Airbnb’s Crystal Ball: Predicting User Bookings and Spend with New Patent

Coffee and A-Frames

Airbnb is planning to patent a predictive system to determine if a user intends to make a booking and how much they might spend. The system will gather user-specific data such as account history, browsing activity, and previous bookings within a certain time frame. This data feeds into a “predictive intent model” to calculate the probability of a booking, and a “predictive value model” to estimate the potential spending.

The viewing history heavily influences the intent model, while price and market data are crucial for the value model. The results of these models can assist Airbnb in strategizing its user acquisition for various home-sharing markets.

This prediction system would allow Airbnb to optimize its ad spending to achieve a higher return on investment. This approach contrasts with the current method of relying on advertiser intuition and market knowledge, which requires human involvement.

This patent is not unique to Airbnb. Other companies such as Walmart, eBay, Uber, and Visa are also developing similar technology for hyper-targeted advertisements.

Airbnb, which posted its first profit in 2022, might face challenges ahead as it predicts a drop in travel levels. This is due to the temporary rapid growth created by the pent-up demand for travel post-pandemic in 2022. Data from AllTheRooms suggested that top Airbnb cities like Phoenix and Austin saw a nearly 50% drop in revenue per available listing in May 2023. However, Airbnb disputed these figures.

Despite the travel industry’s slow recovery from the pandemic, Airbnb aims to maintain its momentum by focusing its advertising spend on users who are more likely to make a booking.