Hotels face a significant volume of daily booking requests, both directly and through online travel agencies. However, only a small fraction of these requests translates into reservations, leading to low conversion rates. This factor introduces additional uncertainty into the hotel demand function and poses a challenge to revenue maximization efforts. Consequently, improving conversion rates is a key priority for hotel managers. This study seeks to introduce an innovative framework for optimizing conversion rates. It involves segmenting stay dates and employing a logistic regression model to predict the probability of conversion within each segment. By doing so, this research contributes a novel data mining methodology that can be integrated into machine learning algorithms to enhance conversion rate optimization.