As customer loyalty becomes a key to the profitability, companies in game industry have shifted its marketing focus from acquisition to retention of customers. The present study investigates the factors presumably affecting reacquisition of the lost customers using actual transactional data. The research interest lies in not just regaining the lost customers but also keeping them. One of the objectives of the current paper is to figure out who would “stay alive” (i.e. keep using the service) after responding to the reacquisition campaign. Since few customers actively respond to a reacquisition campaign, the distribution of the response measurement is highly skewed. To handle this problem, this study uses “quantile regression (QR)” method in estimating the model. For the analyses of gamers’ real behavior, a dataset on one of the most successful online games in Korea, Sudden Attack, was utilized. This study focused on the customers who became inactive, i.e., no log-ins, during the 12-week period of April 19th to July 11th, 2012. Some of them returned when the win-back campaign was conducted beginning July 12th until August 8th (“Period 1”), and others didn’t. And further, some of those who returned stayed active (i.e., logged in) during the 4-week period after the campaign was over (“Period 2”), and others left again. In each of the four cases, a random sample of 1,000 users was drawn for the analysis. The estimation model includes four sets of variables: demographic variables (age, location), RFM variables (recency, frequency, monetary value), behavioral variables (level, experience, number of chats, kill per death ratio), and social variables (number of friends, number of gifts). The analyses confirmed their linear and/or non-linear effects in Period 1(win-back) and Period 2 (retention).