ABC Process-Based Capital Budgeting.

These capital budgeting limitations may be addressed by using our process-based management approach with staged risk analysis for a potential capital investment in a new industry, like the cybermall opportunity. Our approach provides a structured analysis of the risks and returns associated with a potential investment so that they may be considered in reaching an investment decision. It provides insight into the relative importance of various risks in the business processes related to the potential investment and thus provides priorities for risk management. By addressing these capital budgeting limitations, our approach has the potential to be generalizable for analyzing risky capital investment opportunities.

CYBERMALL BUSINESS PROCESS FOR REVENUES

Since the cybermall represented a new business opportunity in a new industry, there were no existing revenue business processes to use. Large cable television system operators, both outside and from within the telecommunications company, were consulted in constructing the business process which relied upon the development and deployment of an interactive television network. Also, traditional television marketers, Home Shopping Network (HSN) and the QVC System, were consulted as indirect or "out-of-market" benchmarks for the cybermall on interactive television. Thus, both external and internal types of benchmarking information were used in developing the new cybermall business process for revenues.

Figure I represents the business process used for the cybermall revenues in the ABC process-based capital budgeting model. The starting point or first input is the number of cable subscribers passed by the new interactive television network (that is, homes connected to the network via cable TV wire). Pro-forma benchmarking information was obtained from both external and internal cable television system operators. The ten largest cable operators were projected to build and deploy broadband infrastructures for interactive television over the next five years with all other operators deploying over the subsequent seven years. This deployment was projected to start in the fifty largest cities or suburbs. These same sources were used to obtain the interactive adoption percentage that was multiplied by the number of cable subscribers passed by this new interactive television network to obtain the number of interactive television subscribers.

The avenue or cybermall shopping percentage was obtained from the HSN and QVC "out-of-market" benchmarks and multiplied by the number of subscribers to obtain the number of avenue shoppers. This generated one of the two major revenue sources for the cybermall project: advertising computed as an annual amount per avenue shopper. The HSN and QVC indirect benchmarks were also used for the following three inputs: purchase frequency, average purchase amount, and purchase return percentage. Multiplying the first two inputs with the purchase retention percentage and with the number of avenue shoppers generated the net shopping revenue for the cybermall. To obtain the company revenue for this project, the net shopping revenue was multiplied by the rental percentage charged to the retail sellers. This was the second major revenue source for the cybermall project: rent revenue from these sellers to use this electronic mall. In estimating these two revenue sources, other inputs (annual advertising amount per shopper, n umber of retail sellers and rental percentage per seller) were also provided by these external and internal sources.

CYBERMALL BUSINESS PROCESSES FOR COSTS

Since the cybermall represented a new business opportunity in a new industry, there were no existing business processes for the project costs. Thus, external and internal cable television system operators and the HSN and QVC television marketers were again consulted in developing business processes for the development, deployment, and marketing of an electronic mall on an interactive television system. Four major cybermall business processes were identified: network operations, product research and development, marketing, and content production. Figure II relates the key portions of the revenue business process to the business process costs

A pro-forma ABC approach was used to develop the capital and ABC costs listed in Figure II. Cost drivers, such as the number of networks, video servers, and fiber loop systems, were based upon the interactive television deployment generated for the revenue process. The capital and ABC costs for this deployment schedule of interactive television were only applicable to the first two business processes: network operations and product research and development.

The traditional ABC approach relies upon historical cost pools, historical cost drivers and historical transaction volumes to calculate ABC historical costs. None of these types of historical information existed for this new cybermall business opportunity. Accordingly, a pro-forma ABC approach was used to analyze this new business opportunity. This approach relied upon both external and internal benchmarks to develop new unit costs, new cost drivers, and new transaction volumes. These new unit costs for the new cost drivers were then multiplied by the new transaction volumes for the new cost drivers to derive both capital and operating (ABC) pro-forma costs as noted in Figure II.

As an example of the scheduled deployment of interactive television, the network operations process can be used to demonstrate this pro-forma approach for both capital and ABC operating costs. A key activity in this process was to deploy interactive video servers that connected to the interactive television network in order for the company to provide electronic mall services. The new cost driver was the number of networks needed. From consulting with external and internal cable television operators, the new unit cost was estimated to be $200,000 for one video server needed for each network deployed to 500,000 homes. If the new transaction volume was determined to be four networks, then capital costs would be $800,000 (=$200,000 X 4).

For an ABC operating cost example, a key activity was to adapt and maintain the deployed, post-production distribution network in order to provide ongoing electronic mall services. The new cost driver was the number of fiber loop systems. The new unit cost was estimated to be four full-time equivalent (FTE) employees for one fiber loop system that connected with each video server. Linking this to the previous capital cost example, there were four video servers that would require four fiber loop systems and sixteen FTEs ( 4 X 4).

For the network operations process, the major capital costs were the deployment of interactive video servers, workstations, and head-end equipment, including fiber loops. The major operating (ABC) costs were for maintaining the deployed equipment and software, adapting to post-production networks, compiling databases, and serving customer contacts. For the product research and development process, the major capital costs were for adapting video servers to emerging technology, and the major ABC operating costs were for maintaining server technology configurations, developing new network head-ends, and maintaining network programming for the deployed interactive television systems.

In Figure II, three variables-- the number of avenue shoppers, the net shopping revenues, and the number of retail sellers-- are used to develop variable cost relationships with this revenue business process. For the network operations process, there were three different types of variable cost relationships with the revenue process: 1) retail services per number of sellers or retailers, 2) access fees as a percentage of revenue, and 3) order processing per number of shopping orders. Also, there were three types of variable cost relationships for the marketing process: 1) mailings per number of total shoppers, 2) specific advertising per number of new shoppers, and 3) general advertising as a percentage of revenue. There were three types of variable cost relationships for the content production process: 1) graphics per number of total sellers, 2) new content production per number of new sellers, and 3) core content production per number of existing sellers.

There were also fixed costs for all four business processes. There were depreciation expenses for the capital costs from the network operations and the product research and development processes. There were full-time equivalent (FTE) labor costs for all four processes. There were various start-up costs for each process and there were other annual costs for the marketing and content production processes.

Uncertainties were noted for many capital and operating costs. Since these costs were tied to various volume levels in the revenue process as shown in Figure II, the related uncertainties can be analyzed by simulating the revenue process. For example, the interactive television deployment schedule is linked by various cost drivers to all the capital and ABC costs designated in Figure II. Also, a link was empirically derived between the number of cable subscribers passed by the network and the interactive deployment schedule as shown in Figure II. Thus, the capital and ABC cost driver volumes and variable costs can also be simulated from the revenue process simulation.

ABC PROCESS-BASED CAPITAL BUDGETING DECISION MODEL

All the revenue and cost processes in Figures I and II were linked together in a series of related Excel spreadsheets. The overall model has been designated as an ABC process-based capital budgeting decision model. Table 1 summarizes the base case projections for the project. Using the company's current cost of capital of 10%, the project had a net present value (NPV) of $405.3 million, including a residual value for selling the business at the end of the project life. Managers expressed concern about the impact of both the choice of discount rate and the assumptions underlying the residual (business sale) value calculation. Since the project was a new line of business, they felt that a larger (but not excessively high) discount rate would be more appropriate than the company's current cost of capital. Table 1 also summarizes NPVs for various discount rates to facilitate management's choice of a discount rate. With a residual value included, NPV was positive for the various discount rates. Concerning the res idual value assumptions, the model assumed perpetual cash flows after the last year of the project in 2004 and discounted those cash flows, resulting in an estimated residual value of $944.9 million in year 2004. However, without any residual value, the NPV at a 10% discount rate was only $40.9 million, and was negative for discount rates above 20.7%.

Print Get the Mag Weekly Updates