Has ABC become a Failure?
– or is Time Driven ABC a solution for small and Medium sized enterprises (SME)?
Several articles suggest that Activity Based Costing (ABC) has failed to succeed in practical use. It is even argued that we have an ABC-paradox. Activity Based Costing has won theoretically in numerous articles in journals, through books, being included in all major Business Accounting textbooks and in curriculum at most business schools. But ABC is not used very much by the companies. Especially in small and medium enterprises ABC is rarely implemented. This is primarily due to two things: First, the ABC model is too difficult to implement. Second, the data foundation for the model is not available in the required quality in the ERP systems used by the businesses. In this article we will explore ways of overcoming these two related problems. We will explore these issues particularly in relation to small and medium enterprises. Through a theoretical analysis and practical implementation in a medium sized manufacturing company we will analyse whether it is possible to create a satisfactory decision making foundation using a TDABC as a less complex cost model in the company.
We have analysed our way to the conclusion that compared to ABC, the TDABC model provides the management with a number of pragmatic solutions that can be used in small and medium enterprises.
Several articles suggest that Activity Based Costing (ABC) has failed to succeed in practical use (for example (Gosselin, 2006; Öker & Adigüzel, 2010b)(Velmurugan, 2010)). Gosselin, 1997, even goes so far as to argue that we have an ABC-paradox: Activity Based Costing has won theoretically in numerous articles in journals, through books, being included in all major Business Accounting textbooks and in curriculum at most business schools, and by being endorsed by a lot of consulting companies. But in fact ABC is not used very much by the companies and this is the paradox. Many large companies, which tried to implement ABC in the 1990s, seem to have abandoned it again. The empirical data is uncertain, but in developed countries it is perhaps only about 20% of large companies who are using ABC to some extent(Gosselin, 2006). There are probably even less among the 2 small and medium sized enterprises (SME1) – maybe there are in fact only very few SME companies using ABC2.
SME’s are particularly interesting in this connection. On one hand, these companies are typically too large for the owner or other individuals to be able to oversee the entire company and ensure efficiency and profitability in all processes alone. On the other hand, the companies are too small to be able to devote large resources to implement advanced cost management or business intelligence solutions.Managers need cost analyses to show the profitability of its customers and products. But they do not have the tools to conduct useful profitability analyses. The major problem is that in small and medium enterprises ABC is rarely implemented. This is primarily due to two things:
First, the ABC model is too difficult to implement. Second, the data foundation for the model is not available in the required quality from the ERP systems used by the firms. In (Lahikainen, T. 2000) an analysis shows that an ERP system can not resolve an ABC task alone but need a separate ABC system.
In this article we will explore ways of overcoming these two related problems. Compared to the complexity of the model, Time Driven ABC has been launched as the necessary simplification of the model (Everaert & Bruggeman, 2007; Everaert, Bruggeman, Sarens, Anderson, & Levant, 2008;
Kaplan & Anderson, 2004; Max, 2007; Öker & Adigüzel, 2010a). Is this true? And how will the company provide the necessary data for the model when required? The ABC literature often claims that the necessary data can be extracted from the company’s ERP system (Kaplan, 2007a), but is this the case? What are the possibilities and limitations of the ERP system in relation to providing the necessary data for the ABC model?
We will examine these issues particularly in relation to small and medium enterprises. Through a theoretical analysis and practical implementation in a medium sized manufacturing company we will analyse whether it is possible to create a satisfactory basis for decisions using a TDABC as a less complex cost model in the company.
The paper is structured as follows: In Section 2 we discuss the ABC theory and the development of the TDABC model. Section 3 is a description of our company case study and a discussion of the practical implementation of TDABC in the company. The model is specified in Section 4. Section 5 concerns data, including a discussion of the ERP system’s advantages and limitations in connection with retrieving data for the ABC model together with an overview of the specific data in our preliminary implementation of the model. In Section 6 we present different solutions in the form of either
a simple calculation model or an outline for a more advanced Business Intelligence solution. In Section 7 we sum up some conclusions and suggestions for future studies.
1 In the EU, SME is defined as less than 250 employees, less than €50 million or a balance less than €43 million.
2 A recent survey shows that ABC is used by less than the half of small organisations compared to large organisations
(Chartered Institute of Management Accountants, 2009).
2.1. Cost Allocation
The last decades have seen significant developments on profitability analyses models, which can support corporate managements making the best qualified decisions. In this context, the managements of all modern, multi-product manufacturing companies need to conduct cost allocations as part of their financial control. Corporate management must be able to determine both the direct and indirect production and customer costs in relation to the cost objects use of resources. It must be done partly in order to include the indirect production costs (IPCs) in the inventory balance in the external accounts3 and partly because of the increased importance of cost allocations for profitability analyses and strategic decisions – along with the indirect costs becoming an increasingly greater part of the company’s total costs. The management must be able to calculate the effect of various cost objects on the indirect costs. This applies not only to production costs (IPCs) but also to how the various products and customers and other cost objects influence the increasingly higher costs on product development, sales, marketing, administration and other support functions, which cannot be included as production costs in the annual account. The company is not able to effectively analyse the profitability of the various customers and products until the turnover of these can be compared with all related direct and indirect costs.
Assigning indirect costs to the cost objects is done most simply by using a full-cost allocation model where costs are either distributed proportionally between the cost objects or allocated following a quantitative scale (e.g. based on the actual working hours). This model may suffice as long as the cost objects are relatively uniform in how they affect the indirect costs. But if there is a significant variation in how the various cost objects cause the indirect costs, a full cost allocation model will result in a distorted image of objects’ cost consumption, and thus be unsuitable for assessment of their profitability4. Then you have to determine the different cost objects’ effect on indirect costs by other means.
The activity-based management accounting models come into the picture now where a major theoretical development starting with Cooper and Kaplan’s introduction of Activity Based Costing (ABC) in the late 1980s (Cooper & Kaplan, 1988). Contrary to the full cost model, the ABC model uses both resource and activity pools with several different drivers, including various driver types in connection with the company allocation of indirect costs to the company’s cost objects. The ABC model should therefore be the solution to the challenges of determining cost object’s usage of indirect costs, but this is not necessarily the case in practice.
After much attention on the ABC model since the appearance in the 1980s, interest gradually decreased during the 90s (Gosselin, 2006). Many companies which had initially introduced ABC, 3 This is the case for all medium and large companies – cf. the Danish Company Account Act, section 82 and IAS 2.
4 As illustrated by Cooper and Kaplan in their famous pen example (Cooper & Kaplan, 1988) and many other places.
gave it up because of various difficulties. However, much has been written on the subject in articles,
books and conference papers, and ABC has been incorporated in most management accounting
courses at universities and related textbooks. There is also an enormous amount of information on
ABC on the Internet. Nevertheless it appears that after more than 20 years, the use of ABC has not
been as widespread as might be expected. On the contrary, we now face what the literature has described
as an ABC paradox : “..if ABC has demonstrated benefits, why are more firms not actually
employing it?” (Gosselin, 1997). It is estimated that ABC probably has not spread to more than
about 20% of larger companies (Innes, J. & Kouhy, R. 2011). Why are more companies not using it
now that the ABC model has shown all its advantages?
What could be the reason for the ABC paradox? Since the introduction of the ABC model, a number
of difficulties connected to applying the ABC have been introduced:
Firstly, there are aspects of the logic of the model. It is so complex in relation to corporate organization,
that it may be difficult to implement. It is a challenge to limit the activities of complex organizations
and to determine the appropriate drivers. In addition, extensive work is required to collect
and record time spent and other data to drivers, which makes the implementation of ABC very costly.
Furthermore, the model provides only a static picture of the organization, unless it constantly is
updated with changes in activities, processes and product lines. This continuous updating of the
model is also time and resource consuming.
Secondly, the implementation and recurring updates of the ABC model involve significant challenges
for the organization. There may be resistance to new ideas and changes, which requires management’s
ability and willingness to complete the process. The management will often choose to
implement the model in local units of the company, and it then requires a well-developed collaboration
across units to implement the model throughout the organization in larger companies. It can
also cause problems if top management is mostly focused on the financial accounts and does not
prioritize the development of cost accounting. One of the organizational challenges is also that it
requires a comprehensive knowledge of the company to develop and apply the ABC model. Any
lack of education becomes a problem here. Furthermore, the continuous recording of time consumption,
as well as a possible inconsistency between performance appraisal and performance related
pay, can give cause to resistance and behavioural challenges within the organization.
Thirdly if both the technical and organizational difficulties regarding implementation and maintenance
of the ABC model are overcome, you might have problems with the practical use of the model.
It is e.g. an important aspect of the theoretical model to be able to determine unused capacity, but
management risk to some extend to ignore this when time consumption is determined by interviews
and surveys. The results will be based on employees’ subjective allocation of their time, which will
typically equal 100% and thus become implicit assumptions of full capacity utilisation. Together
with the subjectivity it will also result in inaccurate cost allocations and associated difficulties in
validating the results. Unreliable results will also reinforce the organizational reluctance against the
model. Coupled with that there probably only are local implementations of the model in the compa5
ny, it means difficulties in making more comprehensive profitability analyses, and thus even using
The challenges of applicability also include that the ABC model is not equally suitable for all businesses.
Its usefulness will depend on organizational and technical factors, company size and degree
of centralization, type of production and product differentiation, company-specific strategies and of
course the question of the indirect cost share of total costs (Velmurugan, 2010) (Kennedy & Affleck-
Graves, 2001). The many challenges are all reflected in the fact that it has been hard to
demonstrated that a transition to the ABC model leads to rising corporate profitability or shareholder
value (Kennedy & Affleck-Graves, 2001).
We find from the different explanations two especially important and interrelated causes to the
ABC model’s lack of success in our opinion:
1. The theory/model is too complex.
2. The necessary data to use the model as intended are not available.
2.3. From ABC to TDABC
Time-Driven Activity Based Costing has been introduced as a simplification of the ABC model
(Kaplan & Anderson, 2007b) both in relation to the complexity and to the data requirements and
maintenance of these.
“In the revised approach, managers directly estimate the resource demands imposed by
each transaction, product, or customer rather than assign resource costs first to activities
and then to products or customers. For each group of resources, estimates of only
two parameters are required: the cost per time unit of supplying resource capacity and
the unit times of consumption of resource capacity by products, services, and customers.”(
Kaplan & Anderson, 2004)
We will now further explore the details in the different parts of the model, and how the model can
be implemented in our case company.
2.4. Direct Costs, Indirect Costs and Overhead Costs
In the basic concepts of a company’s cost structure you often use conceptual dichotomies such as
‘special costs’ versus ‘joint costs’, or ‘direct costs’ versus ‘indirect costs’. When you talk about costs
in relation to cost objects, it is not enough with a division in two. Instead of a dichotomy, it is perhaps
more appropriate to operate with a division in tree of the cost structure:
1. The ‘direct costs’ that can be directly measured at cost objects using ERP or similar systems
which can continuously collect data on the direct resource use per cost object.
2. The ‘indirect costs’, where resource consumption on the one hand is not immediately apparent
(or not counted) per cost object, but for which there on the other hand can be a significant
variation in resource consumption of the individual cost object at the same time. Here is
the ABC model’s relevance as a tool to calculate the cost object’s consumption of indirect
3. The ‘overhead costs’ which is not caused by cost objects such as customers or orders, but
may be strategic and developing resource use for the general management, research and development,
etc.. That means costs of resources, which are not relevant in terms of profitability
analysis of existing customers and goods, but must be considered as overhead.
The relative boundary between the 3 parts of the cost structure is not determined beforehand, but
must rely on the specific assessments in connection with implementation of cost accounting and use
of data for profitability analysis. ABC, including TDABC, is only relevant to the handling of indirect
costs. The direct costs are recorded and compiled in the ERP system, while the overhead costs
should not be involved in cost allocations aimed at profitability analysis.
2.5. Resource Pools
The relevant indirect costs are allocated and calculated in resource cost pools. A resource pool can
for example be a sales department. The cost of using resources of this resource pool is basically
indirect costs compared to cost objects such as products, customers and orders. But even though the
cost consumption for individual products and customers can not be measured directly, they are still
caused by these cost objects – the resource consumption in the department aims to support the production
of the company and supplying goods to the customers.
Cost objects are typically customers and products. The starting point for the ABC model is the cost
objects that we must add the part of the company’s indirect costs, which are caused by these objects.
It may be costs in the service or support departments (see e.g. Bhimani, A. et.al, 2008, chapter 5) as
well as overhead costs in production departments. The proportion of indirect costs (expenses) is
included in the ABC model; initially from the financial chart of accounts to the ‘resource pools’.
Time Driven Activity Based Costing (TDABC) is a simpler and less maintenance-heavy solution
than a traditional ABC. This is particularly in relation to the allocations of the resource pools such
as service departments’ resources. The characteristic of the service departments is that the resource
usage can not be included as production costs, but are on-going service and support for other activities.
The key resource here is the employees who produce services – of course supported by the
equipment, systems, etc. that they use. The resource costs of resources that they have available are
calculated for each resource pool.
The cost objects’ use of resources from resource pools are calculated using drivers. The prerequisite
and core of TDABC is that the measurement of resource consumption happens by using a unique
scale in terms of measuring the time spent. This means that the capacity of a resource pool (e.g. a
department) can be measured in time in terms of minutes/hours/man year of employees who are
committed to engage in resource pool activities. The total number of minutes available equals the
resource pool capacity.
Here we must differentiate between the paid capacity and the practical capacity (Kaplan & Cooper,
1998, chapter 7). The paid capacity in Denmark for a full time employee is normally 52 weeks of 37
hours = 1,924 hours annually. To calculate the practical capacity available to the company, the time
for holidays, public holidays, sickness, training and any other matters, which should not be included
in the resource pool, is deducted. In this model the practical capacity is assumed to be 80% of the
paid capacity5, and is measured in minutes.
For each resource pool, the following is calculated:
– The pool’s resource costs in money.
– The pool’s practical capacity in minutes.
– Unit costs in money per minute (resource costs divided by the practical capacity), ie. the
Capacity Cost Rate (CCR) .
Calculation of cost from the resource pools is managed by drivers. In the TDABC model the measurement
of resource consumption is simplified to a measurement of time consumption – i.e. use of
time rates as a unique units measure of resource utilization. Regarding the definition of drivers, this
is also possible to simplify in the TDABC model.
In the TDABC model only transaction drivers are used. These are primarily time rates, but could
also be other rates6. Transaction drivers are defined by the time rates being determined in advance.
There is no real measurement of time associated with the actual execution of the specific activities.
The Time Driven ABC model will be appropriate to use when resource pools can be limited in a
way where you can meaningfully measure the time consumption of resources from it.
The TDABC differs from the full cost model. The full cost allocation operates with a single additional
rate for indirect costs. Or you could say that full cost allocation method is a special case in the
ABC model, where there is only one transaction driver for allocation of the resource pool expenses.
With such an overhead added rate, the different cost objects’ variable effects on the service department
resources are not taken into account. With the introduction of TDABC the intention is to improve
the allocation of indirect costs to better reflect the cost object’s use of the service department’s
resources. To make this as simple as possible, you basically choose to define the time drivers so that
they can be immediately included in the model by using the existing data from the enterprise information
If you want a relatively simple TDABC model, the key is generally to select and define the appropriate
drivers on the basis of existing data in the ERP system and other enterprise information sys-
The 80% is an estimate. It must be calculated specifically when implementing the model.
The scale need not to be time. It may also be other scales. But specific rates must be linked for each sub-activities, so
that no measurement of each iteration is necessary, as when using the duration drivers in the ordinary ABC model.
tems. It may not show exact results, but certainly give a greater precision in the allocation of indirect
costs compared to just using the same added rate for all the indirect costs as in the full cost
model. Then you can assess whether the result is good enough, or if a possible wish for further precision
justifies further data collection and recording.
Activities are “links” between resource pools and cost objects. It is by specifying activities associated
with implementation of the model that you will determine how cost objects affect resources usage.
You can distinguish between 3 types:
1. A uniform usage: the special cases where the resource pool is only one activity and one
driver. This corresponds to the full cost allocation method.
2. Simple workflow: consists of a number of sub-activities.
3. Complex workflow: the individual sub-activities can not be immediately identified, but must
be constructed through methodological procedures for studying the use of resources from
the resource pool.
The resources of each resource pool are used for one or more activities/processes in more or less
complex relations. In the TDABC model it is shown in the way that a process usually consists of a
number of sub-activities. Such an activity/process can also be described as a “workflow”. However,
not all steps are carried out for each iteration of a process or workflow. An example would be the
receipt and dispatch of an order: here, some sub-activities typically occur every time (recording,
processing, confirmation, etc.), while others only occur in special cases (registration of new customer,
specific delivery conditions, etc.). For each activity/process the possible sub-activities must
then be identified and determined. The various sub-activities are alternatives7 to each other and specific
choices must be made for an alternative to be used in the cost calculation.
An example of an activity as “receipt and dispatch of an order”, would be that you can typically
extract data on: number of orders, number of new customers, number of special delivery conditions,
etc. in the information systems. This data can then be used directly in the TDABC model drivers
without having to make new records of data.
Furthermore, time rates must be included. A time rate is a fixed time for each activity, meaning
each sub-activity in a workflow. Time rates for all sub-activity must be determined through estimations
Based on this, there can in the TDABC model be established a cost function8 for the process. The
cost function will include a number of terms equal to the number of sub-activities. Each term will
consist of: a time rate, driver condition and the resource pool unit costs. See more under “The Model”.
7 In relation to our case company, the sub-activities are implemented as ”Alternatives”
8 In the literature specified as ”time-equations”, but here we prefer using the term cost function in stead. TDABC is also
a version of the use of “complexity index”.
Costs for these processes or activities in the sales department will often be recorded as period costs
in the company’s annual accounts. However, by using cost functions in the TDABC model, the
management will in the reporting be able to allocate these costs to the appropriate cost objects (e.g.
products or customers).
We can distinguish between three types of activities (or links between resource pools and cost objects):
– Uniform usage
– Simple workflows
– Complex workflows.
These activities must be established specifically per company. Therefore, we now proceed with the
business case study.
3. Business Case
Our case company is a family-owned manufacturing company in third generation, with approximately
70 employees. The company is a supplier of components for manufacturing companies in
many different industries. The production is highly specialized and customized with thousands of
different items. Production takes place in Denmark, in an Eastern European country and, as something
new, in China. Half the production is sold on the Danish market and half is exported. The
manufacturing itself is highly efficient, while the company has a challenge with resource consumption
in the various support functions such as: sales, the production engineering department, scheduling,
purchasing, etc. The question is how the resource consumption in these functions are related to
orders and customers. This is relevant in order to make profitability analyses to customers.
3.1. Business Case Activities
Not all activities/processes/resource uses can be that easily identified and described as stand-alone
workflows. Let us take some of the activities in a company’s sales department as example. It can be
their customer contact – and let us assume that they have a fiscal year from 1st October to 30th September:
– The first week of October, you talk with a customer.
– Several dialogues are established with the customer during the period.
– At one point, you adjust the prior agreement with the customer on several product items.
– At a later stage, the agreement is expanded to including new product items.
– At the end of September the following year, you have another dialogue with the customer.
– 30th September, the fiscal year ends.
– 5th October, you have the first contact to the customer in the new fiscal year.
This is how the processes are continuously happening with many customers – across accounting
periods. New customers are added, others disappearing, etc. All customer dialogues take time and
thereby costs in the sales department. In this case it is a challenge to make non-arbitrary allocations
of those indirect costs to the cost objects. The complexity influences the possibilities of dividing
and identifying the specific processes and their sub-activities. In this scenario, we distinguish between
· Resource use by uniform effect
· Resource use by simple workflows
· Resource use by complex workflows
3.1.1. Uniform Use of Resource Pools
Here it is given that almost the same amount of time is used on cost objects for some activities. It
can for example be the accounting department’s handling of order lines, where the same use of the
resource pool will exist for each order line. So there should only be one driver for this resource pool
in the form of an additional rate, triggered for each order line. You can choose to do the same for
smaller resource pools, were the cost objects’ effect is not necessarily uniform, but where it from a
cost-benefit perspective is not worth using the ABC in the form of several different drivers. In our
case study this could be the scheduling department, which is not very large. Finally, it could also be
the case for the departments/resource pools that have not yet implemented TDABC in companies
where you make a gradual introduction of ABC, and therefore temporarily are using a full cost allocation.
In our case study this applies for the purchasing department.
3.1.2. Resource Use in Simple Workflows
Even in connection with production of services, some activities/processes can relatively simple be
– A start event.
– A number of sub-activities in the process.
– An end event.
It can, for example, be receiving and processing an order. Such a process can be described as a
workflow consisting of a number of sub-activities with a start event and an end event.
The time required for the individual sub-activities will be relatively equal each time the process
goes through its cycle. If it is not the case for a sub-activity, you can operate with an average time
or divide it into several alternative sub-activities, each being more homogeneous.
For each sub-activity, a time rate is defined either through estimation or direct measurement, indicating
the number of minutes it takes to complete sub-activities. Furthermore, a driver for each subactivity
is defined. The driver consists of a condition of whether the sub-activity is to be included in
the specific process (e.g.: Is this a new customer? Yes = 1 or no = 0).
3.1.3. Resource Use in Complex Workflows
Some processes (such as sales department customer contacts) are difficult to divide meaningfully
into individual workflows. There are no specific start events defining sub-activities or actual end
events. There is no product flow that can be stocked as “work in progress” or “finished products”.
Therefore, we commonly refer to the annual consumption of resources in such a sales department as
period costs. One might even call them real period costs, unlike the period costs where measure12
ment on workflows enables allocating of costs to products and customers, as mentioned in the previous
If there was an actual time tracking in the department, it could solve the task. But there is not in our
case company study, and it must generally be considered to be too resource intensive to implement
time tracking in the form of hourly project registration or otherwise. There will also be other disadvantages
of introducing time tracking, such as organizational resistance, etc..
In our case company the sales department’s activities are connected to the company delivering
products to customers in approximately 55 different industries. In this case study, the sales departments’
estimate is that customers in approximately half of these industries can be described as
“standard customers” based on the time spent on them in the department. Furthermore, you can
identify an additional 5 different groups of industries in which the time consumption is estimated to
be 1.2, 1.4, 1.6, 2.0 and 3.0 times the time required for customers in the “standard industries”.
Moreover, it has been estimated that the countries can be divided into 3 groups, where local markets
in the Nordic countries can be described as “standard markets/countries”, while customers in a large
group of other countries is estimated to be twice as time consuming, and customers in a third group
of countries is estimated to require 3 times the effort compared to customers in local markets.
With this information cost drivers and associated weighted values for time rates can be constructed9.
A product number for a customer in a standard industry in a standard country weights 1. Other
industries in the standard countries are weighted with specific values from 1.2 to 3.0. Concerning
customers in the second group of countries, the weights are multiplied by 2 and when customers are
located in the third more complex group of countries, the weights are multiplied by 3.
Thus we reach the point at which the product numbers – weighted in accordance with customers’
business and geographical location – can be defined as drivers with associated time rates for subactivities.
In this specific case it is also the sales department’s estimate that the capacity was fully
utilized in the past period for which data are available. Now we can calculate the real time rates for
the individual sub-activities in this complex workflow.
4. Specification of the Time Driven ABC Model
Er Resource expenses – the total costs in the period (including any accrued expenses
such as depreciation) to the relevant resource pool10.
Cr Resource Costs
Tr Practical capacity measured in time
9 This corresponds to the concept of “complexity index”.
10 The cost concept here is similar to the cost concept (Expenses) in the financial accounts.
Expenses per time unit (Capacity Cost Rate – CCR)
tj Time rate per sub-activity
aij Driver (trigger, variable) aij = 0 or 1
i Cost object number – e.g. product number. i = 1 to n
j Sub-activity number. j = 1 to m
r Resource pool number – e.g. (sales) department. r = 1 to k
This means that
shows the costs per time unit in department r.
4.1. The Model
For a total activity (process), a series of time rates tj are established for the sub-activities which can
be a part in the activity (process):
t t t … t
For each sub-activity an if-statement is constructed. If the condition in this if-statement is met for a
particular cost object (e.g. for a specific item number), a variable aij with value 1 is returned. Otherwise
the value 0 for variable aij is returned.
This results in a matrix, where there are m columns with sub-activities, and where there will be n
rows of cost objects of the period:
aaa. . . . aaaa. . . . aaa.. . . . . a
aa.. . . . . a
To calculate cost consumption, the Capacity Cost Rate must be included:
Cost Consumption Calculation
The cost consumption can be calculated by matrix multiplication:
C = t t t … t Å~
aaa.. .. aaaa.. .. aaa.. .. .. a
.. . .
.. . .
.. . .
aa.. .. .. a
The result is a table with cost objects in rows and columns of sub-activities:
… … …
Sums can be added to the rows and columns. Each row sums Σj will respond to each cost object
consumption of resources from the department (resource pool), measured in money. Column sums
Σi, will show consumption costs for the individual sub-activities.
… … …
The last cell with the total sum (i.e. Σij) will show the cost objects’ total resource consumption for
the department (Cr), which can be compared with the department’s resource expenses. Meaning:
Unused resources in department r = Er – Cr
If you do not multiply with CCR in the last step in the calculation, you will get similar tables with
the values measured in time used instead of measured in money terms. This means row sums with
time used for individual cost objects and column sums of time spent per sub-activity. Here, the last
cell will show the total time consumption. Deducting this time consumption from the practical capacity
Tr the result is the unused capacity, measured in time.
Here, the result table relates to a single resource pool j. From the model there will be a number of
result tables – one for each resource pool 1 to k. These are then assembled into a multidimensional
5. Data: The Possibilities and Limitations of the ERP System in Relation to
The next issue will be the extent to which the data foundation for the model is available in the required
quality from the ERP systems used by the firms. ERP (Enterprise Resource Planning) systems
originate from MRP and MRP II systems, which with MRP I were focused on the company’s
product management and with MRPII also its capacity management. These systems have evolved to
be database-based ERP systems that seek to cover several different functional areas of businesses
(Davenport, Harris and Chantrell, 2004) in a way which integrates both the organizational and IT
aspects of the company.
Basically, the central problem in an ERP system is to solve value chain related tasks. The ERP system
must manage operational tasks related to how production aspects are transformed into finished
products or services. It is observed that in the majority of all businesses, regardless of size and type,
a very big part of corporate registrations are recorded in an ERP system. Despite the fact that there
were major difficulties in the implementation of ERP systems (Ribbers & Schoo, 2002; Willis &
Willis-Brown, 2002), they are used in many companies.
There are many different ERP systems on the market. There are systems for small businesses that
do not have a very complex value chain. There are ERP systems for companies where there is a
great complexity in the value chain and where there is also a great complexity of the individual value
chain. One way to solve the multiple versions of how a value chain may occur – and not least the
complexity of each link in the chain – is to allow implementation of various “best practice” solutions
in an ERP system, so the actual system can be adapted to the activities and problems of a company.
The disadvantage of implementing more complex best practice solutions in an ERP system is that
the total solution consequently becomes more complex as well. The consequences are huge costs in
investing in software and not least in the implementation (Andriole, 2006).
The central activity of an ERP system is to solve the task of delivering well-structured and complete
data as the foundation for operational decisions in relation to the relevant value chain in a specific
company. The moment the problem in relation to the ERP system changes from not only requiring
data for value chain related tasks at the operational level, it is a question of the extent to which it
can be done. If data input is needed for solving financial problems that have a more strategic and
decision-oriented view and thus the company management as a recipient, there will inevitably have
to be a different focus on registration and structuring data. Baxendale, 2003, recognizes that ERP is
an essential system for companies and is a system that integrates many of the company’s functional
areas. But related to solving an Activity Based Costing problem, the necessary non-financial data
are seldom available.
In order for an ABC model to be used for its intent and logic, there must be relevant and qualified
data for the following elements:
• There must be data describing the company’s cost objects
• There must be a classification and registration of the company’s direct and indirect costs
• There must be a classification of the company’s resource pools
• There must be relevant resource drivers
• There must be a categorization of the company’s activities
• There must be relevant activity drivers to be divided into different driver types.
Once the relevant data for the above categories are established, data must subsequently be included
in the ABC model calculation logic, so that the desired management information can be extracted.
Very often the cost objects are either the company’s customers or its products. These data are readily
available in an ERP system because they are central to the operational activities of the company.
If you analyse customers and products respectively, we see that the description of the two cost objects
is extensive. The cost description for both customers and products are related to the costs allocated
If the company’s ERP system and other tracking systems can not provide the necessary data, it is a
methodological question whether the degree of data relevance is sufficient for the chosen model to
still be used and implemented in a company. Missing data quality should result in management
providing advice on the precautions to be observed when the model is used for decision making
information. If the data relevance is insufficient for a model to be applied, the model’s theoretical
relationships give the company’s management the possibility, based on the theoretical model, to
define how the relevant data foundation can be improved or established in the company. It can be
done either by future data being upgraded or by taking more extensive decisions of replacing an
existing registration system with a new and more modern one.
The data quality established in an ERP system is the foundation for whether analyses at the strategic
level can be prepared in a company in addition to analyses at the operational level.
A key element in relation to the transactions carried out in each module is that there is a simultaneous
recording of the financial consequences in the company’s chart of accounts. This aggregation of
transactions in relation to an account number in the chart of accounts is entirely consistent with the
chart of account’s central position in the literature, where the chart is described as a central source
(Kaplan & Cooper, 1998) to the transactions subsequently used in various financial models. Both
the Danish literature and the American literature (Bukh & Israelsen, 2003; Kaplan & Cooper, 1998)
say that data to the financial model ABC is to be found in the chart of accounts. However, compared
to the way the chart is structured in an ERP system, it is not fully clear that it must be resolved
in a manner as the example below shows.
In connection with invoicing a product, a registration of direct costs is done on an account for products
sold on the income statement. If, however, the charge is a service defined as a sale of an indirect
cost capacity, no item is registered on products sold accounts. Not even if a service is created in
the product category. The allocation that an invoicing of a service leads to is basically two account
entries: one crediting a profit account and one debiting the debtor account.
However, if a service is included in a parts list, the following transactions will be released in connection
with invoicing the parts list: credit of an income account, debit of customer account, credit
of finished products and a debit of sales. An account entry will thus be made on a product account.
However, this is an entry composed of both direct and indirect costs. Besides the fact that the last
example is a mixture of both direct and indirect costs, there are also no analyses of how and to what
the indirect cost has arrived. Often the indirect costs are charged to relatively few accounts, and if
they are to be identifiable in relation to specific activities, they must have a dimension registered on
each entry. The more accurate the dimensioning is in relation to the activity, the better it is compared
to the analyses and the following allocation task is less arbitrary.
Another way an ERP system can solve a task allocation can be illustrated using the ERP system
SAP Business One11. It is possible to create allocation keys so that a specific transaction can be distributed
between different profit centres. In SAP Business One, it is not possible to allocate an allocation.
The consequence in relation to ABC is that the individual profit centres can be seen as either
activities or cost objects. It is not possible within the framework of the standard system to address
both the allocation of resources to activities and allocation of activities to cost objects.
11 The ERP system SAP Business One’s target group is the small and medium enterprises. SAP Business One provides
the following core modules: Administration, Finance, Prospects, Sales, Purchasing, Partners, Banking, Inventory, Production,
MRP, Service Personnel and Reporting.
As shown above, the complexity of the ABC model is a challenge when trying to get an ERP system
to provide the necessary data. Just as the ABC model needs the relevant data to generate decision
making information, the TDABC model also needs necessary data but not as complex as the
six bullet points shown at page 17. The question is whether the model’s simpler logic means that
the ERP systems can deliver enough data. The answer is ambiguous. Cost data should be extracted
from its chart of accounts and here the issues are the same as for the ABC. The timelines in TDABC
are basically composed of transactional data and a number of relevant transaction data can be easily
retrieved from the ERP system such as the number of invoices and invoice lines, the number of new
products etc. But it is difficult to retrieve transaction data in a simple way in the many cases where
the key transaction data is not directly related to buying, selling, item numbers, etc.. In our case
study it has been observed that an assessment of whether or not a product for a specific customer
must go through a PTA12 unit is made on the basis of the actual product. There is no product or customer
dimensioning or feature in the ERP system, which in advance will indicate whether or not the
workflow will include a PTA.
These examples illustrate that it is not immediately possible to solve an allocation of indirect costs
in a specific ERP system. What can be observed is that the data to some extent can be organized so
that they subsequently can be used in an ABC solution outside the ERP system.
The combination of the correct data and a correct theoretical model makes it possible to create a
concrete solution to how TDABC can be implemented in a company. In connection with our analysis
of the data actually available in our case company, we will provide Excel-based suggestions to
how we have developed a simple solution for the company.
Compared to the theoretical requirement specification that TDABC is, we will also provide a proposal
on which data must be available during the preparation of a more advanced solution proposal
for the company.
6.1. Simple Calculation Model
One solution includes a simple calculation model, where ”simple” refers to the data are available,
how data is retrieved and how data is specifically included in the solution. We have the following
12 Produktion Teknisk Afdeling ( Production Engineering Department)
TDABC Data Availability in the Case
Types of Data
Customer Complexity Customer file in the ERP system Database keys
Activities and Sub-activities Interview and certification material
Capacity Costs Extract from the chart of accounts
Theoretical Capacity Work time calculation Numbers
Practical Capacity Interview Numbers
Cost Capacity Rate (CCR) Calculation Numbers
Time Drivers Interview Text
Time Rates Interview Numbers
A distinctive feature of the data included in the model is that all data is either obtained through interviews
with key persons, or by the financial officer extracting the data and delivering them to us
in a spread sheet. No automatic extracts of the relevant data has been made. Therefore, the above
data is implemented in an Excel solution with ad hoc update of the model data.
The spread sheet model is implemented so that a number of sub-activities with related drivers and
times have been created for each main activity. Depending on how the activity is conducted, there
are various alternative ways in which the activity can be implemented. This is shown in the figure
below, which is from the activity “Offers, Receipt of Order and Sales.” The left side of the model
show which sub-activities are included inthe five concrete alternatives take. Next, a description of
the activities and time drivers and subsequently a calculation area, where “Time Driver Amount” *
“Time Driver Time” specify the time for each sub-activity. The various alternatives are thus summaries
of the “Amount * Time” shown in the calculation area of the model13.
13 There are no records in record number 3, 6 and 10 as they are merged with record number 4, 7and 11.
In another part of the model the Capacity Cost Rate (CCR) is calculated using the theoretical and
practical capacity and the actual capacity costs for the specific activities. The figure below shows
the calculation of CCR for the activities Internal Sales, External Sales and PTA.
These key data elements will be collected in an analytical model using a number of formulas and
data, so that a decision maker in the company can choose between various alternative time rates
applicable for the various activities in order to either calculate the costs of a specific order or the
analyses of future orders.
6.2. Advanced Profitability Analysis
The simple calculation model above is working well, but is difficult to keep updated as the data
foundation is found primarily through interviews with key employees in the company. If the company
desires a more advanced profitability analysis, where each order’s profitability is continually
updated from source systems into one solution where more complex analyses of profitability can be
created – for example in relation to different customers, products and industries, it is not sufficient
to develop a spread sheet based solution. Such a desired solution has different requirements for collecting,
updating and reporting data than what is possible to solve in a spread sheet. A Business
Intelligence solution should be used instead.
6.3. The Necessity for a BI Solution, including ETL, DW and OLAP
The term Business Intelligence (BI) originates from Howard Dresner of Gartner Group in the late
1980s. Business Intelligence is a central concept in order to provide the correct management infor22
mation to the right people at the right time (Watson & Wixom, 2007). BI systems are also solutions
which are widely applied in many different business types (Chee et al., 2009). A Business Intelligence
System’s functional area is located in the central area existing as data in the different source
systems in a company, in the relevant reporting options available to the company, the knowledge
that exists within the organization and the requirements made by the company’s decision makers.
The basic logic of the technological part of a Business Intelligence solution is that data is extracted
from its source systems and stored in a data warehouse, and then used in a multidimensional reporting
tool(Kimball, 2006) (Wixom, 2010) (Azvine, 2006). The complexity of the technological part of
a Business Intelligence solution requires a dedicated IT knowledge. Together with the technological
aspect, knowledge is also a very central element in building a Business Intelligence solution. In this
case, knowledge of TDABC.
In order to establish a Business Intelligence solution, the central parts of the theory behind TDABC
must be converted into a data warehouse model as shown in the figure below:
After an analysis of the case company’s ERP system, we observe that we have limited data available
to be used in the above TDABC model in the standard application. In relation to the tables tdCustomer
and tdCustomerAttributes we find the data about customer, country and industry in the tables
in the ERP system related to the customer files.
Data for the time rate analyses includes the tables: tdAnalyzeHead, tdAnalyzeLines, tdTimeDrivers,
tdTimeUnit, tdAlternatives and tdCapacity. Here it is only possible to identify data on the company’s
capacity costs. These data can be found in the company’s chart of accounts.
Time Rate Analyses are related to the company’s activities, which are described in the tables tdActivityLib,
tdSubTaskLib. It is not possible to find data for TDABC’s activity description in the ERP
system or other systematically updated registration systems.
The organizational relationship between the activities is described in the tables tdDepartments and
tdEmployee. These data can be found in the company’s ERP system.
To develop a multi-dimensional reporting, data from the above relational data model in a data
warehouse is transferred to a dimensioned data model with the following appearance:
The center of the model includes the company’s records at an invoice line level, while the satellite
includes the dimensions that can be used to analyze the individual invoice lines. All dimensions are
organized hierarchically so that management can make analyzes at different levels.
Based on the above dimensioned data model, it will subsequently be possible to develop a multidimensional
7. Conclusion and Future Studies
The challenge with this paper has been to clarify whether it is possible to use TDABC in small and
medium enterprises. It should be viewed in light of the fact that there are several studies pointing
out that ABC is not used very often, despite its great emphasis in research and education publications
etc. Compared to a specific case company, we have analysed our way to the conclusion that
the original ABC model continues to be perceived as very complex. But the analysis also shows that
a major cause of this is that it is very difficult for the management to provide the correct and necessary
data for the ABC model. The main reason for this is that standard ERP systems for small and
medium enterprises are not developed for or configured to solve strategic decision tasks.
We have analysed our way to the conclusion that compared to ABC, the TDABC model provides
the management with a number of pragmatic solutions that can be used in small and medium enterprises.
The main reason for this is on the one hand, that the model complexity is reduced since only
time is being used as a driver. On the other hand, complex workflows can also be handled with cost
functions based on time rates and what-if conditions as drivers for sub-activity costs. In our case
company, we specifically analysed the complex workflow in the External Sales department, where
the consumption of costs has varied significantly depending on which customer, industry and country
the activity is performed for.
Based on our case company analysis we expect that our further studies will involve several different
types of companies in order to develop more simple solutions, where small and medium enterprises
can simulate a TDABC model based on their own data.
Andriole, S. J. (2006). The Collaborate/integrate business technology strategy. Communications of
the ACM, 49(5), 85-90.
Azvine, B. (2006). Real time business intelligence for the adaptive enterprise doi:10.1109/CECEEE.
Bukh, P. N. D., & Israelsen, P. (2003). Aktivitetsbaseret økonomistyring : Danske virksomheders
erfaringer med activity based costing. København: København : Jurist- og Økonomforbundet.
Chartered Institute of Management Accountants. (2009). Management accounting tools for today
Chee, T., Chan, L., Chuah, M., Tan, C., Wong, S., & Yeoh, W. (2009). Business intelligence systems:
State-of-the-art review and contemporary applications. Symposium on Progress in Information
& Communication Technology 2009, , 96.
Cooper, R., & Kaplan, R. S. (1988). Measure costs right: Make the right decision. Harvard Business
Review, 66(5), 96-103.
Everaert, P., & Bruggeman, W. (2007). Time-driven activity-based costing: Exploring the underlying
model. Cost Management, 21(2), 16.
Everaert, P., Bruggeman, W., Sarens, G., Anderson, S. R., & Levant, Y. (2008). Cost modeling in
logistics using time-driven ABC: Experiences from a wholesaler. International Journal of
Physical Distribution & Logistics Management, 38(3), 172-191.
Gosselin, M. (1997). The effect of strategy and organizational structure on the adoption and implementation
of activity-based costing. Accounting, Organizations and Society, 22(2), 105-122.
Gosselin, M. (2006). A review of activity-based costing: Technique, implementation, and consequences.
In Christopher S. Chapman,Anthony G.Hopwood and Michael D.Shields (Ed.),
Handbooks of management accounting research (pp. 641-671) Elsevier. doi:DOI:
Innes, J. & Kouhy, R. (2011). The activity-based approach. In M. G. Abdel-Kader (Ed.), Review of
management accounting research (pp. 243) Basingstoke : Palgrave.
Kaplan, R. S., & Anderson, S. R. (2004). Time-driven activity-based costing. Harvard Business
Review, 82(11), 131-138.
Kaplan, R. S., & Anderson, S. R. (2007a). The innovation of time-driven activity-based costing.
Cost Management, 21(2), 5.
Kaplan, R. S., & Anderson, S. R. (2007b). Time-driven activity-based costing : A simpler and more
powerful path to higher profits. Boston, Mass: Boston, Mass. : Harvard business school press.
Kaplan, R. S., & Cooper, R. (1998). Cost and effect : Using integrated cost systems to drive profitability
and performance. Boston, Mass: Boston, Mass. : Harvard Business School.
Kennedy, T., & Affleck-Graves, J. (2001). The impact of activity-based costing techniques on firm
performance. Journal of Management Accounting Research, 13, 19.
Kimball, R. (2006). The data warehouse toolkit
Lahikainen, T., Paranko, J., & Seppänen, M. (2000). Implementing activity-based costing in an enterprise
resource planning system. Papers Presented at the 11th International Working Seminar
on Production Economics, Igls/Innsbruck, Austria.
Max, M. (2007). Leveraging process documentation for time-driven activity based costing. Journal
of Performance Management, 20(3), 16-28.
Öker, F., & Adigüzel, H. (2010a). Time-driven activity-based costing: An implementation in a
manufacturing company. Journal of Corporate Accounting & Finance (Wiley), 22(1), 75-92.
Öker, F., & Adigüzel, H. (2010b). Time-driven activity-based costing: An implementation in a
manufacturing company. Journal of Corporate Accounting & Finance, 22(1), 75-92.
Ribbers, P. M. A., & Schoo, K. (2002). Program management and complexity of erp implementations.
Engineering Management Journal, 14(2), 45.
Velmurugan, M. (2010). The success and failure of activity-based costing systems. Journal of Performance
Management, 23(2), 3.
Watson, H. J., & Wixom, B. H. (2007).
The current state of business intelligence. Computer, , 96.
Willis, T. H., & Willis-Brown, A. H. (2002). Extending the value of ERP. Industrial Management
+ Data Systems, 102(1/2), 35.
Wixom, B. (2010). The BI-based organization. International Journal of Business Intelligence Research,
1(1), 13. doi:10.4018/jbir.2010071702