With the growing influence of travel review platforms, management responses to online reviews have gained increasing relevance for hotel online reputation management. While previous research has explored the impact of responding on subsequent ratings, there is limited understanding of how response characteristics influence guests’ decision to revisit, particularly for initially satisfied versus dissatisfied customers. This study investigates the influence of five key response attributes — response length, diversity, personalisation (topic consistency between the review and response), linguistic style matching (LSM), and sentiment polarity — on customer retention (revisit decision). Using review and response data from over 1,000 hotels on TripAdvisor, we employ text analysis to categorise response attributes and assess their effect on guest retention. Our findings reveal that response length, diversity, LSM, and sentiment positively influence customer retention. However, while response length, LSM, and diversity have a stronger positive impact on satisfied customers, they may negatively affect dissatisfied guests. Additionally, topic personalisation is found to have a negative influence, particularly among dissatisfied travellers. These insights offer practical guidance for hotel managers on writing responses that enhance customer retention by adopting an effective linguistic style.