will help us immediately? And I'm going to use the example of a data steward. Many of you have heard this. Now, if somebody is in a stewardship relationship with you, they are in a position of trust. So we could have business stewards, technical stewards, project stewards, domains, that, that, that, that, that, that. By the way, if you've got all those others, you need a data audit steward, too, right? And if you've got that many, you need a data steward manager. Oh, my. By the way, David Plotkin's book. A good friend up in the right-hand corner does talk about these in more detail. Gosh, this is not the way you want to start out. Instead, let's look, and what the definition of steward actually is. It's one who actively directs and a data steward then. We're going to finish up the use of organizational data assets in support of specific missions. So let's just start out with some stewards. Again, I could do this this way where I have all of these stewards. Sorry, I'm going to go back over here. All of these stewards, ah, I mean, if you're just starting out, it just makes no sense at all to try and figure this stuff out. These are weight. Let's get down the road a bit. Let's get some experience, years of experience under our belts before we start to fine-tune this. And let's start simply. Start with a steward, and the steward is going to take a look at data. So we'll take a look at it from that perspective. So what is data governance? I've already said it. Managing data with guidance. Now that took us up for an interesting question. Would you want your soul non-defeatable, non-degrading, and durable strategic asset managed without guidance? Most people will say, ah, I don't think so. Let's go a little further. As far as the components here, I like to talk about this in terms of making a better data governance sandwich. I know it's close to lunch out here on the west coast. Many of you guys are taking your lunch break to do this. Let's talk about it maybe making it more appetizing to people. So I think we'll all agree that there is a varied amount and level of data literacy out there in the world and specifically in our respective organizations. Similarly, the data supply is also uneven as is our use of standards. I don't know about you guys, but that's not a very appealing sandwich. Now, if we try to get better with data literacy, better with respect to our data supply, and better with our standards, only then can we take these three pieces and combine them into what I call the better data sandwich. By the way, I put these things out there for you guys to help concentrate and think about ways of improving them. And I had a colleague come up to me at one of these meetings just a week or so ago and say, you know, if you're going to have a sandwich, you probably have somebody chewing on it. Okay, that's an interesting piece. I'm thinking about that. But the more important part is, notice that last little bit, we want to get all these things to work together. These cannot happen without engineering and architectural concepts, which are completely missing in most of your organization. That's a separate topic, but it's something that you do need to be aware of. Similarly, I was on a tea farm in India this last summer, and this was hanging over the cash register. Quality engineering and architecture work products do not happen accidentally. I was not able to ask the person, actually, I asked the person the cash register. They said they didn't know. I don't know why that particular tea farm had this sign on their cash register, but it makes sense. You can't build quality products that are engineered to do high speed data transitions and translations and other types of things without the ability to have accidentally without the ability to do all of these other products. They've got to be done with conscious forethought in order to do this. And I say that because this is a barn where I live in failure, Virginia, very close to Richmond, Washington, D.C., and I'm what's called a horse husband, and so as we were building our barn on the process, the bank gave us exactly this much money to build the barn. You might ask, why would the bank give us only this much money? Because the bank had a pretty good idea for governing the loan that they were giving us to build the barn. They said you build the foundation, and then you stop. Before any further construction can proceed, we want to inspect the foundation to make sure that it has a solid basis. If we don't have a solid basis for this foundation, we will build a good barn on top of a poor quality foundation, and I will spend more in vet bills than I will pay back the bank loan. Now this is a very interesting concept. It's a great governance concept that the banks implement. There is no I.T. equivalent of this, and this is important for everybody to understand, because I.T. is a relatively new profession, whereas we know banking has been around for an awful long time. We need to implement this, and governance is an incredibly important way of starting to put these things in place, because governance gives you a general framework. Now I'm going to tell you what a framework is just for starters. It's a system of ideas for guiding analysis. It helps us organize project data, gives the ability to set priority decisions. Again, I'm going to build the barn next. I'm probably going to put the walls next, and then put the roof on. And then I'll put the electricity in, as opposed to putting the electricity in, and then putting the roof on, which just seems a little backwards. We can use the framework as a means of assessing process, right, and make everything else dependent on continued funding. So our framework for governance here, then, are several different governance frameworks that are all available to you. Remember, you guys get all of these slides at the end of it, but if you Google data governance frameworks on the Internet, this is what you'll see. There's the Dibbock. Here's one from the data governance institute. Here's one from Rob Siner's group. Here's one from the IBM data council. I'm not going to walk through these things with you, because they don't mean anything outside of the context of your organization. I leave them there for you to take a look at, however. And go back and do some more. Here's an interesting one for the American College of Personnel Associations, and they make theirs look like a boat. Okay, fine. Whatever works. The point is, look at each of those seven different frameworks and see what works for you and your organization, and how it can be helpful. Try them on. It's like going to a store and trying on clothes. Play with it for a little bit. It's very likely that none of them are perfect, but that you may say I like some from A and some from B, and we'll put them together in order to come up with our piece for C. Again, Sassett Institute here has a great one. So all of these are fine, but it is frameworks. The question is how useful are they for you in your organization, and that depends on what your organization is facing at the moment. So look through these various levels. Again, I know I'm scrolling really fast through them here. They're all available in the notes. You'll get them all, and Shannon says this out to you in a couple of days, so that you can take a look at them and say, wow, I like this aspect of KIKs, but I really like this aspect of the DGI framework. That's what they're here for. That's why they put them out. IBM's got some, again, mentioned this. Find out what works for you, and then look for a little set of checklists. What sort of decision-making authority are we looking to put in place? Very likely very few people make decisions about your data in a formal sense. I've been in many organizations where there's a single person who goes out. What sort of standard policies and procedures will be useful and help your organization in the immediate future? It's not do I need a complete list of all the policies, procedures that I'm going to put in place. I need a couple to get us started. We'll try those out. We'll evolve them over time, and as we evolve them, hopefully they will get better. I've never seen any company anywhere in 33 years complete a data inventory. Again, that's one of the first things that everybody has asked to do. It's important to understand what piles of data you have lying around, but it's not important to try and be comprehensive about it. Again, we're looking for a specific piece here that says what do we need to get started if the customer data was messed up? Then let's start to work with the customer data and see what things we have there. Content management, records management, what aspects of quality have been noted by either auditors or customers or others who are saying you have some data quality challenges. How do we make data access easier? But don't try and do this for everything in your organization. You will go bonkers. By the way, any of you out there can actually predict the future. Get in touch with Shannon and I. We've got a couple of other side projects that we'd like to talk to you about, because since we can't predict the future, we don't know what's going to happen. I mean, who would have predicted a couple of years ago? We'd be at war with Canada and making peace treaties with North Korea. The world changes quickly. You could tell what day we did this talk. Anyway, another part of all of this. You need a scorecard. You need to be able to say how effective are we doing with our specific data. And the question that comes up is what is data valued at? Well, for Facebook, for each of its North American users, it's about $100 per person. So Facebook lost a lot of users. The valuation of that company would go down. Now that's a very easy case to make a Facebook piece, because all they do is connect things. There should be some costs around the data management group. One of the things that I always encourage the groups that I work with is to say that if you have five people in your data management group and they all get paid $100,000, then you have an obligation to show the organization that you've contributed back to the organization to stop $501,000 worth of value in the process. If you don't, then management has a reasonable question to ask them, what are you doing and why are you doing it? What sort of specific objectives you get under promise over delivery? Let's start out small. Let's try to accomplish a couple of things. Let's try to keep something from happening that happened last year, whatever it is that you're looking at. How many decisions were specifically made around that process? What sort of coverage should we have? Remember, I had 10 or 15 different types of stewards out there. Can you imagine trying to plan that out in advance? One set of stewards covering a small portion, maybe a small but a significant portion of your data. And let's see how that works. Then we can come back and argue for more. What is your professional head cap? How many people are doing this full-time? How many people are doing it part-time off the side of the desk? And how effective are they? And finally, there's a way of describing your data management process maturity. I'm going to deviate here for just a little bit. Most organizations and most management are familiar with something out of the software engineering institute at Carnegie Mellon University. It's actually research that we funded from the department of defense many, many years ago that resulted in something called CMMI, capability measurement integrated. It's a very good process. And all you have to know is that there is a data process. That was the subject of our webinar that we did last month from Bogota, Columbia. Well, it's fun we had on that one. Again, if you didn't get a chance to participate, it's all archived out there at the university website. Let's talk about some worst practices around this. And I'm going to put on a little bit of background here. It's hard to hear. In the hotel of California, it seems to be the different words. Up ahead in the distance, I saw a similar note. My mouth was heavy. My face was young. I took a kind of a needle in the hole. I was thinking about the data governance way. It's a cute little piece. You can go out and get it. I heard voices down the corridor and I heard them say, so what's wrong in data governance? Well, you get a lot of buy-in, but not commitment. Sure, go ahead and do this. If you get to the rest of the song, continue over there by the way, oh, yes, such a dreadful place. There's a difference between real commitment and the real buy-in. One of the things we've noticed over the years of this conference and others is just that we have a lot of trouble sustaining data governance and organizations when the people who start it go somewhere else. Whether it's to another company, promoted to another part of the organization or whatever, you need to have a long-standing commitment in there. Second worst practice is organizations that go out and do ready, fire, and aim. Again, what we want to do there is be careful, call, walk, and run is the proper way to do this. Trying to solve world hunger or boiling the ocean. Again, these are both incorrect. You're never going to do it. I actually tried to do a controlled vocabulary for the Department of Defense for about 5,000 terms. Theoretically, it was possible practically. It didn't make any sense whatsoever. The Goldilocks syndrome. This is not a case where if somebody wants the data to be one way and somebody else wants the data to be another way, half way doesn't make any sense here. Goldilocks. We'll just make a compromise here. That may be dead wrong. Be careful of that. Committee work. Overload. Again, lots of organizations that we've worked with have set up massive data governance initiatives and people are so sick of that. Data governance, people come into the room and everybody goes, I don't know why it's because they take up too much time and they don't really accomplish any proceeds of value. Failure to implement. We'll get lots and lots of plans, but plans are no good unless you actually have a detailed execution behind it. All of these initiatives require somebody in the organization to start addressing the big elephant in the closet, change management, how are we going to make the organization realize it's the way we have been doing, it's harder to do it the old way than do it the new way. Eighth, worst practice is starting with technology. Or assuming the technology by itself is the answer. It is not. Technology is a part of the solution. I like to people to think of a stool. The one-legged stool is not going to help. That's technology. But if you put in people and processes around it and you have a three-legged stool and you're not building poor sustainability, it is absolutely critical that you think not just what's immediately in front of you, but how can I build this for the future. So you may say I'm going to have 100 stewards in three years. Well, okay, what happens the fourth year? Are those 100 stewards going to have any value for the organization? I don't know any organization that has 100 stewards, by the way. And finally, shadow data systems. One of my favorite stories I was working for an organization, a guy who is a CEO, and he said you know, I have a SQL server implementation under my desk. Yeah, we know that. You won't tell IT. We'll continue to perpetuate the fiction that you don't know that IT knows you have a SQL server under your desk. And IT doesn't know about it as well. But believe me, IT understands the value of your SQL server system under your desk because it helps you with your specific decision-making process. And it also helps you with knowledge on to your successor because you've been in the business for 33 years and I know you're going to get ready to retire. These shadow data systems are huge out there. And it just takes a spreadsheet, right? That's sort of our worst practices, the components of place. Let's talk specifically now about how stories can be helpful here. And I want to tell you a little bit about the department of defense. They wanted us to go out and find out what we should learn from industry out there. Actually, my measurements show that government is head of industry by a slight fraction in this area, which is kind of fascinating. On the other hand, it is against the law to not do this in the government. So it's a federal government area. One of the lessons I brought from the government is they were looking for 11 CIOs for their transmission division. There's a couple of things wrong with it. Why is the CIO span of control only over the transmission division? And by the way, why do you need 11 of them? That seems like the wrong number for each chief. Now, from a government's perspective, it was very, very easy to see that they had some challenges around the data in their transmission division and the other piece that we took back though, when Detroit was making things, engines, cars, things like that, they were looking architecturally at what they were doing. And they would say, okay, I have a chassis of a car and I have some things I need to attach to the chassis of the car. Maybe it's an air conditioner, a generator, a motor, oil, coolant, a pump, several things get bolted to the engine. Another measure for doing this was whether you could attach that component to the chassis of the car and not slow down the assembly line. After all production is key there. Success was simply, did it attach and didn't slow down production? If the answer was yes and no, you're good to go. What that left you with were a bunch of different bolt types which meant that to have a bunch of different wrenches you were going to maintain that automobile and different bolt inventories in order to do this. And we contrasted that with Toyota that took one additional step. They would make that first engine attach it with a bunch of different bolts. Toyota would do that and they come back and ask the question how many of those bolts can be standardized? Again, I'm showing one here. We know that's not the answer. We can't standardize everything. It's like boiling the ocean. It's never going to happen. But if I have fewer bolts that means I have fewer bolts in the inventory and fewer types of wrenches that need to be in place in order to do this. Again, there were many bolts, many different wrenches, many different bolt inventories which made for a more complex existing environment. But for the Toyota perspective, if I can use the same bolt more rather than less, then I will have fewer bolt inventories to connect with that because everybody understands car engine. Many people try to work on their car engines and become frustrated by these types of problems. Again, it's not the Detroit always had 100 and Toyota always had one. But there was this extra step and that extra step added quality and maintainability to the imported cars that were coming in in competition with these domestic cars that were there. Here's another aspect of this. It's pretty confusing for most organizations. They've started off in what seems to be a fairly reasonable perspective. Organizations say, hey, I've got a strategy and let's implement some IT to support that strategy. But data becomes sort of an afterthought. You say SAP or PeopleSoft or anything else in that IT project space, it sucks all the information out, all the conversation out, and the only thing we talk about is SAP data or PeopleSoft data or whatever. The problem is to make sure the data is formed around the applications as opposed to around organizational information requirements. The process architecture that you have is narrowly formed around the applications and you get very, very little data reuse as a result. Our data-centric approach is much different. We just flopping, excuse me, flipping the data and IT pieces here. So we start with strategy and next we should define the organizational data information assets that need to be there in order to have everybody using them organization-wide. And only after we've done that step should we then define IT projects. By doing it this way, it makes sure that the data assets are described from an organization-wide perspective, that the systems support organizational data needs and complement the organizational process flows, and we get maximal data and information reuse around this. A little bit of a twist on the traditional process, but we found very effective for most organizations. Now this is not something you snap your fingers and it happens, but if you set this as a vision, everybody will be able to move in that general direction. So let's look at a couple of choices that you have from a governance perspective. Most organizations start to talk about strategy and I've already said strategy drives your governance piece. There are only two dimensions of strategy. The first one is to improve your operations. The second one is to innovate. Anything else is a variant on each of those pieces. And yet most organizations start down here from their data governance efforts, which is not much, and they're not really working at this level. If you move over into this quadrature here, Q2, what you see is that organizations start to say, hey, I want to use data governance to make things more effective and more efficient. And again, we'll use Walmart here as an absolute exemplar of this. Everybody understands that with Walmart's business strategy being every day low price, being efficient and effective is a key part of Walmart's ethos. It's what drives that company. On the other hand, we say Apple is an innovative company here. And we want to do innovation. That makes good sense, too. Apple is trying to innovate to come up with new products and services so they can keep going, keep their stock price high, et cetera, et cetera. Now, I just want you to imagine for a minute here that we take the people at Apple who are innovative and tell them now they need to be innovative and at the same time be effective and efficient. Their heads will explode. And the only thing that will make your head explode even faster than that is if you take the people at Walmart who are increasingly efficient and effective and tell them to be innovative. It just doesn't work. So while we can have both of those pieces in place, it really only makes sense to pick one of them and to start off moving in that demo direction gradually. Let's get the innovative piece down. And by the way, if you use the innovative piece, you can take the savings from that to fund your innovation. So you take the efficiency and effectiveness, drive out some cost savings, use that money to go reinvest in your innovation and you're in a great shape. I told this story yesterday at the conference. There's a company that says chocolate. They're selling lots and lots of things about it. But they had a little problem. They tried to do too much IT transformation at a peak sales period for the organization. There's a lot of unfortunately unsold chocolate around. And we position this for the public business initiative. Many organizations now, including this organization, implement a lockdown. So we're not going to make any major changes to the organization. While we have this. And even now, years from that incident, when I start to talk about this with them, they come back and say, you mean the chocolate story? Yes, Peter. We understand the lessons for the chocolate story. Not too much around this. Healthcare.gov is another instance of governance. Very poorly done. First of all, when I say 55 contractors, I do not mean 55 people that were on the project. I mean 55 contracting organizations. They were six weeks from launch and the requirements hadn't been finalized. And so we're not going to make any changes to this organization while we have this. And even now, years from launch, we have been finalized. And Jim Johnson, who is the guy that runs the Sandish group on this, said the real news would have been if the system actually did work, given this particular piece. You should never have 55 cook trying to make your soup. Another aspect of this as well, though, was very interesting. One of the investigators who went into health diagnosis system said it was obvious from the first look that the system hadn't been designed to work very well. Every single thing that slowed anything down would slow everything down. So I want you to take a look at that. So I want you to take a look at that. I want you to take everything down. So I want you to imagine that your blinker lighting, your car doesn't work, and therefore you can't turn your car on. That's a good thing from a safety perspective, but it certainly doesn't make any sense. Actually, let's take the blinker off. Let's make it your window. If your window won't roll down, your car doesn't work. You can't start your car where your car's perfectly functional with the windows not rolling down. There's one other component to the healthcare guide. I've been talking about big data. Well, this is a fascinating example. A path the group was talking about. So we have all kinds of things that were wrong here. Governance would have helped. They would have said too many contractors. They would have said, six weeks from the launch and you don't have your requirements finalized, we're not going to launch on that time. It doesn't make any sense. They would have looked at the system design and said, oh my gosh, you shouldn't have the fact that your car window doesn't roll down and prevent you from rolling your car down. And most importantly, mixing big data technologies and traditional data technologies is a careful process. It certainly can't be done on the rush or on the chief. Another really good story that we were able to do had to deal with when we were doing governance for the army. And we were working with these guys. One of the easiest projects that we had from a cultural perspective, because as you might imagine, everything in the army is governed. And when we pointed out to them that their data was and they went, ah, we have to fix that. Everybody went, okay, that's great. Now while we were there at the army, we also happened to catch this particular project. The military suicide prevention project. Again, even today, where we are almost eight years away from the start of this piece, we still have more U.S. service members being killed by their own hand than by bad guys. Those Canadians will get to us eventually, I guess, but who knows. So we were working on this. And I had a room. I want you to picture about 100 kernels in the room. I'll show you a picture of them in just a second, but I want to show you this matrix. Each kernel would stand up and say, sure, Dr. Akin, you can use my data from this project for this reason. We would mark it down in the matrix. And by the time the matrix got to be about 30 by 30, I don't know if you've ever tried to work off of a 30 by 30 matrix, but in a real-time interactive function, it's pretty during difficult. So here's my room full of 100 kernels. And I had a favor that I could ask of one of the high army officials in this process. The figure official came along and was listening into this. He said, you know, I see why you asked me to this meeting, Peter. And he stood up and said, ladies and gentlemen, I have an announcement to make. Anybody who wants to tell me why they can't use my data to save my soldiers' lives, they are welcome to make an appointment with me. My door is always open. So of course, this dose of management support really empowered the team, and he gave me full permission to use this with attribution. So I put it in the book called Monetizing Data Management. I'm going to go up the story up, get more detail about it. The important thing here, though, is that while that is a great story, and it was a wonderful way of supporting the troops, I've told that story to more than 100 corporate CEOs and not a single corporate CEO will say to the rest of their group, this data belongs to the corporation. We're all going to share and use it. And this would solve an awful lot of problems. I'm not sure why we don't have that. Back to our suicide story, we were able to develop patterns and understood a little bit more about this. We have not solved the problem, but we are definitely making progress. Another quick story about the military here. We had some organizations that were trying to work with tanks. Actually, I'm sorry, this is not directly a military story. This is actually an oil company story. And they had bought an accounting system. And their package recognized that a transfer of their product from one tank to another tank represented a sale. And they came to me and said, we don't think that all of these sales are right. For example, if I move it with this truck or into these tanks or tanks that float or tanks that fly, those are not sales. And we can't use it that way. So do we change the software or do we change our business practice? And the data governance group put forth the idea that if everybody would prohibit the use of the word tank in its native sense and instead only use it qualified, retail tank, wholesale tank, internal transfer tank, then they could use the existing accounting system without any changes. And just do a simple adjustment, take all of the non-revenue producing sales away as they cost a good sold and leave the bottom result as it comes out of it. Very, very innovative process. Save them literally millions of dollars all due to proper use of data governance around here. Of course, I did say the military here is the military version of tank and you wouldn't want that tank to be confused with other types of tanks. Again, I live in Richmond, Virginia. If you looked last week, we had some poor soldier that got into a tank and drove it down Main Street of Richmond, Virginia. I haven't seen any pictures yet but I'm sure it was a great thing. Luckily nobody got hurt apparently at it. Well, of course we got tanked which is another, okay, just a joke. So if you buy a tank for the military, it turns out that you're also buying three million pieces of data. And exactly one of those pieces of data actually keeps track of whether the tank is obsolete. So for the military in this case, we were able to determine that they had about $5 billion in inventory that they were maintaining tanks that didn't need to be maintained because they hadn't governed the data properly. These stories get repeated all around the world. Barclays had a huge spreadsheet problem when they were trying to buy Lehman Brothers. They actually had put their sales in a spreadsheet and then they hid the lines of sales that they weren't going to do and accidentally a junior associate unhid the sales caused immense amounts of confusion around that whole process. Another final example here, Zuzo Securities had a thing where they wanted to sell one share and they had a trader that wanted to sell one share of stock called JCOM for 600,000 yen. Got it back which he sold 600,000 shares for one yen, $350 million. There's lots and lots of other pieces of these stories that these are the kinds of things that you need to pay attention to. I'm going to finish up here and then hopefully we'll go back to the experience and find the key to it is that data management is awful lot like now as well as hierarchy of needs. You need to get your food clothing and shelter needs met before you can get up to self-esteem and self-actualization. Everybody wants to do fun things with the data but you have to do the foundational pieces first and most people aren't aware that that's just the tip of the iceberg in there. Those foundational practices, clearly governance is one of them, have to be in place in a good way in order to do this because the foundation can only be as strong as the weakest link and data governance is often the weakest link. Focusing on technologies and capabilities, again, if they tell us to do it faster it will take longer, cost more, deliver less and present greater risk to the organization than it's not. The need for data governance is increasing. It's a relatively new discipline and we do not have a lot of ways of going about and figuring this out. We can't tell you that we have 5,000 years of history like we do in banking or accounting to say what is the best way to do this. It's got to be driven by strategy. Comparing those frameworks can be useful and the data governance directs the effectiveness and is responsible for the effectiveness. The data governance efforts, the language of data governance should be metadata and process improvement can improve those practices. Hopefully this will help your organization understand that data is not something tiny between the business and IT that they share responsibility for. It's a big beast and we have to change it and bring it into place this way. With that I'll turn it back over to Shannon. I'm giving you a couple of references there but I'm looking forward to the rest of the conversation around that. Shannon, back to you. Peter, thank you so much for this great presentation as always. We really appreciate it. I'm going to turn it back over to Jenna from Experian to see if we can get a word from our sponsor and make sure we get that great presentation going from Experian. We always appreciate the help in making sure we get these webinars available for everybody. Jenna, are you there? Hi, Shannon. I'm here. Yay! Okay. Thank you. Thank you so much. Peter, great presentation. Hello everyone. Thank you for joining us. Better late than never who you finally made it through. A big thank you to Data Diversity for having me today. Just to give you a little background, I'd like to talk just a little bit about Experian and what we do. Experian's data quality and management business really enables organizations to unlock the power of their data. What does that mean? That means we focus on the quality of our customers information and it really gives them the ability to focus on all of the very meaningful ways that they can use that data. Sometimes that looks like us helping empower our clients to optimize their data for better customer experiences. Sometimes that means enhancing their data governance strategy or sometimes that might be preparing data for improved business intelligence. But whatever it is, we help customers both across the country and really across the world to maximize the value of their data. So just to give you a little bit of a definition about what we mean when we say data quality, just like Peter gave you seven different definitions of what data governance means to different people, data quality might mean slightly different things. But one of the ways that we find easiest to define it is to really just define it through the capabilities that comprise data quality. So we'll just quickly review some of these here for you. So cleansing is the first one and that really means the removal or the correction of mistakes in data records. And the second one there is profiling. This is a really important and really crucial capability in allowing an organization to analyze and identify data quality issues. The third we have here is matching and that's when you have the ability to join two or more related records. So you take one John Smith and you take another John Smith record, they look really similar, same address, let's match those up. Standardization, that has to do with reformatting data to create consistency and uniformity both within fields and across fields. Next one here is enrichment and that's about appending additional third party data to create a more complete record of your customers or what have you. And the final piece here and this one really does play very well into a lot of data governance strategies is monitoring and that is what enables you to have that sort of ongoing maintenance to provide insight and understanding of your data and how high quality it is. So for an experience we love this graphic, we love this analogy about thinking about data as water and when you think of it this way you realize that data governance and data quality really are two sides of the same coin. So data governance, it's a complicated subject, it's not super easy to understand and it's really easy to confuse it with the many different interrelated aspects of a data management program. So we love this graphic because it's a really clear and easy way how data quality and data governance works together but also how they're separate. So if we think about the water system data quality would be that purification process making sure that none of the information is contaminated and then data governance is about having the right people and the right tools and just the right processes in place to make sure that you're maintaining that water throughout garbage in, garbage out and that really for me comes to mind when I look at this graphic because you know if you have contaminated water that enters the system at the beginning it's always going to create more complications downstream and so that's really why we believe that data quality is a cornerstone of data governance. So you know when you think about data governance practices it's mostly about exercising that control and that authority over data governance is not of high quality so while data quality is one of the key pieces to an effective data governance program a strong data governance program really helps to maintain the quality of that data throughout its life cycle and a lot of organizations in the same way that Peter was explaining they think of an IT strategy and a data strategy is two separate entities a lot of organizations follow the same data quality is two really different things but like I said they really go hand in hand and when you pair them together they really do help to maximize the value of your data by helping to ensure that it is accurate and fit for purpose for whatever use at any time. And you know this is the final one so then we'll move into our question and answer session but each year Experian produces a global data benchmark and for the 2018 report you know we surveyed a thousand people from different organizations across the world and we found that 95% of businesses struggle to implement data governance programs. Now that's a pretty profound stat 95% the vast majority of organizations really find this challenging and of those that see it as a big challenge 37% cited consistency as the biggest part of that challenge. And so this again you know really just speaks to why a data quality program alongside that data governance is so important because with a strong data quality program consistency isn't an issue you know your information is going to be validated it's going to be standardized it's going to ensure that you have consistency across and within your databases and then between profiling capabilities and standards is going to be identified and can be easily and quickly resolved. So you know we talk a lot about trusted data here at Experian it's a really big topic for us and a lot of organizations really struggle with knowing what trusted data is and what that looks like and so I provided here a definition of what that might be and remembering that a strong data quality program and a strong data governance program they're able to help for your organization to be able to achieve the accuracy the consistency and the control that you need over your data in order to feel confidence that it can power your strategic decision making you know whatever that may be. So with that I'll turn it back over to you Shannon. Thank you so much for that great presentation so glad we got to work in there and again thanks to Peter for his presentation we had a great discussion with both Peter and Jenna are available and deciding right in here what is the significance of data lineage from data governance perspective. We'll start there. We'll start a question too. The lineage is where the data came from and the question people ask a lot is where did those numbers come from? I can remember working on the Lucent Alcatel merger and I was a chief financial officer and I said so what are we trying to accomplish here? He said well every day I have people to come into my office to give me a report that show me specifically why I should sell all the rest of the company except theirs to the French. He said of course if I believe all of these guys nothing gets sold to the French. He said I don't know where these numbers come from. I think I'm going to move them into a trusted data portal so that any data that has been vetted into that trusted data portal could then be used to make the types of decisions that we're being asked about specifically on there. I'm sure you have a bunch of examples as well but where does data come from can be hugely important. My favorite number for those of you that don't know is the secret, the channel and I know that's the answer to the life of the universe and everything. But Jenna you've got some lineage stories in there as well. Yeah absolutely. From our perspective the really crucial piece about data lineage is just being able to have that record, being able to track everything that's happened to data throughout its life cycle and we see all of a sudden they'll have done all of their due diligence and making sure that the data that they collect at the outset is of high quality, they have validation practices in place, they've done all of the steps that they need in order to make sure that that data is high quality and then all of a sudden a little bit further down the line they've got to use it and it's no longer that accurate information that happens. Often times if you're using a data governance tool or even data quality tool a lot of times they'll have built in lineage capabilities that will really help your organization to be able to map every transformation that has happened to the data throughout its life cycle. Let me push a little further on that too. One of the things you guys do is you keep track of my records, the one that I am, but there's another Peter Akin, actually there's two others out there that I'm already acquainted with, one of them is a divorce lawyer in Florida, PeterAkin.com. You guys keep my records separate from his records which is probably a good thing because he's a lawyer and I'm not. The other Peter Akin that I think is a really fascinating character is that he's the biggest rock and roll promoter in the state of Ireland and I'm proud of that. Thank you for that. Next question here. Should the data governance apply equally from an important perspective to all different types of data such as master data, reference data, metadata, control data, et cetera, et cetera. Do you want to start on that one? Yeah, so I think with this question of course data governance is going to be important across all forms of data and it's all about prioritizing. It's all about really understanding what data is the most important, what is the most essential to your day-to-day strategic decisions and once you understand that that's really where you want your priority to be. If you think of a major retailer what's going to be their most important information, it's going to be that customer record. It's going to be having that single customer view, that golden mark, what their purchasing decisions are and how they can market most effectively to them. If you're a retailer, that's the data that you want to start with. Whereas if you're a financial institution or if you're working in the public sector, there's all different sorts of priorities. While data governance is absolutely important across the board, just like Peter had said earlier, you cannot boil the ocean. The key to this is to really look at it from, first of all, we haven't done a good job of managing data in the past. This is where this particular slide comes from. I'm going to pop it up there. I'll just make it big so you guys can all see it on the screen. I'll share. There we go. Thank you. Thank you. I do say, and I absolutely can show, that 80% of the data in most organizations is rot. Redundant, obviously, per trivial. That's a problem. It means it's in the way. We've got to get it out of the way. Your organization will hear that and it's a true statement and they'll go, oh, my gosh, well, that's just terrible. We need to fix that. We need to make sure that data is treated back as an asset. Again, I think the soil versus oil distinction is a very important one to do. What Jenna was saying is, yes, we want all the data to be there, but what we want to do is find out what speeds 20% of the data that's non-rot and make that data the better data. There's no point in improving the quality of data that's redundant. There's not really best interest of everybody to do all of that. The other part that she emphasized as well was strategy. It's where we want it to be, but we're not going to get there unless we start to measure the stuff in years and maybe even in decades in some organizations. It's going to take a while to get to that. How does data governance influence the regulatory compliance related to data security, the IDFS, credit data, et cetera? Back to priorities. The key to this is that if you've got a regulatory piece that's coming up, there's nothing that focuses your mind but a lot of a hangman's news in the morning. Forget who said that. I'm sure we can look that one up on the internet and see who's better alleged. It does provide you that part. If you are out of regulatory compliance, you're not going to be able to do that. That's a big way to focus it. But that's not the only thing that you need and you need to think of it in the larger context as well. Jenna, thoughts on that? Peter, I completely agree. The 2018 mobile data management benchmark report that I referred to earlier, we ask a bunch of, we ask our customers and we ask them about data management practices and regulation time and again comes up. It's a key way of really providing that incentive and that added motivation to, you know, govern the data, have it high quality and really make sure that it's managed effectively. Perfect. What are the sort of questions do we have on that? Considering that data and information our assets produce are different. Yes, but at the same time, the whole purpose is to govern data that is shared. So you do have to have some, some relative agreement there. If somebody says 42 is this and somebody else says 42 is give a specific event, 42 is the meaning of life, the universe and everything, or 42 is my age 17 years ago. Two different pieces, right? One of them says I'm old and the other one says where do we come from, what are we doing here and where are we going to go and we're done. Those are completely different ways of doing that and yet there needs to be a common shared governance around them at least to say, you know, there's multiple people using this data so when you produce something that's an input, I'll have to tell a quick quality story around this that I like to tell. They were wonderful about knee surgery and they were apparently quite good at this. One third of all our hospital admissions were for knee surgery. Now, we dug a little further into the piece and found out that guess what, knee surgery was the default admission code. So the people who were checking people in were just flashing back on the screen and saying we got this all taken care of. They were optimizing for speed and doing nearly the amount of knee surgery that they thought they were doing even though they thought they were managing with good data. I'm sure you've got a lot of stories that talk about the same types of things. Oh, yes, we do. I will share a quick story but I'll just add a couple of things before I dive into that and that's, you know, that you would set it before Peter and oftentimes one of the best ways of getting that consistency that I had mentioned before to be a part of your data governance strategy is to make sure that people are speaking the same language and to have a glossary of terms or to have data definitions that are uniform across the organization and that's really crucial but Peter, a very similar story to what you were saying, we had a customer, top customer. I mean this person was doing hundreds of transactions a month. It was amazing how this one individual could be doing so much activity. Well, they had data governance practice in place that you had to verify a social security number from your customer in order to access their account and do anything and they were centralizing the record around the social security number. So she was entering her own for every single record. So you can imagine what a sort of nightmare that was from pretty much every perspective from their data management. Shannon and I were on view because we both had belly laughs on that one. I told the story on here as well about vocabulary, right? It is absolutely critical that she gets that right. Speaking of data quality here, I tend to reporting data quality and clean data rather than identify the root cause and revamp some of the poorly designed architecture systems and any transactional systems that aren't entering data integrity. So it's a really good question and it is a perception that we collectively fight as an industry. Most people think if you've got a data quality problem you fix it. Every time I find something in B.I. they fix it in B.I. but they won't go back to the legacy systems and put the fix into the legacy systems which would be a lot cheaper and I said you've got to put the expenses down and make sure the data governance process acknowledges those are the right amounts of expenses that can be done and they can measure millions of dollars for some organizations. I've got several examples in the right now. Okay, this is some extra work. Thanks for doing the extra work. If that extra work is repeated thousands of times by thousands of knowledge workers in your organization it has up to a big number eventually and that's a really good way to keep track of it is to run those things through governance so that everybody understands what the dollar costs are associated with those bad practices. The organizational approach is to build that case across business units. Make sure that it's not just someone in the IT department that understands that this is an issue. It's not just the data administrator but anyone who is experiencing these issues across the business really comes together, build that case and bring it to your senior management team and that's how you're going to affect actual organizational change and not just a manual quick fix every time. If you're always doing quick fixes you're always going to be doing more quick fixes whereas if you can take a step back from it it does give you the ability to go back and add a more systemic approach to the process. I've got the Barclays slide up here right now. I glossed through this because I was running a little short on time but if you work for people that's kind of good. The only bank that I know that has that in place I'm sure there will be others in the near future because Barclays is not the only one that has spreadsheet problems. We all have spreadsheet problems. Should the data governance initiative be integrated with data projects or should it be launched as a standalone? It's a good question. It's going to depend from organization to organization. Data governance is personal to your organization. I don't mean it belongs to you but it does belong to your organization. Your organization uses data differently from your competitors and your collaborators. We saw an example yesterday in one of the sessions where the data governance piece was run out of the chief information security office. Why? That individual understood it needed to happen and he was willing to pull that group into his purview and he said officially it's not going to stay here. I'm going to get ready to hand it to somebody else who's going to be able to move it out in the right direction. What works best for your organization? Again, be careful. Don't give it to a person who's not going to be there in 90 seconds. I agree with Peter. Ideally, data management is always looked at as a holistic practice and your data governance and your data quality and your data architecture and everything speaks to one another. Unfortunately, the reality is there's always going to be limits in terms of budget. There's going to be constraints on resources. I really would ask what Peter said. The real state would be that everything is interconnected and it's interdependent and you can't really have one piece without the other at some point. Again, this is the slide that describes that interconnectivity but I would not show that to management. Keep it simple for them. I love this next question because we get some form of this next question in every conversation. What is the simple way to convince management the importance of DGE than linking that to a dollar threat? One of the examples I gave was live. That was an interesting one. Tell me why we can't use the Army's data to say the Army soldiers make an appointment to see the Chief of Staff. There does have to be a motivator. Part of what this conference is about is about communication. There's a whole session devoted to communication planning around this. One of the things we use is the law of threes which is that you need to have a 30-second elevator pitch that may lead to a three minute conversation that might lead to a 30-minute overview that might then lead to a three-hour deep dive for some people on this. Again, Jenna, that's one of the things I know you guys do, very strong messages around focusing. Yeah. Again, like Peter said, the money piece, whenever you can build the case for how this is affecting your business's revenue, how this is affecting the bottom line, how this is affecting the day-to-day operational efficiency of your workers, that's always going to be the best way. If this is truly impacting your future strategy, if this is going to stand in the way of your growth opportunities, just always having a broader case as to how data governance, the why, as Peter said before, the why of data governance and how that fits into the larger picture. The issue of multiple data owners for the same KDE, I'm not quite sure what the KDE is scanning for. KPI, key performance or key data entry point. Yeah, so if you could clarify that for me, Beth, to ask, when we established the element as a KDE, we are figuring out that multiple business units, for example, deposits, loans, etc., are using it, and in some circumstances have different definitions. So basically, how do you deal with the issue of multiple data owners? One of my favorite examples of this was that we were working with a hospital system, and a hospital system was using a very nice ERP internally. The ERP was such that you could right click on any data field and it would show you the data dictionary definition for that field inside there, nirvana to its data geeks, right? Well, this was wonderful and in spite of that, we found with an audit that there were 12 different uses, right? I mean, here it is, right click and it tells you what admission date is. On the other hand, 12 different uses of it. Now, the 12 different uses was nice, but, yeah, key data element. The 12 different uses was nice, but man, when we tied that back to dollars and showed that the hospital system was losing tens of millions of dollars because of that inappropriate use, that got the attention of people. Yes, we absolutely have that. Again, my records over at Experian are key data on a couple of different things, right? They've got things out there that they've got, for example, in my particular instance, they have what's called a credit watch on my account because I have a lot of people that are very interested in my stuff, or maybe it's the other Peter I can stuff that they're trying to get. Who knows? But they keep my data very, very close to the class of data users that are out there. There's lots of other pieces. I've worked with a lot of different organizations where they're doing different KPI's and things. If you say what does sales mean, you think, oh, I got sales, right? Well, no, no, no, sales to this country means that data comes in after return. Sales to this country means data comes in accumulated with returns. It can be horribly important, horribly confusing, and that's one of the challenges that we all collectively work with. I say to governance professionals to try and make a lot of stories that you've seen there because of the particular central role that experience plays. Yes. So for the conflicting data owners, what we often find is that one of the best ways to sort of mitigate that is to enable people to have a tool that has that sort of flexibility. So if you want to define data in one way in this part of the organization and someone wants to define it in a different way in the other part of the organization, it's not affecting that centralized record. So it enables them to work with the data in their way and then in a centralized strategy, ideally you might even have a chief data officer being the one sort of making that final call, being the one who defines what it will be in your centralized MDM or whatever your centralized database system looks like, and then sort of democratizing that data and allowing the different users and the different owners across the business to maybe have that flexibility if they want to define terms in slightly different ways, again, without sort of compromising the integrity of that centralized record. And for those of you still on the line, we really appreciate your staying with us. I'm going to tease back on a couple of things that Jennifer said just for clarity's sake. You use the word a key decision maker. Yes. There needs to be one person who makes that decision because if you have two people making that decision it's kind of like having two watches. Doesn't really help. That can fact introduce lots of confusion. But at the same time, that's got to be balanced between the word use and word democracy. What is the interest of the organization and how can we best use it? There may be valid reasons for replicating data. If you've got it and you can control it, that's great. But if you don't know what's going on and it's just happening, then you lose the lineage. You lose a lot of the threads that we've been talking about here. You lose sight of what your strategy was. Time for at least one more question here. We are seeing issues that when trying to run data quality rules on the element from business requirements, some business units don't use it in that same way and if such the quality rule does not apply to them. We're glad to find those but we don't want that practice to continue. If I say that the numbers have to be on our age precise to 12 decimal points that may be very important for the actuaries but it may not be for other parts of the group. So that may be exactly the case that I was just describing where from a, there may be a best use that you can do if you're going to do it one way but if there's a valid use for doing it both ways, let's document them, keep track of it and monitor that particular process to make sure it happens. Because if I need data that's precise to 12 digits and everybody says your age is 42 plus 17, that's not going to be very helpful for people. This is another instance too where that data profiling piece that I discussed before could really come in handy. So data profiling would allow for you to look at your data and be able to recognize if a field that normally would have 12 digits in it, all of a sudden you're looking in 20% of your data is not fulfilling that requirement. Profiling will give you that insight. It will allow you to flag it and then even allow you to see what pieces of data that's related to and then again, if you have that lineage piece as well, you can figure out where it's coming from, where the issue is stemming from and then that basically becomes the cultural thing. Technology can only do so much. At some point you need to make sure that the people in your organization who are interacting with that data day in and day out understand the value, understand the importance of it and make it a day to day part of their jobs, adhering to the data standard and adhering to agreed upon standards is just par for the course. Indeed. So I do believe we were just right. We just got a fairly couple minutes left. So let me wrap it up here very quickly. Jenna, thank you so much for joining us today. A thanks to Experian for sponsoring today's webinar and helping us to make all these webinars possible. Peter, thanks to you as always for another great presentation. We hope that we will see you next month in the next month's webinar. Again, these data ed is the second Tuesday of each month. And thanks all of our attendees for being so engaged in everything we do. We just love it. We really appreciate it. And as I mentioned, I will send a follow-up email by end of day Thursday with links to the slides, links to the recording of this session as well. Thank you again, everybody, and I hope you all have a great day. Thank you. Thank you. Thank you.