Glossary

Predictive Analytics (CRE)

The application of statistical modeling and machine learning to historical and real-time CRE data to forecast future outcomes — rent growth, tenant default probability, cap rate movement, or optimal lease renewal timing — rather than simply reporting historical performance.

Predictive analytics in commercial real estate typically draws on the large datasets aggregated by comp databases, building automation and IoT sensor data, and portfolio management systems to build forward-looking models, distinguishing it from traditional market research, which is primarily descriptive and backward-looking. Common institutional applications include predictive maintenance (forecasting equipment failure from BAS sensor drift before a breakdown occurs), tenant retention scoring (estimating renewal probability to prioritize leasing team outreach), and submarket rent growth forecasting used to stress-test underwriting assumptions. The reliability of predictive models depends heavily on data quality and sample size, and CRE's relatively illiquid, low-transaction-volume markets (compared to public equities) mean predictive models are more prone to overfitting on thin historical data, which is why model outputs are generally used to prioritize human attention and flag outliers rather than to make autonomous investment decisions.