This blog post from Airbnb discusses the implementation of Proximity Features to address the cold-start problem in personalization for users who do not have a login or prior history. By analyzing geographic signals from nearby users rather than relying on individual identifiers, Airbnb effectively personalizes search results and recommendations for new users. The post details the methodology behind grouping users into proximity buckets and the adaptive clustering algorithm used to maintain effective personalization while adhering to privacy regulations. Personalization is achieved without persistent user identifiers, transforming how cold-start users receive tailored experiences on the platform.