 Hello and welcome. My name is Shannon Kemp and I'm the executive editor of Data Diversity. We'd like to thank you for joining this month's installment of the Monthly Data Diversity Webinar Series CDO Vision. This series is designed to give year-round education on data strategy topics and for a more in-depth education. We hope you can join us this year in just a couple of weeks here at our CDO Vision 2016 conference, April 19th through 20th in San Diego, California. Just go to CDOVision.com for more information there. And this month Malcolm Tillson will be joining us with Kelly O'Neill and to discuss have an open mic session. I see questions are coming in already. That's fantastic. Just a couple of points to get us started due to the large number of people that attend these sessions. You will be muted during the webinar. For questions we will be collecting them via the Q&A in the bottom right hand corner of your screen. Or if you'd like to tweet we encourage you to share our highlights or questions via Twitter using hashtag CDOVision. As always we will send a follow-up email within two business days containing links to the slides, the recording of this session, and additional information requested throughout the webinar. Now let me introduce our speakers for today. Well known in the Data Diversity Community industry expert Kelly O'Neill Kelly is a founder and CEO of First San Francisco partners. Having worked with the software and system providers to key to the formulation of master data management and enterprise information management initiatives, Kelly has played important roles in many of the groundbreaking initiatives that confirm the value of EIM to the enterprise. Recognizing an unmet need for clear guidance and advice on the intricacies of implementing enterprise information management solutions, she founded First San Francisco partners in early 2007. Joining Kelly this month is equally as well known especially in our data governance community Malcolm Chisholm. Malcolm has over 25 years experience in data management and has worked in a variety of sectors including finance, insurance, manufacturing and government, defense and intelligence, pharmaceuticals and retail. In addition to being a well known presenter Malcolm writes and columns and trade journals and has authored the books managing reference data in enterprise databases, how to build a business rules engine, and definitions in information management. He holds an MA from the University of Oxford and a PhD from the University of Bristol. Hello and welcome to you both. Good morning. Hi. Hi Malcolm. And I'm excited. Today I get to play a special role of moderator to lead the discussion and questions we most commonly see throughout the webinar series. Maybe questions we haven't had a chance to get too much as well as some questions coming in from the audience. Now we've prepped a few questions but I see questions already coming in from the audience. So let me just dive right in and start there because it's along the lines of what we see commonly already anyway. First question coming in from, I'm interested in understanding organization design of various CDO organizations. What data management functions reside with the CDO? I realize they will vary based on industry, business size, et cetera but where can I see some good examples? So I think maybe just in terms of the answering, maybe depending upon Malcolm and I can give an answer to each question just because we will have a slightly different perspective unless of course Malcolm might completely defer to you based on your technical expertise. Anyway, so you started to answer this one Malcolm so I'll let you go first. Ah, thank you Kelly. So, well I think that there are a lot of data management functions. I think one of the problems is that it's not clear what all the data management functions are yet. The sum that are emerging like how do you govern aspects of big data such as if you do columnar databases controlling the names of column families and column qualifiers in these columnar tables. So we say what data management functions reside in the CDO. I think it certainly can vary. We've got to realize it can vary and there's new ones that may be emerging fairly frequently. But I think the sum that are reasonably well understood amongst those are legal privacy and compliance so setting policies and guidelines around what you can do with data is something I think you see in the CDO. That would include responsibility for onboarding new data from external data vendors and also setting the rules for permitted use of data. Many people think that this legal privacy and compliance is just around data security and access controls. That's looked after by data security which sometimes is within CDOs but not often. But all these aspects of what you cannot do with the data itself with its content are run by the CDO. Closely related is the rationalization of data demand which is well if you want data where do you get it from because people will just get data really nearly based on who they know, if they've got some spend available and so on. And controlling that rationalizing who gets what data, what makes the best sense for the enterprise is also a function of the CDO. So there's a couple Kelly do you have any thoughts? Yeah, just to add to that, I tend to see there being delineations across business and technical is one delineation and then the other delineation is between data management and data consumption. So when we're looking at the scope of a chief data officer's role, in general it's a business role. So let's just kind of put that on the table. I have not seen many that report up to the CIO although I'm sure they do exist. Because it's a business role, those data management aspects that are focused on business involvement, business strategy, data strategy participation by the business. So these would be things like data governance, data quality, master data management strategy, reference data management. All of those fall under the umbrella of the chief data officer whereas some of the more technical roles like data architecture for example, many times is outside of the purview of a chief data officer. So I would see that as one delineation. And then the other delineation is does the chief data officer include data consumption in the form of business intelligence analytics and big data from a non-technical perspective. So that's another delineation I've seen in some chief data officers have that analytics function under their umbrella. And some chief data officers don't where the analytics function is a key consumer for their organization. Sure. And Kelly, while we've got you on the line here too, let me back up a bit to the first question that we most commonly get. What's the best way to build a business case for a data program in general? And then how do you fit the CDO into that program? So, you know, I love this question. When we were prepping for this webinar, this category of question came up a lot. And unfortunately the reason that this question comes up a lot is because there's so many different answers for this question. So the best way to build a business case for a data program is to look at existing and known business challenges and how inaccurate and unavailable untimely impacts those business challenges. So if you're looking at, you know, what your CEO is committing to for 2016 and he's got on his or she's got on her list things like increased revenue by a new market segment, for example. Well, there's a data component of that to determine do you understand how, who is in that market segment? Who do you sell to currently in that market segment? What products do they buy? Why do they like those products, et cetera, et cetera. So to me the best way to build a business case is to start with those known business drivers. Sometimes those known business drivers are regulatory requirements. That's a great justification to get started with a data program. It's not your long-term sustainable justification but it is a great way to get started. Malcolm, you've had tons of experiences. What have you seen is a good way to build a business case? The way I have approached some of these is divided into three areas, efficiency, effectiveness, and risk. So efficiency is to show that you could actually save money operationally by doing this. That's not the most exciting but it tends to be the easiest to quantify. So you can come in with cost savings. Effectiveness is I think somewhat more on the lines of what Kelly was going over in part of her answer, which is if you know, if you can find out what the business goals are, look at the business strategy and say, well, in order to do that you'll need to do this and such and such with data, then that also resonates. And business models are becoming more data centric. For instance, if you move to customer centricity you need to have information at hand about your customer to manage all the touch points with your customer. So there's a big data component to that already. And then the final category is risk. How do you reduce risk? That comes into the regulatory aspects and that's increasing in lots of industries. Banking would have BCBS239 which is an industry regulation which is very specific about data and having a data program. But if you can show that you can use data to mitigate risk in general or you can reduce the amount of operational risk in the data itself, that goes down as well. So that's kind of my approach to efficiency, effectiveness and risk. I love it. And it's a perfect segue really into the next three questions I'm just going to bring up because these really all deal with exactly what you're talking about in terms of risk and the monetization of data. As it's listed here, what does it mean to monetize data? How do you monetize data? And then I'll get into the next question from our attendees. If you want to further dive into this, Malcolm, into these kind of questions and expand a little bit further. It looks like you're on mute there. Sorry, I was on mute, Shannon. Sorry. So it should elimination of risk be a reason to invest in data? Well, if we look at the sad story of the Mars climate observer probe that crashed into Mars in about, I think it was 1999 and cost the U.S. taxpayer about $350 million, then the answer would be yes because that spacecraft had its navigational system programmed in metric units and the motor programmed in imperial units and they just sent numbers to each other and the units were different. So when the spacecraft arrived at Mars, it essentially misfire and went slammed into the surface of Mars and was completely destroyed. So there's an example of how a data risk was not mitigated and led to total mission failure. Now, we don't often see examples of that. There's a scale of risk. It goes from just acceptable cost of business all the way through to some really bad event. And you could make the same case about Lehman in 2008. How exposed were people to Lehman? So yes, elimination of risk should be a reason to invest in data. And the regulators are, I know if you can eliminate it but you can mitigate it. And there's ways of doing that and a lot of the regulation in finance is striving for that but you also see it in things like engineering and manufacturing too. And the data in those areas is very, very important. Data coming, data about quality defects. For instance, if you look at Schuhart's approach to data quality, he was very heavy on data. If you say, well, we've had an average of three problems with our aircraft in the last 10 years, that's one statement. But if all three of those problems occurred yesterday, I think you've got something going on that you need to investigate. So the data can be used to find assignable causes for mitigating risks. There's a good body of work on that. And I think it really, you know, you can, if you dig into this, find reasons to mitigate risk as something that needs investment in data. Kelly, as I toss it over to you, maybe I know you have a similar view. But to add in the question from the audience, when working to establish total cost of ownership, aka data management, is there a standard list of menu components that can be referenced? And what should be categorized as a data management expense? Got it. Okay. So I guess just one thing to extend Malcolm statement and then I'll answer that question directly. Elimination of risk is a great place to start with data just because risk resonates with just about everybody in the C suite. So I fully believe that this elimination of risk is a great way to approach it. Now, if you look at how to quantify that total cost of ownership and just quantify the value, as Malcolm said, things like productivity, cost, effectiveness is another way of productivity, risk sometimes when we're quantifying that is avoidance of risk. So it's hard to quantify it and that's better served in terms of an anecdotal story. So some of those anecdotes that Malcolm just told, I think that preempting, you know, the failure of Lehman, people probably would not believe the quantification of risk, right? In the sense that the gap of lack of understanding of exposure, they wouldn't have believed the calculation that came up in terms of how much risk was associated with lack of understanding of data. But I would point to anything that's quantifiable from a productivity perspective, anything that's quantifiable from an additional cost perspective, those additional costs can be in processes that take longer than they should, it can be in what John Ladly will have to use some of his terminology since he's not here, make sure that he's somehow included. He calls it rot, so redundancy, obsolescence and triviality of data and those are costs associated with maintaining data that is trivial, meaning it's no longer used in an organization, but it is being stored somewhere and there's a cost to store it, obsolete data where the data may have been meaningful at one point, but based on potentially the date of that data, it's no longer used in the organization, so it's obsolete, and then redundant, so there's a tremendous amount of redundant data, so there's a specific cost associated with rot that can be calculated. So one of the things I could talk about this a lot, we do have some content specific to this calculation and Shannon, a lot of it is either available on the Dataversity website or via the EDW and other sorts of conference events, so I'm not sure who asked that question, but we can point the audience potentially in a follow-up email to online slide share or content from previous conferences where John and I have gone through a tremendous amount of opportunities to calculate costs associated with data, upside of costs associated with data, and therefore how you get to a total cost of ownership. Sure, and you know, the next question we had, you know, kind of talked about working with a business sponsor, but let me go back to that. I want to stick with finances for a little bit. So how do you put a cost on poor data quality and if data has value, why don't we have data accounts? I'd like those two questions together just to kind of wrap up the ROI of data and then we'll, again, get back to this response in the previous questions. Malcolm, do you have a take on that? I know you work with, you're a strong advocate of data quality especially. Well, right, I rather like the question, what does it mean to monetize data? Because I think that comes into it as well, particularly around data accountants. So if you, what it means to monetize data is to sell your data and people are generating a lot of data that is actually valuable in other contexts. And I've worked with data vendors who do that. Now we're seeing, particularly in Silicon Valley, the emergence of companies whose business, core business model is actually selling data. So monetizing data means, to me, selling it. There's different ways of doing that, but you could, you know, some very simple ways. So if you do that, you can start to have data accountants. There are some specialized companies that put value on data, including data breaches, unfortunately. So that is beginning to happen. But if you say if data has value, why don't we have data accountants? Well, lots of things have value, you know, lunch. My lunch has a value. Why don't we have lunch accountants? I know it's not quite the same thing, but I think that there's, you know, there's how do you set a value on anything is an interesting question. And I think that, you know, I think we will see the emergence of people who are able to value money, data, sorry. How do you put a cost on poor data quality? It kind of gets back to what Kelly was talking about in terms of, you know, workarounds, processes that run for longer than they should do. And I think the inhibition of the sclerotization of the business of the enterprise, so that the enterprise can't adapt or is too scared to because it doesn't know what it's doing because it's data so bad. You can't trust its data. Those things are a bit more intangible, but you can, again, say that these have a cost in terms of something to do with the business. So I think that those are some of the approaches that I've seen. And just to add to that, there's the forensic investigation to go through the cost associated with each step of a process that, of course, comes up with a very specific, quantifiable answer. But what we've seen is that even if you take a general assessment of how long a process takes and how much rework is done associated with validating data that should be accurate in the first place, double checking, root cause analysis. So even if it's not to the absolute accuracy and you can get an estimation, those estimates are pretty accurate within, say, a 10% to 15% delta, and many times just the anecdote or the story associated with that estimate is compelling enough to put that cost on poor data quality. So when we're presenting this information, a lot of times we get questions. Well, Kelly, we don't have a process-oriented culture where I can't quantify what a process is. I have no idea how long it takes certain people to do this. So I think if you can even work with a subgroup and come up with an estimate, it doesn't have to be perfect to be compelling. Sure. And so, okay, so we've built and we're building an ROI. There's a slight disagreement with the dear definition of monetized Malcolm terms of that it's selling data. Certainly, you know, when you're building an ROI, and just as you both talked about even after the comma came in, you know, it's putting a value to it, to build that ROI and justify, you know, how much, like you said, Kelly, you know, how much does it cost when mistakes are made, how much does it cost to manage the data quality, et cetera. So we've built an ROI. So then let's go back to the question, back up here a little bit. So how do you, and we're making the case to executives to our data governance program, our data program in the company. So what is the true role of a business sponsor and how do you get the buy-in from the executives and get that business sponsor to get everything started? And then we'll get into work charts. So Kelly, maybe if you want to start there. Yeah, sure, absolutely. Again, I think this is a great question in terms of the true role of a business sponsor. A lot of times people will be tapped to be a business sponsor and they will accept the responsibility or the accountability, but not really be clear on what it means to be a business sponsor. So I think that it is important for the organization who is trying to get whatever data program justified, whether it's the setting up of the CDO office or whether it's actually just trying to, you know, put together some improved data quality processes. So the role of the business sponsor is to provide air cover, is to provide advocacy and evangelism. It is to help sell up as well as sell down the data initiatives. It is to help get funding. It is to sit down and validate and review and assist with those presentations in which you're asking for budget. I mean, a business sponsor should be someone who is very, very active. And it's important that they understand that if they take on that responsibility, that they do need to be active and they need to be that senior level voice evangelist advocate who is constantly talking about the value of data to the enterprise. Now, not every business sponsor really understands what that means, nor do they understand the time commitment associated with it. Going through that validation process of, okay, if you are the business sponsor, we do have this ask of you. Now, executives are all very busy. I'm sure there's people that are listening to me right now and saying, you know, nobody's going to take on that level of responsibility if they're truly a senior executive. So it's not necessarily asking them to do all of that over time or immediately, but over time you want to get them involved in a way that helps you sell beyond the data program and helps you identify value, helps you identify how you can link your data program to the strategic objectives of the company or the division or what have you, and be active. So a business sponsor is not just someone that introduces a meeting for five minutes and walks away. It is someone who is very active in promoting the program, whatever that program may be. So that's my view of a role of a business sponsor. I don't know, Malcolm, if you want to add to that, because we've both seen what happens when you have a good one and what happens when you don't have a good one. Last, so. Yeah, I think in the early days and maybe some today even in the early days of data governance, there was an expectation by those who were in the data governance office and, you know, tasked with doing data governance. The business sponsor was just like IT. The business sponsor was going to tell you everything you had to do in enormous detail. Well, that's not going to happen because they are not domain experts. And this is part of the reason that IT has not always done too well with data governance because their attitude is, tell me what you want me to do and I'll do it and then I'll go and do something else. And the business sponsors, as Kelly said, cannot be expected to know the details of data governance. But they must do some due diligence to make sure that what they're sponsoring is on the right track. They have that obligation. And I do agree with Kelly that they also have the obligation to socialize with the other executives at the executive level what's going on with data. Now, a way of looking at business sponsors for data governance organizations is like venture capitalists. So if you had to, you know, fund yourself with a startup, you know, you've got funding through venture capital partners, you would have to report back to them on a fairly frequent basis. So you better be doing that back to your business sponsor. They can then use that information to tell the rest of the executives how well you're doing. And just like VCs, they want to make informed decisions about their investment. So it is up to the business sponsor to do some degree of due diligence, get some, you know, expert advice, whatever they have to do to make sure that what they're investing in is going to work. So I think those are just a couple of things I would add on to Kelly's description. Sounds great. So moving from monetization, so we've talked to the executive, we've got a business sponsor, you know, so we've determined and maybe we have or have not determined that a CDO is needed. But also we've talked a little bit about the CDO role and who they report to a little bit already, but how do you build the CDO office who reports to the CDO themselves? And Kelly, you already addressed the question a little bit in terms of what is reporting, the reporting line for the CDO, whether it's a CIO or a CEO. And is it, are you guys typically seeing the CDO as part of the business or IT? So those are kind of three questions built in there. First, you know, let's start with just the CDO, where they're sitting, as we can just kind of go back to that a little bit. And then how do you staff the office? Who's actually reporting to the CDO themselves? Yep, okay, great question. So most effectively they're part of the business. We have seen the CDOs report up to chief marketing officers, chief operations officers, chief finance officers, or to the CEO themselves. So I guess the viewpoint is it doesn't really matter where it sits as long as it has the advocacy and the reason for being that creates a consistent funding model for the organization. So many times it falls under marketing because many businesses, a lot of the use of data is to sell their product and market their product. So that's why it tends to be successful under a marketing or a chief marketing officer. Now a chief operations officer may have responsibility for all shared services across an organization and data being a shared asset falls nicely under a COO sort of structure. CFO many times have compliance and regulatory. So there's a great relationship between data, of course, and regulatory compliance. So sometimes that's a nice fit. So it really doesn't matter and it's a highly variable based on industry and company culture and who has the juice, right? Do you have a really powerful CMO or do you have a really weak CFO? Those are all of those questions. It needs to be someplace where it is supported and funded on the long term. Now who reports to the CDO? Again depends on the scope. So does your chief data office include the data consumption function or the analytics and business intelligence and all of that? So that would be one delineation and so reporting up to the CDO is of course data governance. However you define data management. So the business aspect of your data quality initiative, your big picture metadata initiative and when I say big picture metadata it's everything from the data glossary or sorry the business glossary all the way through to other aspects of metadata. It is your master data management initiative, your reference data management initiative, your data quality initiative all from a business perspective. Now I see that there's a question here around how the CDO group or how the CDO works with the MDM group so I'm going to go ahead and just preempt and answer to that if that's okay. The business strategy associated with MDM many times is under the CDO umbrella and the technical implementation and technical support of MDM is in the IT organization and they need to work closely together in order to make it successful. That same sort of dotted line matrix structure exists across every aspect of data management not just master data management but it's a very good example. And so there is a tight partnership between a business CDO office and the direct reports and the IT organization and it's most successful when people understand that it's a shared responsibility and that the role of the CDO is just calling out the business accountability for data but it does not take away from IT's importance to help to manage that data. So hopefully that answers some of the questions. Absolutely, and you totally read my mind in terms of going to the MDM question. Malcolm, anything to add there? I've seen situations where and this may be getting more common in very large organizations like individual CDOs in particular lines of business and then they roll up to a global CDO so it's time to see that now. In terms of the CDO, I mean they need an office of data governance a data governance office to actually do the data governance work and a lot of that kind of work that I have seen is simply you're sitting there and you're getting governance cases of all kinds of types coming in every day that you have to deal with so there's a need for that kind of staffing to deal with things about related to data that come in all the time. Then you can get in also as Kelly went through I think pretty effectively all the different specializations and the liaison with IT all of which is necessary. I think there's one other aspect which is in terms of the dotted line you have IT and that's fine but I think it's important to establish partnerships between the office of the CDO and areas like legal, vendor management, PMO and internal audit because they have important roles to play with data. Legal can advise you on the laws about data. They don't get involved in managing who can do what with data but they know the laws. The audit will, you need to get them to make sure that when they do an audit that they look at the data policies and see if they're being complied with and if there's a finding they should work jointly with the office of the CDO to cure that. I think there's some other matrix relationships as well. Perfect. Pulling up the next question and then I will go back and then we skipped over the data governance question just to kind of look at the next question we've had in the series. We've answered a lot of this already in terms of building the ROI and getting a business sponsor and executive fly off and how to build the CDO office should it be necessary but anything to add to the question of assuming an organization is starting from a clean slate what would be the sequence of data management areas to implement to get an EIM program off the ground? Malcolm, you want to kick us off there and then we'll turn it over to Kelly. Sure. If you are starting from a clean slate and thinking about data governance I think you want to start at a fairly high level in terms of well, this is asking about data management areas rather than the organization so I think you need to get something for doing data policies in place and data standards because they have wider implications over the different data management areas. If you want to dig into the different data management areas themselves I think data quality is an important one a continuous data quality monitoring program for production environments or at least important ones is very essential actually but that goes hand in hand with you can set up the detection mechanisms and the controls you then need data issue management which tends to be much more of a business function than data quality detection does which is more tool oriented so I think that's another one and then semantics in terms of understanding the data because a lot of the demand we're seeing for data is being driven by analytics areas reuse of data in the old days when everything was its own silo people didn't care so much but the understanding of the data goes hand in hand with things like quality to make sure that you can do it so what I call information knowledge management is very important as an area to get under control then you can get into the more specialized areas like reference data management and master data management which are a little more technical heavier in some ways so those are at least my initial thoughts Kelly? Yeah, I would add to that and Malcolm you started what you were saying assuming you have the organization in place so I would say that's a great place to start based on how you have identified your priorities based on your business need so if you have identified that you've got a regulatory requirement as a priority or maybe you've identified that you have a new market segment that you're going after and you need the data to understand how to access that new market segment or you've identified that your order management process is faulty and there's a lot of costly problems associated with that so whatever your business cases to get started having those people in place will enable you to create the policies, the standards the data quality management, the issue management, all of the stuff that Malcolm talked about in a more efficient way so that once they get started they're not starting and stopping and starting and stopping so that would just be what I would add to Malcolm's statement. So perfect and you know again let me get back to the last question on the slide here. It's a more refined question and a little bit deeper in terms of our narrow and scope. How do you address the belief that data governance and data management are projects that have beginnings and ends and most importantly ends as you, you know Kelly you were talking about projects there not necessarily taking on a whole CDO program that goes on forever of course but how do you address that? So I'll go ahead and start the answer to that. I think that this is a big challenge because many organizations are very project driven and things only get funded through a project lens. So understanding that, getting justification to maintain high quality data as an ongoing operational process is really, really important and what the way that you go about doing that is broken down into a series of projects. So we don't want to have people believe that it is a project that has a start and an end but that it is a series of projects that has the constant ongoing requirement of improving the quality and the value and the consumption of data across the enterprise. I do know that there is a practical problem with funding and sometimes it gets funded as a project in the beginning and the communication is constantly educating the organization that yes, it might be funded as a project but we can't think of it just as a project. Data is an ongoing asset within our organization and an ongoing problem and it may end up that there are periods of greater investment or lesser investment based on how we are addressing those business cases in the sense that we might have a lot of activity for a period of time that data category starts to be managed in a steady state and the investment starts to drop a little bit and then you turn your sites to another category of data or another operational process and so the investment increases again and then it becomes more stable in steady state and it goes down so it ends up being a little bit in ways but in general it does need to be believed and the communication needs to be consistent across the organization that it is an ongoing management of an asset just like HR, just like expenses, just like infrastructure and logistics. It's a management of a corporate asset. So just see there's several projects within the program but it's always a program to manage everything long-term. Just to clarify and there was a comment that came in a program not a project and so yes, there's definitely both involved. Malcolm, I know you have things to add to that. Yeah, I would go beyond a program even. Yeah, I would agree. Yeah, the project mindset is particularly dominant in IT. I mean they're not here to defend themselves so I can beat them up a little. But it is a problem. I mean IT has typically worked in projects, thinks in projects. They think they're concerned about data. They tend to see things in projects. Now as Kelly said, that happens in some of the business too. But to pick up on Kelly's point, I think you say look, people are everywhere in the enterprise. Who sets the rules of the road for how we manage people? Answer HR. It's not a project, it's not a program, it's a function. Money is everywhere in the enterprise. You can't play games in their expense accounts. Who sets the rules of the road for that? Finance does. Again, it's not a program, it's not a project, it's a function. Facilities management is, we all work in office spaces where we don't buy our own desk and furniture and equipment and supplies. Facilities management does that for us. Again, they're a function, not a program or a project. Well, data is everywhere in the organization. Who sets the rules of the road for that? Who says you can't play games with data? Well, that's data governance's job. Again, it's a function. Now, you might have activities in building out the data governance function that you can say are projects such as should we put in a metadata repository or a business glossary? Should we extend data quality monitoring to this particular application and so on and so forth? There are build aspects to it, but that's just the same as human resources, finance, or anybody else experiences too. Those are legitimate projects. You can see them as that, but not the function itself. It's a genuine horizontal function in the organization that has responsibility for data. I think that this is really what we're seeing as a maturity of management, corporate management functions in general. There was a period where HR departments didn't really exist. There's the history of how human capital management came into the subject area that is now widely spread across universities and get master's degrees. Guess what? About five years ago, universities started doing very specific master's and analytics degrees and things like that. I think we're maturing as an industry where the function of a CDO is going to be more common because people recognize that data is an asset, not a byproduct. I do think that back to why we're here and this is all talking about CDOs, that's kind of the first equivalent of the head of HR or the head of finance or the head of facilities like Malcolm just talked about. I think we just need to go through this as an industry and continue to justify and tell each other how we've been successful and the value we've provided back to the organizations to continue to mature the industry. And I just see that this is a growth path that we're on. And I'm so glad you brought it back to the question of value again. The next question coming in is specific to the areas that we have just been talking about. How do you build a rationale to justify the acquisition of data? Oh, I like this. I like data acquisition. I love it. It is a very complicated area and it's an emerging area. So in the old days when, you know, we all sort of had just operational silos, data was just created in them, but today we see the value of bringing in data sets for things like analytics, understanding customer profiles, demographics, and so on and so forth. Data subscriptions are very important in finance for things like pricing, bonds, sorry, financial instrument, static data, market data, so on. So the question is, and this stuff is very expensive, so you need to, you need, you can I think fairly easily justify that you have to do this centrally. Now you do have to work with vendor management as well on this because there's, you know, for instance just one of the things you have to look for in these contracts is atypical clauses to see if you, for instance, did you know that data vendors often put into a contract that they can come in and audit you unannounced? And I have, I have experience of that actually happening where one of the data vendors just showed up one day and said we're here to audit you. So other ones do things like you can't play games with the data. Today people think that you can, any data within the firewalls of the enterprise is just fair game, well it's not. And the vendors have things like data seeding, they employ companies that spend all their time checking up to see if royalties, sorry, contracts are being broken and so on and so forth. So data acquisition is a very, very important area but it goes beyond the contractual legal aspects of it into well what is our data demand? How do we plan for our data demand? What kind of, what is the process for that? You can't have people just going out and getting data which unfortunately in many enterprises just happens today and you're continuing to add bricks essentially to the tower of Babel. How do you know what to do and what about the rational rationalization of the numbers of data vendors that you're going to deal with? We all know it's better to deal with fewer vendors. So all of these things come into this really fascinating area of data acquisition which I think is something that's really sort of up and coming now in data governance and is going to develop quite a lot. Kelly anything to add there? You know I can hardly add to anything about data acquisition from Malcolm's perspective but I was having lunch with an old colleague yesterday and she talked about how she's in this kind of consortium of people that talk about master data management. It's a group of technology companies and they come together in an informal way to talk about how they are mastering their data. And one of the conversations that has come up in this forum is that there's so many ways to crowdsource data now. So it's not just buying data from D&B anymore. There are so many data service providers out there on the market and crowdsourcing options on the market and this is an area that I think is going to continue to explode because there's greater requirements about validated single entity identification, so legal entity identifier in the financial services industry, the unique device identifier in the medical devices industry. So there's a huge industry about validated identification and at the same time there's the opportunity to just crowdsource demographic information and other sorts of shared data. So it's almost getting to be like a build versus buy decision where you're trying to determine based on my data requirement and what do I need to know? What's my gap that I don't have now? Am I looking for eye color? And I don't have eye color. Is there a crowdsourcer that can help me get to that based on relationship with optometrists or something? I don't know. I'm making it up. But it's just a growing industry that I believe it's going to continue to expand and continue to be highly specialized and so this is a question of not just how do you justify the acquisition of data but what are all of the costs associated with that? I mean Malcolm just went through all of these different variables and all of these different costs. It's not just buying the feed. That's just one piece of the puzzle. So anyway, really interesting category to discuss for sure. Yeah. Can I add something to this Shannon? Yeah. The point is that there is a grow in emerging field of data fraud as well where people generate fraudulent data to sell it. They will, you know, these click farms that do the fake click-throughs on online ads, there's people that will sell you, I think a million Twitter followers cost $650. These aspects come into this area too. So it's the buyer has to be aware of what they're buying and that's something that I think is, you know, difficult for people to get their heads around but I think we'll become unfortunately a more important area for data governance as time goes by. Sure. And it sounds like we can almost do a whole webinar just on this topic alone. But we do want to move on just we've got 10 minutes left in the program. We do have a couple of questions that always come up. They're really kind of funny and it's involving, you know, we see this question come up often with established companies of course and people seeing, especially this question comes up when people see a lot of data quality issues that aren't getting addressed. So how do you handle the politics of middle and upper management which prevents you from reaching out to the CEO and just really getting data management issues addressed? Malcolm, you want to start with that one? So let's just go through the politics of middle and upper management which prevents you. Actually, Kelly, can you go, I think you're better able to do this one than I am. Sorry. This is what I deal with all the time. Absolutely. So Malcolm, just like in-depth answer to the data acquisition question, this is something that I deal with all the time, we call this the clay layer because nothing permeates the clay layer going up from kind of the worker bees through middle management up to the executives and nothing permeates the clay layer coming from the executives through middle management to the worker bees. We call this the clay layer. It absorbs everything and nothing permeates it. And it is the biggest challenge from a data perspective because executives generally don't even know what is going on from an operational perspective that adds to cost, decreases productivity, creates problems, let alone how long it takes their poor worker bees to generate monthly or quarterly reports. And if they knew how many sleepless nights and weekends and all of that it takes to generate a standard monthly or quarterly report, they would gobsmacked. So it is a really interesting thing and I think what can happen quite commonly is that the incentives for the middle management layer are very different from the executive layer and from the operational layer and trying to permeate those incentives and determine how to either feed into those incentives and get them recognizing that the data program, the data discipline, the data initiative whatever you want to call it will help them achieve their objectives. That is really the best way to start to permeate this clay layer because many times they are incentive to do things quite differently than the data program wants. Many times they're incentive very departmentally in a very siloed way and for them to allocate resources beyond their silo is in fact detrimental to their success. So I realize this is a long answer but the politics of middle and upper management to me is fundamentally down to what do they need to accomplish from a professional perspective, what do they want to accomplish from a personal perspective and trying to feed the data program into that and helping them to see how their support of the data program will help make them more successful, not will hinder their success. So I'm done with my soapbox. I used to reflect that very often upper and middle management didn't care about data so they didn't get in your way. That's why I like reference data, coattails because nobody else was interested in them and they wouldn't bother you if you went into them but as Kelly pointed out there are a lot of politics today actually to navigate and I don't know if I can add to that so we can move on to another question. I find that when I talk to the most senior executives they have no idea that there's data problems in their organization and I heard executive after executive saying we have no data quality issues. Our data is perfect. I mean I sign off on our financial reports every quarter. Of course we don't have data problems and then you do your investigation into how you get to that perfect data and it's like hamsters in a wheel that are just like spinning and spinning and spinning and it's really it's just a lack of knowledge and that middle management layer they would expose themselves and put their profession at risk by admitting to it. So again I just think it's this layer that is really hard to navigate and hence this is why it's a really good question. Sure. And the next question coming in deals with a little bit different politics in terms there's always been the gap between business and IT and it also brings up another question that we always are another topic that always comes up in terms of metadata. So do you agree that data and information asset management is an area that should be under the CDO specifically metadata management over the years I've seen it handle the entire metadata management approach. So again I think that's a two layer question. Where does metadata management sit and how do you bridge that gap? Well let me take a crack at that one. I think you need a metadata subject area model because there's different classes of metadata. So business semantics I think is in the business actually and could be under the CDO quite comfortably. So that would be trying to understand business information that any thought as to how it's stored as data what are the terms we use? What are their definitions? What are the relationships? Semantic models that could then be used from which you could comparing it to the data used to generate abstraction layers to render the views of the data that you need. So I see that under the CDO. If it's more like data lineage then you typically need more tooling for that and that's more of an IT that's a little bit more in the IT realm I think. Although it can be under the CDO the problem with the technical data lineage is that it has no business understanding around it. So if you're saying well data is moving through this ETL job to this one to this one to this one. Well okay but who's responsible for checking data quality? What's the meaning of the transforms that occur? Why are they happening? Those when you get into that aspect of it then I think it comes back to the more business understanding of data and should more properly I think be at least influenced by the CDO if not having it underneath it. But you're right. I too have seen IT doesn't know how to do definitions. IT doesn't understand really well how to do races around data responsibilities. How to do a subject area, a subject matter expert matrix for data. So all of those I think mitigate are mitigated by having it under the CDO who as Kelly pointed out these days in the business and has the business perspective which they can bring to bear on it and I think get better results. Kelly you want to expand on that? You know what no I think Malcolm answered it perfectly. However I do think we have a webinar coming up that talks about information asset management. So Mark I think you asked that question. I think we've got one of those. Is that the day one? I think this conversation on EIM and data governance is around information asset management. Anyway I could be wrong but I think so. It could be past June 2. I know you just finished working on the rest of the year. Which is exciting. We've got a lot more questions coming in so we've got two minutes left. You know and again we've got a lot around metadata. So just to finish up the metadata conversation I agree that means for metadata should come from the business side. Oh and do we have a webinar breaking through the clay layer possible? That's certainly an interesting question and possibility. I don't know if we've built that into this here Kelly. I am rapidly looking at our abstract so I apologize. I put myself on mute to try and see if we do. Yeah you know what if not I'm just trying to think I agree Michelle that it's a great conversation. Let's consider it and we'll let you know. Sure. Well thank you Kelly and Malcolm for this great conversation and thank you so much to our attendees who are so engaged in everything we do. It's so nice to have a webinar like this because there are so many questions which come in every webinar. We just love the open discussion and conversation going on throughout the community. So thank you all for your participation today and Malcolm thank you for joining us today for our webinar. John's again on vacation but we'll see him next month. And next month as it's showing here we'll talk about compelling statement to corporate leaders why you must address EIM and data governance. I hope everyone can join us then and if you're going to be at CDO vision in a couple of weeks we hope you'll stop by and say hi to each of us. I hope everyone has a great day. Thank you so much. Thank you.