Adding Time to Activity-Based Costing

BSE | NSE 10/02/12

To improve any business, managers need to understand how much it costs to produce a profitable product. It seems a simple task, but the process of securing and analyzing the data can be incredibly complex and organizationally taxing.

                   

Robert Kaplan is a Baker Foundation Professor at the Harvard Business School.

In Time-Driven Activity-Based Costing , Harvard Business School professor Robert S. Kaplan and Acorn Systems founder and chairman Steven R. Anderson (HBS MBA '95) introduce a system that calculates product, customer, and regional or branch P&Ls quickly and inexpensively. The work marries the Activity-Based Costing system developed by Kaplan and Robin Cooper in the 1980s, with time dimension modifications introduced by Anderson for clients of Acorn Systems.

We asked Kaplan to describe the problems with traditional costing approaches, the improvements made by TDABC, and how it works with the Balanced Scorecard developed by Kaplan and David Norton.

Sarah Jane Gilbert: What inspired you to develop the concept of time-driven activity-based costing?

Robert Kaplan: When Robin Cooper and I introduced ABC in the mid-1980s, we were strongly influenced by the approach taken by strategy consultants, who used interviews and surveys to assign the cost of employees to activities. We also wrote several cases where other companies, independently, relied on this interviewing approach.

These time surveys turned out to be the Achilles heel of ABC. They were time-consuming and expensive to perform, irritating to the employees who had to fill out the forms, and error-prone by using employees' subjective time estimates. Also, in the mid-1980s, we failed to fully understand the critical role played by capacity when estimating cost driver rates. The insight about the central role for capacity did not occur until about 1990.

Thus, while ABC still represented a definite advance over the arbitrary overhead allocations done by traditional standard cost systems, many companies abandoned ABC or updated their model only infrequently, because of the high time and cost associated with re-estimating the model, and its subjectivity and latent inaccuracies.

In the 1990s, many companies introduced ERP systems that captured data at the transaction level. It was natural to think how we could modify ABC systems to benefit from the ready availability of transactional data about orders, products, and customers. The stage was set for a major advance in the conceptual foundations of strategic costing systems.

The breakthrough came by combining an idea I had while writing Cost & Effect with an innovation developed by Steve Anderson, an HBS MBA, who was a student in my second-year elective, Cost Measurement and Management. My insight was that you could build cost systems with only two parameters. One is the cost rate of supplying resource capacity (such as cost per minute for people and machine-driven processes, or cost per cubic meter per day in a warehouse). The second is an estimate of the demand made on resource capacity (typically time) by transactions, products, and customers.

Anderson, after graduation in 1995, worked at McKinsey where he got the idea for using time equations to estimate how the capacity demands by a transaction vary based on all the contingencies that introduce diversity into transaction time processing. Anderson left McKinsey to found Acorn Systems where he developed software to incorporate time equations into ABC.

The software modeled how, for example, the time to process a customer order would vary depending on whether it was a standard or a custom order, a rush order or standard delivery, a domestic or an international order, and so on. Time equations handle complexity by simply adding terms with Boolean logic that test for the presence of a particular feature that adds or subtracts time to overall processing time, a simple and elegant technique. In one of the cases described in our book, a company used quite complex and multiple contingencies to reflect the factors that caused a particular transaction to be either fast or time-consuming to process.

Q: What is the relationship between TDABC and enterprise resource planning?

A: We probably could not have implemented time-driven ABC in the 1980s even if we had devised this approach then. The legacy information systems at the time could not supply the TDABC model with the detailed information on individual transactional and their special characteristics.

Now, with most companies having some form of ERP systems, the time-driven ABC software easily imports from ERP systems the extensive data on all transactions, which includes the special features of transactions that we can exploit in time equations. With this data feed, the software calculates the product, customer, and regional or branch P&Ls quickly and inexpensively.

Typically, time-driven ABC projects pay back in less than a year.

It's quite impressive to take companies with millions of transactions and customers, and be able to prepare detailed and accurate profit and loss statements, at various levels of aggregation, within days after the end of the monthly close. This would not be possible without the power of ERP systems. Many companies invested tens or hundreds of millions in their ERP systems but found it difficult to identify a return on investment from their spending. Installing analytic software, such as to implement time-driven ABC or the Balanced Scorecard, gives managers substantial benefits that would be difficult to realize without the ERP IT infrastructure.

Q: Must a company have an ERP system in place to use time-driven activity-based costing?

A: Clearly, having an automated data feed facilitates the calculation of product and customer P&Ls. But companies do not need a full-fledged ERP system to get most of the benefits. The TDABC system can import data from separate systems, such as the monthly general ledger, a production scheduling system, and a customer order file, and still do the cost and profit calculations. But having to integrate the data from these diverse systems makes the initial install somewhat more time-consuming, and the monthly and quarterly runs less automated.

Q: Does company size matter?

A: Size is not the critical parameter for the benefits a company gets from installing and acting upon a TDABC model. We have seen companies with less than $10 million in annual sales get enormous insights about products and customers that were losing money and that could be quickly transformed into profitability once the executives saw what was driving the losses.

Companies whose Balanced Score Card describes a low total cost strategy need ABC for accurately measuring the costs of critical processes.

I feel the critical parameter is diversity of products and customers. If you have only product and sell to only one or two large customers, you don't need much of a cost system to learn where you are making and losing money. But companies typically have hundreds of different products or product variations and hundreds or thousands of customers. In this situation, traditional cost systems will not accurately trace a company's expense base to each product and customer, leading to a highly distorted view of the company's economic model. Even a simple time-driven ABC model will fundamentally change the way the company manages its process improvements, its product variety, and its individual customer relationships.

Typically, time-driven ABC projects pay back in less than a year as companies increase their net profit margin by 1 to 2 percent of sales within months. These improvements are significant for both small and large companies.