2013 46th Hawaii International Conference on System Sciences

A Prototype for Service-Based Costing Jürgen Dorn and Wolfgang Seiringer Institute of Software Technology and Interactive Systems, Vienna University of Technology Favoritenstrasse 9-11, A-1040 Wien, Austria [email protected], [email protected]

Such a transaction can be modeled easily by a change in a database. In a state-based knowledge representation this can easily be modeled. Either we have two states or if we do not assume the two actions to take place at the same time, we have three states and two state changes. In a conceptual accounting framework such as REA (Resources, Events and Agents) [4] two events are modeled when two agents exchange resources. If this accounting is done in real-time, an agent can determine at any time the amount of its resources (money and goods) by a database operation. Proponents of REA claim that with such a computational model traditional financial balance computation is obsolete because no double entry accounting is necessary [5]. Services are different to goods in several aspects. Services are co-created by provider and customer and thus the customer has to invest at least effort in its cocreation. If the customer is a company with employees of different degree of competences and an employee is needed for the service consumption, the success and the efficiency of a service provision is also dependent on the selected employee at the customer side. It is said that the provider makes only a value offering to the customer. If we account costs of services, we have to consider costs at both sides: at the provider and the customer side. If we only want to estimate at which price a provider shall offer a service, we might consider only costs at the provider side. However, if we want to sell services we have to compare different value offerings on the market and then we have to compare full prices including the cost for the customer, because the customer has also to consider the offered price and its own costs in co-creation of the service. In a service system, the cooperation of agents also requires to consider the common costs. Moreover, as stated by S-D Logic, the most important resources in services are the competences of the human resources. Thus, we especially have to account the costs of human resources and need a detailed cost model for levels of competences. If an expert is required for a service, the service must be more expen-

Abstract Service systems consider the co-creation of values in a process between provider and consumer with a win-win situation for both. Service-Dominant Logic demands a new view on economic activities where competences of provider and consumer are the most important resource for value creation. Given these assumptions, we investigate whether costs in a service system shall be accounted in a different way to reflect the change in view. If service provision is seen as a process, activity-based cost models seem to be appropriate. We argue that an extension is necessary, because a service is co-created by resources of service provider and consumer. Furthermore, activities performed by the consumer are not under control and therefore additional uncertainty has to be considered. Our research question is whether we are able to compute service costs with a higher degree of accuracy than with traditional cost models.

1. Introduction In most economies the service sector is growing, however, the productivity in this sector is typically much lower than in the first two economic sectors. Service science tries to understand and to address services to improve the efficiency of creating new services and to improve the productivity in this sector [1]. A service system is the main abstraction in service science to investigate phenomena in service science and to foster innovation in the service sector [2]. Service-Dominant (S-D) Logic is a new approach to marketing [3]. Motivated by the growing importance of services for our society, Vargo and Lusch present an approach being in contrast to traditional Goods Logic. The traditional logic is based on the exchange of goods or respectively the exchange of goods for money. If a good is sold from a provider to a customer it is already produced, potentially stored in a warehouse and conceptually the transaction of exchanging good and money can take place at the same time. 1530-1605/12 $26.00 © 2012 IEEE DOI 10.1109/HICSS.2013.56

1298 1300

2.2. Scenario 2: Repair at costumer’s site

sive than a service for which only an inexperienced agent is required. We apply and evaluate our approach in a manufacturing company producing high-tech medical machinery mainly on demand. Today, the company has to provide more product-enhancing services then before due to a strong competition on the market. However, the sales department has problems in forecasting costs for these services. Costing methods designed for goods are not directly applicable to services. But the increasing economic importance of services demands for a costing method for services also in the context of manufacturing companies. In the following we describe two scenarios of a service derived from the manufacturing company. These scenarios are used to show the problems and later one is used to illustrate our approach. In the third section we elaborate on characteristics of services and present current approaches to cost modeling. In the fourth section we introduce our model and the developed prototype application. In the fifth section we describe our evaluation results applying our prototype to the scenario. Finally we summarize and conclude.

If the defective device cannot be shipped and the customer accepts the repair costs or has a valid SLA the repair process is started at the provider’s site with the activities production order, plan materials, stock receipt, commissioning. Then the technician drives with required spare parts to the customer. The process step production order is used to create a production order, to plan the required personnel resources and organize the necessary spare parts. Then the repair is carried out at the customer’s site. After a positive quality check by the customer, the technician returns and the repair process is finished. The service consumer provides besides defective device and problem description, human resources during the repair process to help and supervise the provider’s technician. For both scenarios the company needs a method to compute service costs before service co-creation (precalculation, e.g. for making an offer) and compare them afterwards with the real service costs (post-calculation) for controlling the efficiency. The obtained costing information about the profitability can then be used for future economic decisions.

2. Scenarios

3. State of the Art

Service provider and service consumer provide input and must participate in the process of repairing a device. The depth of the involvement of the service consumer into the process of the service co-creation depends on the service and the required activities. When customers have a defective device a repair service is requested. We distinguish here two scenarios depending on whether the device can be shipped to the provider or must be repaired at the customer’s site. The device may be repaired as part of an active ServiceLevel-Agreement (SLA). If the SLA is valid the service provider has to bear his costs for the service otherwise the customer must pay for the whole repair costs.

Accounting of costs has a long history. There are internal reasons as well as external reasons to compute costs of a company. Based on estimated costs a company may decide on prices and on offered products and services. Externally, costs are used to evaluate the value of a company. There are always reasons to evaluate costs higher or lower than the real costs, but in principle, decisions are easier if costs are estimated as accurately as possible. A further reason for estimating costs of services in a service system, is to identify drawbacks (high costs) that could be eliminated by innovations (e.g. an automation by computers). When considering costs of services, a company is confronted with characteristics of services not known from physical products. In a manufacturing company principally everything is centered around physical objects like raw materials and the main task is to organize the flow of these materials. As a result, the used costing systems are usually oriented to account material of finished goods. Traditionally, costs in production theory are computed by determining the direct costs of material, machines and humans that are required for a certain product. These factors are also called cost drivers. With such a method of course we cannot consider easily costs for management. To consider also such indirect costs, overheads are used. These are often only estimated by a fixed amount or percentage.

2.1. Scenario 1: Repair at provider’s site The customer has a defective device and when an active SLA exists or the customer accepts the repair costs, the repair is carried out at the provider’s site. The defective device is shipped to the manufacturer and the following process steps are carried out to repair the defective device: stock receipt, quality assurance, production order, plan materials, commissioning, production where the device will be repaired, quality assurance and shipment of the repaired device to the customer. In this scenario the customer provides the defective device, a problem description and the shipping for the process as external factors.

1301 1299

service. Because of the dominant role of the external factor it is necessary to integrate it into the costing of services. During the co-creation different activities are carried out to combine the internal and external factors (factor combination) and to obtain the desired output. This means also that both sides (consumer and provider) are integrating resources into the co-creation process. To get the total service costs this means we have also to account the costs for the consumer. The final output of the service co-creation can be something tangible like a repaired medical device or just the information about the failed repair [14], [15], [16]. Three kinds of uncertainty in the context of service co-creation are disposal uncertainty, production uncertainty and integration uncertainty. Disposal uncertainty occurs when the service customer knows the quantity, quality, time and location of the external factor, but the service provider is not sure whether the consumer can provide it. Production uncertainty describes problems creating the external factor. Integration uncertainty regards problems of how to integrate the external factor into the production process of the service provider [10].

3.1. Service characteristics Compared to goods services have some special characteristics such as perishability, inseparability, heterogeneity and uncertainty [6]. Perishability describes the intangible nature of services. As a result it is not possible to store a service in a warehouse. For a service provider it is difficult to manage and plan the necessary capacities and resources for the service cocreation. Inseparability describes the simultaneous production and consumption of services. As a consequence, the service provider and consumer can determine only together when the service process shall start. The service provider can only offer the willingness to perform the service co-creation. Heterogeneity describes the difficulty to standardize the service process of service co-creation because a service is often created only once. In connection with services uncertainty plays an important role and there exists different reasons for uncertainty. For instance a service consumer cannot check the quality of a service before it is produced. Additional information about the described service characteristics can be found in literature like [7], [8] and [9].

3.3. Activity-Based Costing 3.2. Service co-creation Activity-Based Costing (ABC) is a well-known approach to account service costs. The concept of ABC has its origin in the manufacturing industry [20]. Since traditional costing systems have used only roughly estimated overheads, companies required a method which enables calculating the overhead costs in a correct way and map them to customers, products, services or other cost objects. ABC allows adding a proper amount for the overheads to the cost object. Based on the activities performed in a company, the activities and business processes are identified and analyzed. Usually this is done with interviews. Different departments are part of the value chain when producing overheads. Therefore the identified business processes are often inter-divisional. The total costs for the selected activities are computed by considering the required personal resources and the respective salaries. Additional cost drivers must be specified, e.g. the number of inquiries, offers or bills is counted. In order to get the unit costs or cost driver rate, the total costs are divided by the cost driver quantity. These results can now be integrated as overhead costs into product cost calculation. Due to the similarity between overhead costs and service costs, ABC is also applicable for the service cost calculation [23]. Because the implementation and operation of a conventional ABC System is very complex and time consuming, the so-called Time-Driven Activity-Based Costing (TDABC) was developed in the last years. The

In the context of services we use the term co-creation instead of production because it fits better to describe the integrative nature of the service co-creation process between the service provider and the service consumer. The process of service co-creation is influenced by the customer integration and the external production factor. Customer integration describes the mandatory participation and integration of the service consumer during the process of service co-creation. The integration is expressed by an external production factor provided by the customer [12]. To describe service co-creation more detailed it can be separated into three different parts. These are the production factors, the factor combination and the form of the output [13]. The production factors are all the “entities” which are necessary to co-create a service. The production factors are split up in internal and external factors. Basically a production factor can be something tangible or intangible for example a defective medical device, information, or the customer himself which is necessary to create the service. The internal factors are entities that are provided by the service provider. The service con-sumer supplies the external factors. It is not possible for the service producer to produce or buy the external factor on the market. This means the external factor and also the associated knowledge is under control of the service consumer. The service provider is forced to integrate it into the process of service co-creation to provide the demanded 1302 1300

basic idea of analyzing the activities performed by the employees is the same. But to make it easier to implement and operate such a costing method, time equations are used. Time is regarded as the leading cost driver. Most of the supplied and consumed resources like employees and machines can be measured using time. For these equations the cost of capacity supplied and the practical capacity of resources supplied are required. The first one is the costs of a department and the second one is the sum of the available personnel resources of this department [23]: p

In contrast to raw materials and produced physical products. With inseparability the costing method must consider the customer in form of the external factor even if there is only vague and imprecise knowledge about the external factor available. Heterogeneity requires a flexible costing method that allows an easy adaption of the modeled service and service costs. The kind of uncertainty that we investigate into our costing method is expressed by the external factor and the implications to the service process. From the costing domain the most important requirement is to use the origin of the costs for the service costs calculation to achieve traceability. This will enable transparency about the costs origin and the related activities this will help to achieve process and cost transparency. These are two main goals of service costing and controlling. To get a complete costing method for services the proposed solution must support the calculation of the service costs before the service co-creation in form of a pre-calculation and after the service co-creation a postcalculation is made to compare the results. In practice a possible application field for the proposed costing method is the sales process when information about the costs of a service is required. At this point a pre-calculation is used to get the required costing information. This may be the case when the customer needs price information for the repair and to decide about the economic efficiency of the repair or to decide which level of a Service-Level Agreement (SLA) fits best. In any case it is necessary to integrate and evaluate the degree of customer participation to get service costs calculation results which are precise as possible This requires to integrate the external factor because it determines the necessary activities during the service co-creation and therefore also the used resources – in particular human resources. Since only the customer is able to provide the external factor and the knowledge about it the service provider depends on the customer’s knowledge.

Formula 1: Calculating the capacity cost rate [23] In contrast to ABC the approach of TDABC is easier to maintain and to keep it up to date, because the existing data of an ERP-System or a CRM-System (Customer Relationship Management) can be directly integrated and it is not necessary to prepare and process them with an additional application. Another advantage is the possibility to keep the underlying ABC model up to date and to integrate changes in the associated business processes. This adaption can be realized by simply changing the related time equations. Also different alternatives of an activity can be formulated with a time equation. Thus it is relatively easy to model a packaging activity of a warehouse with different versions for different customers and also the described scenario is relatively easy modeled using TDABC. Already existing time equations for an activity can also be used for another business process. It is also no longer necessary to maintain and update a complex activity directory. The general concept of ABC is accepted by academics and industry [23], [24]. The huge amount of available literature and the large number of described industry projects substantiate the acceptance of this concept. However, ABC and TDABC do not cover uncertainty of the external production factor. In the next section we propose the concept of a costing method to solve this problem.

4.1. A Cost Model for Services The proposed approach is called Cost Model for Services (CMFS) and is based on two main parts. The first part is service modeling with the target to model the service process. The output of this step is the basis for the service costing part. In the service costing part, the service model is extended by costing data. See also Figure 1 where the main parts of the CMFS are illustrated.

4. Costing Model Requirements To develop a model for service costing we have to consider requirements from service science and the cost accounting domain. From service science we have to consider the four described service characteristics perishability, inseparability, heterogeneity and uncertainty. Perishability means we have to model something whose output and costs are difficult to measure.

1303 1301

Figure 1: Main components of the CMFS The output of our approach are the service costs including the Customer Integration Factor (CIF) which is a rating of the customer participation and it is calculated by evaluating the impact of the external factor on the service costs for the service co-creation. The CIF rating has been developed based on our problem analysis. It is difficult and expensive trying to model all the different external factors required during the process of service co-creation. In our repair service example, the staff working for the customer is an external factor. We also have to consider that different customers have staff with different competences and further, different activities may be required for different customers. For the described approach we regard a service as a process between service consumer and service provider, which is carried out to fulfill customer satisfaction with a tangible or intangible output. We also assume that the required process steps depend on the provided external factors and that our costing method reuses the process model several times for different customers and service provisions. The modeled standard process like in our described service scenario covers the complete service process with all the required activities for the provider and the consumer domain. The usage of a common service model for different customers makes sense because it is not always possible to change the complete process model. Another reason is that the required information for a new process model is not

always available and the modeling process is a time consuming task. For the service modeling part the first step is to identify the relevant service. A service (Si) provision consists of different activities (Ai). Certain activities depend on the behavior of the customer. These external activities are not under control of the provider and thus uncertainty about the performance and the service costs has to be considered. In the described repair service scenario, the customer provides the defective device, information about the defect and different employees with different competencies. The service provider integrates these cost inducing external activities during service co-creation. Using activities as the smallest modeling unit allows the identification of the service costs and the traceability to the costs origin. To calculate the costs for an activity a time equation is created estimating the time required to carry out this activity. This step is analog to TDABC where the time equation for a service is the sum of all activities of the process realizing the service: A1+…+Am. Time is used because it is possible and relatively easy to calculate the consumed resources of humans and machines. For services human resources are the most important cost factor [23]. Our Cost Model for Services integrates the external factors of the customer in form of the Customer Integration Factor (CIF). The CIF is used to calculate and

1304 1302

rate the customer participation and expresses the uncertainty introduced by the external factors during the process of service co-creation. A central point of the CMFS is the correlation between the CIF and the service activities, see Figure 2. The value of the CIF parameter is designed to be independent from an activity but it is related to the activity and it is used to measure the customer’s impact on a single activity or a complete service. The pre-calculated activity utilization represents the expected resource consumption and consequently the costs for an activity. The sum of the involved activities corresponds to the pre-calculated service costs. During service co-creation data about the performed activities are recorded from which the real value of the CIF parameter and activity utilization can be derived. The recorded data are used to make a postcalculation of the real service costs. The results of the pre- and post-calculation can then be compared and represent valuable costing information including a rating of the customer’s service participation in form of the CIF value.

ple linguistic variables are used to classify the four cost effecting parameters of the CIF. In the CMFS a maps values like very low, low, medium, high,... to real values. The central formula of the CMFS is: m

Si = ∑ (A0A i + T jA i ) * ((IdepA i * Iint A i ) * IfreA i ) + IdurA i i=1

where the variables have the following meaning: ,

,

, Formula 2: Calculating activity time The CIF parameter is calculated per activity by multiplying Idep, Iint and Ifre and adding Idur, see also Formula 2. The same formula is used for the pre- and post-calculation of the service costs, which enable the comparison of the results and use the obtained data for the purpose of cost accounting. For the post-calculation, the real amount of consumed resources is accounted during the service co-creation. The required data are service process, activity, start time, end time and employee type. This data can then be assigned to the co-created service and a post-calculation using the same cost model is possible.

5. Prototype for the CMFS Figure 2: Correlation between CIF and activities To evaluate our proposed costing approach we developed an application and tested it with data we obtained from the manufacturing company. In this section the architecture of the prototype is described and the results applying the prototype to the described repair service scenarios are presented. The developed prototype consists of three components. The Community Edition (CE) of the Adonis Process Modeling Toolkit1, a relational database in our case the Microsoft SQL Server 2008 R2 Express2 and Microsoft Excel 2007 with Visual Basic for Application (VBA)3. The Adonis Process Management Toolkit is used to model the service process based on the Business Process Model and Notation (BPMN) Version 2.04 and the CMFS costing data for the pre-calculation. Modeling of the cost data is not available in Adonis and was therefore implemented for the CMFS proto-

The CIF consists of four parameters: influence depth, influence intensity, influence frequency and influence duration. Influence depth (Idep) measures how deep the customer’s external activities are integrated into the value chain of the service provider. Influence intensity (Iint) quantifies how intensive a resource related to an activity will be utilized. The influence frequency (Ifre) indicates how often an activity will be required during service co-creation. The influence duration (Idur) is used to evaluate if more resources (time) will be required than considered in the standard service process model. Resource costs are mainly the costs of the human resources depending on the type of the resource (i.e. expert or novice). The CMFS extends the basic TBABC formula to integrate the CIF parameters and we consider that the knowledge about the external activities is often vague and not precise. This assumption is based on service characteristics, which complicate the standardization of the process of service cocreation and the dependency on the external factor, which is not under full control of the service provider. As a consequence and to simplify the CIF rating, sim-

1

http://www.adonis-community.com http://www.microsoft.com/germany/express/ products/database.aspx 3 http://office.microsoft.com/de-de/excel/ 4 http://www.bpmn.org/ 2

1305 1303

type. In Adonis only access to the SQL database is available and not to the source code. Consequently different workarounds to add the CMFS costing data are necessary. The provided database structure of Adonis was extended with stored procedures, tables and views to store additional data, which is required for the CMFS. The main stored procedure generates for each activity a time equation and a CIF value based on the modeled process and costing data. To get postcalculation data a table was created to store these data consisting of the attributes cost carrier, process, activity, human resource, duration, Idur, date and Idep. Based on these database extensions, it was possible to create an Excel based VBA application for querying costing data and for providing a detailed pre- and postcalculation comparison. The main screen of the CMFS prototype is illustrated in Fig 3. With this main window the CMFS model can be parameterized for a customer and service and the pre- and post-calculation is created. The user gets immediate visual feedback if positive (green) or negative (red) differences between the forecasted and real service costs are calculated.

5.1. Prototype Evaluation To evaluate the developed prototype the following objectives have been formalized. Some of them are more technically oriented to check the feasibility of the CMFS in form of a service modeling and costing application. The objectives are: 1. graphical modeling of service and cost data, 2. store CMFS data in a relational database for further cost analysis, 3. automated comparison of pre- and post-calculation results, 4. cost traceability (activity to service and vice versa), 5. individual CIF values per cost carrier (customer) and common standard values and 6. evaluate the soundness of the approach. Point 6 concerns our hypothesis to generally evaluate the CMFS using our test data: An increasing CIF value indicates increasing activity time utilization and human resource consumption but must not imply increasing service costs for the service provider. In the next paragraph the results of the technical implementation of our CMFS prototype are described and discussed. The results of the cost accounting based evaluation of the CMFS are described and discussed in the subsequent section.

5.2. Results Technical Implementation The implementation of the CMFS prototype was difficult because the source code of the Adonis framework was not available which forced us to reuse existing fields of the user interface and the database. Regardless of this restriction the selected prototype architecture (Adonis, SQL database and Excel) enabled the realization of a fully functional CMFS prototype application. The usage and extension of Adonis enabled a complete graphically modeling of the evaluated service process and CMFS costing data. This was an objective for the prototype. The only data not modeled within the Adonis framework is the post-calculation data about the real activity utilization. For the current prototype version the post-calculation data were added using the CMFS prototype Excel user interface, see Figure 3. The used SQL database structure was designed to enable the integration of costing data from an external information system like an Enterprise-Resource-Planning (ERP) System. It is also already possible for an external costing accounting system or ERP System to query the stored CMFS costing data for further cost analysis. Consequently the first two points of our prototype evaluation are fulfilled.

Figure 3: Main screen of CMFS prototype All costing data (CIF parameters, human resources, cost carriers, time equations and cost drivers) required for the CMFS pre-calculation are modeled for each activity in Adonis. A screenshot of an Adonis process model is illustrated in Figure 4.

Figure 4: Adonis process model example

1306 1304

The automatic comparison of pre- and post-calculation results is realized using Excel with VBA for our prototype. The provided SQL data can directly queried and represented using Excel. To get access to the SQL database VBA is used in combination with SQL and (Transactional) T-SQL5. The fourth point about cost traceability from (service to activity) top-down and (activity to service) bottom-up is enabled using BPMN 2.0 and the database structure of Adonis. The individual and common CIF values (point five) for Idep, Iint, Ifre and Idur were implemented by the possibility to assign them to an individual activity, a process step or the complete service process. A process steps usually consists of several activities. On top of a modelled service a cost carrier must be defined to get specific values for a customer, a customer class or also a project. Customer dependent CIF values were possible due to the ability to assign customer individual CIF values down to an individual activity. The successful technical implementation of the CMFS prototype is additionally confirmed by the CMFS service cost results we got applying our test data in our CMFS application. These results are presented in the following paragraph.

for the customer domain. Especially the CIF value measures how intensive the customer participates in the process of service co-creation. Consequently we also include process steps of the customer and the provider into a modeled CMFS service model. In the presented example for the first five process steps (CD Customer Complaint to Commissioning) and for the steps from inquiry to production order more resources have been forecasted then actually required. In contrast for the processes engineering, repair at consumer and technical feasibility more human resources were consumed than expected.

5.3. Results CMFS Model Evaluation To check our hypothesis we compared the pre-calculated CMFS activity time and CIF values with the associated post-calculation results. As already mentioned our hypothesis is: An increasing CIF value indicates increasing activity time utilization and human resource consumption but must not imply increasing service costs for the service provider. We applied the CMFS prototype in the two in section 2 described repair service scenarios. Due to space limitations we only present in this paper the results of applying our costing approach in the second scenario (repair at the costumer’s site). In the following figures the pre- and post-calculation results generated by our prototype application are illustrated. We took data from the manufacturing company for the pre- and postcalculations. For the scenario we have analyzed and modeled 90 activities, 50 cost drivers, 10 human resources, 23 processes and 34 time equations. The number of modeled activities, cost drivers etc. depends on the complexity of a service. The first three process steps are part of the costumer domain, the remaining belong to the provider domain. The processes of the costumer domain are part of the result because our costing approach is designed to include also the costs

Figure 5: CMFS results for scenario Figure 5 shows the CMFS results (i.e. the activity time per process step including the CIF rating value). The activity time in hours is the time it takes to carry out a specific service process step e.g. CD Customer Complaint and includes the CIF rating. In Figure 6 the same results but only for the CIF values are shown. The most important difference between both figures concerns the process step engineering. This process indicates that the customer was less deep integrated than forecasted (pre-calculation CIF value is lower than post-calculation CIF value. But in total for this process step more resources were consumed than scheduled. When comparing the remaining process steps in both diagrams it is obvious that there is no greater divergence. The CIF value represents the specific CIF rating for a service process step. A service process step is a combination of different activities.

5

http://msdn.microsoft.com/enus/library/ms189826(v=sql.90).aspx

1307 1305

costing approaches for services especially in the manufacturing sector the CMFS provides service-costing data that were not available with this precision. The CMFS enables detailed comparison and tracking of pre- and post-calculation data and consequently provides valuable information about the uncertainty about the customer’s behaviour in the process of service cocreation. The calculated CIF value allows a cross-customer comparable customer participation rating with different levels (service to activity and back). The CIF allows assigning and comparing customer/customer class CIF ratings. Additionally the CMFS is easily adaptable for different service scenarios and customers.

6. Conclusions Figure 6: CIF value results for scenario

Service systems are system configurations where actors cooperate in order to provide services to other service systems. On one side the cooperation in the service system demands a win-win relation between provider and consumer and on the other side service marketing demands a similar relation between provider and customer. Consequently, we ask for a cost modeling approach that takes into account costs at both sides. Especially for services the most important cost driving factor are human resources. For the quality and efficiency of a service, the competences of both actors in the service provision process are most important. Additionally, the uncertainty about the customer’s behavior is a relevant factor in any cost model applied by a provider. We have analyzed the service costs at a real company and made computations with a new model that considers these problems. Computations with our developed CMFS prototype application have shown that we can reach a higher accuracy in forecasting of service costs. This evaluation must, however, be repeated in the practice and compared with more actual data. Here we have to evaluate whether the difference between forecasted and actual costs are really smaller than in the actual practice. In our experiments we have only measured costs for the service provider. For a service system we have argued that costs for provider as well as consumer have to be taken in account. In principle this costs of the consumer can be computed in same way. However, then the activities of the provider must be associated with CIF values and the total service costs would be the sum of the providers service costs as computed in our example and the consumers costs. The four components of the CIF are ratings for the customer integration. We have assumed standard values which can be adapted for each customer individually by considering past service co-creations. Additionally, we could establish an explicit rule frame-

To check our hypothesis it is also necessary to consider the service costs. An increasing activity time does not always imply increasing service costs. In the CMFS for each forecasted and actually used activity a human resource with the corresponding personnel costs is available. The activity “process engineering” is an example where an increasing activity time does not imply higher service costs, compare also Figure 7.

Figure 7: Service cost results for scenario This means the required human resources were twice as high as the forecasted but our CMFS prototype shows that the actual service costs are lower. We can justify this with lower salaries provided by the post-calculation data. Based on the presented example we conclude that our stated hypothesis is tested positive. But with the restriction that additional tests are necessary to reprove it. The prototype evaluation leads us to the following assumption: The developed CMFS service costing approach and prototype provides precise and useful forecast information about human resources and the associated capacity utilization. Compared to other

1308 1306

work to derive the values from customer attributes (e.g. if the customer is a small company the uncertainty may be higher than for company with many experts). The attributes may also be part of a service level agreement. In further experiments we want to simulate hundreds of customers with each having many service invocations. In this experiment we want to study the learning of the ratings. The new accounting/costing method may be too complex. Thus a next step in our approach is to investigate how our approach can be integrated most efficiently in an enterprise resource planning (ERP) system. Here additional information about customers has to be managed and the integration with human resources data has to be realized. We also plan to carry out additional statistical analysis of our evaluation results to refine the proposed approach.

7. References

[11]

[12] [13]

[14]

[15]

[16] [17]

[1]

Spohrer, J., Maglio, P., Bailey, J. & Gruhl, D. (2007) Steps Toward a Science of Service Systems, IEEE Computer, January., pp. 71-77. [2] Vargo, St.L. and Lusch, R.F. (2004) Evolving to a new Dominant Logic for Marketing, Journal of Marketing 68, p. 1-17. [3] Spohrer, J., Vargo , S., Caswell, N. & Maglio. (2008) The Service System is the Basic Abstraction of Service Science, Proceedings of the 41st Hawaii International Conference on System Sciences [4] Geerts, G.L., McCarthy, W.E. (2002) An Ontological Analysis of the Economic Primitives of the ExtendedREA Enterprise Information Architecture. International Journal of Accounting Information Systems 3, 1–16 (2002) [5] McCarthy, W.E. (1982) The REA Accounting Model: A Generalized Framework for Accounting Systems in A Shared Data Environment. The Accounting Review, 554–578. [6] And de Vries W., Kasper H., van Helsdingen P., “Service Marketing Management – An International Perpective”, Pages 8-20, ISBN 0-471-98490-6, John Wiley & Sons, West Sussex, England (1999) [7] Berry L. L., Parasuraman a., Zeithaml V. A., “Delivering Quality Service – Balancing Customer Perceptions and Expectations”, The Free Press (1990) [8] Edmond D., O’Sullivan J., ter Hofstede A., “Service Description: A survey of the general nature of services”, http://www.service-description.com/papers/trdapd-osullivan.pdf (2002) [9] Zeithaml V. A., Bitner M. J., „Service Marketing – Integrating Customer Factor Across the Firm“, third edition, ISBN 0-07-247142-5, McGraw-Hill/Irwin, New York (2003) [10] Erkoyuncu J. A., Roy R., Shehab E., Cheruvu K., “Understanding service uncertainties in industrial product-service system cost estimation”, The International Journal of Advanced Manufacturing Technology,

[18]

[19]

[20]

[21] [22]

[23]

[24]

1309 1307

Volume 52, Numbers 9-12, Springer-Verlag, London (2010) Xu X., Wang Z., “State of the art: business services and its impacts on manufacturing”, Journal of Intelligent Manufacturing, Springer-Verlag, (2009) Wattanakamolchai S., “Managing Customer Participation in the Service Production Process”, 2009 Flandel G. Blaga S., “Elements of the production of services”, Modern Concepts of the Theory of the Firm – Managing Enterprises of the New Economy”, ISBN 3-540-40509, Pages 175-189, Springer-Verlag, Berlin Heidelberg (2004) Helm S., Kuhl M., “Quality Uncertainty and Customer Integration – The Vendor’s Perspective”, Interaction, Relationships and Networks in Business Markets, Proceedings of the 13. IMP-Conference, Pages 239-261, Lyon (1997) Bruhn M., Georgi D., “Service Marketing – Managing The Service Value Chain”, ISB 978-0-27368-157,1, Pages, 5-27, Person, Essex (2010) Frisch R., “Theory of Production”, ISB 9789048183340, Pages 3-5, Springer Netherlands (2009) Frauendorf J., “Customer Processes in Business-toBusiness Service Transactions”, ISBN 13 978-3-83500601-0, Pages 60-62, Deutscher Universitätsverlag, Wiesbaden (2006) Kucera D., “Costing methods (manufacturing)”, http://www.referenceforbusiness.com/encyclopedia/Co s-Des/Costing-Methods-Manufacturing.html, last call 15.01.2011 Myers J., “Traditional versus activity-based product costing methods: a field study in a defense electronics manufacturing company, Proceedings of ASBBS, Volume 16, Number 1, (2009) Antos J., Brimson J. A., “Activity-based management for service organizations, government entities, and nonprofits”, ISBN 0-471-33158-9, John Wiley & Sons Inc., New York, USA (1994) Cooper R., Kaplan R. S. „Measure Costs Right: Make the Right Decisions“, Harvard Business Review (1988) Copper R., Kaplan R. S., “Cost and Effect: Using Integrated Cost Systems to Drive Profitability and Performance”, ISBN 978-0875847887, Mcgraw-Hill Professional,1997 Anderson S. R., Kaplan R. S., „Time-Driven ActivityBased Costing: A Simpler and More Powerful Path to Higher Profits“, ISBN 978-1422101711, Mcgraw-Hill Professional, 2007 Szychta A., “Time-Driven Activity-Based Costing in Service Industries, Social Sciences/Socialiniai Mokslai, Nr. 1 (67), 2010