Brandon Taubman Shares the Role of Data Science in Business with Stablewood Properties
When it comes to applying data science and analytics to business decision making, there’s perhaps no one more adept in the industry than Brandon Taubman. His commercial real estate firm, Stablewood Properties, is making waves by combining subject matter expertise and technology to make more precise and efficient investment decisions. This effort was spurred by Taubman’s 15+ years of experience working with data science to guide investments from Wall Street to the baseball diamond.
Taubman’s career began with a love for using numbers to develop informed decisions, prompting him to pursue a Bachelor of Science in Applied Economics Management from Cornell University. This gave him the opportunity to embark in speculative investment and equity derivatives for both Ernst & Young and Barclays Investment Bank. While at Barclays, Taubman rose through the ranks to a role as assistant vice president on the equity derivatives team.
A Long Island native, Taubman continued to nurture his love of Mets baseball and fantasy sports by developing his own data analytics driven modeling for daily fantasy baseball team management. The platform caught the attention of those in Major League Baseball, which at the time was beginning to turn toward more analytics-driven decision making in the tradition of the recently popularized Moneyball mindset. Taubman took the opportunity to join the Houston Astros, where he helped revolutionize their approach to talent management and scouting using a blend of eyes on the field and data science in the front office. The result of these efforts was the Astros first franchise World Series Championship in 2017.
Most recently, Taubman has brought his data science expertise to the field of commercial real estate, demonstrating the power that analytics can have in any industry when applied properly.
Brandon, can you tell anyone who might not be familiar with your firm what Stablewood Properties is all about and your role as chief information officer? And what excites you the most about this role and opportunity?
Stablewood is a private equity real estate shop. We invest institutional capital in commercial real estate assets, but what makes us truly special is the technology and the data capabilities that we have. The company itself was formed by a team of data scientists: half data scientists and half real estate veterans. We’re trying to make every step of the real estate investment process better, faster, and more efficient for the investor. We’re currently in what is called the net lease space, but we’re actively working on new strategies since our technology is truly broadband. As a young company, we have a lot of growth potential because our technology can do so many different things.
What excites me the most in real estate is the same thing that challenges me the most. It’s the data quality. Unlike in baseball, where you have immaculate records of in-game events, the data in real estate is highly fragmented. For example, in baseball, we have roughly 25 million pitches thrown since the creation of pitch tracking technology. We have decades of performance records. It’s rare that you don’t have the data that you need to answer certain questions. Conversely in real estate at Stablewood, private transactions dominate deal flow, and data aggregators are well behind the curve. The data is incomplete and inexact, and that means building accurate and reliable systems and models is more difficult. It also means there are higher barriers of entry for the competition. There’s opportunity for us to figure it out and get a huge head start.
What is it that originally sparked your interest in data science?
My desire to be a better investor is the short answer. If you think about my 15-year career, I’ve strived to be and maybe hopefully I’ve already become, a great investor. Whether it’s derivatives, such as the ones that I transacted on early in my career at Barclays Investment Bank and at Ernst & Young, or it’s baseball players or real estate, I see an opportunity to use data to make better decisions.
To put it another way, data in my view is a structured interpretation of history. As we all know since days of grade school, history is likely to repeat itself. With data, you can build models that help you understand what events or outcomes might be coming in the future. If you do that well, you’ll be ahead of the competition.
Prior to Stablewood Properties, you worked as the assistant general manager for the Houston Astros. Why is data science crucial to managing baseball operations?
Data science was very much at the heart of our operations across baseball functions, player development, scouting, and research and development. When we started the rebuild, data was most relevant in the creation of our draft model. Much like Billy Beane and the A’s during the so-called Moneyball revolution, the idea for us was to source historic structured data to better predict the future of amateur players, including players coming out of high school, junior college, and college, so that we could better understand whether we should draft them and what their career outcomes would be.
For example, in the international amateur space, we used data from tools like pitch tracking and a host of wearable physiological monitors to better understand and project the career outcomes of young international players. In this way, the Astros’ data revolution was very similar to what A’s had done the decade prior.
The difference is the Astros took it a step further. We said, “What if we can use data to proactively change and improve the career outcomes of our players?” Rather than taking a passive role in player selection and evaluation, what if we empower the players directly with world-class data and tools so that they could be the best versions of themselves?”
I remember traveling around the country to all of our affiliates to meet with every player in our organization and to talk with the players directly about the player development opportunities that we saw in the data. Initially, there were both good and bad conversations, but ultimately it led to a tremendous amount of success for the organization.
In my view, our willingness to empower the players directly through data was the main reason that the Astros had as much success as they did over recent history. This is an organization that’s produced so much talent from its farm system and has helped to turn a bunch of lesser-known players with lesser expectations into world-class athletes who, to present day, continue to contribute wins at the major league level. It was truly a team effort among players, analysts, and coaches.
When you joined the team back in 2013, the Astros had averaged a 54–108 win/loss record over a three-year period of time. By 2017, Houston won the World Series. What technology did you pioneer during that time and what role did your analytics crew have to do with this turnaround?
Quite a lot. I had the opportunity to work with a lot of great people at the Astros as I rose from analyst to assistant general manager. With the growth of our analytics effort over the course of an eight-year period, we eventually outgrew the nerd cave, and it became more like a nerd trading floor. It is fair to say that we went from truly pathetic, to be frank, in terms of the quality of the baseball operation, minor league farm system, and major league team, to one of the most successful sports organizations in history.
Analytics was truly embraced by the organization leadership, but at first, not so much by the players and coaching staff. Over the course of time, as our analytics initiatives yielded positive results, the trust built up. One great example is with the infield shifts. In 2013, we were shifting more times and more dramatically than any team in baseball. We created roughly four wins a year extra with our shifting, which is just a tactical decision about where to position infielders defensively. At first, the pitchers were uncomfortable with it, and there was a lot of push back. But I think that as they saw more of the ground balls that they yielded converted into outs, they bought in, and now every team in baseball shifts as much as the Astros.
Another example is the utilization of this pitch-level data. We discovered applications to use data points such as the velocity or vertical and horizontal break of a pitch to help pitchers be the best versions of themselves. At first skepticism came out of the analysis, which was performed on 20 million pitches thrown over history. But then we had some continued success with the likes of Collin McHugh, Charlie Morton, Gerrit Cole, and Justin Verlander. With that success, the buy-in went up like a hockey stick.
Lastly, how does data science aid your decision-making on commercial real estate investments, and what excites you the most about applying machine learning to traditional investing?
What excites me is that there’s a wealth of data out there that most institutions and market participants are not making great use of, and so there’s an opportunity with data science to get ahead of the competition and make better investments than the competition.
That’s a short answer. Here’s an example. On the buying side at Stablewood, when we actually go to buy assets in our net lease space, it’s probably a $20 billion market per year. We have a proprietary system that utilizes what we call AI-assisted underwriting to underwrite every deal on the market using data and basic deal terms to let us know what the prospective returns are on any given asset. We’re the only player in our space doing that. Other shops will selectively look at and underwrite deals that may fit their investment criteria. We look at everything.
With technology, we can take a boil-the-ocean approach. With all of our predictive analytics, we are able to understand where land value is likely to increase or where income and population are likely to increase in the future. So we can look at all of the assets out there and pick up the top 5 percent in the market at lightning speed, and with the opportunity for investors to have outsized growth potential or returns because we’re picking them better. That’s an example on the buy side of how we are using data science in commercial real estate investments, but we’re also closing on, managing, and selling these assets. You can imagine that we have different software and applications that support us doing business better, cheaper, faster at every step along the way.
Be sure to follow Brandon Taubman on Twitter
Originally published at https://techbullion.com on January 19, 2022.