The massive amount of capital chasing commercial real estate deals today has forced investors to seek out tools that give them an edge over the competition. One of the most effective tools is non-traditional data such as foot traffic, point-of-interest (POI), cellular movement, online reviews, walkability scores and school ratings, just to name a few.
In fact, some of the industry’s largest and most successful investors have embraced non-traditional data, layering it over data from traditional sources including CoStar, Real Capital Analytics, RealPage, REIS, Trepp and Yardi.
“Non-traditional data sets are a complement to the existing data sets,” says Shilp Shah, director of portfolio management, Americas, for Nuveen Real Estate. “Integrating traditional and non-traditional datasets will be key to making investment decisions in a competitive market and driving performance.”
Data journey just beginning
While other industries have been using non-traditional data to make strategic investment decisions for decades, the commercial real estate industry is just now embarking on that journey. In fact, the amount of non-traditional data has dramatically increased over what was available five to 10 years ago, according to Allan Swaringen, CEO of JLL Income Property Trust.
“As a whole, we believe the real estate industry is behind on using new technology and using non-traditional data sources,” Swaringen says. “This is the reason we have made a large push into the space and have spent considerable resources to try and expand the datasets we use and technology platforms we have established, as a way to differentiate the way we invest.”
Non-traditional data is crucial to JLL Income Trust’s underwriting and investment review, providing a broader understanding of the assets it’s pursuing for acquisition. The addition of these datasets gives the daily NAV REIT the ability to look at assets differently from its competitors. “We have found that analyzing non-traditional datasets can paint a different picture, help explain why certain headline risks may not be as relevant to a specific asset, or why there may be certain tailwinds to an asset long term that may be overlooked,” Swaringen notes.
In an effort to mine non-traditional data more effectively and efficiently, a number of commercial real estate companies have hired data scientists from other industries. For example, California-based investment firm Archer brought on a data scientist with a hedge fund background and built its own proprietary technology platform to help identify acquisition opportunities.
“In just a few short months, we’ve already seen how that expertise in hedge funds with non-traditional data provides insights that are at the forefront of acquisitions analysis,” says Fred Canney, co-founder and COO of Archer. “It’s early in this marathon, but there are pockets of really creative and great things happening with data in commercial real estate.”
Nuveen also sees the value in building a team of experienced data scientists to take advantage of non-traditional data. In 2019, the institutional investor launched an in-house data science group called Nuveen Labs, which helps “provide innovative ways to generate alpha and give Nuveen a competitive edge through application of artificial intelligence, machine learning, analytics and Big Data,” according to Shah.
Holistic sense of potential
While commercial real estate firms continue to look to traditional data sources to gauge asset occupancy, rental rates and valuations, they’re looking to non-traditional sources to provide greater insights into specific markets and properties.
“We’re also seeing a wave of data getting adopted in real estate that provides underwriters with a more holistic sense of a property’s potential, not simply its rent potential over the next few years,” says Nima Wedlake, principal of Thomvest Ventures, a San Francisco-based venture capital firm that invests in a variety of technologies.
For example, real estate investors are using cell phone data to track visits to retail centers and to monitor work-from-home versus office commuter trends, according to Archer’s Canney. Some firms have started to monitor credit reports to identify recent migration trends, as well as the real-time impact of the pandemic (and related stimulus payments) on renters’ ability to pay rent at various apartment communities across the nation.
“Non-traditional data sets are becoming more valuable as groups like Archer move beyond currently common datasets used in real estate to more up-to-date, live data sets that are common in other industries,” Canney points out.
Beyond cell phone data and credit reports, Wedlake points to Habidatum as a unique non-traditional data source. It provides rankings of a location’s commercial potential relative to other locations in a selected geography considering its centrality and catchment area.
Foot traffic and point-of-interest (POI) data providers like SafeGraph and Unacast are “really popular these days,” according to L.D. Salmanson, CEO and co-founder of New York-based Cherre, a real estate data management platform that connects disparate data sources in a single location.
Nuveen utilizes foot traffic data from Placer.ai to augment its traditional datasets on the retail sector. Ethan Chernofsky, vice president of marketing at Placer.ai, says the commercial real estate industry was the first to recognize the value of the company’s data.
“The commercial real estate industry sees itself as backward, but it’s not as bad as people think,” Chernofsky notes, adding that Placer.ai has grown to over 600 customers since coming to market in 2019 because of the initial buy-in from property owners and investors. “You have to give the commercial real estate community credit for embracing non-traditional data.”
Benefits of non-traditional datasets
Non-traditional data is fairly new to the commercial real estate world, so most investors are in the “trust but verify” stage, according to Canney. But there’s no denying that these datasets can be very helpful, especially when layered over traditional datasets.
Canney estimates that traditional datasets are about 60 percent accurate, due to the relative opacity of the information. However, the collective combination of various datasets, augmented with non-traditional data that can validate or invalidate certain data elements, can vastly improve accuracy to upwards of 90 percent in some cases, he notes.
Because non-traditional datasets provide real-time information that allows investors to make better and more timely decisions, they can be very beneficial when analyzing markets and assets, according to Nuveen’s Shah. However, the accuracy and usefulness of the non-traditional data source is dependent on the data and its application to specific sectors.
Nuveen, for example, has found success with causaLens and StratoDem Analytics. Based in the U.K., causaLens is an AI solution used to optimize organizations across many different industries. The solution understands cause and effect, which is a major step toward true AI, according to Shah.
StratoDem, meanwhile, uses a massive reservoir of government-sourced economic and demographic data to build predictive analytic models. These models help investors across the real estate universe prepare for the macro forces that shape market outcomes. “StratoDem has been a great non-traditional data source for Nuveen when making multifamily investment decisions,” Shah notes.
Nuveen’s core multifamily sector fund, the U.S. Cities Multifamily Fund, utilized StratoDem to incorporate migration patterns and demographic trends to evaluate acquisition opportunities in suburban Austin, Texas, as well as the Dr. Phillip’s submarket of Orlando, Fla.
Both markets have continued to see tremendous growth post-acquisition, Shah notes, adding that non-traditional data sources like StratoDem provide additional insight into a certain market and/or asset projection, which further solidifies the underwriting and investment decision.
Helpful or hype?
One of the biggest challenges related to non-traditional data relates to collection and processing procedures. Different types of data require different collection approaches, and depending on the industry, each data source requires its own set of collection procedures.
“If proper collection and processing procedures are not followed it would cause serious impediments to the utilization of the data sources,” Shah notes.
The sheer volume of data available today is also a challenge, and many real estate firms are not equipped to utilize such data, according to Wedlake. He points out that it’s also challenging to navigate through “a sea of novel data without a deep history of utilizing that data and back-testing its effectiveness.”
And adding more information, even from non-traditional datasets, doesn’t always clarify a situation. Simply identifying what data truly matters and determining out how to use that information can be difficult.
“Figuring out what is real, what is hype, and where data scientists should spend their time when it comes to non-traditional datasets is critical,” Canney says. “That can be the deciding factor whether value is truly gleaned from the data.”
