Data-driven Real Estate: BGO’s Shift to Local Market Focus
BGO, a real estate investment firm, discovered that the success or failure of its investments was overwhelmingly persistent by the specific local market chosen, reinforcing the industry’s long-held “location, location, location” mantra. This realization prompted a significant shift in strategy, prioritizing a deep understanding of local market fundamentals over broader property pricing trends and national economic indicators.
While numerous research firms offer rankings of local real estate markets, BGO found their assessments inconsistent.Instead, the firm turned inward, analyzing its own ancient performance to build a predictive model. This model backtests the factors that drove both its best and worst results, incorporating a wide range of localized data points – including demographic shifts and unique supply trends. Artificial intelligence was then leveraged to amplify the model’s capabilities, processing a considerably larger volume of data at a faster pace.
“We have taken thousands of data inputs, many freely available from government sources, and others purchased from providers like telecom companies. We’ve identified the key drivers,” explained BGO’s Carrafiell, emphasizing the model’s proven accuracy through rigorous backtesting.
This data-driven approach recently informed a prosperous investment in an industrial growth in Las Vegas, undertaken in partnership with Northpoint Development. Despite conventional research suggesting only mediocre performance, BGO’s model predicted substantial growth. The firm initially projected rents of $5.88 per square foot, but ultimately secured leases in the $9-per-square-foot range – a result Carrafiell attributes not to luck, but to the model’s insights.
The model identified a shift in logistics, noting that the Inland Empire of California was becoming prohibitively expensive. It then analyzed transportation routes and determined that Las Vegas offered a compelling alternative, providing significant cost savings in rent, taxes, and labor. “You had an extra two-hour drive, but you saved like 60% on your total cost, and that’s what the model saw,” Carrafiell stated. The tenants attracted to the Las Vegas development serve a regional market, not just the city itself.
BGO has replicated this analytical process for investments in Florida and the Rust Belt, consistently achieving strong returns.”We think our performance has materially increased as a result of this model,” Carrafiell confirmed.
However, he acknowledged the inherent limitations of any predictive model, stating that unforeseen events - such as a major company relocating – could disrupt even the most accurate forecasts.
BGO employs a dual-modeling approach: the investment team focuses on upside potential, while the lending team concentrates on downside risk assessment. Future iterations of the model will incorporate asset allocation strategies, aiming to identify optimal portfolio mixes across different commercial real estate sectors.
Carrafiell emphasizes that AI is not a replacement for data science, but rather an “enhancer and accelerator.” He describes the firm’s dedicated data science team as an integral component of the investment process, working directly alongside the CEO, asset management, and acquisitions teams.
