Sequential pattern mining is a powerful data mining technique for finding time-related behaviour in sequence databases. In this paper, we focus on mining sequential patterns in the business-to-business (B2B) environment. Because customers' sequences in the B2B environment are very long, and almost all items are frequently purchased by all customers, using traditional methods may result in a large number of uninteresting and meaningless patterns and a long computational time. To solve these problems, we introduce three conditions (constraints) - compactness, repetition, and recency - and consider them jointly with frequency in selecting sequential patterns. An efficient algorithm is developed to discover frequent sequential patterns which satisfy the conditions. Empirical results show that the proposed method is computationally efficient and effective in extracting useful sequential patterns in the B2B environment.