 Hello, everybody, and welcome to our next EDW session. We are in a session called The Most Common MDM Business Case Mistake and How to Avoid It, which is presented by Bill O'Kane, the VP and MDM strategist at Prophecy. All audience members are muted during these sessions, so please submit your questions in the Q&A window to the right of the screen. And our speaker will respond to as many questions as possible at the end of the talk. We also want to know that there is a linked form at the bottom of this page where audience members are able to leave evaluations for the session, and we'd really appreciate it if you were able to do that. So let's go ahead and get started. Thank you, and welcome, Bill. OK, thanks, Natalie, and thanks, everybody, for joining and for your time today. This is actually my favorite topic, The Most Common MDM Business Case Mistake and How to Avoid It. And for those that don't know, prior to joining Prophecy about a year and a half ago, I led Gartner's MDM and Data Governance Practice for eight years. I was the lead author and creator of the current version of the Magic Quadrant for MDM. Decided to make a change and decided to join Prophecy a couple of years ago as sort of a leading MDM vendor in the space. The reason I bring that up is just because the reason I'm so confident in saying what I'm about to say is that after about a half dozen pretty successful MDM implementations before that, at Prophecy, I had a documented 3,000 client interactions of 30 minutes or more over eight years. They actually tracked that, so that's how I know. And this particular topic took up about 80% of those interactions easily. And Prophecy still interacts with Gartner as an MDM vendor. And so we check with them pretty regularly. And they've told us that this proportion of this problem showing up or this issue showing up or this request showing up over the past couple of years since I've left is still consistent. So, you know, they're happy to have anybody talk about it inside or outside of Gartner at this point. So, again, this is why, how I formulated this, I didn't make a point of making slides about how it comes about, but it's a very, very common thing. You've probably heard it before, this is just sort of a really strong version of it. If you're already encountering this, you're gonna get a lot of confirmation and suggestions. They won't all be easy to implement, but certainly if you can sort of have the catharsis that this is meant to bring about at some point during it, during the session, then you'll certainly find this helpful. Just move forward, there we go. So, out of all those thousands of interactions and the half dozen programs I worked on, some sort of seemingly sensible trends and reasons for doing things like MDM and governance and data management, always seem to rear their heads over and over again. And they seem pretty logical. And we'll get into some specifics about what they are. I've seen really, really smart people, some of the smartest people I've ever met actually make these mistakes. And if they don't stay on top of the programs, they're in charge of, they actually revert back to those mistakes if they don't take an active role in steering people away from them. So why should you care about this? Well, there's some very good reasons. If we do just a few sort of non-intuitive, they're not rocket science, they're fairly simple, they're just sort of difficult to implement even though they are simple. And there are habits that you have to sort of... Habits and levels of fortitude that you have to develop in these programs, which can take a while as we all know. Why should you do them? They help you maximize delivery of business benefits. And the dirty secret is you end up doing less work. Okay, so this isn't all about getting everything done. In fact, that's one of the mistakes is trying to do everything. People use the cliche, boil the ocean. That's certainly something that comes up. And it comes up as a result more of what I'm gonna talk about. So if you can do what I talk about here, it'll naturally sort of lead to not boiling the ocean and delivering value very, very quickly. Something we're pretty fanatic about at Prophecy, which is one of the reasons I joined. Again, you avoid unnecessary effort and expense. In fact, you don't clean up data that's not really gonna add any value. And you'll be able to prove what you did. That's another thing that often happens is you do something very good. It's not out front because you're managing the data layer in the back. And you just never really get the sort of political capital that you need to keep moving in an effective way. So hopefully this helps with that as well. Some guy at Gartner was quoted over and over again as having said this years ago. And it's still true. That's a pretty good number still. By failing, when MDM fails, it fails. It's not like when a data warehouse fails and it's there and nobody decides to use it. They keep doing what they're doing. MDM, when it fails, doesn't get installed, doesn't get put into production, doesn't get fully modeled and populated. And when they fail, the biggest reasons are again because they've been justified by things that sound right and they're actually, they are technically correct, but they have a fatal flaw in them that we'll get to. And that flaw causes one of two things. And even this is not really intuitive. The first thing that happens, I like to say this, sometimes in certain situations, there are two bad things that happen when you go for funding for MDM. The second worst is actually you get kicked out of the room and you don't get funding. That's the second worst. Someone says, you know, we're talking to an ERP or a CRM vendor and they say if we buy it and just install it and get some data into it, our revenue is gonna go up 15%, our costs are gonna go down 25%. May or may not even be accurate. It may be accurate. It may not be, but you don't get paid. You don't get any money. And they say, come back next year. We'll talk about this again. We're not comfortable with it, that kind of thing. Keep talking to us. It sounds interesting, but we don't wanna do it right now. That's the second worst thing. The absolute worst thing is you get funding for one year and you justify it in the ways we're gonna talk about today and eventually someone's gonna say, what did I get for all that? And I had clients say things to me in the past like at Gartner as an analyst. You know, I got asked what business value I added and I heard myself respond with data architecture. And I knew I needed to call someone because I knew that wasn't gonna be very helpful to me or to anyone. Even though everything I was saying was correct, I could tell by the looks on the faces that I was kind of done for a while in terms of getting any additional funding to complete what I wanna do. And these things, I've seen them go 80, 85% complete and have this problem and literally get stopped at until another funding ground where somebody can give whoever this is some advice as to sort of what went wrong even though they did everything right. The idea is to have this discussion before, again, before you get any funding and before you spend it. I like to say this, you will have this discussion the one I'm about to describe. You will have it, it's just best to have it before you spend any money because again, it'll keep your scope down. It will ensure that you have buy-in. Not so much get you buy-in but show you whether you have it or not or you need to do more work because eventually if you don't have this discussion now a year later when you go back or maybe sooner someone will say, gee, what did I get for all that? What you wanna say is, what will we get for all this? What is it you want out of this? And there's some logic to not, people always say to me, how do I sell MDM? I would say more than half the organizations I dealt with shouldn't be selling MDM because they're business at all. I use the analogy, if you decided as an IT architect that you needed an enterprise service bus to stop doing a lot of point integration because they were counterproductive would you explain that to the business? How far would you go? And of course it's not very far. And I like to say, think about this you're doing that with data because you think everybody understands that. When in fact, the way we're talking about data in this context is just as infrastructure centric as that enterprise service bus I described. Somebody will see the direct results of it because they will see the cleaner, more trusted data when you're done but they don't really understand the idea of how you're trying to get there. And it may or may not be a mistake to try to bring them to that understanding. I like to say, the best MDM project I ever worked on was called new account opening for the 21st century. The business had no idea we were doing MDM all they knew was on that Monday morning when we went live, here's what was gonna happen. If they tried to enter a duplicate a pop-up window would come up and ask them if they really wanted to do it and make suggestions and things like that. That's what they cared about. They had no idea what master data management was and had I tried to explain it I'm sure I would have been fired. So again, just to reiterate the problem the failure to use a structured framework I believe that we actually did a survey of prophecy that these are quotes from forces ratification of the financial benefits and ratification is a critical thing that I'll get to as well. And failure to do this often leads to program failure. Our own surveys show that almost two thirds of MDM projects, no matter what vendor or if there's no vendor at all failed to go beyond piloting and experimentation and there are a number of reasons for that but one is, this looks very big no one's gonna pay for this over the long haul and the idea is the long haul is not really necessary. So common dysfunctional business outcomes I said people say things that are logical and even correct and they just don't work over the long haul, better data quality. So when I would get sessions with people while at Gartner I would ask them when they would say you know we're about to fail I can't get any more funding. I did all the things I said I would do my project plan looks complete for phase one something's wrong, I'm not at the top of the stack of prioritizations anymore. And I would always start the same way tell me what you said the benefits of this would be last year when you campaigned for funding in order of increasing frequency and popularity here the answers I would get they were always up the same one of the same four or two better data quality, better decisions single version of the truth this is usually where people start to feel like I'm getting personal with them because it's getting more frequent and finally 360 degree view of something and it's usually the customer sometimes it's product but usually it's the customer and these are all great things they're all things we should strive for but they don't make a good business case if you spend 500,000 or three quarters of a million or a million dollars or heaven forbid even more over a year and someone comes back to you and says what are we getting out of this again if you say any of these you're probably gonna have a tough week or month or year and I speak from experience and again talking to just thousands of people these are great things they are things to strive for absolutely I call them sometimes architectural goals but they are not business reasons to undertake data governance or MDM or even data management in general there's also a new one in the bunch monetize our data very forward thinking this is a great thing I've been, I was involved in a program to monetize data in 1995 that went pretty well from a data warehouse and so I certainly appreciate this I think it's great it's a great way to judge the value of what you're doing but again, if you don't have specific business use cases of who you're gonna sell it to and what you're gonna sell them and why they should buy it this is just as bad as the others better data quality, better decisions, 360 view single version of truth or even worse because nobody really knows what this is as much as they know what the other four are so if this is the only thing you're hanging your hat on and I'm gonna come up with an exception that's based on this in a little while again, you're just increasing your risk and again, I wanna say I would never say to anyone if you do X you're gonna fail we're talking about decreasing risk here and decreasing workload and decreasing risk by decreasing the workload that you have to model, have to manage have to implement in order to deliver value so just a quick saying here cliche means to an end is not the same thing as the end that's a brief way of saying all of what I'm about to say so how do you recognize a real business outcome? This is a great trick that I came up with years ago it trips up a lot of people that appear to be or that believe they're on the right track try to state why you're doing all of this or want to do all of this and don't use the word data now I said there'd be an exception for data monetization if you're in the business of monetizing data and selling data or trading data then this doesn't apply to you that you're gonna have business cases with the word data in them so it might be a little trickier for you to see to see one that may not be great but you can use the word data if so what do I mean when I say this? Things like this I want to hear sort of operational and analytical metrics that again don't really use the word data that it's a trick of the trade to sort of help you avoid the mistakes of the previous slide if you force yourself to try to do this it becomes it does a lot of other things for you in terms of mindset and the people you're dealing with on the business as well if you can't do this I hope this isn't too melodramatic we have a huge red flag and I say that from again talking to thousands and thousands of people about this who some said I wish I could go back because this was the beginning of the issue I actually I said earlier you're eventually gonna have this conversation where they're looking for these answers after you spend some money and for a year or two try to try to get something going based on those sort of ambiguous architectural concepts I had one client I'll never forget him he said to me once after a year I know you said to me last year to get them to give me what they wanted and to ask these questions they wouldn't give me an answer they threw budget money at me they said to get going and I got going and I did all the architectural stuff now they're asking me the exact same questions that I wanted to ask them last year that they wouldn't answer for me the exact same questions so again I usually stop here and say this seems melodramatic but over and over again for the last 10 years this is the thing that would have made the difference is if you couldn't and that doesn't mean you stop in your tracks dead if nobody can do this it means let's stop until we can come up with a bunch of these things because and usually two or three are plenty in the real world I've seen one be enough but I usually two or three are good we do a thing at Prophecy where we try to come up with two or three and prioritize them and get ratification and all that is called the business impact roadmap but that's the idea in the real world two or three is usually plenty for the accounting or financial horizon of most companies and what this means is if you can't do it you try to stop until you can have those conversations because like I said this is gonna get you less work because not only we're only gonna talk today about identifying these but this then informs whatever data model you're gonna use you should only model for the stuff that's gonna deliver these three things most MDM vendors now are multi-domain platforms Prophecy is as well they're model agnostic so they only have to model things from domains that are gonna make a difference for whatever outputs you want and they can always be grown sort of picture concentric circles growing outward as you get bigger and bigger but again I really wanna emphasize this is a good sort of milestone here if you really have trouble doing this if you can't have these conversations and I'm gonna try to help with how to have them in a couple of slides and how to measure them as well but again really think this through because it might be a year later it might be 18 months later but you're gonna have this issue you wanna be in front of it when it's still an idea and something that might be undertaken before so it's something that costs a lot of money and time and resources so this is one way to talk to the business I came up with this a few years ago and just never had a chance to really write a paper on it it is a quadrant the horizontal axis is what I call the financial axis everything below it is the classic below the line accounting stuff everything above it is the classic above the line accounting stuff so down below is saving money up above is making money the thing that's sort of new for me is the vertical axis and to the left are better things we do now and that we're really not that good at we're not accurate or efficient but we do it whatever it is maybe we use spreadsheets or pencil and paper or calculators but we do it and we get away with whatever it is that we're doing in the real world on the right or the interesting things those are things we can't do at all and maybe there are people that work for us that used to do these things at a competitor and wonder why we can't do them or they just assume our data is so terrible that we're never gonna be able to so why is this helpful so I will speak as a former worst offender at this so I'm allowed to say it most good IT people have some understanding of business problems to some degree or another the issue is and again I was one of the worst offenders I've ever seen at this early early in doing all of this is IT people left to their own devices will generally feel that the safest thing they can do is to save money by fixing current things that are not optimal because we already have them everybody knows what's wrong with them I know what's wrong with them I know what technology will fix them MDM is a big part of it so I'm just gonna do that that's perfectly fine and stuff that probably needs to be done and it should be done and it should be measured but they're not usually enough to make these things to push these things into positive financial or operational territory so what is so what is it what we wanna do is go sort of this upper right and that new processes that make us money that we're not making now the other quadrants are all important this just happens to be the one that usually where MDM and data management and data governance pay off where you can prove that something happened that was very good it wouldn't have happened if we didn't do all of this stuff behind the scenes and the issue really here so why doesn't everybody do this out of the gate to make a joke that's because people in IT have to meet new people to make this happen they have to go out find business partners take them to lunch find out what it is they really want or what they're unhappy about either because they don't have it hopefully or because they have it and they don't like it if they're really gonna stick to the left side of the vertical there so just a way to partition the thinking and that upper right is so critical you know the most difficult conversations I had I think in my career as an analyst and even before that as a MDM program manager and architect was my business partners aren't gonna talk to me like that they just wanna get to retirement they just want me to leave them alone actually I've had business people look across tables at me and say you're gonna be gone and I'm still gonna be here and they were pretty much right but I was able to do some things before I left that they probably weren't too happy about in terms of disruption if you will but that upper right is so important that if that's the conversation that's going on my response to that for people that were in trouble trying to get this stuff done was then go get some business consultants that have worked with your competitors and scare your business partners because we could never share it because of NDAs but I know people that are doing things that are gonna put you out of business if and when they're successful and if you don't do something you're gonna be in a world of hurt when they get to critical mass of what they're doing so I can only do you a service by saying I can't tell you because I'm under NDA but this is so critical that you should instead of spending money on trying to do all this stuff you'd be better off spending some money on someone telling you what other people are doing because you have no vision is the word I used to use you're all strategy but no vision and you need to get some because it's again very risky to pursue this any other way another trick if anybody's ever worked in HR or with HR or been measured by this there's this thing called the smart goals I won't go through all of them and some of them apply to a lesser extent but this is a way to think about your MDM initiatives and your data management initiatives what is it you know what is it you wanna do can you measure it better data quality there are ways of measuring it but they don't work in terms of business value usually they're quite technical is there a beginning a middle and an end MDM is a program it's a series of projects each of those projects has a beginning a middle or an end but the program probably doesn't the most successful people I've met that do this in the recent years will tell you I've still got five domains left to go but I've got an operating model everybody trusts me I'm gonna get funding I know I can deliver on the data I committed to I've got a whole risk mitigation program that's considered successful they don't think they're ever gonna be finished they know why they're doing what they're doing and that next thing that they wanna do they can tell you in business terms why they wanna do it and what the benefits are gonna be so again if you're not familiar with smart goals Google it it's a good thing to apply to things like data management just as a check a sanity check that says am I in the right place am I doing something that can't be measured how will I prove that this fixed anything or enabled something for the first time so people always ask how do I measure this the preferred way is obvious and we strive for this a prophecy in terms of helping people do this and we've been pretty successful I'm really kind of gratified by how many people step up to this in terms of prospects and clients for us but there are people that view this as risky if I put a dollar value to it and we don't save or make that much what's gonna happen to me kind of thing how will I look what will happen to my job I have to say that in the years I've been doing this I used to tell people take one third of it do you think you can you definitely deliver one third of what we're talking about or one half oh sure absolutely well then say that don't say all of it because I believe you're gonna deliver all of it and no matter if you just get out of bed in the morning you're gonna deliver half of it so again dollars is best however if you can't workflow metrics are effective so throughput rate of customer onboarding meantime to introducing a new products to market that kind of thing I'm happy with that if they can't be converted to dollars that's not optimal but certainly it will work and what's key there is consistency which is another great thing that if you use an ROI model like ours or somebody else's you'll get a yardstick that not only can you compare this to other things when you're trying to kick it off but when someone comes along after you're in business for 18 months and says I wanna buy a CRM system and I need the MDM money to do it and they say we're gonna make 15% more money if we do that and I pick on CRM and ERP because those are the ones I see the most but you have something to compare it to to say well wait a second we have something that everybody ratified that said the benefits of this were three times that are we giving up on that now and you wanna have a real discussion in effect what you're doing is replacing subjectivity I just know this will be great for the business with objectivity here's how great it'll be and here's how long it's gonna take and here's what it's gonna cost. Just a little sort of real life note there operational versus analytical MDM analytical MDM has real benefits and a lot of our prospects and clients start with that because it's less invasive to just populate a bunch of data warehouses in March and other analytic platforms than it is to send data back to operational systems at first that's often sort of the strategic goal in the second or third years to become operational to integrate real time with CRM and ERP prevent duplicates during entry and all of that good stuff and trigger business events like overextension of credit risk or fraud or whatever it might be those are easier to quantify of course they're harder to implement so this is the plan of there if anything that's really gonna help you it's gonna be harder to do you know analytical MDM it's harder to prove and I'm happy to have this discussion with anybody because I could be wrong but analytics the value of them unless they're operationalized which is sort of a hybrid here it's harder to say how valuable improving them is if you're already sort of getting the right answer and you're able to act on it getting it from 80% accurate to 100% I've had people tell me I'm just not interested in that but if you say you know I can stop you from extending credit risk where you shouldn't as soon as you implement this as provided you've got a credit reporting service that you already subscribed to or are willing to then all of a sudden they're interested of course again that takes longer to do but easier to quantify right so what do you do next? Where do you take all this? So go back to that benefits quadrant and there are a couple of other tools that people use and get a framework to capture, categorize and ratify I'm really happy I italicized that business benefits for strategic management programs so typical candidates are internal ROI methodologies and tools some companies I've worked for some very big enterprises they have things you have to use they're not all that friendly to data management but you have to do them I've used other tools like prophecies or Gartner has one as well that I worked on for a few years we used to use those as preprocessor sometimes to a client mandated ROI methodologies and tools because they just didn't have the kind of metrics generators that worked for data it was more of a throughput kind of thing which we could eventually talk to but we had to do sort of a data centric thing first consulting firms often do this and again software vendors like Prophecy we do it as well we use a Deloitte value map which is public domain by the way we're not stealing it from Deloitte I find it quite effective and again Gartner has a framework if you have a Gartner license that's very, very good as well and so I italicized the ratification step there because that's something that people enterprises often skip they generate some document that says how well this is gonna pay and they don't get anybody to sign off and say these are the right things to shoot at to aim for so what we want among other things is that sort of buffer against external events that these things provide so they not only result in less work but they give you that yardstick I refer to that if somebody comes along it doesn't have to be somebody that wants to take your money and buy a business application it could be a merger or an acquisition and someone says we're already doing this we're gonna shut you down well how much value is yours providing how much did your tools cost how far along are you how difficult is it you're paying a lot in implementation services compared to what we're paying who's got more data all of that kind of stuff and I've been through a few of those and they're made much easier when you can immediately show up with this thing that says a couple of these people have moved on internally or externally and we're buying a company or we're getting bought but here's the thinking and it still looks good to me what does everybody else think and at least you have a discussion a set of discussion points rather than we don't like you and we don't like this vendor so we're gonna move on and just absorb you which very often isn't the right thing to do so again you wanna replace that subjectivity with objectivity look we all did this work everybody agreed that this was it most of these people are still around not all of them and we should discuss this this is another sort of thing if you can do it if you can't you can't but if you can you should larger things like I get this all the time we're doing an ERP re-engineering or consolidation project should we do MDM now yes because you're gonna do most of the work anyway and then you're gonna let it decay if you don't put something in place to maintain the data quality I've written articles and articles about that about business apps don't get paid to manage data CX and CRM same thing if you're doing a CRM replacing one consolidating doing one for the first time if you've got some sort of one customer initiative or one enterprise initiative these things usually get funded pretty easily the idea there is instead of proving just the standalone value of what you're doing you have a little smaller scope there you can prove the value that you're gonna add in terms of not allowing further technical debt to accumulate after you go live with whatever the business app is ERP CRM something else and again you should have real numbers if you've actually done this work in whatever way you choose to do it again this is the ratification step get business and IT stakeholder agreement that these priorities and these measurements are the valid ones doesn't mean they can't be added to later which we'll talk about next but without broad based ratification the effectiveness of the framework is fundamentally compromised again if you get bought or somebody decides they need your funding if you lift this thing up what you don't want is everybody saying yeah I think I remember that but I never really believed it or I never really bought into it these vendors are taking me out to lunch and saying all I need to do is install it and get some data in there and we'll be doing great or I don't even have to put data in and I can just start using it and let people enter data manually as we move along as they adopt it you can short circuit those conversations and it's not about spinning and getting your way it's that's probably the wrong answer and you probably have the right one in terms of business value again you reflected in tons of conversations I've had over the years if we had just done the right thing three years ago we'd be so much better off now well there's a reason you didn't do it and it's probably gonna happen again unless we change our approach and again treat the completed framework as a living document again I've had people do this I alluded to this at the top of the session had people do this and then they're on the right track and you don't see them for 18 months and you get a chance to talk to them and they're in trouble again and you say what happened why are you doing this now oh 360 view of the customer wait a second we had three business use cases everybody agreed that that would be success what happened well we got entranced by integrating stuff on the IT side and it worked and we started the scope got bigger and bigger and now it's a 360 view thing and guess what we're not going to deliver on time so I need to go back to the well and there's not a single business use case that's been enabled by this thing the data looks better in the current ones and they were really the business were really psyched up about that in the beginning but now they're starting to kind of have the same problems and they're not too happy with us so again that can be avoided by going back every few months or if something happens that says I need to go look at this could be a management as simple as a management change if your executive sponsor changes you should certainly be going back and looking at whatever justification entities you put in place so I hope this was helpful we got a ton of questions I'm going to go down the list here somebody asked how to come up with actual numbers for the example so that's actually a great question and at least something I didn't say and I probably should have one of the biggest mistakes enterprises organizations make when they do this is baseline whatever these metrics are before you do anything to them to improve them that sounds obvious but I've had tons and tons of terrible things happen on my watch where people improved things and couldn't prove how much better everybody knew it was better but they just couldn't say you know it was worth spending this much money because we don't really have any documentation for how bad it was before so trap your onboarding rates over time even if you have to get acquire some tools to do it there's some very good tools out there now to do that operational if you decide those are your metrics operational productivity if you can quantify costs of you know I come from a lot of financial services non-compliance is a good way to say how bad your analytics are non-compliance with regulations so you know there are ways to come up with actual numbers we have a whole team of prophecy that does this I'm sure others have it too they're quite quite good at this and they work with with your business people and say you know why do you think that's a problem what do you think it's costing how can we get to a number how can we trap that how can we track it before we make the improvements which we know we can make and again we've had things come up where we find out the priorities are actually different than everybody always said they were we we worked at a large manufacturer who said you know my problem is really that my suppliers are duplicated and I can't really get any get the discounts I should be getting from them and it's really killing me and we went through a process with them and we discovered that yes that was a problem but they actually had a much bigger problem this was a big manufacturer with on-hand parts inventory they just had tons and tons of factories with parts sitting in them that they just didn't need to hold on to for that long and we literally changed the program sequence because of that so again you have to work with the business to get numbers you have to get baseline numbers it's it can be difficult there are people out there that can help we help there are consulting firms that help but they should be real and they should be agreed to again by the stakeholders that govern them someone said I think it's focusing on use cases for justifying MDM but does the total cost justify the limited use cases I'm going to go back to something on this slide what I should have said I hate to use buzzwords but that business IT collaboration up on the top right there I had somebody attend a webinar I did who interrupted me and said that's digital transformation that's what that is that's where it is that's what it is if you don't do MDM and I've done some writing on this it's out there somewhere either on our website or in some trades you're not going to do digital transformation if you don't do MDM unless you only have one CRM and it's completely clean and one ERP and it's completely clean and I'm pretty confident saying that nobody does based on what I've seen you can come pretty close if you're very small but if you do that successfully you're going to grow to a size where that won't be possible anymore and you know also down to there's another question similar how do we justify something that helps everyone but that doesn't help anyone buy a lot so again if it doesn't if it helps everyone it's helping them by some amount and again there's you know what if I do nothing there's there's the the old PMO PMP thing PMI thing of you know what if we don't what's the cost of doing nothing that's another good exercise I should actually put that on one of these slides that's you know there is a cost to that play out what's happening to your organization the trends that are going on on the business side over the next three years I've had two clients in my career who were able to prove for example that if they didn't consolidate their master data and leverage it as an asset and that's a big buzzword but it's the shortest way to say that if they didn't do that then every time they introduce a new product from six months on into the future it would cost them money no matter how much revenue that product generated because the fixed cost of adding that bit of technical debt in systems in whatever else they they had to do that dealt with data in analytics was going to be more than they would make because they just repeated infrastructure over and over and over again to support data and there's some of that was operational stuff too but most of it was data and I had two clients that were able to do that to say look we have to do this there's no more economy to be had and just continuing to expand our IT footprint in support of expanding our product set or our lines of business or whatever it might be so again it has to help somebody buy a lot or in the aggregate it should be a lot or I'm going to say something inflammatory you probably shouldn't do it then because it is it does take time but what I say you probably shouldn't do it in order to be provocative because what you should do is get help to do this because I've never seen it come out that it wasn't worth doing once you add up all the things now there's a difference between that and getting someone to pay for it and you know here's sort of the bonus answer I believe that data governance organizations that are usually virtual not always and not always totally but that are virtual should have their own budget to pay for the things like this that no one person is going to take out their wallet and see enough benefit to pay for if you limit yourself to only being able to pay that way and I understand why somebody would these things are going to be even more difficult to do for the reasons that were brought up in the question so my big answer to that is put in a data governance organization owned by the business supported by it that you know and I have templates for that that I use that are sort of logical models for organizations for to do governance that will actually have money to do this and they can charge back if it turns out it works out they can charge back if not they can eat it but again you know there are ways to do this it just again takes some shifting and I don't want to trivialize that uh how do companies measure the post MDM benefits again it's it's operational and analytical metrics analytical is harder you can do it so for example it's easier in financial services and maybe healthcare if you have regulatory requirements that you're better able to comply with and more efficiently able to comply with those are pretty easy to quantify because of the fines that you pay if you don't do that then I speak from experience on the financial services side there are many banks and other institutions that regularly pay fines as a cost of doing business when they really don't need to and you can pretty easily eliminate them with things like MDM so you know it is worth it to do it I will say that I've taken surveys in my career where the number of people that actually do it is very low I I think maybe there's they don't want to spend the money even if they're successful to go back improve it everybody knows quote unquote I saw somebody at a gaming company the best one I've ever seen is they actually did two time in motion studies pre and post implementation and were able to prove that every person sitting in their customer in their call center was you know like 50% more efficient per day those are very few and far between I think I saw one other one that sort of approached that but wasn't the same so again you know it should be done I've only seen a few do it I saw one do it unbelievably well and that was great but that's a very good question it's a good one it's kind of the dirty secret of the data management world almost no one goes back and I really think even when people are successful it's kind of a self-serving answer but the success is so evident that they don't really want to spend the money to prove it and you know that maybe that's okay it's probably unfortunate but if you are able to do it someone asks you know should we keep measuring and publicizing that I think again back to that governance organization and if you're trying to do MDM and you're not putting a governance organization together you're going to have other issues and some of them are going to be related to this but it should be there it should be related to the business in fact another sort of thing that's out of scope of this is while you're building this business case those stakeholders you're talking to are probably the same ones you're going to want to say I need somebody to be to run the governance council I need somebody to be data stewards and you know that it should be happening in the same discussion you know they're talking about what the benefits are going to be you want to have in some proximity sort of what you need back from them in terms of management and ongoing meeting and and prioritizing changes that might need to be made or additions that are going to want to be brought in once you see how well this works so again I believe I hate to say big governance but I like this idea that this organization does other stuff besides adjudicates fights between people arguing about what a customer is which they do need to do pretty early on but I really believe in this idea of they have money they prioritize stuff they don't pay for everything but they pay for things sort of like the question stated where no one no one area or person sees enough benefit to pay and so again take that concept away from you as well with you as well we have another question how do you justify data management for AI programs so I'll give you a long answer to this right before COVID hit I did a presentation interactive presentation for 60 CIOs in Phoenix Arizona and you know they were pretty much allowed to ask whatever they wanted and one of the things they want they were that they said they were the most afraid of was the inability to do root cause analysis and that was in so I'll tell you the way they said it to me you know this AI stuff and we do use some AI like prophecy we use AI to do matching we use some machine learning stuff we don't publicize it too much because we've been doing it for a long time and you know we we just don't think it's that cool but you know the other part of AI and data management is reversing the polarity you know how much why is why should we do DM if we're going to do AI well all 60 of these guys it was pretty much unanimous they said you know I want to say it exactly the way they said it if you're lucky you'll get recommendations from your AI engine or your models that are patently ridiculous and that you can tell are wrong if you're unlucky ladies and gentlemen you'll you'll get things that kind of look right but they're not and they end up being very costly and you could incur reputational risk or or any number of things they were they were pretty much they said this is the thing that keeps me up at night is we're low we're loading up data lakes or swamps or lake houses whatever you want to call them and then we're loading that into these AI models and we're acting on the stuff that's coming out and I have no idea if we're doing the right thing and you know I we have data scientists we ask them how to prove it to us and they always start with the same assumption that the data is trustworthy and I'm pretty sure it's not is what they were saying and you know the good news is sometimes it says something that's so crazy we stop other times we we sort of get a bad feeling and we stop even though it looks right the problem in both cases is it's it's very very very difficult to virtually impossible to do the kind of root cause analysis you do on regular sort of rules engines and data quality and integration rules so again it's going back to some sort of risk analysis that says you know what if we don't do anything and we just start doing this stuff what's happened to other people that might happen to us you know the problem with AI is and it's a little bit of a cop out on my part so I apologize the problem is it's still it's not new but it's it's broad based attempted usage is relatively new so the horror stories that people are willing to put out there and say don't let this happen to you are very very scarce but I can tell you just being around these 60 people is really stuck with me almost to to you know almost to a person in this crowd you know everybody there my response to them was ladies and gentlemen you're absolutely right and I don't have a good answer for you other than to do MDM I we just I don't think they're published yet but I just wrote a couple of magazine articles on this topic that say you know look you're playing with fire and you're not even going to know where the fire is or how to put it out the only way to effectively fight it is to stop this stuff upstream which is what MDM is designed to do conveniently for me but that's that's really the only way that's practical you know theoretically you could go up there and clean up everything but if you don't put in some sort of technology that keeps it that way it's immediately going to start to decay and you might be okay for two or three or four months but then that technical debt builds back up because again the things that are generating this data don't get paid to keep that data clean and you can put some things in like okay the company name has to be at least two or three bytes but that's not going to do it because that's not going to stop you from creating duplicates that kind of thing and anybody here that manages CRM and ERP knows what I'm talking about there's virtually nothing there and you know up until recently most of these vendors and again to their credit they didn't even pretend to take care of that stuff now they're starting they everybody says data is important so they're starting to say some things that sound sort of good but I haven't seen a lot of real execution and even if they do it it's only going to be within their universe regardless of what they say they still almost never do a lot of work on how to integrate external things with their data models for example because their data models work for them as well they should so again I can't offer a concrete suggestion other than that sort of zero-based risk analysis that says okay what if we don't do anything what do we think we can accomplish and what risks are involved there what's what could happen let's look and see what our anything happened and again the issue with that is there's not a lot of stuff out there like there is with data security you can find security breach documentation everywhere on the internet historically going back years you can't find that for AI yet because nobody really wants to put that out there I think eventually they're going to have to but they don't yet to be perfectly blunt and I wish I could be a little more helpful there but I can describe the problem for you and the fact that again I was with 60 CIOs before we went on lockdown who all said they had this problem and we're very very afraid of it so I think if you can make the argument that look the only way to fix this is to fix it up front so that we trust it 100% so that that's not part of the equation if we see something funny downstream we can assume it's a problem with the AI models whether it's ML or something else and work on it from that standpoint because if if it's could still be the data figuring out what data it is and what what what was wrong with it or anomalous about it that caused this is almost impossible and again I'm going by with 60 people with boots on the ground told me all at once in unison pretty much so I hope that's helpful kind of a downer but again it's a legitimate point that's the best way to do it today I hope it gets easier as people are more open with things that have gone wrong but I haven't seen a lot of that unfortunately about one minute left any final questions or thoughts stump the panel or they're trying to hear me talk which could be all right then thank you so much Phil for that wonderful presentation oh thank you guys and thanks for the great great questions really good yeah everyone is encouraged to continue networking and chatting within the spot me app as well when this video is posted probably by the end of the day the q&a will be live so we can go in and keep chatting if you want to do that I just wanted to remind everyone that our virtual exhibition booths are open until 1 30 p.m. Pacific today and tomorrow and we look forward to seeing you in the next session thanks so much thanks everyone