This project consists of implementing a k-link clustering algorithm to optimize the allocation of product category amongst a set of warehouses to minimize the splitting of e-commerce orders and lead to logistics cost reduction. This has a wide range of different business applications particularly in data mining to find patterns from network optimization to market basket analysis, and in the world of finance for portfolio optimization. Below are a few paragraphs to give you context about the project.
In e-commerce, many customers favor companies that offer shipping costs. Indeed, the latter can represent a significant cost for the customer who can therefore cancel his order at the last moment. Surveys confirm this: according to a study conducted by PayPal and ComScore, 43% of French Internet users do not make a purchase, despite having put together a cart because of high shipping costs. We can therefore say that offering shipping costs allows to get rid of the main obstacle to online purchasing. In order to be more competitive and to attract more and more new customers, it is therefore interesting and judicious for a company to offer these costs related to delivery. But caution, this means that these costs will be charged to the company, which can impact the profitability and therefore implicitly the sustainability of the business. The decision to offer delivery costs should not be taken lightly and must be based on a solid, successful, and controlled strategy.
So, how to offer shipping costs while remaining profitable?
In this project, we addressed this issue by studying one of the major problems related to delivery in the e-commerce domain, the "splitting order". This phenomenon occurs when an order includes several items, often of different categories, stored in different warehouses. In this case, it is impossible to avoid splitting orders and therefore the multiplication of warehouse-to-customer shipments, thus considerably increasing the delivery cost. If this phenomenon is not mastered and studied upstream by the company, two cases appear:
- Impossible for the company to offer delivery to its customers leading to a significant loss of earnings.
- The company will become less profitable.
Both cases inevitably lead to an increase in costs and a decrease in customer satisfaction, two key notions for the survival of a company and its development.