This paper discusses the systematic application of the latent class model on business analytics. The latent class model is one of the effective statistical model classes on business analytics to represent essential statistical structures by learning the sparse and high dimensional data. This model class is useful for the purpose of reduction of feature dimension and cancellation of sparseness of the data. This is because many practical data can be assumed to consist of several unobserved subgroups. For example, a customers set, which is a target in the field of marketing analysis, consists of several subgroups with different characteristics and preferences. In addition, data clustering can also be realized by estimated belonging probabilities to latent classes of each data.This paper gives a general form of the latent class model and discuss how to construct the model structure and apply to a real problem in business analytics. After describing important points to be noted in analysis based on a latent class model, several practical examples are also shown.