The Entrepreneurial Approach to Data Management

Marc Alvarez:

I think people have been doing things a certain way for a long time. It's really in the past few years that we've just started to see a wholesale shift in the economy.

Nate Nelson:

Hey everybody, and welcome to the Data Masters podcast. I'm Nate Nelson, sitting with Mark Marinelli from Tamr, and he's going to introduce the subject and the guest of today's show. Mark, how are you doing?

Mark Marinelli:

Doing well today, Nate. How about you?

Nate Nelson:

Good.

Mark Marinelli:

All right. Let's talk about this episode. As Vice President of Data Management and Operations at Thomson Reuters, Marc Alvarez applies an entrepreneurial approach to data management. During his career, he's found that treating data less like an IT project and more like a business asset requires fundamental changes in the way that we go about data management.

Mark Marinelli:

In this episode, Marc is going to talk about what's driving change about how companies are using their data, the role that technologies like cloud computing and automation are going to play and continue to play in this shift, and how organizations can just get better business results from their data.

Nate Nelson:

Okay. Let's listen in to Marc Alvarez.

Nate Nelson:

Marc, you've a somewhat different approach to data from some of your colleagues. Am I correct in saying that?

Marc Alvarez:

Oh, certainly a different approach. You know, I come out of the capital markets and finance world where regulation is the primary driver to too many aspects of data management. I think it's been really the catalyst for a lot of change. Whereas in Thomson Reuters, we've divested ourselves of those businesses. We're now looking at a much more integrated and holistic program to how do we shift to a 21st century business operating in a digital model. And we started to see digital as vital to our future. That calls for a different approach to things, that calls for planning to service the organization with a high level of quality timeliness and just down and all around data capability, to really power what the business needs is that business becomes increasingly driven by quantitative and statistical analysis. We're taking a slightly different approach. We really focus around some basic data management aspects, like cataloging our data and documenting it, but also making it available for reuse across your organization. And I'm trying to get scale and really driving it much more from an entrepreneurial approach and from an IT based approach.

Nate Nelson:

What is data fundamentally for? Is it a technology asset? Is that a business asset? What are we not understanding here?

Marc Alvarez:

That's a big question, very philosophical question. I mean, my view is, it's the business asset. Data is a record of something that has happened. It represents considerable cost from any firm to produce and maintain. It carries both value and in some cases responsibilities and potential liabilities when it comes to regulatory requirements. My personal view and the approach I take is, it is a business asset, and we look to leverage automation and technology to really exploit the value of that data and really move it on to a value creation change, where we're not just using the data for one purpose to maintain the books and records of the firm, or maintain a view of our customers, we're actually looking to see what the value is as the whole, because that's really at the heart of our business.

Marc Alvarez:

Without it, we can't be focused on our customers. We can't be challenging ourselves to improve how well we execute and how well we deliver for our customers. And lastly, I think for the business in general, it gives them visibility. Gives us visibility into how we are operating, where there are areas for improvement or confirming that we are operating at industry best in class.

Nate Nelson:

If data is a business asset, as you just mentioned, what metrics can you put around it to demonstrate its ROI?

Marc Alvarez:

Well, that's an emerging area of interest. That's for sure. I mean, the metric I uncovered a few years ago and I continue to rely on it quite broadly, was a study done by Experian an IBM across the Fortune 1000 Companies where it's found the companies are spending up to a quarter of the revenues on data and how it's used in the organization. If that's not a business asset, I don't know what is. The real question is, are you doing it efficiently? Does it really need to be involving as much of an overhead as that? And can you repurpose those resources to exploit it for more value? That's kind of in my lodestone that I followed. But I think a lot of the use of data content in organization should be put to the rigor that we would put into an underwriting a hundred million dollar capital investment program or something.

Marc Alvarez:

And that in the past has never happened. No one has sat down and said, "Look, if I get this much data, I'm going to generate this much more velocity and sales, and I'm going to therefore raise our top line revenue growth by 2% or whatever." I think that type of economic rigor is now starting to focus on this. I think the message is there, but at the end of the day, what distinguishes data is, you can't turn the lights on without it. You have to do that in an environment, which is both running a business, as well as starting to figure out how to exploit this asset. And I don't think we're alone. Everybody in industry, any industry, is starting to ask themselves this question right now.

Nate Nelson:

When I'm now considering your more sort of business perspective on data, it occurs to me that you have a degree in economics. Is it your academic background that helped form this perspective that you now have on data? Or was it something else that sort of knocked you onto this path?

Marc Alvarez:

I think the background is in economics was enough for me to help me analyze the scars I've received over the years of managing data content and technology and exploiting value in a number of industries. I think it's served me very well to take a step back and lift the concept of a data or an information service and what it means in the firm and really look at it from that perspective, rather than just viewing it as a series of IT projects to service a capability need, or enhancement, or rationalization on some part of the business. What you start to see is, in today's modern world, one firm can't acquire another without the aspects of integrating data to their operations. The phrase is overused, but it's the new oil. It's the way in which you align the rest of the business assets and control them and bring them forward.

Marc Alvarez:

I think that image is starting to be seen across the industry. There are certainly some new books that have been published recently that are quite good in the topic. I think we are increasingly starting to see the case that monetizing data doesn't necessarily mean going out and selling what you have out of the exhaust of your business truck, it's rather about how do you take advantage of this asset and how do you gain the efficiencies and economies of scale? I have to admit having come at this from the point of view of economics and econometrics, that has given me a very different insight from most of the chief data officers and other executives in the data space that I liaise with.

Marc Alvarez:

Again, we're looking to drive this much more as an entrepreneurial program, not really constraining ourselves to any one technology or any one approach, and looking really quite honestly, looking to leverage the best technologies that are out there. And I think in that thing starting to move to cloud services is a big advantage for us because it's introduced a lot of dynamism and rapid time to market, so that we can produce this analysis, and we can start to ask ourselves these questions and we can start to build these ROI models. And that I think is eventually going to feed back and make this look into much more of a science and much less of an art.

Nate Nelson:

Listening to you talk now, it doesn't seem like anything that you're saying is particularly inflammatory and yet, correct me if I'm wrong, this is a rather controversial point of view. Why is your point of view not shared, we'll say, by some of your colleagues in the industry?

Marc Alvarez:

I think people have been doing things a certain way for a long time and it's really in the past few years that we've just started to see a wholesale shift in the economy. I actually think this is, what Malcolm Gladwell likes to call, the inflection point. Companies are looking to drive their business by using their data assets. Companies recognize that the use of data goes beyond simply addressing the use cases as a finance organization. This is introducing wholesale change, the notion of how companies are structured and how they structure themselves to deliver the services for their customers. And quite honestly, having being a veteran in this space and having been through many of these exercises, that's a fundamental change because technology is advanced to where we can drive data now, so that many, many more users can take advantage of it.

Marc Alvarez:

And the old style method of feeding data as an IT function just isn't set up to support that size of a user base. I look at my own firm where we have 20,000 odd employees, across our technology organization. They're really geared to driving certain deliverables on a certain roadmap. That introduces this whole new democratic aspect. That's what you'll see a lot in the industry referred to it. I think that's a fundamental change. And I think companies and technologies are all coming together now where we're starting to open this Pandora's box and it is a Pandora's box. I think, with things like CCPA and GDPR there's outright liability on the company and its officers to get this right. Those pressures weren't there before and equally, that the whole move to digital methods and cloud and open services, that's just made all industries hyper competitive now, especially coming from a world of capital markets and finance. It's hyper competitive to achieve, you know, incremental improvements at the margin in your bottom line.

Marc Alvarez:

It's really tough, and there's thousands of new entrance to any vertical at any given time. I don't think we've ever seen that before in the industry. I think trying to compare what's going on in this space today to what's happened in the past and other sectors, can only go so far. I think you're seeing a lot of people like myself, just having to think on our feet, and they've never had to do that before. And re-examine it in light of a much, much broader user community. To do that, just to bring it full circle, it's obvious conclusion I came to after working through many of this is, your best service by viewing, making data available to your organization as a true information service, rather than trying to do these as one off projects that are siloed and will not share any common underpinning resources.

Marc Alvarez:

You need to develop a whole new skill set around how you manage and govern this content. And by govern, I mean, governance, not government, which very [inaudible 00:00:11:59]. I think that's what's driving this change. And for those of us who've worked in the world of finance and capital markets, we're well familiar with this. Firms like Thomson Reuters and Refinitiv have been publishing data services to their clients for decades. I think this is a shift. We're not broadcasting this data outside the firm. Now we're doing it inside the firm. That means you have to tool yourself to do it. That's just not there as a firm. We've progressed from a world of databases to data warehouses, to now with cloud and on demand services. I think those are big fundamental changes. I think there's plenty of scope for people to do and make achievements and get results using different approaches, but we're going to go down this path and try and keep our eye on that Northern Star of, this is a business initiative, not a technology initiative.

Nate Nelson:

Now I want to challenge you a bit. Is it that your approach is just another way of tackling these issues or is it that you believe that this more entrepreneurial angle is effectively more valuable than a more technology centric approach?

Marc Alvarez:

I'm glad you credit me with thinking it's actually a discreet and final model. Thank you for that. I think we're going to learn by example. I think, different industries, different firms have different ways of dealing with things. You see this in banks and financial institutions and how they manage their operational risk. There are plenty of ways to skin the cat. I think if at the end of the day you really have to put your economist hat on, and figure out what's going to give you the most bang for your buck, and not just do something because it's the shiny new bit of technology that everyone's using.

Marc Alvarez:

Must have technology. I think it's must have solution and the solution will comprise business processes. It will comprise talent. It will comprise technology decisions. And I think in today's day and age, the pace we need to move and scale we need to operate out to meet our customer's expectations flows right back into the firm. We have to be making these investments to improve our game, because if we don't improve our game, the firm's going to struggle to keep up with its customers. Our customers are looking to us to break trail here for them and solve some of these problems.

Nate Nelson:

Let's get specific here. There may be folks listening who disagree with what you're saying, what evidence do you have Marc, that might help convince them? How has this approach helped, for example, the companies that you've worked for?

Marc Alvarez:

Oh, that's a really great question. I'm very active on the CDO circuit. I tried and I solicit feedback from other people. I have my own network of contacts that I'll go and throw my stupid ideas up and get their feedback, so it's not like I'm doing it in a vacuum. But I will tell you this. When you get down to the actual metrics of cost, time to deliver reliability of delivery, I'm feeling like I'm on very solid ground. You know, we are able to assemble a discreet roadmap for year by year. We are able to resource against that roadmap. We were able to budget against that rule match. It's not perfect, but it's better than anything I've done in my career. And quite honestly, when I look at our portfolio of work that we're doing and the number of projects we're running in parallel to achieve what the company's trying to achieve, it's quite impressive.

Marc Alvarez:

Mount of work for an organization of 60 odd people, I'm responsible for. Those metrics are definitely pointing in the right direction. And the other thing is something credit Thomson Reuters. I've only been there a couple of years, but they take very seriously the health of the organization and they have been measuring it. This is showing up in our organizational health index scores. People are starting to really see and value that. Again, I think it's as much art as it is science at the moment. I think you have to be willing to break some trail in your organization. Certainly in this area, what I found, was putting together a proper business plan and laying out the steps to ROI and what we could do with those returns over time and never been done before.

Marc Alvarez:

I find that that was viewed by my senior management as very welcome. First they thought I was speaking Greek, but when we sat down and we went through it, they could actually see where the journey was taking them, as they would with any other investment. I think it makes sense to look in this way, but I'm not going to be anyone to say, here's the prescription pad and every company should do it. There's a very organic element to this that every firm has to consider. And again, you're doing it in increasingly demanding environments where budgets and returns on equity have been put under a lot of pressure. I think it's incumbent to think on feet. And the other thing is, don't throw anybody's good ideas away. It's not like I've come up with all of these ideas.

Marc Alvarez:

I've quite unashamedly borrowed them from some of my other colleagues who are in the industries. I've worked [inaudible 00:17:25]. We're forging much the same trail and in different industries, different firms. I think, I look at best practice a lot. I look to adapt and implement best practice. So far, I'm pretty pleased with the results. I think as a whole we're fairly pleased with the results. I think we'd like to go faster, but you know, there are trade offs to these things. But I think we are on a healthy trail and we're picking up pace and we're picking up scale. I think we're heading in the right direction.

Nate Nelson:

So when you enter a new company and you take along this perspective with you, I imagine that not everybody is necessarily on board right away, right? If what you're saying is true, and some people do disagree with your approach or are just used to doing things in a different way, than your job isn't just to use data how you want to use it, right? But to get everybody else to approach it that way as well. Mark, how do you go about changing the culture around data among your colleagues or in your company?

Marc Alvarez:

Slowly and with pain? It's a different way of thinking. I have focused a lot of my efforts on telling the story. Really focused on what to expect and how we're going to get there. And I've really focused on short term, low risk wins, where you can demonstrate what's going on. That has involved considerable amount of change for the organization. One of the changes we introduced was to take our operational sides of things and completely separate them from the transformation side of things. I have direct oversight of the transformation program that we're in and we've been able to demonstrate results. We've been able to build out our first product master solution in about a six month period, and now we're building out additional datasets, all in the same integrated environment. It involved bringing in some entirely new ideas.

Marc Alvarez:

If you go and read any of the data management organizations out there, their first recommendation is almost always unilaterally, you have to catalog your data. You have to know what data you're talking about. You have to know what the landscape of your data is. We didn't just talk about it, which is what most firms do. We went on and actually did it. So now when we talk about our data or our master data in particular, it is a completely defined universe. That universe is completely modeled and documented and under change control. And these are things that were never done before. Now, the process seems to be feeding on itself. People are saying, "And I wouldn't want to do it the old fashioned way of sloshing spreadsheets back and forth ever again." And I have growing list of constituents who want to take advantage of that catalog and leverage the work that's been done and extend it in many cases. We get, we've got a lot of requests to extend.

Marc Alvarez:

I think the approach has been right, but you are spot on, and there are several books I've read over the years in this area. Data has a cultural component to it. People get very passionate about data and how they use it. And they're not particularly generous in their thoughts of how other people choose to use it. So you need to get to the lowest common denominator, right? Something that technology people call it normalizing. I think data standards helps. That helps gets us there, but you need to set it up so that everybody's benefiting. You have to set up a parade of optimal outcome here and that's not easy and that's not cheap. I can tell you if it's something you have to budget for and invest in and find the people for. People component of data is not to be overlooked. Some people are thriving on the new challenges this presents and the opportunity to use state-of-the-art AI tools to assess and manage data content. These are groundbreaking areas for a lot of careers.

Marc Alvarez:

I'm quite honestly, I'm quite excited about it and quite passionate about it. I think we're making some really good progress, but we're running into some problems as well. It's not all bright sunshine and chocolate cake. It's not this stuff. This stuff is tough. You're dealing with, in the case of Thomson Reuters, decades and decades of legacy systems measured in the scores across the firm, resulting from hundreds of acquisitions. This is not a trivial thing to unwind, and yet here we are, we're getting there, and we're working it through and we're building our methods of working, and we're building constituency. I think we're heading in the right direction. I think we're starting to see some of the real results.

Nate Nelson:

You mentioned it earlier that harnessing the power of data is as much an art as it is a science. We hear a lot about the science behind data. Could you talk about the art of it?

Marc Alvarez:

I think one of the lessons I've learned is, in the last 10 years, there's been a big rush for everybody run out and hire data scientists, right? We now have a critical shortage of data scientists up there, so that's been an interesting finding. But look, if you talk to people who work in data science as professionals, as graduate and postgraduate level researchers, they complain incessantly that their biggest challenge is, they're spending well over half of their time, simply sourcing and massaging data and normalizing it. There's an art to doing that. It shouldn't be reduced to very expensive, very bright people having to do that legwork. That's where you should be looking to gain economies of scale by doing that normalization for them. Their value is in analyzing the data and writing forecasting models and identifying predictive patterns and unlocking areas of causative relationships, because these are all things that happen within your firm.

Marc Alvarez:

I've been really lucky to work with some absolutely brilliant data scientists, but you're constraining them if you're forcing them to go back to every one of your legacy systems and haul all that data out and get it into a data lake and massage it. There's so much more you can do to make those people more productive. And that's the art side of this. That's knowing what an information service is. That's being able to anticipate, how are these guys going to work with this data? Are they going to use Power BI? Are they going to unleash Python or Aura at it? Are there some other tools out there? We're certainly using a lot of AI tools just to analyze the data so that we can improve the quality, timeliness, that service level of that data. I don't think it's an either or thing. A lot of what we do requires the help of data scientists and for data scientists to be successful they need us to be successful.

Nate Nelson:

You mentioned AI. Could you talk about what you think is next for data in terms of driving business? Are there certain skills or positions organizations need to see value in? Is there any technology that they should be making more use of?

Marc Alvarez:

I think that's a really exciting area. It's certainly an area I think is right for a lot of innovation. By definition, the data we produce and consume in our firm is produced daily on a routine basis in response to business transactions, in response to sending proposals to customers. If you unwind it all, it's very much this regular recurring model. We create over a thousand accounts a day on average in our firm for our firms for various reasons. If there's ever an area that I think is going to be really valuable for us, is we're now using machine learning and setting up machine learning models to really identify the patterns of our data, really identify the distribution of the data and areas where we see the most activity and correlating that to areas where we see... Right now the focus is on quality, so we want to find candidate errors and put in workflows that can stop bad data or erroneous data, or even potentially erroneous data, from being dropped into a financial analyst or a salesperson's view of the world. And we want to capture those.

Marc Alvarez:

We're seeing really good results here. I have a saying, which is, if there's ever an application for machine learning, it's in the management of your data because it's this repetitive ongoing flow, this is not static, and machine learning thrives on the fact that the data is in motion and moving through the chain. And it's heuristic, the more we can run them, the better and better results we're seeing. We're using one example as we use to identify outliers of corporations that may or may not be classified in our customer hierarchies correctly. Hugely important to us. Hugely important to equipping our salespeople with the fact that yes, you have this account and here are all the subsidiaries of these accounts, and here are other locations of business, and here's all the key contacts associated. Incredibly important, mundane sounding, but incredibly important to being a successful business and a successful sales organization.

Nate Nelson:

Lastly, what would you recommend to folks in your position who understand data, but maybe hearing your point of view for the first time? What can they do to get better results from their data operation?

Marc Alvarez:

I can tell you what's worked for us. We started, we went right back to the basics. We continue to work from that premise where we were like everybody else. We were harvesting data passively from all our systems and passing it through the organization on a batch basis, not particularly timely, and consolidating it and standardizing it, and basically running our data operations after the fact. The observation I had coming in to this job, when I first started taking a look at, was we kind of got it a little bit backwards. What we should be doing is producing the content independent of any one system. We should be ensuring that the state of the data and that when it is published is that it has been registered as part of a universe. And that has been indexed and collated and validated and enriched.

Marc Alvarez:

It's not just a value. Someone typed into Salesforce somewhere. To me it was no, we have to become this... In order for this to work, for this strategy to work, we really do have to turn these things around and we can't be passive. We have to be active. When I started thinking about it in those terms, it really made a big difference to me because now I'm not constrained by the 200 odd legacy systems which are kicking around in our various back offices around the world, it's much more one of, okay, how do we make sure that our general ledger has the most up-to-date register of products and pricing that we have? How do we make sure that the view of our customers and our CRM systems is consistent across all of them in the firm? Those are the problems we need to solve.

Marc Alvarez:

Then I think once you look at it that way, then the tactics are yours to decide. They're yours to figure out and move. Basics would work for me. This was new. I inherited five separate groups who had never talked to each other before. They were all doing different things. They're all doing data quality work. They're all doing data governance work. They weren't doing it to the same level. Consolidation was one of the things. Consolidation of the operations was one of the key things that worked for me. The other thing was cataloging. Getting our data catalog properly and putting in place the means to manage that has opened up huge avenues for us because now we have a very discreet definition of what's there. And it's growing. It's growing both in terms of the data objects that we manage to that catalog. It's also growing in the richness of the metadata we capture for each of those data items, to the point now where we have a whole new domain within our data catalog, which is what are the data quality parameters we run against each data item we have in our inventory and how frequently.

Marc Alvarez:

This is all going to feed back, and just when I turn the data scientist guys loose on that, they're going to start predicting where we're going to have data problems before we have them. I think it's doing these things that does it. Up to a few years ago, these capabilities weren't there. These were big dollar, big, big, expensive projects involving major services firms because it wasn't core competence. Now, I think we're turning it into a core competence. I think if you look at it that way, I think it can take you a long ways. But I won't lose sight of one fact, I haven't worked in other industries is, I don't believe there's one size fits all solution. I think this requires the organic nature of the firm to be part of and you need to be sensitive to what constitutes a success for the firm.

Nate Nelson:

Mark, that was my interview with Marc Alvarez. What are your thoughts at the end of all of this?

Mark Marinelli:

Well, Marc clearly believes that the future of data is in its use as a business asset, but data for the business is easier said than done, as we all know. Getting to the point where you can leverage it effectively requires substantial commitments and how the data is managed in an organization. I liked what he had to say about what technologies can be brought to bear there, but also around who can actually access the data governance being, being really important here. As Marc talked about shifting how things have always been done is not easy. I think the key here is putting a business plan around the data, the chose ROI over time for this business asset. Like Marc said, the idea was well received when he put this in front of his business peers at Thomson Reuters, and we've heard similar messages from our other podcast guests. If data leaders want a seat at the executive table, they need to adopt a business mentality around their data.

Nate Nelson:

That'll just about do it then. Thanks to Marc Alvarez for speaking with me. And thank you, Mark Marinelli.

Mark Marinelli:

Thank you, Nate. Take care.

Nate Nelson:

This has been the Data Masters podcast from Tamr. We'll catch you all next time.