AI Doesn't Replace Human Judgment in Retail Real Estate - Where We Buy #375

In this episode, James Cook interviews Paul Sill, head of JLL’s Visionary Insights Group, to explain how sophisticated predictive analytics are reshaping retail real estate decisions. By mining internal e‑commerce, reseller, and sales histories, the team builds models that forecast top‑line revenue for prospective sites, allowing clients to calculate NPV, IRR, and risk probabilities before signing leases. The discussion highlights a global running‑shoe brand that leveraged these tools to secure five letters of intent and launch its first stores in 2026, illustrating how data transforms gut‑based expansion plans into quantifiable business cases.

Sill emphasizes that raw correlations can mislead without context. He cites the counterintuitive finding that excessive traffic volumes often depress sales because congestion deters shoppers, and the classic ice‑cream‑shark‑bite analogy to stress correlation versus causation. Foot‑traffic and mobile geofence data, while valuable for existing locations, fail to predict performance for brand‑new sites lacking historical patterns. Consequently, successful site selection blends instinct with analytics, using data to validate or challenge intuition and to uncover hidden risk factors that might otherwise be ignored.

The conversation also tackles the limits of AI in retail site selection. Although machine‑learning techniques and emerging social‑media behavior datasets enrich the analytical toolbox, they cannot compensate for the fundamental bottleneck: a retailer’s relatively small sample of historical sales points. AI models require massive data volumes to produce reliable recommendations, a luxury most brick‑and‑mortar operators do not possess. Instead, firms rely on a hybrid approach—combining traditional demographic, competition, and site‑attribute data with newer sources—to craft the most accurate forecasts possible. This balanced strategy underscores why human judgment remains essential, even as technology continues to evolve.