Research on the Behavioral Classification of Consumer Brand Transformation Based on Big Data of Cigarette Consumption

With the emergence of big data technology, it provides new solutions for consumer behavior insights and innovative marketing strategies. This paper constructed a consumer behavior classification model based on the perspective of brand transformer from two dimensions: brand switch probability and price trend tendency, which was called CBST model. The NMF model was used to calculate the probability of consumer brand switch, and the ARIMA model was used to predict the trend of price change when consumers converted brands. Furthermore, the model effect was verified by taking the data of cigarette consumers' scan code brand record in a certain city for one year as an example. Specifically, consumer behavior can be divided into four types: high switch probability and high price tendency; low switch probability and high price tendency; low switch probability and low price tendency; high switch probability and low price tendency. For each type of consumer behavior, targeted marketing strategies could be formulated from three aspects: consumption trend, value of conversion, and attention degree, which could provide suggestions for effective customer relationship management and efficient allocation of marketing resources for enterprises.