 Hello and welcome my name is Shannon Kemp and I'm the Chief Digital Manager for Data Diversity. We want to thank you for joining the latest in the monthly webinar series, Data Architecture Strategies with Donna Burbank. Today Donna will be talking about data governance and data architecture, alignment and synergies. 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 and we very much encourage you to chat with us and with each other throughout the webinar to do so just click the chat icon in the bottom middle of your screen to activate that feature and for questions we will be collecting them via the Q&A section in the bottom right hand corner of your screen or if you'd like to tweet we encourage you to share highlights or questions via Twitter using hashtag DA Strategies. And if you'd like to continue the networking and conversation after the webinar and learn more about Donna just go to community.dativersity.net. As always we will send a follow-up email within two business days containing links to the recording of the session and additional information requested throughout the webinar. Now let me introduce to you our speaker for today Donna Burbank. Donna is a recognized industry expert in information management with over 20 years of experience helping organizations enrich their business opportunities through data and information. She currently is the managing director of Global Data Strategy Limited where she assists organizations around the globe in driving value from their data. She has worked with dozens of Fortune 500 companies worldwide in the Americas, Europe, Asia and Africa and speaks regularly at industry conferences and with that let me give the floor to Donna to get today's webinar started. Hello and welcome. Hi Shannon, thank you always a pleasure to join you folks and thank you for a lot of you who on the call who often come to these as a monthly event and if this is your first time with Dataversity or this series it is a series and all of the common question is is this going to be recorded can I catch it again send it to my friend but yes everything is recorded and is stored in perpetuity on Dataversity so if you missed any of the previous ones that you see there on the list and Master Data Management was last month you can catch them and this one will be on demand as well. We hope you can also join us upcoming you'll see there's one each month and hopefully some of those topics are of interest to you. So today's topic is data governance and data governance is complicated it can be very valuable but partly why it can be complicated is that it's very nuanced and it consists of a lot of things and often I you know full disclosure I'm a consultant in my day job and often what can be some of the complexity and doing data governance is a misunderstanding of what people are talking about. Well some people they're talking about data governance at a very technical level in terms of data architecture and naming standards and business rules other people are talking more about committees and data stewardship roles and they're all right as what's complicated in the world right the and condition right there's not nothing is only one thing so we'll talk about both aspects of that because I think when data governance things and as well both of those can work in harmony so that is the topic of today if you've joined these sessions in the past you'll be familiar with our framework this is what we use in global data strategy in our practice and I think it sort of touches on the holistic nature of all of this so what we spent a lot of our time doing by our name you may guess it is a lot of strategy right and even when someone comes in brand new and I want to do a strategy yes obviously we need to start at the business layer and see you know sort of the why well why are we doing why are we here but then pretty quickly after that it ties into the what so what sort of data is being stored and you have to look at every piece of this and a key part of this always is data architecture and then I like to call it almost the blue that the reason why it's across all of these is it really is that layer between the business and tech so you'll see that it's the people process policy and very importantly we'll talk more about that the culture I've been doing this for a long time probably we're 25 years now not specifically governance but data management and the hardest thing and I think the most nuanced thing of now when I go into a company is to really look at that culture that's often what can make or break data governance you don't want to get that wrong you know a military organization is different than a startup is different than a nonprofit is different than a hospital and the culture of those can be very different and you need to look at that carefully so we'll talk more about that and then you know if you're in data management you're familiar with a lot of these these are still all tied together because you can't look at data governance or data without data quality right we can't be looking at that without master data so these are all tied together security obviously the big part but we'll specifically be talking about data governance and the architecture component of that so since we're data folks we like facts and figures this is some figures from a survey we do yearly now this is last year survey on trends and data management for those of you who just know who might have done the survey for this year's it is in progress so hopefully by later the summer or early autumn you'll see the results of the 2020 version but this is last year's which is still fairly covered current and I thought this was great news and probably not so surprising everybody and their sister now is thinking of being a data-driven business or data-driven organization it isn't always just a for-profit company could be nonprofits could be universities right over 76% have data governance in place I think is great and over 50% now just be careful how we read this not over 50% who have a data architecture which I'm sure is also true this is 50% that specifically said they're able to collaborate better because of their defined data architecture which I thought was a great finding not surprising to me I do a lot of that I think a conceptual data model for example is an excellent way I worked with two clients just this morning on a conceptual data model and two very different industries where they were using that as really the roadmap for their entire data-driven initiative it's a nice way to communicate that's not the only architecture artifact we'll talk about a bunch more but that's not a surprise to me and hopefully if it is a surprise to you by the end of this session you'll instead of understand more of how you can use architecture to really be a communication tool as much as a governance tool because I see those as linked very closely together which leads me to oops sorry Shannon can't say your own name I can't move my own slides to the next slide which really is that that balance right so if you look at what do we even mean by data governance I took this from the data management body of knowledge which is a great resource if you're not familiar with it in the inbox we would like to call it talks about governance as that exercise of authority control and shared decision-making which I see is almost the opposite right you have authority which is sort of you're maybe wondering why there's the yin and yang at the bottom my hippie side comes out so if you think of the yin and yang which is sort of the Eastern philosophy of that the two sides to every coin which might be a if hopefully I don't mangle this definition for folks right but on one side you do need to be very strong and precise and the other side you might want to be a little more flexible and yielding so that's very similar to data governance right I think a lot of us when we're thinking of data governance we think of the stick that authority and that control but just as much or perhaps especially if I get older and more experienced and learn a little bit about the world often it's that carrot that really gets things going or that shared decision-making but you just can't tell people what to do I think a lot of us in data management who are very sort of fact and figures oriented and well we all are it would be a lot easier we could just have people follow the rules but people are people and often those rules may not be the right ones if you really don't get the holistic view in that shared decision-making we'll talk a lot about that throughout this session it's almost the theme of the whole session if you have to get that balance right that's the whole point of yin and yang is that it's a balance we all have both sides of those in us and when you are in harmony you have both of those right and I think that really applies to data governance as well you really need to balance both of those so if we look again where data management folks we love definitions data architecture what do we mean by that so again reference the data the embark and I like their definition as well and we'll talk about this is representing the organizational data at different levels of abstraction so they're going to be understood and that to me is key to the communication and don't get that one wrong right so you're talking to a business person a conceptual data model a business process model great ways to communicate physical data model probably not so much right you're talking to a database administrator yes very much is a physical data model or DDL or data definitions so just make sure you get that right and I've seen again hopefully I mean what a lot of folks have shared with me and I always appreciate that after the webinar people sort of say what was helpful and a lot of people say what's helpful but these is kind of that real-world experience and I think I'll try to share those throughout often when I do see these maybe go in a wrong direction it's maybe picking the wrong tool for the right job of you know are we are we picking a two-technical tool or on the other sense I've seen people buy a very business-friendly tool that looks great and demo as well but isn't technical enough to really do the governance you need so think of that what level abstraction are you going for because these artifacts are critical to governance they're describing the existing state talking with a future state getting your requirements the integration all of those things at the bottom that really define that data strategy that I mentioned in that framework above so all of those things are things to consider definitions are can be dry and boring I'm going to share this theme might be dated but I know for a while around the Internet there was this sort of meme that went around with you know what my friends think I do what my mom thinks I do and I think that applies to governance right we have to be careful about this so what my friends think I do when I say I do data management and data governance probably not the life of the party you know Donna's that nerdy person that sits on a computer a lot well they're kind of not wrong right but probably not the most flattering definition of what I do because I think what I do is cool what my mom and my parents think I do they do know the word metadata that they were very impressed that they could they remember the word metadata but I think they think I'm some sort of librarian type of information management kind of like library stuff right what society thinks I do you're one of those data scientists right you're this mad data scientist that again sit in your laptop a lot and look kind of nerdy well not quite wrong but I'm not a data scientist what my co-workers think I do now be careful on this one are you the one that just yells at them about stuff and you're going to come in the government is going to have an image problem I think I actually had a presentation on that it's just the word governance you know we'll say it but you know Harvard Business Review article that data science was the sexiest job the 21st century they kind of didn't say data governance right that governance just by definition seems to be what you're telling people what to do so be careful of that people are already going to have that perception so think of that shared decision-making slide the yin and yang don't be you know to yang and just yelling at people it's just too easy to do my favorite one but I think I do I'm the governator I'm saving the world from mismanaged data I actually sort of stole this from just a sound check Shannon you can hear right we can Shannon yeah took me a minute to unmute myself yeah you're fine wanted double checks them went on the checks and they couldn't thank you folks um so the governor actually stole from a client is one of our clients outside of London and she was actually looking for a head of data governance and she says I want to put it as the governator because that's a sexy title and we said you know I don't think you're gonna get past that by management that was your internal view of the role of the head of data governance because you sort of are the governator right I mean you are saving the world from mismanaged data you often feel like that and you really have to work across the organization to really save the world without people knowing right the you're the Clark Kent really during the day but at night you know you're the governor tour but really what you actually do in all full seriousness what maybe people forget is that data governance is really driving the success of the business and I think more and more people are getting that so I love what I do I think it's really fun because I get to work with so many different companies and what I find interesting is I would say five of our new data governance projects this year were either driven by AI machine learning business you know the new emerging tech wanting to be more data-driven and we have two startup companies that basically at the very beginning of their startup tech startups wanted to get governance right which I think is awesome because I think so many people think of governance as being on the left you know data management is kind of the old school you know boring stuff and I'm not seeing that anymore it's not that I have rose colored glasses I'm actually you know seeing people getting that you can't do governance without you can't have a successful business without governance sort of like I always think of governance a lot like finance you know most people even a startup realize you really can't be successful without a finance and financial governance data is the same thing so anyway kind of a facetious slide but hopefully thought you thinking I've got the point across a little bit another way to look at data governance and I've used this with some of my clients as well it's kind of the so what and the sort of two aspects of that you're going to reduce risk or increase opportunities maybe the more simple way of looking at it often that's kind of called the defense or offense model so and again maybe partly because you know the history of data governance or data management too many of us often come in on that too much on the reducing risk well you're going to go to jail if you don't do governance well are they really going to go to jail and if they are maybe that's not the thing you want to lead with or is it more like I was saying so many of our customers now are doing governance because of the offense data is strategic of course I want to manage it I want to be a data driven organization I want to be data centric of course I need good quality data so there's both and most every company has two aspects of that yes there's GDPR yes there's PCI and you want to make sure you get that balance right and I'll talk more about that as well you don't want to go into a health care company that's really trying to protect patient data and say gosh there's so much opportunity you make so much money off these patients which may be true it's just little tone deaf and at the same time you don't want to go to a startup that's all talking about making all this money saying well guys you might go to jail if you don't get your PCI and that's true too but probably not what that audience wants to hear right so just kind of think of that as you go through the other one that I think is not stressed enough and I again I'm confident I think people are starting to get that is the one in the lower left the improved proving efficiency you know what your parents always said if you don't have time to do it right do you have time to do it again kind of famous thing you know a lot of folks think oh gosh why don't we just skip that complicated data management stuff and the data governance stuff and let's just do quick wins and and fail fast and again you will fail fast they can guarantee that if you don't have any governance because really when things are organized and done well things just move more quickly so you know there's a lot of famous statistics out there of that sexiest data scientist that's hired with a very high salary to do all those great analysis and they spend 80% of their time or more cleaning up the data maybe because the data isn't good in fact one of the clients just this morning we were on a data governance call using a data model and their biggest driver was that very expensive data scientist hired to do data analysis and they basically said you've got good data but you don't have good data quality you've a lot of interesting stuff you need to get data governance in place and that launched the entire initiative so good to think of doing it right is more efficient and then that collaboration and accountability and I put those together again that yin and yang so yes you can tell people what to do yes you have a responsibility but maybe I'm naive but I'm usually not proven wrong on this one that I think most people in companies or adults and very few people come to work and say I want to you know mess up the company's data if that's the case get rid of them there's just no question there but it's usually because people don't see the big picture or I didn't realize or I'm using this data in this way didn't know the downstream effect so that's why getting in a more collaborative way and having these committees that are cross-functional a lot of folks don't like the word committee but gosh they are so much more efficient with some of this just look at all of the impact before you build and get sales and marketing and development in the same room and understand before you build can go a long way so hopefully that's kind of a helpful framework to sort of thinking of that and as I hopefully appointed already it really is both the stick and the carrot and give that some thought so a lot of it we think of governance is avoiding risk that's true but governance can also be an opportunity driver you know think of that as getting that collaboration together and getting your data right so that you can you can really be more effective again hopefully in these webinars I can give you some kind of that real-world experience and gotchas that I've been gotcha with or someone who looked like me I mean I've never made a mistake but other people with a similar name to me might have so this one what style of governance fits your organization I think the hardest part of data governance and I have been super techie I've written data governance tools and I've written the code myself I know how hard it is but even with how hard data lineage and all of that is the people side at least for me I think for everybody it's hard getting that culture right is probably the biggest nut to crack so think of that what is the style for your organization that you're working with or is it more that offense the people want to hear about profitability revenue customer satisfaction competitive advantage that is what's driving governance or is it more about we want to be compliant we want to avoid audits and fines we've had a fraud we've been audited we want to be secure and it's probably not all or either one so but just think of that in terms of a spectrum and maybe give some pondering time right now are you a full-on red your startup and they don't want to hear anything negative about compliance they just you have to sneak it in or is it completely defensive and we're just trying to not be sued we're just trying to protect our patients whatever that is or more likely it's somewhere in that purple spectrum but just give that some thought and that's where you can go either really well with your especially if you're trying to present up to management think of what they're worried about did they just get audited and you're going to go talk about you know increasing revenue and profitability of course everyone cares about that they probably worried about the audit or you know are they talking about their great new marketing launch and you're talking about being audited and you seem like a buzzkill just give that with some thoughts and get that one right on on a similar note I would say the same thing with architecture you may have seen these lines before I like this one it's just a good level fat of where you are in your organization because neither one of these is right just like totally offense or totally defense is not right are you so academic that and I've seen both of these all of the above that you spend so much time getting every data model right and full data lineage right that it takes two years to even launch a product I I would have seen I see less and less of that now although I shoot lives we're working with now that really wanted to inventory every single data element in the entire organization before even starting governance and I tried to convince them not to do that because I think that is too academic and you're not going to be agile enough to keep up the business that's just too much to you know and so that's on one side are we so academic that we seem like the thinker and we're so set in stone that we never get anywhere or is it ah let's skip that whole architecture stuff we're too busy for that let's just get stuff done we're going to move fast and as you may know or have experienced that's almost even worse because nothing gets done either because it's just chaos you keep redoing the same thing so the right balance between that there is no one perfect answer and there may be a spectrum depending on the project or the stage you are in your organization the right answer is that business value what is going to give the most business value are you developing a heart monitor that's going to save the life of a patient and there's a data IOT version of that yes please be a little more on the academic side because you know that's my heart I'd like you to get that right right or are you doing just a startup proof of concept that's going to be thrown away and you're you know maybe you don't need everything as academic right so just give that some thought and there's no right answer it's really dependent on that business value but give that both of those some thought are you a Wild West type company you don't want to be too academic or are you too academic and maybe need to loosen up a little bit so you may have seen that this kind of two aspects of that there's a technical side there's the business side which can be good or bad the good part of that is there's something something in it for everybody with data governance some people again think of data governance as the committees and the stewardship and the collaboration with the business and selling up to management and down to folks in operations absolutely a great part of governance kind of the touchy-feely side perfect place for you some folks really get very excited about getting data lineage right and setting data standards and doing data inventory and creating data models and creating the database the techy nerdy side great there's a place for you if you're a smaller organization you have to do it all and you're one of those unicorns that can do all of those well and if you are that's great because I think the perfect data governance person can sort of do both of those or at least understand enough of both of those to really make that thing but that's also where there's some confusion sometimes when one person is talking about the committee and one person is talking about data lineage and they're kind of two sides of what data governance is we often use this framework for data governance we've touched on some of that and really you need to look holistically around governance to make that work so we've talked a little bit already about the aligning everything with the business goal or objective doing data governance just because it's a good idea it's not the bet you know there's a lot of things we could do because they're a good idea and there's limited time and budget so focus on something that's going to be of value to the business and really focus on their issues challenges or again opportunities if they're more of a you know the carrot type of organization and then look holistically across all that what is the vision for governance and I we spend a lot of time on this where we do projects even create a marketing plan for your data governance what's the vision what's the strategy why why do people want to do this which ties into your business goals and then how do you set up the organization and people will talk about that you created stewards data owners data custodians you call them none of that you have committees do how do you how you do that is a big part what are the processes and workflows and that's kind of meta level there there's the data governance process and workflows how do you log an issue how do you correct deficiencies how do you publish standards those are processes and workflows that are important but there's also the business process and workflows you know how does the business manage and touch data because that's often where the problems go wrong I'm entering data as part of a sales cycle or I'm onboarding a no provider onto my hospital did I check the credentials you know all of that type of thing that's true governance often because that's really right at the coalface and then I can't manage what you can't measure so what are the KPIs and measures you have in place are we looking at quality are we looking complete in this are we looking for linears which of those are the most important because again you can't manage everything so you know how are we going to manage it and then if I haven't beat this over the head enough this idea of the culture and communication how do you communicate this vision and strategy and almost like I did a stint in marketing for a while and the techie side of me cringed at that and the rest of my body decided that was the one of the best experiences I had because everything is marketing so I almost see this as a launch plan right what what are the objectives what's our vision strategy what's going to get people bought in how do you organize it and then how do you continually communicate and communicate and then tell people again why this is great and highlight your successes along the way so the tools and technology I don't want to mislead that it's most important to this at the bottom it is a foundation but I see too many people start with the tools but yes you need tools plural probably there is no one data governance tool I'll get into that and really the data to manage all that so but this needs to be holistically across the board another kind of matrix you put together and don't worry I won't read every cell here but hopefully if you these slides are available after the webinar just to kind of give you if for each one of those buckets you know what what do we mean by vision and strategy what do we mean by organization and people so yes it's your organization structures but also give some thought about who's the key stakeholders who's producing the data who's modifying it you know just kind of some ways to look across all of these questions to make sure you've kind of covered all the bases because that's often one of the risks of governance you don't want to boil the ocean and do everything but often you didn't talk to the right group and there was a key area or key business process that was left out so looking holistically without going too far into the weeds it's kind of that again that you know balance to get right so when we talk about the people in the process of the organization again referring to that trends and data management report I thought this was interesting so this was data management not specifically data governance but they are related and data governance had come out strongly in this so when we asked who is driving data management in an organization it was multiple choice but you'll see it was a fairly wide range so a few things were I thought heartening to C level titles the CEO was driving data management and again this was multiple choice I don't think the CEO is setting data standards in the database be really surprised at that one but the fact that they're championing it and supporting it awesome that's a great thing to see that that many CEO really part of this but then you'll see across the board business stakeholders again I don't think they would be creating the data lineage but they are completely bought in the value it should be data stewards and data owners as well as some that probably are supplies data architect I would hope would be part of data management right but the the breadth of that and the number of roles I think is spot on because I think to get gate governance right you need all of that and then the top one other that you see was pretty high their right in vote that we probably should listen to be clear on that one was the data governance lead and I would say that's true because data governance lead is that sort of again that unicorn person that can motivate and champion and really drive this initiative for those of you are data governance lead I often see that a data governance lead can be a nice pathway up to Chief Data Officer because you're really understanding all of the aspects of the business you're coordinating or championing and you're doing and you're really you know helping data be part of a data-driven business so but that was sort of interesting when we get to the data governance lead but you'll see there's a lot of other roles particularly sort of business stakeholders that typically in data governance land we call everyone has their favorite term but a common one or kind of your data owners or data stewards or data stewardship is probably the more generic term that encompasses all of those that's important and I think especially when people are starting that can seem daunting seriously we're going to all these business people are really going to want to talk about business definitions and do this with us us being the IT people and generally when we come in as consultants and maybe do a strategy or do a data governance initiative we start with a lot of interviews which might be weird formally if you work for the company but doesn't mean you can't do that informally talk to a lot of different people across the organization and what I I would almost guarantee I'm not a betting gal but I would bet on this one you will find people who are passionate about this stuff and we call them the hidden hero who maybe had never been asked about their data quality and probably I have seen across the board I've seen people build their own data models people who have done their own data quality remediation and profiling and are often very pleased to finally be given a voice to actually have some responsibility yes but be part of the story so I wouldn't don't be surprised that there are people on your side they get this and often it's the business to get governance much more than IT because they're feeling the pain point every day they're seeing what bad data can do so don't be surprised that said be careful there so it isn't just it's great that you have much people that are excited about this but you do want to have some sort of structure around that again where sometimes I see data governance having a hiccup or a misstep is oh great I know Mary John and Michael and we've been talking about this forever where we're going to be the community and we're just going to start because we've been doing this stuff give it some thought how are you organizing it by business area by subject area you have the right levels is their direct level manager level you know a mix of tech and IT and don't do it based on Joe likes the stuff and so do I so we're going to lead it really look more you know or in an organized way here's one way and again I this is not prescriptive at all every government organization is different yes there are excellent tools out there the data management body of knowledge is probably the best first place to start it really goes through the basics of what a governance committee is what you know a data owner is etc but then just use that as a guide it is not a how-to manual it's a guy right and so the most important thing is to customize this to match your organization and your culture but here is a sort of common one that you know we tend to see generally and I would say everywhere there is is and should be a executive sponsor an executive sponsors you know the whole I would assume the whole executive council would be bought in but there's one person that really is the champion for this when the rubber hits the road and this will always happen everyone's bought into governance until they need to change we're all human beings oh that's no one else is going to do the data quality mean I have to change my business process yes and that's often the executive sponsor they can help aid champion it say hey guys we're doing stuff differently this is awesome you want to be a part of it and also when those hard decisions are made they're going to back you up guys yes we do we can't just do it the old way it didn't work before it's not going to work now and then in terms of the business side you have the business data owner that I would say that the difference between that and a business data steward and again names names everyone has a different but kind of a common way to look at this if the owner might be more of the strategic or the manager or director level that sets the direction understands the high level business rules and then the steward is someone more at the operational level that really probably understands the pain points day to day maybe we were more subject matter expert that's entering the data using the data etc and both are super important but you don't want to get them someone too high level and asking them again we've seen this all go wrong I've had seemed VP level people being asked to do kind of master data match rule brew mediation all day long and funny they never wanted to do that again or you have people too low level being asked strategic questions and both of those are wrong you want to get the right role for the right responsibility and then a technical data steward separate those a system is not the same as the business area and often that's an easy kind of mistake to make and maybe you do that as the first step oh that's our CRM data well CRM is the system are you talking about your customer data that's different and so often an organization or it'll be something oh yeah the x29 system you know oh yeah we know that's our product database well that's not your product that's a system and systems change data remain right you'll have products hopefully the whole whole history of your company if you're a product company but your systems will change so getting separating your data from the system super important that said someone who understands those systems is pretty critical because that's where the data often lives so data governance lead again some asterisk in a perfect world that's a full time role if you really want to make this critical that person would be able to spend a full time really championing understanding setting up running the meetings all of that isn't always feasible that would be ideal data architects similarly or plural but generally they would be a lead data architect that's almost the corresponding party on the technical if you think of data governance is probably more of a business type role or someone who understands technology but it's more in the business side data architect is probably that the opposite who's more technical but can speak business and they can take those requirements and create data models and things like that and then data security obviously is very important is a different thing than governance often this kind of confused between them but should be part of this group as well because those rules and regulations are key a couple of comments coming through ELT is often acronym for executive leadership team every company seems to have their own acronym that's kind of a common one we love our acronyms but probably you're right with governance we should have a glossary to find that so my bad so that's one way again please don't just take this and say yep that's how we're going to do it because that may not fit you that may not fit your organization but just kind of some thoughts what's so important if you go away with one thing this is a good one to remember is when you create your governance structure map with your organizational capability and your organizational structure are you a top-down organization is it more federated distributed is it rules-based that people hate rules that people like committees are not like committees so much of that is what can make or break this because you want to make people want to be part of this data governance in a way that makes sense for them so define the organizational capabilities that's also going to help you to find who your data owners are again have someone from finance or someone from product development not someone from a system and there's different ways to do ownership that's just one but think of that and then how is your your organization organized so they want a little that's one example on the right of a data governance organizational structure which is pretty top-down you have your executive strategic tactical operational a lot of them sort of turn out that way but absolutely that is not the only way to do it for example this next slide is one that was interesting this is a anonymized version of a governance structure we put together for ironically it was a manufacturing company a big auto manufacturer and I would have thought that that would have been a very top-down we think of manufacturing I think very process oriented very kind of rules and regulation we need to get things done the culture there was very different and we should have started with a structure that was more sort of top-down like the one you see in the right you have a committee and even though yes we had committees their culture they just they did not like boxes they didn't like hierarchy and they really sort of said no we're more of a distributed overlapping circles kind of people they wanted to see things flowing again this was an engineering company which really surprised me like we're at a nonprofit that said that or something or a you know counseling company that would have made more sense but this was an engineering company but they said we want pastel colors and we want overlapping circles and we do things as teams and also they want they were using agile and they said we need to do things fast so the idea was we would have these overlapping councils where it was more focused on data innovation in collaboration but then we also built things quickly so it doesn't mean collaboration slows things down again I think it speeds things up so but I thought this was another helpful one another one that's actually just came up this morning so I'll share it with you I thought it was why often we show this idea of this work groups right so I've got my data governance committee and then maybe we create work groups to solve a particular issue and some we often kind of show it this way as a result of the steering committee we have a work group this particular client said you know that makes people feel like they're at the bottom and I hadn't really thought of that how you visualize these things so we should have showed along the top because really these working groups are making very strategic decisions and giving recommendations to everybody and they weren't at the bottom and that's not what we were trying to show at all but it kind of looked that way so again how you show this and does it fit with the culture is actually super critical you want to get people psyched to come to these meetings and they are so if you're skeptical rethink of your model because most of the time once you get these right they are so for example what you call it you'll see here that this was called the data innovation council and they had data innovation teams was governance a big part of that yep but they were also very agile very forward thinking company and they they didn't want to make it seem like you know we're just going to tell you that person with megaphone telling you what to do we actually what this is going to be driving the business and we want to innovate and move in an agile way so just even what you name it do the insurance company we work with up in Canada and this was like such a dramatic version of that we had a data governance council and we're really having trouble getting people to be excited we called it the data innovation team and people were kind of fighting to be on it again that was literally just a naming which we could have because we actually were doing a lot of innovation but it was just kind of how you look at it so again all of this is sort of meta there is a organization structure for your governance there's also an organization structure of the organization that you need to align with for your governance so again next month's webinar is on enterprise architecture so I don't want to go too deep into these but a lot of enterprise architecture tools are great for governance so this is kind of a business capability model which is an EA enterprise architecture tool but what we often do is kind of overlay data domains onto that where is customer data used across the organization well that one probably across a lot of places marketing and sales and product development and human you know etc etc have a nice again a lot of these high level overlays I mentioned there's a lot of ways to do stewardship and I'll fight you outside the of the hallway I often get an argument especially a lot of the conferences about this almost the classic academic way to do data stewardship is I've seen too many people say well I have a steward for customer and a steward for product and a steward for you know employee I do not like that and so I'm the minority there but I'm right because I'm me I'm right but here's the reason why I'm right who the customer nobody owns a customer right maybe yes maybe in the data architecture there's someone that should be managing the customer database or even the customer data model but there's a whole customer lifecycle and journey and often what goes wrong is that there's one person in a customer for someone who owns the marketing side of customer it's not the same person who owns the sales side who owns the customer service side data is used differently and there's different types of data maybe the easiest way to think of that is who owns patient right there's someone that when you go into the doctor and they ask for your address and your insurance and your medical history it's not the same person talking about your diagnosis which is your doctor and it might be easy to say oh that person at the front desk owns all your customer data I do not want the receptionist owning my diagnosis right so give that some thought so anyway aligned to your business organization another thing to think of again so much of what makes governance thing is that touchy-feely nuance and getting that just enough data governance and this is where it also starts to go into the technical side so this is touchy-feely meets technical I've seen it go wrong in both ways you certain things in the organization must be highly governed patient data for customer data that's probably your master data that yes it's highly shared it's highly governed and it should be very much locked down other things like you have a sandbox for some data exploration on social media analysis shouldn't be you're gonna slow people down too much by having it overly governed yes you shouldn't put credit card information out on that but other than that with the right guard rails in place let people be and it's a progression so yes core enterprise data maybe that's your warehouse should be governed but probably not as closely as master data which should be very carefully governed functional and operational data can be a lot of things it could be your operational systems it could be a department data mark again should be governed but maybe the departments themselves can govern it maybe doesn't have to be enterprise wide and it can be what we're also showing up evolution so maybe things that started with exploratory we well we found a variable that's actually super important did you know that I don't know hobby is really important in an insurance company we want to start tracking that actually is master data because it affects your rates okay well now we're now that should be meant you know up there in the master data level so it's a and if you have master data I'm sure people doing data science will love to see it so it isn't an either or a progression but you want to give careful thought to that one another thing to think about is getting that right balance and I see this sort of the difference maybe this is hard to read whether it's human or whether it's automated and technical and whether it's sort of reactive or proactive and whether you resolve it kind of the source system or post-processing so kind of a lot to put on one slide maybe doesn't work but proactive on the business side would be you know if bad customer data is going in maybe we should train the people putting the data in to do it right and let's look at the systems we have the right policies and procedures I'll give a case study at the end where it was the sales team putting in the data and they didn't know where the data went they didn't care once they found out that the data fed the leaves they got to sell the customers they were a lot more careful putting the data in probably the best data governance we could have done get people to care and fix it at its source right if you don't do that then you need to do more reactive which maybe the data quality cleanup it's not an either or right but the more you do up front you don't have to have a data cleanup and keep cleaning up cleaning up and then somewhere in the middle is your data stewardship right because this is an evolution of yes this data stewardship that looks at what we need to clean up and what we can do proactively and this is this is also a valid thing I saw one of the comments went through about critical data elements not everything is critical conscious disregard maybe some stuff doesn't need to be governed as carefully so give that some thought so in the technical side so how are we going to change a business process well let's actually change the business rules in the application so I still see this I still see that for so many years in the industry drop down lists actually my funniest one actually tweeted it was a data quality webinar and they asked for in the US your state and there wasn't a drop down and they felt me like male or female there wasn't a drop down which all right for a data nerd that's actually very funny because that's one of the easiest things to fix if there's only two values in your form have a drop down and that's going to help your data quality right get the business rules up in the front and that can be a very you know some of the data quality tools can do address validation of the source etc etc because if not you're going to do it reactively and do your data cleansing and ETL into the warehouse and then again in that middle side can you be auditing it with dashboards can you augment with external data sources that's kind of that that middle part of doing the data stewardship as you go so now that we're kind of getting into more in the technical side the question I love to hate is and they vendor the good news is that they now data governance is hot and everybody wants to be on the yes is more to spend two values gender but or or is sex two values and gender more than one but the point is have dropped down so data governance is hot and because it's hot now every tool the data governance tool which is probably not too incorrect but what I get frustrated with is we'll have a data governance project people jump right to what tool can I buy to fix it and that's the wrong question to ask the tool should come secondary all that other stuff we talked about really should be the primary but that said tools are important is a lot of this you cannot do without a great tool and so give some thought to it one way to look at this is you know some of the functionality first what are we trying to do are we trying to get a tool to help with the organizational structure with the process and workflow logging issues glossary that's kind of more in the business side and or are we trying to get a tool to do more of the data lineage data quality trusted data sets and that sort of thing and then do a heat map the other part of that is how technically important it is and what the business need is and then on the bottom is who your key audiences and I've seen very good tools go very wrong because people pick the wrong one and I've seen a customer that wanted to do really data detailed data lineage for GDPR and they wanted you know your almost classic source to target mapping for your warehouse and where every field was and how it's populated and they picked a tool that was very user-friendly for more of the processes and glossary and things like that but didn't have the guardrails it wasn't on that spectrum of the technical side that they needed so it was a great tool but they're sort of lulled into the nice pretty front end and it didn't do what they need on the back end also seen folks that maybe go too far on the data lineage and mapping but they don't need to go that detailed and really when you try to sell it to business they need something more simple like a glossary and again a perfect governance would have both but just give that some thought into the relative importance as you look at a tool so a nice thing about these tools that can really help and you can do them in a nice targeted way so here's an example of some of the data architecture tools that I can provide a roadmap so you can do these in small targeted projects and I'll give an example of that you know what data do we prioritize can I do a data model to really start to align some of those grid business data elements can I have process models to show where that data is used what's the full data architecture the system data architecture what are the business rules what's the policies around that and do I have a data quality dashboard to really see the effectiveness of that that seems like a lot but if it doesn't have to be if you do it in small chunks right what if we're just looking at a small project about what this is an insurance company what what data would need for our brokers what do we need for customers try to pick a one question you know we're trying to do credit history what what small piece of this can we do to make these artifacts real to the business what you don't want to do is start with an enterprise logical model that's going to take you two years that no one cares about you could even do a conceptual model in little pieces or a logical model in pieces or just do the business process that people care about and do these in small chunks so you don't try to boil the ocean I'm a big fan of metadata and the way I look at that is that metadata can really make data governance actionable so these tools are good and you can sort of take these policies and then implement them in the database so again the more you can automate so that whatever the business rule is you know gender codes or state codes or whatever it is those are your only choices make that easier and then you can sort of audit that with a lot of these data tools metadata management tools we have and will and could do a whole session on this but but think of the right tool for the right job so do you need and all of these are good in the right in the right sense that enterprise catalog metadata repository that stores everything together with a common metadata could just so maybe even your data modeling tool repository be enough if you're really looking for glossary and some definitions that may be already in your law school model or and or are you needing to share this externally with things like XML and external registry again give that a lot of thought because again you could probably save some money I used to one in my earlier life do a lot of consulting and product development and one of these big metadata repositories and people would spend literally millions in a full implementation that in some cases made sense some people they could have just done it with their data modeling tool because really they were looking at with tables and columns and definitions so give that some thought so you know if you've heard me speak before I'm also a big fan of data models and they're a big part of data governance just give some thought of what level again if you're talking to the business look at a conceptual model it's a great road map to that and then you want to turn those rules into more of a physical data model there are other models that are sisters to that business process model is a great way to really get to the show what so what what what is the business process what data is used across that business process and that's a great part of governance because again if you can catch it during the business process that really is where governance really starts to sing kind of the modern sexier version of a process model and again especially if you're engineering or these are still great I use them all the time don't get me wrong but when we're talking about things like customer journey or I've done them for patient journey I've done student journey it's kind of a customer journey map and again when you're talking governance you cannot forget this customer data is used across an entire life cycle as this patient you know it's employee and really getting all of the actors in that of what data is used what data we care about and who's touching it this could be a nice interactive way to use design thinking and workshops and whiteboard and sticky notes which is a great way we did one of these in January with a big marketing company and it got it got the steering committee for governance so bought into the process it's been one of the most more successful governance right at the beginning we had a brainstorming workshop with customer journey maps and sticky notes and kind of fun ways of working and we overlaid the data and not only did we have really clear direction but we got some really great requirements in a very quick way very quickly we had our critical data elements we had our use cases and it was one of the more effective ways kind of using some of these modern methods you can mix modern methods with some of the old-fashioned stuff that's still really valuable one of my favorites is a great tool with a terrible name the good old crud matrix sounds like something is in the bottom of your shoe but crud that's why maybe we thought it drunk I don't know if I can rename it but there has to be a better name it shows where data is created read updated and deleted and something as simple as that when you're doing master data you're doing governance can solve so many problems the data might be created somewhere but you realize that someone updates it downstream or or is reading it in a certain way again a lot of these are very simple tools you can do in small pieces that have tons and tons of value and and even just getting an inventory of what systems are touching this or what processes are touching this again you can do a lot of these in small chunks which leads me to one of my favorite use cases because we in the data architecture data governance can get kind of a bad rap this sounds so complicated and so hard so this was a very fast-paced US based sales and manufacturing company they had a product they manufactured and sold so they had the full life cycle from sales to actually delivering and shipping the product customer support so huge and they had a lot of issues with their customer data so they had a very high-end product that was expensive and their customer loyalty was through the roof but they also had some really embarrassing you know you bought one or two of these in your lifetime and you were in kind of the executive club and you expected to be treated as such to their credit they sort of were trying to ship a product and they didn't know the person's name address or phone number actually called the board address they called the person and the person was flattered that they called them but they said seriously you don't know my address I gave it to the sales guy because they didn't have that full lineage so what we did because again fast-paced this was not a company that was going to stop business for months to set up data governance we did interviews this is one where we talked to the sales people went to the stores and it found out how they were gathering the data and what we did with four weeks print we did all of those artifacts that most of those artifacts I showed you we had a small data model did a customer journey map credit matrix data flow system market all of that with a small piece and we told the story of why that address wasn't right awesome results so those are real quotes the chief marketing officer she was a trip she said I never thought I'd use the word data flow diagram sounds really nerdy but I love it no one explained to me why my data was wrong before because we told it in that story with all of those artifacts which is a small piece and my favorite one was the head of sales said ah shouldn't I have my sales guys he actually used the word govern how they're putting the data in yes trying to get a salesperson to do that is hard but he saw he reasons you saw where the data went you saw why that day was valuable so you can do all these this one I worry I like this story is it sums up everything I talked about in the presentation we aligned with the corporate culture we did it fast we use their words we found a pain point that everyone could relate to that the customer address was wrong everybody even the customers felt that one and we we did the architecture right but we did it quickly in small chunks and really saw value and that was a great way to kick off governance and they're still doing it but they did it with a quick win so hopefully that kind of and it's a real world thing so hopefully that pull things together so without further ado because I saw that there are some questions just summarizing hopefully that did summarize it but you know the governance is people process tech all of that tools are important to pick the right one for your use case and no matter what it is pick some quick wins that also deal with architecture it's not an either or and you should see success as Shannon gathers questions just remind you next month we'll be talking about enterprise architecture and my standard sales pitch if you need help we do this for a living so Shannon I'll pass it back to you for questions Donna thank you so much and just to answer the most commonly asked question just a reminder I will send a follow-up email by end of day Monday to all registrants with links to the slides and links to the recording of this session so diving in here Donna what are your thoughts on data governance around data reservoir and our data lakes what have you seen that works and are there any good references you would recommend so yeah data governance I think maybe one of the school for data lake or data reservoir whatever people want to call it these days I think a good one to reference was this one of reminding just enough data governance no data governance for a lake is the wrong answer like that but getting just enough and kind of going so one of the comments on kind of key data elements you know what needs to be managed I had a really embarrassing example not for me but for the first new set it was a junior intern as a big insurance company that we had in his they were talking about data quality and he said so I shouldn't be putting the credit card data on the lake in the cloud and his boss talked to him after that right it was a it was a data lake I speak exploratory but there's certain PCI you cannot put out but once that was in place he should have had the freedom to do whatever he wanted so just you know there's certain things that are nonce you know you cannot mix with if it's customer private data you cannot that is still a human being but then don't overgovern it either so that's kind of a high level answer but kind of keeping this pyramid in mind might kind of help with that data lake versus non data lake and Donna do you find that executive executive leadership terms our teams are including or excluding the data stewards pro and concept so you'll see I would just terminology tend to see a more executive person would be a either executive sponsor or maybe a data owner at the high level but if you're talking to executive like CEO CMO I would say they're a little too high love they would be more your executive sponsor they should be championing it probably not the person going to get down and talk about business rules I would say yes they should be involved absolutely but I would probably their title might be more of an executive sponsor and probably not the data owner that you're going to be kind of rolling at their sleeves and doing things I can try and slip in one more question here don't do you typically take all rules to find in the business rules and policy documentation and also added into a metadata catalog any tips recommendations on how to capture those in a major agile implementation style I love that idea I think that was one of these slides here of that policies are fine you need them they're important but really to make them actionable you need to put those rules in your MDM tool or your data catalog or ETL scripts or all of the above and I think that's where the prioritization with the business case is important and getting those data owners that steward's involved so you probably just don't want to take the entire policy to start unless that's really easy but probably just base that around the business use case that we're talking about patient data what are the policies we need to worry about just protect the patient privacy and let's put those in the tool for example because if it's in a policy it's not being implemented in a you know automated way which is really what makes it happen well Donna that does bring us to the top of the hour here thank you so much for this great presentation as always and thanks to all of our attendees for being so engaged in everything we do and all the great questions again just a reminder I will send a follow-up email by end of day Monday with like suicide and the recording everybody thank you so much really appreciate it and help you have a great day and stay safe out there thanks Donna