Considering the rapidly improving restaurant market, a customer’s intention to revisit particular restaurants have been consistently addressed. Because the intention to revisit is one of the most effective predictors of service success, marketing and management primarily focus on intention. For instance, Kumar et al.  explored customers’ intention to revisit online food delivery applications through two theoretical frameworks: the stimulus-organism-response and pleasure arousal dominance theories. Based on 446 responses, the mediating roles of pleasure and arousal as well as the indirect determinant of aesthetic formality on revisit intention were confirmed with good fit indices [RMSEA (root mean square error of approximation) = 0.06, CFI (comparative fit index) = 0.92].
Rajput and Gahfoor  studied customers’ intention to revisit fast food restaurants through the responses of 433 customers. Considering the conceptual research framework including food/service/environment quality and customer satisfaction, the structural results confirmed the direct motivations of customer satisfaction (0.528) and word of mouth (0.312), indirect factors, and the three quality dimensions.
Han et al.  investigated the relationship between revisit intention, perceived satisfaction, consumption emotions, and switching barriers through structural path analysis and qualitative approaches. With 406 validated samples in the United States, the moderating effect of customer satisfaction (\(\beta\) = 0.71) and the significant roles of comfort and annoyance on revisit intention were examined.
Meng and Choi  introduced a research model that analyzes customer’ behavioral intention to revisit theme restaurants, based on the theory of planned behavior. Based on the results of an on-site survey with 357 customers, it was concluded their attitude and involvement play a mediating role in determining their intention to revisit.
Although several scholars have focused on addressing customer revisit behavior , there are several limitations. One such limitation is that there are a huge number of factors, which can have notable impacts on the academic generalization of customer revisits. In addition, it is too difficult to address customer ‘actual revisits’ with traditional data-driven approaches.
Due to the rapidly improved data-driven approaches and technologies for analyzing customer behavior, several recent scholars have addressed the “exact” and ”direct” meanings of customer revisit. Furthermore, recent data analytics, as well as machine and deep learning approaches have allowed researchers to analyze customer revisits. For example, Hwang et al.  collected the responses of 133,872 airline service customers, extracted user experience features from customers’ responses, and investigated whether each customer revisits the same airline. User experience dimensions are based on machine learning approaches. They presented that an 83.42% accuracy is achieved in predicting customer revisits in terms of airline services.
Kim and Lee  also proposed a systematic framework for predicting customer revisit intention for seven flagship stores. Based on three feature groups—upcoming events, group movements, and store accessibility with approximately 3.75 unique customers in the stores—67–80% accuracy was achieved in predicting customer revisits using the XGBoost classifier.
Kim et al.  also proposed a deep neural network framework to address customer repurchase behavior. Considering a 2-year survey including 119,923 (the first round) and 74,088 (the second round) respondents, and the framework integrated by the three sub-modules of long short-term memory (LSTM) (customer comments), convolutional neural network (CNN) (model images), and deep convolutional neural network layers (evaluation rating with brands), 95.13% recall, 94.18% F1-score, 93.25% precision (same brand), and 90.71% accuracy were achieved in predicting whether each customer purchases a smartphone from the same brand.