Data Discovery to Data Intelligence with Heidi Lanford, Chief Data Officer, Fitch Group. Thomas J. Lee, Managing Partner & Head of Research, Fundstrat Global Advisors. Charles Poliacof, Chief Executive Officer, Knoema.
Moderated by Marc Lopresti, Co-Founder, BattleFin.
SPEAKERS
Heidi Lanford
Chief Data Officer
The Fitch Group
Thomas J. Lee
Managing Partner & Co-Founder
FundStrat
Charles Poliacof
Chief Executive Officer
Knoema
MODERATOR
Marc X. LoPresti
Co-Founder & Board Member
BattleFin
TIMESTAMPS
EPISODE TRANSCRIPT
Marc LoPresti: (00:07)
Hello, everybody. My name is Mark Lopresti. I am the co-founder of a little company called BattleFin. You may have seen our pavilion down the hall on the fourth floor. I co-founded it with Tim Harrington, sitting there in the front row, about eight years ago or so. We were really early in the alternative data space. Very proud of what we've done since we founded the company. We've developed what we believe is one of the most compelling and robust platforms for alternative data. And we're going to get into that. This panel is of course all about data, from data discovery to data intelligence.
Marc LoPresti: (00:50)
But before we get into that, I want to say a big thank you to Anthony, the entire SkyBridge production team, John, Joe, Kat, everybody that makes this possible. We have an events business at BattleFin. I know a little bit about what it takes to put a production on of this magnitude. And I think you'd all agree this was an absolutely fantastic event to attend. We've loved every minute of it. And what a way to cap off my Salt experience, but with this unbelievably distinguished panel of experts that I have here with me.
Marc LoPresti: (01:25)
To my left is Heidi Lanford, from Fitch. To her left, a face you may be familiar with if you watch CNBC, Mr. Tom Lee from Fundstrat, one of the smartest people I know. And Mr. Charles Poliacof from Knoema. I have known Charles for a very long time, even before we were in the data business. So it pays to be nice to everybody because you never know when you may bump into them again. So, why don't we start? Heidi, a little bit of your background and what you're doing in Fitch with data?
Heidi Lanford: (01:58)
Sure. Sure. So nice to see everybody here, and thanks for having me. I started off my career as a data scientist and spent most of my formidable years working as a consultant and a data scientist. And then I spent my time before joining Fitch as the chief data officer, I had spent years in the technology space.
Heidi Lanford: (02:23)
And I recently joined from Red Hat open source software company that was acquired by IBM recently. And I have been really building strong, competitive data organizations as the latter part of my career. So really excited to be here. And I'm excited to represent Fitch.
Marc LoPresti: (02:44)
Thank you. Well, we're happy to have you. Tom?
Thomas J. Lee: (02:48)
My name is Tom Lee. I'm the co-founder and Head of Research for Fundstrat Global Advisors. It's a research advisory firm. We've really got two businesses. One is providing macro research and digital asset research to institutions in 22 countries. And we have a family office, a RIA business called FSInsight, and it's largely the same focus, which is education.
Thomas J. Lee: (03:16)
I'm very interested in doing this panel because I've been doing research for almost 30 years, really the first 15 as a wireless analyst. And when I started in '93, there were only 34 million cell phone users versus six billion today. So a lot of the business and knowledge that we needed to do to make equity calls, was gathering alternative data.
Marc LoPresti: (03:38)
And you touched on something. I want to put a pin in and come back to, the digitization of the modern economy and what that means for the data space. But that's a little teaser. But before we do, Charles, wonderful to have you on stage with us. Tell us a little bit of the company that you've built in your background.
Charles Poliacof: (03:54)
Yeah. So I am Charles Poliacof, the CEO of Knoema. For starters, I just want to thank you and Tim, and of course, SkyBridge and everybody else for putting this amazing event together. It is fantastic to be back, after having spent two years or so in this two-dimensional space. It's been great to be around people in a three-dimensional form again, and seeing a lot of folks that I haven't seen in a long time, and having the opportunity to interact with prospects and clients. It's been great. So, thank you for that.
Charles Poliacof: (04:21)
So a little bit about Knoema, a little bit about my background, although I'm probably not as interesting as these two folks here. So Knoema, we're a data technology company focused on making data accessible and usable. One thing that people know that are users of data is that there's far too much time that's probably spent making data useful. So we focus on three main pillars of capability. The first is discovery.
Charles Poliacof: (04:46)
So making sure that folks have access to data that they need, or new data sources that are interesting, that could potentially impact policy decision and investment thesis, or any sort of data-driven insight. The second is just around managing data pipelines. So data is the lifeblood of digitization. So somebody needs to manage that lifeblood, manage that infrastructure, making sure that everything is arriving in the formats that they're expected in, so that folks can consume that data on the other side.
Charles Poliacof: (05:14)
And lastly is workflow integration. As we all know, anybody that is a knowledge worker, works in their tool of choice. So looking at the color of my hair, you will guess that my tool of choice is probably Excel. But there are others that use Python, R, Tableau, Power BI, their own native applications. And that's really a core of what we do, is being able to integrate in those native applications so people can use their workflow or productivity tools of choice, so they can make those insight-based decisions.
Charles Poliacof: (05:42)
My background, I spent 15 years on the buy side, and then was part of the management team over at Novus, where we built a portfolio analytics solution which was focused on measuring portfolio manager skills. We'll talk about that a little bit later on in terms of what top line alpha and bottom line alpha actually means. And then part of the management team over at Visible Alpha, when it was a very early stage company. It's now a data stalwart. So, that's my background.
Marc LoPresti: (06:07)
So almost all of us on this panel are a good combination of TradFi, I think we're calling it now. I've heard that phrase used a few times this week. And DeFi and the data revolution. It probably makes sense just for a second to contextualize what we're talking about here. And what exactly is alternative data?
Marc LoPresti: (06:29)
And at BattleFin, we think about alternative data into 42 or 43 separate categories, but it is essentially that which you cannot derive from traditional sources like your Bloomberg Terminal, although that's obviously changing now as well. Satellite imagery, geolocation, credit card, point of sale receipt, email receipt data, sentiment, and many, many others.
Marc LoPresti: (06:56)
And this data, this resource, we often talk about alternative data as a resource or as a commodity, being used in ways that five or eight years ago we, when Tim and I started the company, we never could have anticipated. So Tom, maybe to have you start us off, you've been a data junkie, a self-proclaimed data nerd, part of why I love you. Tell me how you've observed the evolution of the use of alternative data in context of Fundstrat's work.
Thomas J. Lee: (07:28)
Yes. Well, first of all, alternative data has been a tool used by the best investors for decades. Many of you guys are probably familiar with Phil Fisher, and he wrote the book Common Stocks and Uncommon Profits. And his original thesis was visit companies, go visit warehouses, talk to sellers and vendors. That, he was the original.
Thomas J. Lee: (07:54)
That's the original alternative data model. And he had extraordinary returns, many 100X stocks. And I think in today's investment world, stock investors, the ones who really find the big opportunities are the ones that are exploiting alternative data, especially where they find variance versus either consensus or conventional views.
Thomas J. Lee: (08:15)
And I think in the past year with COVID, it's been enormously important to use alternative data to really have informed views on macro and investment decisions. Because in the past, people might have only looked at the bond market or VIX to have an understanding what the market is saying. But I think in the past year, there's been so much uncertainty created with COVID that the alternative data was critical for people who did well.
Marc LoPresti: (08:41)
And we were talking the other day about inflation by way of example, and how looking at these various traditional metrics to try to anticipate inflationary trends. How has using alternative data help refine that and make its predictive capabilities enhanced?
Thomas J. Lee: (09:01)
Inflation is a great example of why investors need alternative data right now, because in the past, CPI and understanding what the CPI print would be, has huge ripple effects across credit sector positioning, single stocks, and even commodity prices. And in this current context, we've got supply chain glitches, we've got shortages, and we've got demand build-up, and then explosion of demand. It's been very difficult for people to know what the real trajectory of inflation has been.
Thomas J. Lee: (09:35)
So I would say right now, if people are using alternative data to be informed about the trajectory of inflation, they're going to have a huge edge because it's the difference between thinking we're in ... As you know, there's a lot of people hyperventilating about inflation. I think a lot of our clients who've been using alternative data have realized a lot of these things are transitory, and you're seeing it play out. Even Stevie Cohen yesterday mentioned it. The bond market is telling us this is probably transitory.
Marc LoPresti: (10:01)
Right great. So Heidi, maybe turning to you.
Heidi Lanford: (10:04)
Yeah.
Marc LoPresti: (10:04)
So you've got a big job on your hands, building up this whole new division. Tell us how you are viewing the integration of alternative data into Fitch.
Heidi Lanford: (10:13)
So for us, I think this is a transformational shift in how organizations start thinking about data as a strategic asset. Whereas in the past, it has maybe been something that you go to IT or your tech team, and you ask them to provide you with data or give you access to things. As we start to see this data organization being a strategic asset for the firm, that's where over the past 10 years, a relatively new role, that chief data officer has emerged, because it is that strategic asset.
Heidi Lanford: (10:51)
It's a member of the C-suite. It is helping to influence new product innovation. And with that in mind, we have to recruit people who are not just great at the technology, data warehousing, data lakes, and data mesh. We're also looking for people that think about data as a product and how we can productize that, whether it's for our internal analysts that are consuming it, or our external customers that want to buy data from us.
Heidi Lanford: (11:21)
And so the talent and the recruitment is top of mind for me, as I've been building out my organization, because thinking about strategic players, great at technology, but also have that innovation and product mindset, this is not for the faint of heart. This is about moving resources and organization. This is a little bit about data is powerful. Data gives people in some ways, control. And to see that shift in mindset, it requires grit and determination from your leadership team and your company.
Marc LoPresti: (11:54)
And there are some practical challenges associated with it, too, I would imagine. Divergent datasets, different departments across the organization in different file formats and servers, and things of that nature.
Heidi Lanford: (12:04)
Absolutely. Yeah.
Marc LoPresti: (12:06)
That's a good transition. I know I see Charles chomping at the bit. That's a good transition to you and Knoema. That's a big part of what you guys do, right?
Charles Poliacof: (12:13)
That's a huge part of what we do. It's all about empowerment. What Heidi was talking about, data productization, the types of data assets that people used to rely upon 10 years ago, 15 years ago, channel checks, speaking to management. I know in the '90s, we used to send folks to the mall and count how many people were actually walking in and out of stores.
Charles Poliacof: (12:36)
And now all of a sudden, you have data that you can get in near real-time, maybe by looking at Carvana or some of these other inputs, near real-time visibility into what's going on in used car pricing, which is of course, influencing some of the inflation numbers. The Fed actually looks at those numbers, at least according to our friends over at M Science, which is one of the data providers we work with.
Charles Poliacof: (12:56)
But when you think about the challenges that organizations are going through right now, to be data-driven, to understand what that actually means, to be able to access different data products, there's a high variability problem. So there's a lot of data that comes from a lot of disparate sources. You have a subset of aggregators that collect a certain amount of data, but then you have a long tail of data that may or may not be relevant depending on the regime that you're in.
Charles Poliacof: (13:25)
And then there is the pain that you just surfaced before, Marc, which is this idea that folks need to build these connectors to constantly have this influx of products that could continue to inform their views, and then what it means to build an organization that supports that. And is an organization saddled with legacy technology? So it limits what they can actually do.
Charles Poliacof: (13:48)
Is an organization, to Heidi's point, thinking about data as an actual product, and an asset, and as an input that influences their digitization or digital strategy as they move forward? You see this in the insurance space. You see this in retail. You see this now in policymakers. The state of Texas is making all of their data available publicly, so they can attract capital to come in.
Charles Poliacof: (14:11)
So people are viewing data as a strategic asset. They are viewing data as an internal product. And I think those that continue to think that way, McKinsey just wrote a piece recently about this, are going to have a substantially greater advantage than those that do not. And again, I'll end on this point. I think it was Stevie Cohen and Dmitry Balyasny's panel when they talked about their firms that have enormous resources, and markets are fairly efficient.
Charles Poliacof: (14:36)
So understanding how to find those differentiated datasets and being able to unify them in a way where you can contextualize those insights. So you're not just relying on one or two datasets. That's going to be a big differentiator going forward, not just in our space, but in all industries.
Heidi Lanford: (14:55)
And use them over and over again.
Charles Poliacof: (14:55)
Sorry. Go ahead.
Heidi Lanford: (14:55)
And use them over and over again.
Charles Poliacof: (14:55)
Yeah. Yeah.
Heidi Lanford: (14:57)
All those great nuggets that you're talking about, or that Tom talked about, you don't want them to be one-offs. And that does require some discipline and getting into a data organization, and maybe doing some of those traditional things that we've done in terms of storing data.
Marc LoPresti: (15:13)
The challenges remaining are abundant. And that's one of the reasons why companies like Knoema and shameless plug, BattleFin are such an important part of the data industry. It's about aggregation, organization, curation, purification, and visualization.
Charles Poliacof: (15:30)
And I think one thing to think about also is I think a lot of firms, particularly in the financial services space, whether it's on the asset management side or even wealth management, there's been a very DYI mindset.
Charles Poliacof: (15:41)
And I feel like DYI should be left to like Home Depot and Lowe's. And folks really need to start thinking about what it means to leverage trusted partners. There's a lot of great new technology out there that can be leveraged to introduce ROI that didn't exist just two, three years ago.
Marc LoPresti: (16:00)
So Tom, maybe bringing it back to you, with Fundstrat's organization and structure, you really thought about it from the beginning as a data-driven organization.
Thomas J. Lee: (16:09)
That's right.
Marc LoPresti: (16:10)
Tell us a little bit about that process for you.
Thomas J. Lee: (16:14)
Well I would say Fundstrat, when we started the firm in 2014, we did come with a mindset that we wanted to do evidence-based research, and really help investors navigate markets not with opinions, but the idea of helping them understand future probabilities based on what we can observe.
Thomas J. Lee: (16:36)
And so capturing data and aggregating it, and ingesting it and cleaning it up has been really important. I think one of the challenges that we found with a lot of data is that one, there is a lot of noisy data and a lot of errors in the data. And sometimes things look like contemporaneous indicators, but they're not necessarily helpful with future.
Marc LoPresti: (17:02)
Give us an example. That's a great point. Give us an example there.
Thomas J. Lee: (17:05)
Well, some examples of things like ... Here's something interesting, because a client pointed this out to us once. So he felt that there were some things that really helped explain the PMIs on a real-time basis. And I would call that an observable relationship, but it didn't work influentially. We weren't able to then say, "Okay, well, six months from now, what are the values of these different things?"
Marc LoPresti: (17:31)
Right.
Thomas J. Lee: (17:32)
So now you need to make an inference. And if you don't know what those are, then you're just making it up. And now it's just an opinion. Yeah, so that's a challenge. And I think another thing we found was that sometimes, you can get a very complete picture from alternative data, but it only just tells you what everybody else knows.
Thomas J. Lee: (17:47)
And so it's difficult to find something that gives you an edge that's actionable. Like something, can it give you an idea of how a data point might look? Or is it going to help you understand positioning? But obviously, the goal would be something that helps you know where things are six months from now.
Marc LoPresti: (18:04)
So, what has worked? So we touched on inflation before, CPI, consumer confidence, very volatile in a COVID exit, whatever the heck that means, I'm not even sure at this point that I know. What's worked?
Marc LoPresti: (18:16)
Have you taken things like sentiment, which is one of the most popular categories on our platform, I think on yours as well, Charles? Has that helped with predictability, with actually extending the usefulness of these insights in the way that you've described?
Thomas J. Lee: (18:32)
Yeah. Marc, this is a great question, because we've found some really interesting things over the past year that are actually helpful. But then it's not clear to us if it's because of what's happened during COVID. Because in COVID, the world has become very digital. Really, last year, everybody lived a digital life.
Marc LoPresti: (18:51)
Right.
Thomas J. Lee: (18:52)
So things in a digital world are much more measurable. In fact, I think one of the most interesting conversations I had at JPMorgan was our economists had said that at the time, he said if you looked at the previous 15 years, 50% of all global GDP was pure digital. But the idea is that the next 20-year interval, it'll be 75% digital.
Marc LoPresti: (19:15)
That's an incredible statistic, right?
Thomas J. Lee: (19:17)
Yeah. And it probably goes to 95% in the following 20-year interval.
Marc LoPresti: (19:21)
Yeah. Yeah.
Thomas J. Lee: (19:22)
Which means alternative data becomes way more comprehensive than what the BEA collects, or what you can see in weekly claims. The cadence of the data is so different. So I think to us, what we think, and we brought in a new guy, Adam Gould, who's from Empirical Research, but he's going to be doing a lot of machine learning work, is that the shelf life of a lot of the products we develop might be quite short, too.
Thomas J. Lee: (19:47)
And of course, that's why we would want to use guys that you guys have on your platform, because it saves us the homework. Because now, we can just spend more time trying to curate it or understand it.
Marc LoPresti: (19:58)
Charles, so how are you positioning Knoema to help your clients address these challenges?
Charles Poliacof: (20:03)
So I do want to just address one point that Tom just made. I would say that on that last point, data tends to be a bit regime dependent. So you will see data assets become popular during a certain time period. And they ebb and flow. I would say three years ago, macro data was probably more focused for macro strategists. Now, macro data is a part of just about every strategy that's out there. So you'll see these data assets become more relevant or less relevant, depending on what regimes and themes are being surfaced.
Charles Poliacof: (20:34)
With respect to our clientele, we work with a really wide range of clients. So it's not just buy-side firms. It's buy-side firms. It's sell-side firms. It's corporations. It's wealth managers. It's government agencies. We work with data providers as well. So I would say there's two fold. I think one, for large organizations combining legacy data or first-party data assets with third-party data assets continues to be a challenge. And folks are really started figuring out how to optimize for that.
Charles Poliacof: (21:07)
You have legacy infrastructure, legacy technologies, and to some extent, some legacy thinking. So for example, in the wealth management space, I think a lot of people are starting to really try to understand what it means to have a customer 360 view. And what are those inputs that they can start to use? So, what are the primary datasets about my customer that I've already collected, that has to live inside some kind of a clean room? And then what kind of third-party data assets can I collect to have a more informed view? So investing preferences.
Charles Poliacof: (21:36)
Does my customer care about ESG, or socially important companies, or environmentally conscious companies? So if they do, I want to make sure that I start sending campaigns that are going to be interesting to them. So you have folks that are starting to build these data-driven practices. You're seeing this a lot in retail. You're seeing this a lot on the supply chain side. People really just trying to understand how they can react to supply chain disruption on the corporate side. So it goes on and on, and on, and on.
Charles Poliacof: (22:04)
With respect to the data providers side, I think data providers need to start to think what it means to offer a data product that is consumable. So, how do I create a data product? What does it mean to tickerize that product? What does it mean to start thinking about omni-channel distribution, so that my data could be consumed through various exchanges, through various portals, through various tools? You want to be a gateway and not a gate keeper. If you're a gatekeeper you're ... Sorry.
Marc LoPresti: (22:34)
No, please. No, it's really about making that process of discovery and ingestion doable, because you mentioned tickerization. We've been talking about that since the beginning of the data industry. Nobody's done it yet, right?
Charles Poliacof: (22:49)
Or even entity resolution.
Marc LoPresti: (22:50)
Right.
Charles Poliacof: (22:50)
So if I'm looking at Athleta, I want to know that that maps up to Gap. And somebody is taking the time to actually do that work. So I think on the data productization side, there's still some work to do. And that's what we found that we've been doing work to help customers. For example, on the Snowflake marketplace, we're a partner with Snowflake. And we're one of the largest contributors of public data over there.
Charles Poliacof: (23:10)
But we're also working with alternative data vendors to actually productize their data on the marketplace. So, what does it mean to actually create a data product? Great metadata tables, source data tables, all those things. It's not sexy work, but it's necessary work so that people can have access to these insights.
Marc LoPresti: (23:29)
And as we expand, as we've seen with our business at BattleFin, from when we started it to today, that evolution beyond just the hedge fund as the main consumer of alternative data, or even the financial services world, generally, as we expand out into more corporate use cases, government and NGO use cases, that prospect or that challenge of ingestion, standardization, and consumability and visualization, which we could do a panel just on that, becomes much more relevant. Heidi, challenges.
Heidi Lanford: (24:03)
Yeah.
Marc LoPresti: (24:03)
There's been a lot, and I'm going to give you a little hint on something I want you to touch on. There was huge, huge news in the data industry yesterday. There was an SEC settlement, the first time ever, ever. And I've been talking about it. For those of you that have the misfortune of listening to the regulatory and legal panels that I do for BattleFin over live or in person, or over Zoom, which are a real snooze if you can't fall asleep, this was big.
Marc LoPresti: (24:30)
We were talking about this. Is the Gensler Administration going to be the first to actually bring an enforcement action against a participant in the data industry? How do you see regulation, changing regulation as one of the challenges that you face in this massive project that you've undertaken?
Heidi Lanford: (24:46)
It definitely affects us because obviously part of our business on the rating side is highly regulated, and then part of our business on our solution side is not as regulated.
Marc LoPresti: (24:57)
Right.
Heidi Lanford: (24:57)
And I guess with everything, I think there's pros and cons to both. One of the benefits that a lot of people get from regulation if done well, is standardization and consistency.
Marc LoPresti: (25:10)
Yeah. Right.
Heidi Lanford: (25:11)
We've all read every week, an article in the paper about ESG, and how ESG scores are sometimes challenging to interpret and understand like for like, because they're done differently.
Marc LoPresti: (25:24)
There's no standardization.
Heidi Lanford: (25:25)
And I'm not suggesting that we regulate that, but Fitch just released a sustainable Fitch product actually today. And so we're all doing a lot in the ESG space, and we think we've got an innovative way to look at that. The con though, is if you over-regulate it and you've not necessarily got the right people who understand data who are making these regulations, becomes really challenging then, too.
Heidi Lanford: (25:52)
It's all about, we want to get the benefits of data. We want data to give us insights so that we can make better decisions. And if it becomes too bureaucratic in making those laws, that can hinder us. But I did want to actually touch on a topic that we've been talking about. And that's Tom talked a lot about, inference. And it's the classic causation and correlation thing.
Heidi Lanford: (26:16)
I'm not trying to steer things a different direction, but I think this really gets down to building a culture of data literacy within your organization. So again, and I don't know if data literacy has been a big topic as you've talked about in the past couple of days here, but data literacy is not creating a bunch of PhD data scientists in your organization.
Heidi Lanford: (26:41)
What it is, is educating those consumers of information so they can make the best decision possible and take action on it. That might mean offering some training and education on how to deal with missing values or data that's a little sketchy, or understanding the difference between causation and correlation and when to phone up a data scientist in your organization.
Heidi Lanford: (27:08)
And that's another part of my job, is to actually build out a data literacy program for our entire company, to get folks essentially comfortable and confident enough to work with data, and make those data-driven decisions. And it's not just a learning and development program. It is a cultural mind shift. And it's a thing now. People are relating data literacy to, it's like data as a second language. There are various dialects within data. There are levels of proficiency.
Heidi Lanford: (27:37)
There could be a fluent I dream in data versus I have conversational knowledge of data. And that's okay. But that's what this shift is about. And all these challenges we've been talking about, about integrating data and the platforms, they are going to be here for the next foreseeable future. It's just, are we going to get better at it using awesome tools like knowledge graph, and AI and ML, and RPA, and things like that?
Marc LoPresti: (28:05)
And challenging in particular, talking about the regulation or regulatory perspective, when we have an alphabet soup or dog's breakfast of regulators. Whether it's the FCC from the consumer privacy, from the SEC.
Marc LoPresti: (28:23)
It's not just one entity. We've got unfortunately, like a minute, 20 seconds on the clock. So in 30 seconds or less, starting with you, Tom, 2022, biggest thing you expect to see changes in alternative data. What's in store for us?
Thomas J. Lee: (28:41)
Wow. In 30 seconds? I'll just say, I think that the next 12 months are going to be really critical. And again, Stevie Cohen mentioned it yesterday. It's no longer going to be a macro market. This is going to be single, stock winners and losers, operating leverage, people connect with customers. This is really important to get the alternative data sources correct.
Marc LoPresti: (29:06)
Fantastic. Fantastic. Well, I want to thank ... I was going to do for all of you, but unfortunately I don't think we have enough time. I'm getting the nod from backstage. I want to thank all of you for joining us this afternoon and listening to this unbelievable panel. Charles, Tom, Heidi, thank you for agreeing to do this. It was absolutely my pleasure.
Marc LoPresti: (29:24)
We hope that we taught you something about alternative data and the evolution of the space. I'm sure all of my fellow panelists would be available to answer any questions that you have after the panel. And I would be remiss if I didn't invite you all. If you haven't already done so, well, hell if you have go again, come visit us at BattleFin on the fourth floor.