In the conditions of digital transformation of the catering market, restaurants need to know the typical behavioural patterns of consumers of restaurant services in order to choose the most promising target audience and build the right strategy of interaction with it. The goal of the present paper is to introduce a taxonomy of models of restaurant customers’ behavior. The taxonomy is based on the Weber’s method of ideal types. This study proposes a classification of these models according to the criteria of customer's willingness to visit a restaurant, willingness to spend time in a restaurant and willingness to order food in a restaurant. According to these criteria, the potential customer audience is divided into eight ideal types, each of which is characterised by its own model of interaction with the restaurant. A detailed description is given for each type and general recommendations for managing the restaurant's relationship with customers of different types are formulated. A review of the problems associated with building interaction between catering businesses and customers in the digital transformation of the market is given, and the need to solve these problems is noted. A promising direction for further research is the cluster analysis of the client audience on the basis of the proposed classification of consumer behaviour models.