Major department store retailers process tens of million of invoices every year to pay their large and small suppliers. To manage these huge amount of invoices many retail companies have developed and implemented complex invoice processing systems to accept, process, monitor and pay their suppliers in a timely fashion to meet the terms of trade with their suppliers.

The Challenge:

The Accounting Operations group of a major national retailer that was processing millions of invoices every year had discovered through a simple manual audit that, in spite of the sophisticated vendor accounting system that the company had in place, the company was exposed to two potential problems: over-billing by suppliers and overpayment to suppliers by the company. Over-billing was often the result of partial shipment of orders by suppliers when the stores placed orders. A store may order one hundred dozens of a particular merchandise. The supplier will ship a partial order, say sixty dozens, bill the retail company for one hundred dozens, with the balance of forty dozens to be delivered at a later date. Sometimes the supplier will run out of items and is not able to supply the retailer with the balance of forty dozen items. This situation results in over-billing by the supplier and over-payment by the retailer. The company wanted to develop advanced statistical scoring algorithms to identify invoices likely to contain errors (over-billing) with the goal of recovering the over-billed amount from suppliers and working with the suppliers to correct and eliminate the problem from their side.


The Solution:

A team was set up to clearly identify the key customer requirements. The process flow of the invoice through the Accounting Operations was mapped to develop a statistical sampling plan for selecting and auditing different items on the invoice (merchandise type, store type, suppler size, etc.).
 

	Predictive Invoice Auditing Models were developed to score and identify invoices likely to contain over-billing with a high probability of yielding a sizable amount of dollar in recovery.
	An automated system was developed that enabled the clerical staff to download invoice data from the operating system, perform a simple audit and feed the result of the audit into the Predictive Invoice Auditing Model to determine the opportunity as well as the potential size of dollar recovery from suppliers.

The Benefits:
 

Instead of using clerical staff to use their judgment to decide which invoice to audit as the Manager of Accounting Operations indicated, “ We are now able to scan millions of invoices with less clerical staff and intelligently select and audit invoices that yield dollar recovery with a high degree of certainty”.

The Accounting Operations Center recovered several million dollars in a very short period of time with the aid of the Predictive Invoice Auditing Model. The Predictive Model helped the Accounting Operations Center to improve their invoice process flow and forced suppliers to implement changes in their billing systems. As the Manager added, “ The predictive model helped us recover significant amount of dollars and enabled us to implement changes all the way to our suppliers”.