Viewpoints

Desire Lines in Real-Time: How Non-Real Estate Data can Predict Real Estate Market Trends

Posted on June 19, 2015 in Viewpoints by admin

By Melissa Reagen
Head of Research & Strategy, MetLife Real Estate Investors

For real estate investors, an expanding number of richer data sets can be harnessed to help predict commercial property performance well ahead of existing market fundamentals. Census data produced with a lag can hinder an investor’s ability to get ahead of market moving trends. We highlight how bike sharing data and satellite imagery can help predict real estate demand ahead of traditional market fundamentals data. We also predict how cell phone location data and data from sensors may help investors stay ahead of secular changes in demand for apartments, offices, warehouses and retail properties.

 

The Profitability of Big Data

Consumers and businesses are storing enough new data annually to fill 60,000 U.S. Libraries of Congress, according to McKinsey’s 2011 report, “Big Data: The next frontier for innovation”. According to Cisco, the amount of data flowing through networks will increase from 62 exabytes of data per month in 2014 to 132 exabytes per month in 2018 as technology allows for more information to be gathered on everything from consumer spending patterns to healthcare records.

 

Non-Real Estate Figure 1

Figure 1 — Source: Cisco VNI: Forecast and Methodology, 2013-2018 and The Zettabyte Era—Trends and Analysis

 

The retail and healthcare industries have already made considerable progress in gathering and analyzing data on customers and patients to improve retailers’ profitability and healthcare treatment outcomes. Retailers are collecting customer data from online transactions, social media, loyalty programs and radio frequency identification (RFID) tags in order to identify shopping patterns and segment customers. Beacons track customer purchases to tailor each individual consumer’s shopping experience in the store. Similarly, RotaryView and RichRelevance are attempting to recreate the in-store shopping experience online with a personalized touch by harnessing customer data. Walmart’s mobile app allows users to create shopping lists by scanning bar codes while they shop and are directed to the appropriate aisles. Amazon is using the data they gather from their package and vehicle tracking devices to implement smaller local sortation facilities to reduce the amount of time needed to ship goods to consumers.

In the healthcare industry, IBM researchers are developing “Dr. Watson”, a robot that can help doctors better diagnose and recommend treatments by analyzing big data sets on diseases, their symptoms and treatments. According to Harvard Business Review, companies making data-driven decisions are 5% more productive and 6% more profitable than their competitors who are not harnessing data.

 

‘Build the Bike Stations and the People will Come’

Existing data from real estate fundamentals or the U.S. Census Bureau have been immensely useful in understanding and forecasting trends in real estate performance. Much of this data, however, is released with a time lag, making it difficult to be ahead of market-moving trends. In our view, the Washington, D.C. bike-sharing data could have helped predict the rise of neighborhoods such as the H Street Corridor and NoMa well ahead of traditional data sets, providing investors with a profitable first-mover advantage.

The bike-sharing data shows that a handful of stations in NoMa and the H Street Corridor were the ones experiencing the largest growth in ridership between 2010 and 2011. These stations saw four to five times more riders in 2011 than in 2010. By the end of 2011, the number of rides starting in NoMa and the H Street Corridor were in the top half of the most highly trafficked bike stations. It was also a time when, not coincidently, NoMa and the H Street Corridor were starting to see demand from the capital’s young population.

From a forward-looking perspective, gathering more information on bike sharers such as age, their residential zip codes and purpose of the trip could help predict whether strong demand for urban apartments will continue or not, a much debated topic in real estate circles. A significant decrease in bike rides in current residential hotspots for the young could help identify a shift in urban living preferences before these trends show up in the market fundamentals or U.S. Census Bureau data. Further, gathering data from apps such as Uber or location data from cell phones and combining it with the bike share data could help real estate investors understand where the rising hotspots are in the urban and suburban areas.

Data from satellite imagery also holds promise for informing real estate investment decisions. According to the Wall Street Journal article, “Startups Mine Market-Moving Data from Fields, Parking Lots”, Orbital captures real-time satellite images of items such as cornfields to predict crop yields and parking lots to predict retail sales well ahead of the market. For the parking lot data, Orbital bought one million images from satellite companies. Early signs suggest Orbital’s data set hold predictive power. The company accurately forecasted Ross Stores Inc. would beat 2014 earnings based on an increase in the number of cars in parking lots across the Ross portfolio. For real estate investors, understanding retailers’ sales performance in real time could provide insight into how well a retail center is performing in order to inform either an acquisition or disposition decision.

 

Integrated Analytics: A Sophisticated Investment Intelligence

Understanding current data limitations, apartment owners and operators are implementing business intelligence systems to track detailed tenant information such as the tenant’s current employer, title, pay and reason for moving in and out. Aggregating this information at the portfolio level can help identify reasons for systematic out or underperformance of apartment properties. Perhaps even more important, apartment owner and operators will take their cues from retailers and start capturing more data on tenant preferences to help them stay ahead of demand trends. Gathering more data on tenant preferences could help foreshadow whether millennials will stay renting in urban areas, for how long, under what familial circumstances and based on what income level.

The data sets mentioned above are just the beginning for real estate data analytics. The International Data Corporation (IDC) reports that RFID tags, sensors and supercomputers are generating the majority of data that can be analyzed by businesses. They also estimate that by 2020, 37% of this data could generate business insights, if properly collated and analyzed. From a futuristic view, the sensors currently attached to trucks, trains, and ships could produce real time data on the flow and types of goods being shipped and could identify changes in logistics. The changes in logistics could help foretell, among other things, how goods being manufactured by 3D printers are affecting the amount and type of warehouse space needed. Further, the ability to track what types of goods are being delivered to what types of consumers could be powerful. It could help determine which goods consumers purchase from their local physical stores and which they purchase online in order to tailor neighborhood retail shops to local consumer preferences.

In a nascent but telling trend, Shutterstock figured out that moving their office space from the Financial District in New York to the Empire State Building would result in 4,500 minutes per week of saved time in commuting by gathering data on where their employees live. If more office occupiers choose to locate along optimal commuting paths, real estate investors could potentially figure out where future demand for office space lies across urban and suburban areas. Bike sharing or cell phone location data could be useful in helping figure out workers’ commuting patterns and thus useful in determining future demand for office space.

The sheer size and quality of data coming from technological advancements in the next five to ten years will allow investors to discover trends that may have otherwise gone unnoticed. For real estate investors, gathering and analyzing this data could help predict commercial property performance well ahead of what existing data sets can provide. More importantly, gathering of larger and richer data sets could help investors stay ahead of changes in demand for commercial property of all types.