The combination and application of big data and machine learning in the offline consumer market makes the formulation of marketing strategies more scientific. The use of machine learning can make statistics and analysis of users’ consumption behavior and classify customers according to users’ consumption behavior, so as to realize the personalized promotion of marketing content. In this context, this paper carries out research on enterprise digital precision marketing strategy based on big data. Starting from the actual application, this paper analyzes the current problems faced by consumption data sharing, as well as the characteristic needs of precision marketing for consumer groups, introduces homomorphic encryption technology, completes the structural design, process design, and algorithm design of the scheme in combination with the actual scenario that enterprises need to carry out precision marketing for customers based on consumption data, describes each design link in detail, and verifies the feasibility of the algorithm in the scheme.