 Live from Cambridge, Massachusetts, extracting the signal from the noise, it's the CUBE, covering the MIT Chief Data Officer and Information Quality Symposium. Now your hosts, Dave Vellante and Paul Gillan. We're back, this is the CUBE and we're at MIT in Cambridge, Massachusetts. The CUBE is SiliconANGLE's live production of MIT IQ, the Information Quality and Chief Data Officer Symposium. We're pleased to have Laura Hahn here. She's a senior data governance consultant at TD Ameritrade. Laura, welcome to the CUBE, good to have you. Thank you. So we've been talking about the panel that you moderated today. We want to get into that. But before we do, what does a senior data governance consultant at TD Ameritrade do? Well, I actually have a recent update to that. I'm actually the senior manager of Enterprise Data Governance. Congratulations. Thank you. Come to the CUBE, get promoted. That's right. It's causing a fact. So I've assumed a leadership for Enterprise Data Governance at TDA. But I work also with metadata, reference master data, and data quality. And just making sure we're effectively utilizing all those tools. So we're going to end rolling out Enterprise Data Governance. Metadata went mainstream a couple years ago with the whole Snowden reveal. People really didn't think about it outside of our industry. But is there going to be more metadata than data? Some days. When you actually look at the academic definitions and all the different components of metadata, it does become really, really huge. Yeah, absolutely. And I think you have to determine what you're going to go after to make it more manageable. What do you mean? Unpack that a little bit. Well, I think once you start to look at all the different types of metadata that are out there and all the things you could capture and start to document, it becomes really, really big. And you can feel like you're spread thin, a mile wide and an inch deep. And I think we have to start thinking about what parts of what metadata can we capture that's the most valuable to our firm right now to try to start creating that repository of valuable metadata? Because going after everything might not be a great use of your time. So maybe it's critical data elements. Maybe it's just maybe it's lineage and figuring out where the data is. But you can get into a business process metadata, technical metadata. We've had people internally at our firm approach us about several different things and we've had to kind of pick and choose where we're going to devote our resources first. One of the themes that we've been exploring at this event and the Cube and certainly in previous years is the balance between data as an asset and data as a liability. Where do you spend your time? Financial services, you could be targeting customers, you could be dealing with fraud, you could be looking at new ways to get competitive advantage, you could be protecting your corporation. How do you sort of slice that pot? I'd say I'm definitely on the data as an asset side and making sure we're getting all of the value out of it. I think there's so much more to do in that space. Certainly there's steps that we have to take to protect our data so that we're not exposed to certain amounts of risk. But I think that the benefits far outweigh the liability side. We just have to be responsible citizens with our data. So definitely focused on the value. So your panel earlier today, maybe talk about that. It was sort of the pulse of the CDO, right? It was and creating actionable knowledge which is in my view is really the win of the CDO office is if you can take all of that raw data and convert it into something actionable and serving that back to the business and getting results. So we had a few different topics. We spent a little bit more time on data governance than I expected, which is near and dear to my heart. But it seems that different companies are figuring out what it means in their firms or their companies to own data. What does that accountability structure look like? Who's the decision maker? How do you identify them? How do you acclimate them to that role? And there's a lot of different strategies, but ultimately I think we discussed that the CDO does a lot of influencing, a lot of communicating, a lot of educating to get the rest of the company to understand the part they play in being an owner and a steward. And one of the issues that came up is who owns the data. And in the case of a medical, a healthcare company, the panelist strongly believes that the customer owns the data. On the other hand, a retail company might not be so inclined because there are things that they want to do with the data. How do we resolve this issue? Is this an industry by industry issue that needs to be addressed? I would suspect I'm a big proponent of paying attention to culture and organizational change management. So I really think it could be a company by company difference. I think you have to figure out who the influencers and key decision makers are in your organization, along whatever lines they fall. And leveraging that existing culture to get the most effective decision making. And it may end up being similar across an industry, but I think trying to go against your organization's culture is ultimately going to make it very hard to identify ownership. But don't all organizations now need to be data and analytics oriented. I mean, can you really be competitive if you don't do that? No, I mean, I absolutely think you have to. But I think you also find some surprising allies. And those are the people to align with. There will be people who get it immediately and they are raising their hand and they're signing up to be accountable. They want to be accountable because they see the value. And I think even if it maybe isn't the model you originally wanted to deploy, I think you have to bring those allies in and work with them and find those opportunistic things that you can work on with them. So in that case, isn't the CDO more of an evangelist than a practitioner? That'd be a great way to put it, yeah. But that's not the case in all areas. You see that the CDO in what you're seeing is becoming a very diverse set of skills. For a CDO to have? That depending on the company, the CDO role maybe can play it completely different. Yeah, and it's interesting is one of the questions I asked today was for a next generation of talent aspiring to be a CDO, what are the skills you need to have? And speaking with a gentleman about people coming from the business into that role, people coming from IT into that role, and I think it is such a diverse set of skills. And I have a finance and accounting background and I think we need to be open minded and really look at the leadership qualities. Because ultimately I think that's what they spend the most time on is influencing that senior leader peer group. So CDO is evangelist, let's key on that for a second. So, and I know there's a diverse set of examples, but let's just take that one example. If in fact he or she is the evangelist of the organization and their job essentially is to create awareness and catalyze action and they succeed, they're pretty much out of a job, right? Is that fair or what do you see that? That role long term going. Well, I would place my bet on change. The changing economy, a changing industry or a changing leaders move on too. The peer group changes over. So I really think the job will never be done. I think there's always competing priorities and I don't see it ever going away that you have to continue to advocate for it, lest it get trumped by something else. Well, you hosted a panel and there was no agreement on this. Yeah, I was about what the future of the CDO role and I think we had trouble even defining what the present state of the CDO role is. And that goes back to your panel. I sensed some frustration on your panel with CDO not really being able to get to that bigger strategic role, the potential of analytics. A lot of them still down in the weeds dealing with governance issues. Do you sense that same frustration or is that simply part of the job? The frustration that they're not getting to the analytics side? That they're having trouble getting to that next level where they're really providing strategic value to the business. And a lot of what CDOs are doing today is just finding data, normalizing it, fixing quality problems, categorizing, creating metadata, really kind of plumbing stuff. Is that just a natural state of the business right now, given the quality of data that's out there? I think so. I think there's a lot of heavy lifting to do in the data. Something I've seen several times. There's a certain amount of hard work that you have to invest in correcting your data and getting it to a place where it's usable. And there's really, I think there's some people coming up with ways around that or looking at ways to innovate around that. But I think for the most part, most of us just see that there's a lot of hard work that maybe an organization has avoided for a long time. And you're left with no other choice but to just dive in and start doing it. And so if that's the case, then yes, a lot of CDOs are going to be busy with that. Because that's really, it's the foundation of the house on which if you try to build analytics, you're eventually going to have less valuable analytics because you question the quality of the data. I don't want to put you on the spot as far, I don't want you to name names about employers, but you went from, you've been two very different industries. You were a target which is an acknowledged leader in analytics and customer profile. You've gone to a very heavily regulated business now with Ameritrade. What have you found there in any difference culturally to how people look at data? Well, I think one of the dimensions that's definitely different is time and speed. So the industries have a different time component. So retail customers shop every half hour or every hour or they're buying something every day. And so the cycle to deliver insights in retail is extremely fast because a customer can leave very quickly. In something like financial services, that customer has an account with you. And they've probably intended to have that relationship for a while. So the time component and your risk of losing that customer is longer. So that's the main thing I've seen is that the pressure that we as data management professionals are under to deliver that information, the cycle time is definitely different. How does that affect the way you use data? Target is obviously offers coupons, promotions, the timing is very important. Ameritrade, you're not under that same kind of pressure, but they still use data strategically, don't they? Yes, we definitely do, but I would say, in terms of something like data quality. And figuring out what your data quality capability needs to be, you would adjust that depending on what that demand is for data to be accurate. So in one scenario, I would be getting phone calls at six in the morning. Because the reports from the prior day were off. It was that fast. Whereas in a slow removing environment, I think that demand for data quality, there's more tolerance and it can take a little bit longer before the business is really demanding something like a data quality fix. So just, there's a little bit different priorities. Well, and there's a lot of obviously real time aspects of financial services, particularly in trading. But when we talk about real time, we define real time as before you lose the customer, and so in retail, that's a good way. But in retail, you're giving me an incentive to come into the store and buy something. If you do that in trading, it's illegal, you're going to jail, right? Right, but at the same time, every customer interaction in retail is a new transaction, it's a whole and complete order in financial services. They have an account and it's a sub-transaction of that account. So you may not lose the relationship on the basis of one trade. Now, you don't want that to be sustained over time. You want to shore that up, but on the retail side, that customer may never come back again. They don't have a standing account with assets in it. Every sale is an opportunity to win or lose them. So they're both valid, they're just different cycles. Your financial background, is that the perfect fit for the role you're in? Are there things that you wished you'd studied or brought in from people out there that are interested in this type of role? That's a great question. One thing I appreciate about it is that I understand a P&L. And it would be my wish for everyone in an organization to have a great understanding of a P&L and what it takes to manage that. Because it really helps me understand what the business cares about. Because they're probably held accountable to some financial number. And I can better align what we're doing in data management to those goals. Something I wish I had was more of the technical skills. I've had to learn by osmosis and just jumping in a lot of the technical terminology and being on different IT projects. I've had to pick that up along the way. So that's something I definitely wish I had more of. So that's more parlance really, exposure to technologies and what people mean by, there's a gazillion. I wish I had more exposure to that too when I've been at this almost 30 years. So what about statistical capabilities? I think, yes, now I would say so. So there's more of a, I mean, we always say, what's the perfect mix of for data science, it's math, it's programming, statistics, data hacking. I would think that the more stats the easier it is to communicate with a lot of the folks that are in the organization. Absolutely, and I think it depends, if I were giving someone advice, it would depend on what level they were at in the organization and where they aspire to be. I would say you get to a point where as long as you can interpret the statistics. So there's a difference between producing the statistics and interpreting them and telling the story of them. So I think someone would need to take a look at where they're at and where they want to get to. For me, I'm at a point where I'm really working on interpreting and telling the story. But you want to make sure you can also build a team around you then that can produce and actually work with the math. There's a lot of talk about this concept of the citizen data scientist. One of the failings of the business intelligence initiatives over the last decade or so has been the inability to put those insights into the hands of business users fast enough. You have to go through a big process and it's complicated, it takes forever. So we've been talking about citizen data science now. Does that scare you as a governance professional or is that an attractive proposition? And when you say citizen, do you mean does that anyone can do it? You know, business users and folks that are not trained necessarily in data science, but are everyday salespeople, marketing people, you know. Yeah, I think, honestly, it scares me a little. And it probably comes from my experience managing KPI reporting. Because it becomes, if not managed well over time, you can end up with multiple versions of the story about your company's performance. And that's eventually going to have to be rationalized so that everybody's on the same page. Making sure if someone's saying business is up, you don't have another person saying business is down. Because that affects your decision making. So I think it has to be managed a little bit and not you can't let it get too out of control. Otherwise, you will eventually start seeing various tentacles of decision making that don't always align. And then you're going to have to do the hard work of aligning them again. My governance is important. You were a speaker last year at this conference on a topic related to privacy. What aspect of privacy do you address specifically? I was talking about, it was big data and privacy. And the element I wanted to add was the human behavior element. And it was based on some experience I had with web and browse data and working with Hadoop environments. And kind of going back to your citizen data science, we had a lot of people who were really interested in data. And we're starting to move it around. And I think you have to be careful. Because the opportunities to leak data are, you always think it will never be you. I'm not in that type of a role. It'll never happen to me. But a surprising number of teams I think are in that position to maybe share data where they need to be more careful about sharing data. To where I think we need more education and we need stronger stewardship out there amongst those groups. So that people are more aware when they're asked to share data that they know the right procedures. And it's that tug where we've been talking about in terms of access versus security. People want the data, they want access, they want to bring together pieces of data or maybe work with a third party. But we've got to think through, and notice I'm not saying no. It's not that we can never share data with third party. But we have to be a lot more thoughtful. And I'm not speaking about any company in particular. But you think it's not going to happen to you? But then you're asked. And you have to know when to stop and think and know who to call and make sure it's okay. I don't want to put you on the spot. But you were a target during a very important event related to article in the New York Times regarding target's use of data to target promotions to an expected mother whose family didn't know she was expecting. That was, many people, including me, thought that reflected very well on target. Just technically what they were able to do. But certainly there is also a countervailing privacy issue there. What have we learned from that incident? Or was there anything, you're looking back at that incident, what changed? Or how did your own perspectives on privacy perhaps were affected by that? Right, it's a great question. I think it was the first time it became more widely acknowledged that personalization was happening. I know I had several conversations with friends and family about the fact that personalization has been happening for a while. And that as long as it's within our personal guardrails, we accept it and we leverage it and we use it. But once it crossed a line, I think we all realized that there's a point of taking it to a level where it can make people uncomfortable. And so, I think there's, I've seen studies quoted since then on how many points of personalization are acceptable to a consumer and beyond that it does start to feel uncomfortable. And so companies, I think, are more mindful of how personal to get. But how do you determine those guardrails? I mean, isn't that unique to the individual? I think it can be. But I'm sure they do focus groups and things like that. And it's interesting. I'd love to continue to watch it because there's also been a lot published recently on choice. And when consumers ultimately have too much choice or too much noise, it becomes difficult to find things, especially in the retail space. You have to make it easy for them. So you're trying to balance making it easy with still letting the consumer have choice. And so that, it's tough. It'll be interesting to see how companies answer that as they get more and more customer information. I think it's Malcolm Gladwell that said, he goes to the supermarket and there are 50 different kinds of salad dressing on the shelves. His reaction is not to buy salad dressing. I know the feeling. You'll stand there sometimes and just walk away. I have four-year-old kids. I've learned that you don't give them three items on their plate because they can't process that, even two at most. Right. And the studies say things like when there's too much choice, people are actually unhappier. The fewer choices, the happier they are because they're more satisfied. So how does that play into the role of the CDO? Because the natural tendency of the business side is to give people, test a lot of things, give people a lot of choice. But maybe the data officer should be responsible for advocating for the customer. No, you're going to have exactly the opposite effect. Yeah, well, I think it's ultimately the responsibility of if it's the people deciding on the products or the people deciding on the user experience of the website or it's really up to the business, I believe. But those might be some questions that a CDO asks just to maybe pressure test the ideas and figure out, do we really need this data or that data? But ultimately, the CDO's job is to connect the data supply to the data demand and it's up to the business. How do you avoid being in that position of saying no? Which I think was a brush that a lot of CIOs were tarred with. They were not enabling the business. They were too busy guarding things to enable the business. How do you avoid getting into that same vicious cycle as a CDO? Well, that's a great question. Impossible question to answer. I know I've been in that position of saying no when people ask for data. And I think you have to follow it up with, I call them seek to understand conversations, seek to understand what they're really asking for. What are they trying to do? I've had a lot of those. People would come to me and say, I need this data, and I'd say, well, especially when I was doing more reporting in BI and they would just ask me for the data, a lot of it was just educating them and making them aware of other tools I can give you more than just raw data. Maybe all you really need is one metric and you don't need the raw data. So let's talk about that. Let's talk about how we get you the metric or the trend line you're interested in. And then you're not necessarily pulling raw data out and putting it out there. So I think it's seeking to understand what are they really trying to do. And then sharing more with them about how you can help them with that problem instead of saying no to the initial request. So go ahead, Dave. Well, follow up, please. I mean, so you're really talking at the level of the business then. You have to be smart about what the business, how the business operates. And you have to be seen as an ally on the business side. Right, and that's where, for example, for me, that financial acumen has been so helpful. So if I understand someone's trying to manage a shipping expense line on the P&L, I know exactly what they're trying to do. And I can maybe already guess what they're trying to do. Or if they're trying to manage net new assets or something like that, I know where they're going. So let's talk about where they're going. Well, so that ties into the concept of the citizen data scientist. So a business user might just want that killer chart, the trend line. And so if you can give her or him visualization tools on a corpus of data and allow them to create charts easily and mix and match things and train them a little bit on how to do that, that is going to more than suffice oftentimes. Right, right, and it depends on how an organization is set up. Some organizations have centralized business intelligence functions, some have federated. And so it's about having the right, I think, really good relationships with whoever those people are and understanding their capabilities and when they want to roam free and work on their own. And when maybe they're getting into something that's beyond their expertise and they need more of that, that COE help because they're now pursuing something a little more sophisticated and just providing them whatever help aligns with what they're trying to do. COE Center of Excellence, is that right? Yes. You're making a bit of a career out of data governance. Where does that career lead? For anyone in data governance or for me? Well, no, for any, I'm asking you to be general. I want a job, I think you're great in the queue, but we'd love to have it. That's great to hear. At the panel I just did, this is one of the issues that came up, is where does the CDO job lead to? Is this a CEO track or is it a dead end? I mean, where do you see your career in data governance going? Well, my personal opinion, I think data governance, one of the things I love about data management, I would say data governance leads into data management, provided you can have sufficient experience with both the business and the technical components. I love it because I get to see all the components of the business. I get to see the P&L Alliance, I get to see the revenue side, the expense side, operations, sales, HR and payroll. I mean, you can learn so much because you have to understand all the data, you have to understand all the parts of your company. And so I think there's great opportunity to focus on one of those that you find that you really enjoy and move into that. I don't know that I would say a CDO is a straight to CEO move, unless that CDO had spent some time in a couple of other pyramids within the firm. But it certainly could lead to a strategy role too. I mean, what you just described is a perfect strength in the chief strategy officer is making a comeback with all these disruptions that are going on. So it used to be strategy. If you had strategy in your title, you were dead. Right. But now strategy can create Ubers and Airbnb. Yeah, exactly. Awesome. So in your new role, what kind of things do you want to get accomplished in the next 12 months? What are your objectives? Well, data governance takes time to implement. And I'd say we're still in the nascent stages. So we're in the mode of educating and doing initial pilots on different things and really demonstrating the capabilities and building on some initial momentum to work towards that tipping point where data governance is pretty widely accepted. So we're firing on all cylinders, but trying to really hone in the projects we work on with current strategic priorities set by the CEO and by our leadership. So what kind of outcomes do you look for when you're doing these pilot projects? What kind of outcomes do you aim for, sort of the short-term proof concept or the longer term? I'd say there's shorter term proof of the concept. Although we want them to align with longer term goals. So for us in TD Ameritrade, obviously we're in a regulatory environment. So we have a couple of things that we're working on there to support the compliance team with some things that they're interested in. But we also have a large operations function because we clear trades and we have a big, very well-oiled operations machine and their bread and butter is efficiency. So we think there's some ways we can look into the data and help them save time, save manual effort. So those are definitely two and then I think everybody's always focused on the client experience and looking at ways we can support a clean and positive client experience and teams are focused on that. So those are probably the big three that we have to leave it there. Laura Hunt, thanks very much for coming to theCUBE. Congratulations on the new role and really appreciate your insights. Thank you so much. It's good to be here. All right, keep it right there, everybody. Paul and I will be back with our next guest and to wrap day two, MIT Information Quality Symposium. We're back.