Since the accounting function is essentially a cost center, it can be extremely difficult to justify new systems - they do not lead to additional revenues and may appear to contain only a few new features beyond those contained in the current system. Do we actually need new systems? Upon what factors should we base this need? And how can we justify their purchase?
What many managers do not understand is the extraordinary amount of time required to track down and correct errors that arise in many parts of the transactions for which the accounting department is responsible. Not only does this require many times the effort needed to process a normal transaction, but it also ties up the time of the senior accounting staff, who are the only ones with sufficient experience to make the corrections. Transaction error reduction is an excellent reason to support the acquisition of new systems. The trick is to determine which errors will be eliminated by which systems.
The types and volumes of transactional errors will vary considerably by company, but the main factor they have in common is that they tend to arise at the manual gaps between automated systems (which in turn are generally located at organizational boundaries). For example, the purchasing department likely has a computerized system for creating purchase orders, which they then print out and hand to the receiving and accounting departments for further processing. The payables staff then runs the risk of losing their copy of the purchase order (if they even receive it - ever had problems with interdepartmental mail?), and must then undergo the tedious and error-prone chore of manually matching it against incoming supplier invoices. The errors resulting from this manual communication of purchase order information could be eliminated by either acquiring an automated interface between the purchasing and accounting systems, or of acquiring a fully integrated system that links the two departments.
Consequently, it is useful to categorize errors by the system boundaries where they occur, and then determine if they arise in sufficient volume to justify a new system acquisition that will bridge the system boundary. Clear, this approach will tend to shift organizations in the direction of enterprise resource planning (ERP) systems that encompass nearly all company systems into one integrated package. However, there are lesser system improvements that can also benefit from this approach, such as the construction of Web-based invoice entry pages for suppliers, or on-line credit applications for customers. These types of interfaces are also positioned at system boundaries, and are designed to shift data entry chores to those entities that have a strong interest in entering the data correctly - customers and suppliers.
The same logic can be applied to something as simple as the acquisition of a scanner. If an employee sends information to a customer's fax machine, there is a significant risk that the transmission will fail, or that the fax will never be routed to the intended recipient. This is an error caused by the method of shifting information across an organizational boundary. This is a classic justification for purchasing a scanner, so that information can be digitized and sent via e-mail directly to the customer, thereby avoiding the fax transmission process entirely.