Measuring Your Business, Pt 4: Time-Driven Activity Based Costing

Scott A. Whitmire
April, 2019

Part 1 of this series on measurement began near the middle and described several structural performance measures that can be used to measure just about any process. We’ll take a closer look at one of them in this article. Part 2 circled back and gave the motivation for why we would even want to measure our processes. Part 3 described the most common operational measures, from measures of financial solvency, to fixed v. variable costs, to other common measures of operational performance. This article introduces a wonderful tool for measuring processes: Time-Driven Activity Based Costing (TDABC). We use TDABC because it greatly simplifies the cost calculations that serve to allocate general ledger costs to specific departments and processes. TDABC provides a number of back-of-the-envelope calculations that allow us to quickly model the cost behavior for a business process.

In my experience, the best tools are conceptually simple, born of frustration with existing tools, and useful well beyond their original intent. TDABC is such a tool. Developed in the mid-1990s by a couple of M.B.A. students at Harvard Business School (see reference below), TDABC became a good way to model the costs of products and customers at medium to large companies. It replaces traditional activity based costing, which was developed to address the inaccuracies and bad decisions that came out of standard cost accounting. Allocation of costs to products and services has always been an important goal of management. Until you know what it costs to produce a product or provide a service, you cannot plan to make a profit at it, no matter what price you charge. At best, you can hope there is money left over at the end of the year, and hope is not a valid strategy (although I’ve worked for at least one company that very clearly operated this way).

The main issue with standard cost accounting was that it started with costs and applied them to standard times and actual times based on actual business activity. Cost values can be easily manipulated to make the results look better than they actually are. In addition, standard costs when compared with actual costs were not always comparing apples and apples. Actual business activity can materially change the actual cost per unit of a production run because of the maxim “In the long term, all costs are variable; in the short term, all costs are fixed.” Changes in business activity change the unit cost but are due to many things, none of which are related to the cost of production.

The main problem with traditional activity based costing was its starting point. To make it work, you have to allocate the effort of people and processes to specific products and services, usually as an estimate of the percentage of effort. Sometimes, this is easy, but only when a group of people and equipment is pretty much dedicated to a small set of products or services. This is, however, rarely the case, and almost never the case in small to medium sized companies where everyone wears many hats. Users often found these estimates were wrong, or they were correct only some of the time. The data did not allow for variation and variability. It also turns out that systems built on traditional activity based costing not only require a lot of manual data maintenance (the allocation estimates), they can take several days to extract the cost information out of the general ledger and compute the activity costs. This issue is a relic from the fact that mechanisms for traditional activity based costing were developed before the theory of using capacity-based drivers came along.

TDABC addresses these issues by using the time it takes to perform discrete activities as its basis because, in the words of its creators, time is common to all activities across a company. While true, there is a far more important reason to use time that is never mentioned by the creators: The time required to perform an activity is built into the way that activity is performed. To change that time, you have to change the way you do the activity, change the process. To change the process, you have to invest in it. In other words, time required to do an activity works as a basis because you have to take deliberate action to change it.

The cost part of TDABC uses macro-level budgets and labor rates to convert time into dollars. Like other methods of costing, these dollar rates can be manipulated by management action without actually changing the way things are done. Because of this, we define the true cost of an activity as the time it takes to perform it, hence the activity cost measure described in part 1. Further, we express capacity in the same units because investing in capacity really means either making more time available to perform a set of activities or changing the activity so less time is required. Both require deliberate investment in the process.

TDABC Requires Just Two Inputs

There are two inputs into a TDABC model. The first is the time cost of performing activities in the process in question. Once computed, these activity times become properties of the process and should be maintained as part of the process model. Activity times are built up from readily available or observable information; often, it is as simple as watching people work. Each activity has a standard time, plus additional times for variable activities or variations in the process. As an example, take receiving orders into a warehouse. The computed activity time might look something like:

Order Receipt = 5 minutes to verify the paperwork
+ 5 minutes to verify the goods received against the order
+ 5 minutes per item to stock (for an existing item number)
+ 10 minutes per item to stock (if a new item number)
+ 10 minutes (if hazardous materials are involved)
+ 10 minutes for customs form (for international orders)

Thus, a domestic order with 10 different items that already exist in the warehouse would require 10 + (5 * 10) or 60 minutes to process and stock the items.

A similar calculation is made for every activity, and every variation of those activities, which are then stored as part of the process model. These activity costs are borne out of the nature of the way work is done, so reducing that time requires investment in the activity to change the way work is done (yes, this is worth repeating).

The other input is the capacity cost rate, or the cost of providing a unit of capacity, expressed in dollars per minute. At a macro level, the capacity cost rate is the ratio of departmental costs to practical capacity, where practical capacity is the sum total of time available for the resources that perform the work. For most processes, this practical capacity is driven by people costs. For fully automated processes, the practical capacity is that of the automation tools.

The practical capacity of a person is the number of minutes per unit time, typically a month or quarter, they are available to work. A simple example, that also happens to be widely applicable, serves best to illustrate.

A typical person works 20 days per month and is paid for 7.5 hours per day, or about 450 hours or 27,000 minutes per quarter. From this, you have to exclude activities like training and standing meetings, so the normal capacity of a person in a quarter is 22,500 minutes. You can use this as the standard per person capacity per quarter.

The capacity of a department is 22,500 minutes times the number of front-line employees in the department. If a receiving department has 30 people, but 7 are managers or administrative assistants, there are 23 front-line employees, for a total capacity of 23 * 22,500 = 517,500 minutes. This value become the denominator of the capacity cost rate.

The numerator is the total budget for the department, including full salaries and benefits, depreciation, and overhead and support costs allocated in the general ledger. For this example, assume the budget is $500,000 per quarter.

Combining these gives the capacity cost rate:

$500,000
Capacity Cost Rate = ————-———- = $0.97 per minute
517,500 minutes

This rate of $0.97 per minute can be applied to every activity performed by personnel in the department, including activities where the department provides only part of the capacity.

We used a department in this example, but we could just as easily used any grouping of people. The important idea here is that we can readily identify from the general ledger the total costs allocated to the group. If you want to fine tune the cost rate, you can use different rates for different types of resources. This implies that standard labor rates for classes of employees will work just as well. For example, a law firm has partners, associates, paralegals, and legal secretaries that all have different billing rates. For a given matter, the firm computes the time required for each of these resource classes to perform the various activities involved in the matter. It might, for example, require a paralegal 30 minutes to prepare a standard court motion, with variations for known differences among types of motions. In case like this, each resource type has a separate capacity (two paralegals can prepare only so many motions per month), and you can use the billing rate or the internal cost rate to compute the capacity cost rate. The particular number you use in the numerator is much less important than the fact that you consistently use the same number (use the billing rate for everyone, don’t mix rates).

It is also possible that the key resources for a process don’t use time as the basic unit. When that is the case, simply compute the resource capacity using the unit that is native to the resource. You will get dollars per whatever that unit might be, but it is equally useful for computing the costs of any activities performed by that resource.

By applying this simple tool to processes, we can readily compute the cost of executing the processes. The calculations truly are back-of-the-envelope. I have made these calculations on a whiteboard while conducting a meeting, and several times while teaching a class. TDABC needs to become one of the standard components in your architecture toolbox. Not only is it easy to use, but computing a gross-level process cost during a conversation with an executive often impresses the executive and can go a long ways toward increasing your value as a business architect.

Reference

Kaplan, Robert S. and Steven R. Anderson, Time-Driven Activity Based Costing, Harvard Business School Press: Boston, MA; 2007.

Part 3: Operational Measures
Part 5: Capacity