Intelligent aggregation of purchase orders in e-procurement

A large enterprise generates millions of purchase orders (PO) each year buying various types of goods and services. Each PO has a cost associated with it. This cost comprises multiple elements including the price of the good or service, the shipping and handling of the purchase, and the overhead in initiating, generating, tracking, and managing the PO. To reduce the cost of doing business, it is imperative to reduce the total cost of POs in enterprise e-procurement in an automated fashion. One way to reduce enterprise procurement cost is to aggregate demands so that the total cost of a bunch of POs is reduced by a better price, a lowered shipping and handling fee, and a reduced overhead. The cost of goods and services often depend on several factors including volume, timing, and other business objectives. This paper describes an intelligent aggregation approach for automatically aggregating demands to reduce procurement cost in enterprise e-procurement. Our aggregation approach for e-procurement consists of an information model for representing products (goods or services) and representing purchase orders for such products, a corporate agreement system, a negotiation engine, and a rule-based aggregation engine. The information model is based on an extension of the classic entity-relationship model. The extension enables association of rules and constraints with and among attributes. These rules and constraints must be satisfied during PO aggregation and thus ensure the aggregate PO to be consistent with original individual POs. A rule-based aggregation engine examines POs as they arrive and interact with other decision aids to determine whether aggregation of a particular bunch of POs makes any business sense. Aggregation can happen in two business scenarios, one for POs constrained by existing corporate agreements and another for POs to be refined by online negotiations. The aggregation engine interacts with a corporate agreement system to obtain supplier policies in the first scenario. For the second scenario, it interacts with the negotiation engine to obtain supplier's policies during iterations of the negotiation process. Relevant policies are those that define product pricing, shipping and handing, and post-sale sendees as well as warranties and returns. Examples are given to demonstrate how automated intelligent aggregation of purchases is performed and how it reduces cost in enterprise e-procurement.