Nightly pricing in short-term rentals can be informed by both a property’s position in the market and its own price–demand history. However, for new listings, property specific operating histories are unavailable, creating a cold start pricing problem. This study develops a hedonic pricing model using London Airbnb listing data to provide an initial market pricing benchmark. Statistical and machine learning approaches are compared, with XGBoost providing the strongest predictive performance. Forward predictions from the hedonic pricing model are subsequently compared with predictions from independently developed property specific dynamic pricing models for two London short let properties in active commercial operation. The degree of alignment between the two models differs across the two properties, suggesting potential property level heterogeneity. More broadly, the two models are interpreted as complementary: hedonic pricing provides an initial market benchmark when property specific information is scarce, while accumulated pricing and demand outcomes allow dynamic pricing to subsequently inform property specific listed price adjustments in practice. However, neither model should be considered to provide an identified revenue-optimal price.