Churn management in hospitality
摘要
This research paper presents a novel churn management approach tailored to the hospitality industry. Focused on enhancing customer retention, the proposed model consists of two integral pillars: churn identification and proactive action. Leveraging the BG/NBD model as a foundation, the identification phase is innovatively extended to accommodate multiple interactions, ensuring a more accurate prediction of customer churn. The proactive strategy, utilizing a reinforcement learning framework, dynamically establishes optimal churn thresholds through analysis of responses of guests who received a targeted email with a fixed discount. A multi-week testing phase is included within a prominent North American hotel chain, yielding empirical insights into the model’s effectiveness. The results indicate that guests with a churn risk of 75% or higher should be targeted with a 20% discount email. The effectiveness of sending out the email from this churn risk becomes increasingly apparent.