Operations involve multiple discrete events that are represented as documents in the ERP system. For example, our customer selling cycle includes a sales quotation, a sales order, a customer delivery, a sales return, a sales invoice, and a sales credit memo. Our supplier purchasing cycle includes a purchase order, a delivery, a return, a purchase invoice, and a purchase credit memo.

Although we can create a transactional fact table that allows us to analyze each discrete event, we are interested in analyzing the relationship between the events more than the events themselves. For example, we want to know how much time it took to deliver a product after receiving its originating purchase order. It would also be insightful to compare the quantity that we delivered with the quantity of the originating purchase order. We would have to work between multiple rows in a transactional fact table to discover this information; and just like a row-based database, we would find it...