At Compound, we believe that there are actually only a few relevant data points from which an investor can build a real estate investment strategy that is likely to outperform the broader market.  

More often than not, data creates a lot of useless noise.

The only thing more confusing than the sheer number of points is the number of real estate tech startups promising to weed through and make sense out of them.

But if you can get past the overuse of buzzwords (AI-driven data science enhanced by machine learning--with data mining robots!), there is some actual brilliance brewing. Today’s newsletter takes you on a scenic tour of this well-funded arms race.

Here are some of the companies we’re watching closely, categorized loosely:

Companies that clean widely available data feeds but make them easily digestible (fit for technophobe consumption):

For the residential sector:

  • Realyse - A UK-based real estate data startup with global ambitions and a really thoughtful founder, Gavriel Merkado.  They provide detailed search and comparisons for UK residential property investments. They’re rolling out across Europe soon.
  • HouseCanary - From what we can tell, this is a way to peel back the layers behind home valuations--almost a way to do a deeper dive into your home’s Zestimate.
  • Perchwell - An app for the NYC residential real estate market with tons of data about current listings, current and historical market trends.
  • Localize - “Reveal the truth about any home in NYC.” All the dirt you might want about a building you’re considering buying into, but also probably lots of things that brokers would prefer to sweep under the rug.  It’s a mobile app that is designed to ultimately become a feature on other brokerage and listing websites.
  • Zillow / Trulia / Streeteasy - No conversation about residential data would be complete without a nod to the forebears who wrote the playbook (and whet investors’ palates with dreams of billion-dollar exits in the real estate data space)

For the commercial sector:

  • Reonomy - Real estate data for the commercial industry, trying to take on industry stalwart Costar head-on but primarily pitching as a way for real estate-related companies to do lead-gen.
  • Cherre - The name is meant to allude to “cherry picking,” everybody’s favorite real estate pastime. They’re selling their data to the likes of homeowners’ insurance policy writers and real estate investment companies.
  • Bowery - A tech-enabled appraisal firm.

Specifically for the retail sector:  

  • Spatial - Transactional location data like census data and traffic data, specifically designed for retailers--who need all the help they can get right now.

If you need help visualizing data with pretty colors and maps and stuff:

  • Habidatum - An extremely compelling & powerful time-series urban data visualization platform -- also one of the only companies profiled that promises to use “space-time” which is a phrase I haven’t heard much since the decline of “space-age polymers.”  They promise to assist policymakers and businesses in understanding the “hyper-dynamic urban environment through advanced analysis of data in space-time.”  
  • Stamen - A SF-based consultancy specializing in real estate data visualization.

Companies using data to make investments:

Have we missed anybody?  If so, drop us a line.

At Compound, we use data when it actually enhances return--and not just to make us sound smart.

Researchers and data providers have identified thousands of factors that purport to predict real estate returns. However, very few factors provide actionable information and fewer still inform investment strategies likely to outperform the market.  

Compound’s investment strategy is based upon the following core principals:

  • Most real estate returns come from overall market performance and not an individual manager’s ability to generate alpha.
  • A very long history of returns, covering at least several decades, may provide a defensible forecast of a strategy’s return, but the forecast is still essentially useless.
  • Certain submarkets will outperform the broader market. Accurately predicting that future disequilibria is about identifying non-obvious factors and not from chasing past performance.
  • Although computers are better investors than humans, real estate transactions still require human interaction and the investment of human time. Reducing human bias in investment decision-making occurs when a data-driven investment filter (also called an algorithm) reduces thousands of investment opportunities down to a shorter list which can then be analyzed and negotiated by humans, preventing the selection biases that occur after a human has devoted a lot of time to a specific transaction.

We turn real estate investing into a science.

Fundamentally, real estate investing is a numbers and data game.  And yes, that means using data in a really intelligent way.