Investing with Origin

How Origin Uses Artificial Intelligence in Real Estate Investing


Location, location, location: That’s the mantra for success in real estate. For us, location is just the starting point for multifamily real estate investing. We talk frequently about our approach to markets, properties and fund management because we want our investors to understand how much analysis and expertise backs up a successful investment. And we want them to know we’re using the best data and tools available to support our real estate investing decisions. More and more, that includes artificial intelligence.  

When we study a market, we look at macroeconomic trends, national and local employment, population shifts, and supply and demand. We consider what types of multifamily buildings would work best in a particular location. And we study comparable existing properties and analyze market forces that could impact selling prices years down the road.  

Our dealmakers also spend time in neighborhoods and leverage our local contacts to learn as much as we can. They ask questions like: What are the shopping, entertainment and school districts like? What type of asset would work best on this block versus that one? We ensure that we are working with trusted, experienced partners. And we work out costs and potential returns using a broad spectrum of scenarios and a wide variety of data.  

For us, that “location, location, location” mantra is a multi-dimensional challenge to enhance communities, optimize value and generate outsized returns. But over the years, as our experience in developing and managing private multifamily real estate grew, so did our dissatisfaction with available data. It just wasn’t transparent or granular enough to accurately identify high-potential markets. So, we did something about it, with the help of a team of in-house data scientists. 

Introducing Multilytics: Origin’s Proprietary AI Research Tool 

The result is Origin MultilyticsSM, our proprietary suite of machine-learning models that forecast rent growth more accurately than the industry standard. Multilytics aggregates and analyzes more than 2.7 billion data points per month, from hundreds of sources, to generate alpha in multifamily real estate investment. Its predictive analytics allow us to create and study multifamily-specific scenarios. And combined with our own experience and local relationships, it allows us to identify and act on deal opportunities sooner than our competitors.  

Here are some of the main features of Multilytics:  

  • It incorporates variables specific to or geared toward the multifamily real estate sector, such as public and private sources that monitor economic and demographic trends, real-time data from property managers and spatial data such as local points of interest  
  • It slices submarkets smaller than a ZIP code to pinpoint high-growth clusters 
  • It identifies average annual rent growth down to the property level using artificial intelligence and modeling 

Back-testing Confirms Multilytics’ Accuracy 

The Multilytics model has been back-tested from 2008 onward, allowing us to better develop another critical dimension—time—in identifying outlier markets. For instance, its predictions about rent growth have proven to be accurate to within $10 to $15 annually, according to outcomes back-tested over a five-year period in the 150 largest metro areas. Also through back-testing, the model showed it had the ability to select high-potential markets like Phoenix years before others predicted that city’s boom. 

By zeroing in on granular trends rather than focusing more broadly on general ZIP code-level data, Multilytics finds clusters of high-potential submarkets, including areas that are beginning to gentrify. Short-term rent predictions and property-specific features help determine an asset’s value and support disposition decisions to maximize an investment’s value. Data points from hundreds of sources provide longer-term rent predictions after renovation to avoid overdevelopment.  

How Origin Uses Multilytics 

Multilytics is a key component of our strategy of using artificial intelligence to help find the highest-potential real estate investing opportunities. It also was the foundation of our Multifamily Markets to Watch 2022. For this report, Multilytics studied data from about 150 markets across the U.S. Leveraging the first-hand expertise of our acquisition officers in the markets in which we invest, we narrowed those markets down to our final list of five cities we expect will outperform historic rent averages for the next few years.  

We use it in our research of specific deals as well: We leaned heavily on Multilytics during our pursuit of Sawtell, a 40-acre tract in Atlanta that we determined could accommodate up to 2,000 units of development. In reviewing the spatial analytic output from Multilytics’ machine learning models, we learned that outsized rent growth relative to the broader Atlanta market was possible. We could see that neighborhood redevelopment was gaining momentum, but that data further corroborated our beliefs and increased our confidence in our decision to invest. Today, Sawtell is part of the portfolio of our QOZ Fund II.  

As our investors know, we never make an investment simply because it looks good on paper. We approach all potential investments with a conservative, risk-manager mindset. And our local expertise and specialist teams are critical in helping us find, evaluate and develop the right sites. Multilytics is a powerful artificial intelligence tool to give us deeper knowledge and insights in our pursuit of alpha in real estate investing.  

This article is intended for informational and educational purposes only and is not intended to provide, and should not be relied on, for investment, tax, legal or accounting advice. The information is provided as of the date indicated and is subject to change without notice. Origin Investments does not have any obligation to update the information contained herein. Certain information presented or relied upon in this article may come from third-party sources. We do not guarantee the accuracy or completeness of the information and may receive incorrect information from third-party providers.