Data Tools and Trends in underwriting Retail Real Estate

Due to the COVID-19 pandemic, there has been a decrease in demand for retail, resulting in the foreclosures of many stores nationwide. However there is still hope, retailers have recently begun to move forward with store renovations and negotiating rent reliefs. With new funds launched, many seek to capitalize on the expected wave of COVID-related foreclosures, although only a handful seek to capitalize on retail. New funds have launched and many seek to capitalize on the expected wave of Covid-related foreclosures, but only a handful seek to capitalize on retail. 

Tremendous disruption has come to the real estate sector in the past, creating uncertainty. That uncertainty generally leads investors to examine their risk profile, but it also generates new opportunities—a strategy that retail real estate investors can be banking on these days. 

The retail acquisition strategy in current times would only work with carefully selected high-quality markets with strong foot traffic, whereas it should be based on long-term rent collection, not asset disposition. For that tenant credit risk underwriting matters a lot, thus machine learning machine learning techniques can help with that. The main premise of machine learning is that asset values are affected by an unlimited number of macroeconomic and space-investment market factors.

The main machine learning methodology for underwriting retail acquisitions in the environment of high uncertainty consists of: 1) Determining a set of factors that correlates best for each individual retail property and 2) predicting the value of the retail asset with the most accuracy based on said list of factors. In general, these factors can be grouped into three data categories related to: location, tenant/consumer profiles, and property itself.

Some novel techniques and data ideas that might aid in retail property underwriting: 

Spatial – GIS mapping and location intelligence. Analyze data streams (e.g. MSA average income, foot traffic, drive times, neighborhoods profiles, shopping preferences) to optimize retail stores footprint and the tenant mix for each retail property, predict retail sales for that tenant mix and underwrite revenue (and store closings) potential, assess competitive threats. Check if proximity to certain retailers enhance property value to a higher degree than others. 

Social media data – create neighborhood and customer profiles (e.g. use Natural Language Processing to understand customer interests from Meetup platform in each neighborhood for the optimal tenant mix). Analysis might suggest it makes sense to prioritize specialty retailers over traditional ones.

Heatmaps and location trends – look at standard metrics such as lease expirations, rental income, property values and sales, occupancy rates per MS. But then add another layer of trends analysis to score locations and identify which MSAs to go after. 

Machine learning can become an indispensable tool and a competitive advantage in helping landlords evaluate potential acquisition opportunities.

Feel free to reach out if you want to discuss machine learning implementation in your investment strategies.