The influencing factors on implementing a
Total Cost of Ownership approach
Master’s Thesis
Kevin van der Veen, s2606267
MSc Technology and Operations Management Faculty of Economics and Business
University of Groningen
The influencing factors on implementing a Total Cost of
Ownership approach
Master’s Thesis
First Supervisor
dr. Ir. Wilfred Alsem
Second Supervisor
dr. Jasper Veldman Faculty of Economics and Business Department of Operations University of Groningen Nettelbosje 2 9747 AE Groningen Company Supervisor Willem Dieterman Groningen Seaports Handelskade Oost 1 9934 AR Delfzijl
Kevin van der Veen, S2606267
Abstract
The Total Cost of Ownership is identified as a promising tool to support asset lifecycle management by considering the costs of an asset throughout its lifecycle. However, the adoption of this methodology seems to be limited, due to a variety of influencing factors. These factors include the lack of a standard model for practical implementation, and high complexity. This research provides a discussion of the TCO methodology and these influencing factors, and aims to answer the main research question:
Which factors influence the practical implementation of a TCO approach?
The research is based around the Action Research methodology. Action Research allows the researcher to actively influence the research process, which is focused on a practical problem (i.e. a case study). The practical problem in this research is the investment decision of a floating pier, in the port of Delfzijl. The asset is a low-tech pier, designed to dock inland ships. Its adverse environment (salt water and air) makes the asset unique. The TCO methodology has been chosen by the asset owner to (1) determine the total cost of ownership, and (2) identify whether a tradeoff exists between investment costs and costs to maintain and operate the asset.
The research results show that some factors identified the literature review were indeed found in the case study. First, the identified factors are related to data gathering (unavailable, difficult collection due to aggregation and allocation, and lack of quality and credibility), ICT systems (absence of ICT-system, and ICT system not designed for TCO analysis), and resources (lack of time to develop, implement, and maintain models).Second, the research identified the operating procedures which determine the allocation and aggregation of data to influence the accuracy and availability of data. In addition, the research suggests that the uncertainty introduced by the need to predict costs incurred in the future also influences the accuracy of the data. As such, standard operating procedures and lifecycle uncertainty are added as factors influencing the practical implementation of a TCO-approach. In contrast to the literature, the research found the model type (standard, hybrid, unique), and the model execution (formal, informal) are prescribed by the purpose of the model, rather than selected as input variables by the researcher.
Preface
In order to finish my master’s degree Technology and Operations Management at the University of Groningen, I started work on this thesis at Groningen Seaports in March 2015. Now, almost five months later, it has developed into a body of theory that promises to enhance the current literature, and provides critical insights into the future implementation of a Total Cost of Ownership approach to support the Asset Lifecycle Management at Groningen Seaports.
First, I would like to thank all employees at Groningen Seaports for their time and effort to support my research. In particular, I would like to thank Willem Dieterman for his supervision and guidance throughout this research. Furthermore, I would like to thank Sebo Kroes, Erwin Flikkema, Wim Hanraads, and René Meinders for their introduction into the activities of Groningen Seaports, and the Port of Delfzijl, and their support that led to the findings in this research. Also, I would like to extend my gratitude to Theo Smit for his referral, and Henk Blaauw, for providing me the opportunity to work with Groningen Seaports.
Second, I would like to thank my first supervisor, dr. ir. Wilfred Alsem, for his support, feedback, and extensive practical and academic knowledge. I would also like to thank dr. Jasper Veldman for his role as my second supervisor, for providing another perspective through his comments and feedback. Finally, I would like to thank my family and friends for their input and patience. Indeed, without the stability, support, and kindness of home, I would never have had the opportunity to finish my master. Performing this research at Groningen Seaports the engagement of their employees, their focus on excellence, and their open and welcoming attitude, have made this educational endeavor a successful one, and a memorable and pleasurable experience.
June 22, 2015
Table of Contents
Abstract ... 5 Preface ... 7 Table of Appendices ... 10 Table of Figures ... 10 Table of Tables ... 10 1 Introduction ... 11 2 Theoretical Background ... 132.1 Cost Based Lifecycle Approaches ... 14
2.2 Total Cost of Ownership... 14
2.3 Model type and execution ... 15
2.4 Cost Estimating Methods ... 17
2.5 Influencing factors of Total Cost of Ownership ... 18
2.5.1 Approach ... 18 2.5.2 Complexity ... 21 2.5.3 Data ... 21 2.5.4 ICT Systems ... 22 2.5.5 Knowledge ... 23 2.5.6 Resistance ... 23 2.5.7 Resources ... 24 2.5.8 Management Structure ... 24 2.6 Conceptual Model ... 25 3 Methodology ... 27 3.1 Research questions ... 27 3.2 Action Research ... 27 3.3 Interview ... 29 3.4 Validity ... 29
3.5 Research Setting and Case Description... 31
3.5.1 ‘Drijvende Steiger’ ... 31
3.5.2 Investment decision ... 32
3.5.1 Stakeholders... 33
4 Research Account... 35
4.1 Identifying the Need and Purpose ... 35
4.2 Identifying and Selecting Cost Drivers ... 36
4.3 Gathering Cost Data ... 37
4.4 Model Building ... 39 4.4.1 Model Structure ... 39 4.4.2 Model Analysis ... 40 4.4.3 Informal Model ... 41 4.5 Model Use ... 41 5 Analysis ... 43
5.1 Model type and model execution ... 43
5.2 ICT systems... 44
5.3 Data accuracy and availability ... 44
5.4 Cost driver identification and Selection ... 45
5.5 New Factors ... 46
5.6 Summary ... 46
6 Results ... 47
6.1 Adapted Conceptual Model ... 47
6.2 Sub questions ... 49
6.3 Main research question ... 52
7 Discussion ... 55
7.1 Implications ... 55
7.2 Limitations and further research ... 56
Table of Appendices
Appendix I : Questionnaire ... 61
Appendix II : Interview ... 65
Appendix III : Interview Coding ... 71
Appendix IV : RAW-Specifications and Conditions ... 75
Appendix V : Total Cost of Ownership Model ... 79
Appendix VI : 2013 Account Ledger ... 87
Table of Figures
Figure 2.1: Asset Life Cycle (Adapted from Blanchard & Fabrycky, 1998) ... 13Figure 2.2: Estimating methods versus program phase ... 18
Figure 2.3: Implementation Process ... 25
Figure 2.4: Conceptual Model Implementation Process ... 26
Figure 3.1: Action Research Cycle (Coughlan & Coghlan, 2002: 230) ... 28
Figure 3.2: Port of Delfzijl with the Floating Pier ... 32
Figure 3.3: Stakeholder Map ... 33
Figure 4.1: Research process with iterations ... 35
Figure 4.2: TCO-model excerpt ... 39
Figure 4.3: Cost category costs as a percentage of the TCO ... 40
Figure 6.1: Model type and Execution ... 47
Figure 6.2: Relations between Data accuracy and Data availability. ... 48
Figure 6.3: Cost Driver Identification and Selection ... 48
Figure 6.4: Adapted Conceptual Model ... 49
Table of Tables
Table 2.1: Cost Estimating Methodologies (Adapted from Fabrycky & Blanchard, 1991; Pohl & Nachtmann, 2007) ... 17Table 2.2: Influencing factors found in literature ... 19
Table 3.1: Sub-questions ... 27
Table 3.2: Reliability and Validity in Case Research (Adapted from Karlsson, 2009: 182) ... 30
Table 4.1: Questionnaire Results (n=4) ... 41
Table 5.1: Comparison of model properties depending on purpose ... 43
Table 5.2: Availability, accuracy, and source of data per cost driver or category ... 45
1 Introduction
Modern asset management has refocused management efforts to consider the entire lifecycle of an asset (ISO 55000, 2014). In the early stages of an asset’s lifecycle the most critical decisions are made, both financial and operational. A promising tool to help assess the lifecycle costs is total cost of ownership. However, research suggests the adoption seems to be lacking (Ellram & Siferd, 1993; Korpi & Ala-Risku, 2008). This study aims to find the factors influencing a practical implementation of a total cost of ownership approach.
As global competition increases, the need to reduce costs increases as well, so that asset management takes on an increasingly important role. Although research has identified possible reasons for the slow adoption of Total Cost of Ownership approaches (Ferrin & Plank, 2002; Muchiri & Pintelon, 2008; Schuman & Brent, 2005), little evidence exists on the enabling and inhibiting factors of the practical implementation of lifecycle approaches. The identification of enabling and inhibiting factors adds to the current state of the art twofold. First, the identified factors may provide an explanation for the slow adoption of lifecycle approaches. Second, the factors may enhance the usability of frameworks by providing obstacles that can present themselves during implementation (Bastl, Grubic, Templar, Harrison, & Fan, 2010).
This research focusses on a case study within Groningen Seaports. Groningen Seaports is responsible for the development of the Eemshaven and port of Delfzijl. The asset under consideration is the floating pier located in the port of Delfzijl. This pier is considered ‘End-Of-Life’, and must either be replaced or completely overhauled. Since the Total Cost of Ownership methodology has not yet been adopted by Groningen Seaports, the company provides a real world perspective regarding a practical implementation of a Total Cost of Ownership approach. The research is conducted using an action research case study, which allows a hands on approach to identify the factors influencing this implementation.
This research will test current literature, and aims to find new factors relating to the implementation of a Total Cost of Ownership approach. As part of the implementation procedure a TCO model is developed. Given the practical and theoretical inputs for this research, the research question is defined as:
Which factors influence the practical implementation of a TCO approach?
implementation of a TCO model, and possibly identify new factors that influence the practical implementation.
This study aims to contribute to existing knowledge by both testing current factors related to TCO implementation, and finding novel factors found in practice. By linking theory to practice, and vice versa, the study aims to expand current and future research to include potential obstacles during implementation.
2 Theoretical Background
Asset management has become increasingly important, and has been merged into international standards over the last two decades (ISO 55000, 2014; preceded by, PAS 55-1, 2008)
Based on the PAS 55-1 (2008) standard, asset management is defined as “Systematic and coordinated
activities and practices through which an organization optimally and sustainably manages its assets and asset systems, their associated performance, risks and expenditures over their lifecycles for the
purpose of achieving its organizational strategic plan” [Emphasis added].
As emphasized by the definition above, asset management aims to optimally manage assets over their lifecycles. Consideration of the lifecycle may provide valuable insights in the distribution and allocation of costs (Ellram & Siferd, 1993). This research follows the six phase lifecycle suggested by Blanchard & Fabrycky (1998), shown in figure 2.1.
Figure 2.1: Asset Life Cycle (Adapted from Blanchard & Fabrycky, 1998)
2.1
Cost Based Lifecycle Approaches
Literature suggests a large number of lifecycle approaches, which differ in detail, and sometimes in application (Ferrin & Plank, 2002; Goh, Newnes, Mileham, McMahon, & Saravi, 2010). Korpi & Ala-Risku (2008) identified Life Cycle Costing (LCC) and Total Cost of Ownership (TCO) as two important concepts. Other research adds Through Life Costing (TLC) and Whole Life Costing (WLC) (Goh et al., 2010), and Total Cost (TC) and Product Life Cycle Cost (PLCC) (Ferrin & Plank, 2002).
Life Cycle Costing, Total Cost of Ownership, Through Life Costing, Whole Life Costing, Total Cost, and Product Life Cycle Cost all consider cost as their main driver. Researchers argues their differences lie in the costs considered, the scope of uncertainty included in the analysis, and the intended use in the lifecycle (Asiedu & Gu, 1998; Durairaj, Ong, Nee, & Tan, 2002; Ellram, 1995a). For example, Ellram (1995a) and Ellram and Sifered (1998) argue Life Cycle Costing should be considered as a subset of Total Cost of Ownership, as the only difference is the exclusion of pre-transaction costs in Life Cycle Costing. Other research argues Through Life Costing, Whole Life Costing, and Product Life Cycle Cost are focused on estimating costs during the design phase of a product or service (Asiedu & Gu, 1998), whereas Total Cost of Ownership is considered a purchasing methodology designed for supplier selection. Despite these differences, researchers argue the names of these approaches are interchangeable, and the concepts can be used regardless of their intended life cycle phase, as they all focus on identifying the costs associated with the design, purchase, manufacturing, operation, and disposal of a product (Ferrin & Plank, 2002; Goh et al., 2010; Settanni, Newnes, Thenent, Parry, & Goh, 2014).
As suggested by Ellram (1995a), and following the arguments of Ferrin & Plank (2002), Goh et al. (2010), and Settanni et al. (2014), the approaches are considered either subsets of Total Cost of Ownership, or interchangeable names for the same concept. As such, the broad scope of Total Cost of Ownership is considered to be superior when considering the lifecycle of an asset. Therefore, it will be the focal point for the remainder of this approach. The following section will give an introduction to the Total Cost of Ownership approach.
2.2
Total Cost of Ownership
Siferd, 1993, 1998; Ferrin & Plank, 2002; Korpi & Ala-Risku, 2008), and empirical research on the overall use of the concept (Bastl et al., 2010; Ellram & Siferd, 1993; Ferrin & Plank, 2002).
Ellram & Siferd (1998: 56) defined the total cost of ownership concept as a “Purchasing tool and
philosophy aimed at understanding the relevant costs of buying a particular good or service from a particular supplier”. This includes the ability to influence decision making in the design and
engineering stages of the purchase by identifying tradeoffs between the costs incurred in different lifecycle phases. This is especially relevant as research suggests that most costs during the life cycle, between 70 and 85 percent, is defined during these stages (Asiedu & Gu, 1998; Dowlatshahi, 1992). In addition, Ellram & Siferd (1998) argue that TCO adds strategic value through its ability to support supply chain analysis and cost driver analysis, ultimately impacting the financial bottom line. Further research suggests the TCO methodology is useful beyond enhancing (purchasing) decision making. Both Cavinato (1992) and Caniato et al. (2014) use Total Cost of Ownership as a tool to analyze and optimize the supply chain. Similarly, Ellram & Siferd (1993) state the analytical approach provided by TCO allows the company to identify wasteful activities, and identify areas for (continuous) improvement.
Within Total Cost of Ownership costs are identified by analyzing the flows and activities within each process related to an asset. Ellram (1994) suggests the TCO analysis should include the pre-transaction, pre-transaction, and post-transaction costs (see also Ferrin & Plank, 2002), and includes both direct and indirect costs.
As the above literature suggests the TCO methodology has different uses, and any application of the methodology starts with a perceived need and a purpose for the analysis. Subsequently, a model is developed that will support the need and purpose throughout its use. Finally, the model is used and maintained throughout the purchase. Furthermore, identifying and selecting cost drivers is seen as an input for the model building step. These three steps together can be seen as the implementation process, and will be considered the basis of this research and the underlying conceptual model.
2.3
Model type and execution
the practical use of unique and standard models. Ferrin & Plank (2002) found 65.8 percent of their sample believed a combination of standard and unique models are necessary to estimate the Total Cost of Ownership of the wide variety of products and services their company purchases. Moreover, within their sample, 67.6 percent of the respondents were able to identify a set of core cost drivers that apply to every product or product category. Their research suggests modular models (i.e. standard models that are adapted into unique models) are common, but no further typology is found (Ferrin & Plank, 2002). For the purpose of this research this type of model is described as a Hybrid, given the ability to transition from a standard model to a unique model, and vice versa.
Second, empirical research shows practice differentiates between formal and informal execution of a Total Cost of Ownership model. Ellram (1995a) defines a formal model as “as a written, documented
method for determining the total costs associated with the acquisition and subsequent use and disposition of a given item/service from a given supplier”. To the best of the author’s knowledge, only
Ellram & Siferd (1993) empirically researched the use of formal and informal models. They found only 18 percent of their sample uses a formal approach, 58 percent uses an informal approach, while the remaining 24 percent uses no TCO approach. A possible explanation for the absence of information on this subject is also given by Ellram & Siferd (1993: 172); they state “Without a formal approach, it
is difficult and time consuming to determine cost figures. Such difficulties create inaccuracy and harm credibility”. As such, research focusses mostly on the formal implementation of TCO models (for
example, Caniato et al., 2014; Ellram & Siferd, 1993, 1998; Ellram, 1995a; Tysseland, 2008; Zachariassen & Arlbjørn, 2011)
From the literature we can conclude the model type (Standard, Hybrid, or Unique), and model execution (i.e. formal vs. informal) are factors influencing the practical implementation of a TCO model. Given the apparent difficulties associated with both variables, sub-question one is defined as:
2.4
Cost Estimating Methods
In order to identify and select the cost drivers, and estimate their associated costs, research suggests three methodologies, estimating by analogy, estimating by engineering procedures, and parametric estimating methods. An explanation of each methodology is shown in table 2.1.
Table 2.1: Cost Estimating Methodologies (Adapted from Fabrycky & Blanchard, 1991; Pohl & Nachtmann, 2007) Type
Description
Advantages Disadvantages
Analogy
Estimating cost based on an analogous activity or asset.
Can be based on historical data
Quick Subjective
Based on actual data Accuracy depends on similarities Requires few data Difficult to assess effect of
design change
Known origin Requires case database, similarity measure, adaption functions and case indexations are required
No requirement of full understanding of the problem
Does not handle innovation Accurate if differences between
analogous case are limited Engineering Procedures
Estimating cost by specifying each task, equipment, and material needed to complete the product or asset, and allocating cost to each element
More accurate than analogy Slow execution Detailed breakdown useful for
negotiation
Data may not be available Suitable when all characteristics
are well defined
Inappropriate for estimation at design stage
Inaccurate allocation of overheads.
Parametric Estimation
Estimating cost based on a functional, statistical relationship between changes in costs and the factors upon which the cost depend
Quick Parameters not included can
become important
Repeatable and objective Useful in combination with other methods
Less information required than engineering procedures
High uncertainty as
specifications are not available Good for budgetary estimates
or baseline assessments
Figure 2.2: Estimating methods versus program phase
2.5
Influencing factors of Total Cost of Ownership
The Total Cost of Ownership literature already identifies several barriers to successful TCO use. The factors were listed, and assigned two or more keywords according to the subject of each author’s statement. Where possible statements were combined and categorized according to their first keyword. The categories, and corresponding issues, will be discussed below, and are summarized in table 2.2 (see page 19).
2.5.1 Approach
as ter’ s T he sis Kev in v an d er V een 19
Table 2.2: Influencing factors found in literature
Factor Description Reference
Approach
Context specific nature (Ellram, 1995a; Korpi & Ala-Risku, 2008)
Lack of a standard (implementation) model (Ellram & Siferd, 1998; Ellram, 1995a; Ferrin & Plank, 2002; Waeyenbergh & Pintelon, 2002)
Complexity
General complexity (Ellram & Siferd, 1998; Ellram, 1995a)
Complexity of costing models/methodology (Bastl et al., 2010)
Complexity of internal operations (Bastl et al., 2010)
Data
Unavailability or non-existence of data (Bastl et al., 2010; Ellram & Siferd, 1993, 1998; Ellram, 1994,
1995a; Gluch & Baumann, 2004) Uncertainty of forecasting demand and product lifecycle (Waeyenbergh & Pintelon, 2002) Difficult collection of data(aggregation and allocation) (Ellram & Siferd, 1993)
Lack of quality and credibility of data (Bastl et al., 2010; Gluch & Baumann, 2004)
Resistance to identify and share data (Bastl et al., 2010; Ellram & Siferd, 1993)
Difficulty in identifying cost drivers (Bastl et al., 2010; Ellram, 1994; Ferrin & Plank, 2002)
Difficulty in selecting cost drivers (Caniato et al., 2014; Ellram, 1994; Ferrin & Plank, 2002)
ICT Systems
Lack of (ICT) systems (Ellram, 1994)
ICT Systems not designed to support TCO analysis (Bastl et al., 2010; Ellram & Siferd, 1993)
Knowledge
Difficulty in determining when and how to use models (Ellram & Siferd, 1998; Ellram, 1994)
Difficulty in defining the scope of the model (Ellram, 1994)
Lack of understanding the methods and models (Ellram, 1994; Gluch & Baumann, 2004)
Lack of understanding of the benefits (Ellram, 1994; Gluch & Baumann, 2004)
Lack of financial background (Bastl et al., 2010)
M as ter’ s T he sis Kev in v an d er V een 20
Table 2.2: Influencing factors found in literature (continued)
Factor Description Reference
Resistance
Resistance to change (corporate) culture (Bastl et al., 2010; Ellram & Siferd, 1998; Ellram, 1995a)
Resistance to change from ‘price’ orientation (Ellram, 1994)
Resistance to change (i.e. procedures, authority, etc.) (Bastl et al., 2010; Ellram, 1994) Resistance due to lack of knowledge
(‘Not-Invented-Here’-syndrome (Ellram, 1994)
Lack of internal alignment (functional silo’s) (Bastl et al., 2010; Zachariassen & Arlbjørn, 2011)
Misalignment between internal and external use (Bastl et al., 2010; Tysseland, 2008)
Resources (Time, Financial)
Lack of resources to develop (Ellram, 1994; Waeyenbergh & Pintelon, 2002)
Lack of resources to implement (Ellram, 1994)
Lack of resources to maintain/support (Ellram, 1994)
Management Structure
Lack of top-management support (Bastl et al., 2010)
Mismatch between managerial incentives and long-term
2.5.2 Complexity
Here, complexity refers to the overall complexity of a TCO approach. Ellram & Siferd (1998: 69) identify several factors that refer to this complexity, such as the large initial time investment to develop, difficulty in understanding and explaining the TCO concept, and difficulty to develop a standard system that is easy to use and provides structure. Their refinements return as detailed aspects throughout the literature, and are considered among the categories listed in table 2.2. Thus, the general concept of complexity aligns with the approach category, in that they are overarching problems and dispositions regarding the TCO approach. As such, it is not considered a part of the conceptual model. Another refinement is suggested by Bastl et al. (2010). Their research focused on the application of Inter-organizational Costing Approaches (IOC), and they consider Total Cost of Ownership as one of these approaches. They argue that the complexity of the costing models (i.e. the calculation models) has a significant effect on the implementation and use of the models, suggesting that (overly) complex models are abandoned. This is consistent with thoughts from Ellram & Siferd (1998: 70), who write
“[a] TCO analysis is not the proper approach for all buys. The potential benefits of developing the TCO analysis in support of strategic cost management must exceed the cost”.
In addition to the complexity of models Bastl et al. (2010) suggest the complexity of the company’s operations may hinder the use of both IOC and TCO models. Logic dictates that these are somehow related, as complex operations will result in complex models when these models are derived from the operations.
2.5.3 Data
Total Cost of Ownership, similar to most Life Cycle based methodologies, is data driven. Researchers argue that the availability of data is a prominent issue influencing the development, implementation and use of TCO models. (Bastl et al., 2010; Ellram & Siferd, 1993, 1998; Ellram, 1994, 1995a; Gluch & Baumann, 2004). The unavailability of data, however, does not necessarily mean this data does not exist. Research suggests the availability of data is often limited by the ICT systems used (discussed in section 2.5.4), or is only available as summarized, aggregated data (Bastl et al., 2010; Ellram & Siferd, 1993; Ellram, 1994, 1995b). However, the availability of aggregated data is often based on the detailed cost data, hidden in the system. Indeed, Ellram & Siferd (1993: 173) suggest “This information does
exist, but some detective work and painful data collection may be necessary to establish it”. It is this
The availability and accuracy of the data are considered to be factors that influence the ‘Model Building’ phase of the implementation model, and are added to the conceptual model. Sub questions two, three and four relate to the availability, accuracy, and source of the data.
SQ 2: How does the availability of data affect the TCO implementation? SQ 3: How does the accuracy of the data affect the TCO model?
SQ 4: How does the source of the data affect the TCO model?
Another issue is the credibility of the data, especially when TCO approaches are used during negotiations with suppliers (Zachariassen & Arlbjørn, 2011). Furthermore, the collected data must have both a high quality (i.e. be the right data), and be accurate. This accuracy may be affected when stochastic models are used to predict future behavior. Waeyenbergh & Pintelon (2002) relate the uncertainty to maintenance concepts, suggesting that the performance of a machine and its corresponding maintenance policies may be difficult to estimate due to this uncertainty. As such, it is important to consider the effect of the lifecycle on the accuracy of the data. Sub question five is defined as:
SQ 5: How does the uncertainty of estimating future costs introduced by the consideration of the asset lifecycle relate to the accuracy of data?
A survey conducted by Ferrin & Plank (2002) identified a total of 135 cost drivers, spread over 13 categories. The large amount of possible cost drivers make it difficult to identify the relevant cost drivers, and subsequently select these cost drivers (Bastl et al., 2010; Caniato et al., 2014; Ellram, 1994; Ferrin & Plank, 2002). Given the difficulties with assessing costs and cost drivers, sub question six is defined as:
SQ 6: How does the identification and selection of cost drivers influence the implementation process?
2.5.4 ICT Systems
the business conditions and the use of information systems (Welker et al., 2008: 718). They suggest the internal focus of ICT systems may be an explanation for the limited external use. This seems consistent with the findings of Bastl et al. (2010) and Ellram (1994) that systems are not designed to support data intensive methodologies.
These factors support the inclusion of ICT systems as a factor to data accuracy In addition, the effect of the ICT system on the implementation process and data will be considered with sub question seven:
SQ 7: How does ICT influence the availability of data, and the implementation process?
2.5.5 Knowledge
Knowledge was found to be a considerable factor influencing the implementation of Total Cost of Ownership approaches. Ellram (1994) and Ellram & Siferd (1998) identified difficulty in how and when to use TCO analysis, and subsequently difficulty defining the scope to be a limiting factor for the use of the methodology. These factors are linked to a lack of understanding of the method, and the inability to perceive its benefits (Ellram, 1994; Gluch & Baumann, 2004), and both issues cause a reluctance to use the methodology. In addition, understanding of TCO models and calculations is influenced by the users financial knowledge (Bastl et al., 2010)..
Both Ellram (1994) and Tysseland (2008) identified the lack of training to be an inhibiting factor. Indeed, research outside of the TCO literature emphasizes the need of end-user training to overcome knowledge and understanding related problems. Examples include the implementation of ICT systems (e.g. ERP, MRP, CAD-CAM systems) (Dezdar & Sulaiman, 2009; Nelson & Cheney, 1987), the use of lean practices (Nicholas, 2011), or the use of costing and performance measurement structures (Merchant & van der Stede, 2012). Reviewing the knowledge related factors, and at the risk of stating the obvious, employee training has the potential to overcome all these inhibiting factors, and should therefore be part of the implementation process.
2.5.6 Resistance
Another limiting category is resistance to change. This includes changes in organizational culture (Bastl et al., 2010; Ellram & Siferd, 1998; Ellram, 1995a), changes in operating procedures (Bastl et al., 2010; Ellram, 1994), and resistance to implementation. Neither of these factors are easily avoided, however, research exists on common aspects relating to resistance (see, for example, Robbins & Judge, 2014; Umble, Haft, & Umble, 2003).
due to misconceptions about the TCO approach. Ellram (1994) further suggests the resistance caused by this factor can be reduced through education and training.
Bastl et al. (2010) and Ellram (1994) identified resistance due to changes in authority and function after the implementation. They suggest that employees are unwilling to step down from their current level of authority, and will use whatever power they have to influence the implementation to their preference. In addition, the implementation of the TCO approach requires alignment of the various departments within the company (Bastl et al., 2010; Tysseland, 2008). Internal alignment, or ‘breaking down the functional silos’ ensures all departments work together to support the TCO approach by, for example, accurately tracking and allocating costs. Such alignment is considered an antecedent to success within modern approaches in supply chain management (see for example Lambert & Cooper, 2000)
Since use of the TCO models is the ultimate goal of the implementation process, resistance is considered an influencing factor. As such, sub-question eight is defined as:
SQ 8: How does resistance to change influence the implementation of a TCO model?
2.5.7 Resources
Researchers also considered a lack of resources to be an inhibiting factor. Their focus was mainly on the availability of time and financial resources to develop, implement, and subsequently maintain the TCO models (Ellram, 1995a; Waeyenbergh & Pintelon, 2002). The availability of time and financial resources can be linked to the complexity associated with the development of the models and calculations, as discussed in sections 2.5.1, 2.5.2, and 2.5.3.
Given the research setting, the relatively low complexity of the asset under consideration, and the purpose of this research (see also chapter 3), this factor cannot be accurately researched. Therefore it is excluded from the research and conceptual model.
2.5.8 Management Structure
Management support is considered to be an important aspect for the successful implementation of TCO models. Bastl et al. (2010) suggest that management support reinforces the commitment of employees to the implementation. In addition, it is suggested to increase the knowledge and cooperation of the employees.
control. Because the Total Cost of Ownership approach is based on the tradeoff of costs between lifecycles when considering alternatives, this means departments are no longer in full control of the costs associated with their activities. For example the tradeoff between a lower purchasing price and higher operating costs to achieve the lowest total cost of ownership directly impacts the minimum costs the manufacturing department must make. As such, cost reduction may not be the best performance measurement for the manufacturing department. Instead, more process oriented measures may be used to assess manufacturing performance, such as product cycle times or number of defects (Hopp & Spearman, 2011; Merchant & van der Stede, 2012; Nicholas, 2011; Slack & Lewis, 2011).
While management support is clearly of great importance to the success of any TCO implementation, the support for the methodology has already been granted. In addition, the use of a pilot project (see also chapter 3) limits the potentially disturbing influence lack of management support may have. As such, this factor falls outside of the scope of this research.
2.6
Conceptual Model
Figure 2.3 shows the conceptual model generated from the literature. As suggested in section 2.2 and 2.3, the implementation process starts with the identification of the need and purpose, the design of the TCO model, and the model use. The model design step is decomposed into the identification of the cost drivers, gathering of the necessary cost data, and building the model. These steps are shown in figure 2.3.
Figure 2.3: Implementation Process
Based on the factors identified in section 2.5 the factors influencing the practical implementation of TCO models are combined into a conceptual mode. The conceptual model is shown in figure 2.4. The expected links between the previously identified factors and the implementation steps are also shown, while the scope of this research is indicated with the dashed line.
Figure 2.4: Conceptual Model Implementation Process
To focus the research the ‘knowledge & experience’, ‘management structure & support’, and the ‘continuous improvement’ are excluded. Both gained ‘knowledge & experience’ and ‘continuous improvement’ are long-term effects, which fall outside the time constraints of this research.
3 Methodology
First, the research questions developed in chapter 2 are summarized, and linked to the most appropriate research method. Then the research methods are discussed in more detail, the validity is addressed, and the research setting and case study are discussed.
3.1
Research questions
As suggested in the introduction the main research question is defined as:
Which factors influence the practical implementation of a TCO approach?
In order to further examine the factors found during the literature review, sub-questions were derived, as shown throughout chapter 2. The sub-questions are summarized in table 3.1.
Table 3.1: Sub-questions
Sub-question Methodology
SQ 1: How does model type and model execution influence the implementation process?
AR & INT
SQ 2: How does the availability of data affect the TCO implementation? AR
SQ 3: How does the accuracy of the data affect the TCO model? AR & INT
SQ 4: How does the source of the data affect the TCO model? AR & INT
SQ 5: How does the uncertainty of estimating future costs introduced by the consideration of the asset lifecycle relate to the accuracy of data?
AR
SQ 6: How does the identification and selection of cost drivers influence the implementation process?
AR & INT
SQ 7: How does ICT influence the availability of data, and the implementation process? AR
SQ 8: How does resistance to change influence the implementation of a TCO model? AR & INT
AR = Action research; INT = Interview
Table 3.1 also shows the proposed methodology to answer the sub-question. Overall, a mix of action research and interviews are used to answer the question. The proposed methodologies are explained in more detail below.
3.2
Action Research
considering tradeoffs that may include commercial judgement. Action research further differs from other types of research in that it develops a solution to a practical problem, while using this process to enhance the theory (Coughlan & Coghlan, 2002, 2009; Westbrook, 1995).
It is this dual role that makes it especially suited for this research, as it allows participation in the implementation process in order to confirm and identify factors at influence that process. Coughlan & Coghlan (2002: 227) argue “In general [Action Research] is appropriate when the research question
relates to describing an unfolding series of actions over time in a given group, community or organization”. Indeed, researchers suggest the methodology is often used to study implementation
processes. Action research is conducted based on the action research cycle, shown in figure 3.1.
Figure 3.1: Action Research Cycle (Coughlan & Coghlan, 2002: 230)
The action research cycle consists of a pre-research step to identify the context and purpose of the action research. Then, the six steps shown in the figure are followed, and repeated until a solution is found. The monitoring step ensures the process is monitored and documented, to aid validity (Coughlan & Coghlan, 2009; Westbrook, 1995).
Despite the apparent pitfalls of the methodology, Action Research is still considered to be an appropriate form due to its depth and ability to find novel insights. Indeed, Bamberger & Pratt (2010) argue knowledge often flows form practitioners to academics, and suggest these findings may focus academic efforts in the search for knowledge.
3.3
Interview
Interviews with stakeholders will be performed to acquire the insights necessary to perform the action research, while providing the guidance of a well-documented methodology to enhance validity. Where possible, the interviews will be semi-structured. Whenever relevant information was gathered using informal, non-structured interviews, either quantitative proof (i.e. financial data, documentation) or a follow-up interview will be scheduled.
Semi-structured interviews allow the collection of qualitative data, such as the problems encountered and the perspective of the informant, while also providing the opportunity to check aspects already found in literature (Blumberg, Cooper, & Schindler, 2011). Further, the qualitative aspect of semi-structured interviews captures the feedback and viewpoints of the departments linked to the lifecycle necessary to identify areas for future improvements. Open ended questions allow the informant to elaborate on the subject, adding to the depth of information.
Prior to the interview, the predefined questions are shared among the interviewees. This allows them to prepare, and gather the necessary data to accurately answer the question (Karlsson, 2009). Finally, the questioning of multiple actors involved in the implementation process allows triangulation of observations, adding to the reliability of the research. The interviews were held in the native language of the interviewees, to prevent a language bias. The data is subsequently translated, analyzed, and categorized. The questionnaire is attached in appendix I, the answers can be found in appendix II, and the coding of the interview is attached in appendix III.
3.4
Validity
Researchers suggest the principles of case studies are similar to those of action research, and as such, the validity concepts applicable to case studies also apply here (Coughlan & Coghlan, 2002, 2009; Westbrook, 1995). A discussion on validity and reliability can be found in literature, such as Karlsson (2009, chapter 4 and 5). Validity in action research is discussed here.
Another concern regarding the validity is researcher impartiality. Due to the participation of the researcher the interventions may become biased. Literature suggests larger sample sizes (i.e. interviewees) may mitigate this effect. Furthermore, the type of questioning used may limit the subjectivity. Coughlan & Coghlan (2002) suggest the combination of advocacy (i.e. explicitly stating the goal, asserting and option, perception, feeling or proposal for action), with inquire (i.e. questioning participants to understand their perspectives and views) forces the illustration of inferences with data. Both concerns are addressed in the construct validity and reliability aspects of the case study methodology (see also table 3.2).
Table 3.2: Reliability and Validity in Case Research (Adapted from Karlsson, 2009: 182)
Test Tactic Phase of research
Construct validity Use multiple sources of evidence Establish chain of evidence
Have key informants review draft case study report
Data collection Data collection Composition
Internal validity Do pattern matching or explanation building
Data analysis
External validity
Reliability Use case study protocol Develop case study database
Data collection Data collection
3.4.1 Reporting Action Research
As the action research methodology allows the researcher to actively participate in the research, reporting action research is different from reporting case based research. For example, Coughlan and Coghlan (2009: 259) provide a possible structure for an action research dissertation.
3.5
Research Setting and Case Description
The case study was conducted at Groningen Seaports. Groningen Seaports is charged with the exploitation of the industrial area and ports around Delfzijl and the Eemshaven. Their core business consists of the management of assets in and around these areas. As such, they sell and lease the properties and provide and maintain the infrastructure (roads, railways, and waterways) in their area of operation. Additionally, the company hosts the Water Traffic Control, which is responsible for guiding the ships moving through the ports.
The goal of the TCO model within Groningen Seaports is twofold. First, the total cost of ownership methodology is seen as a support to lifecycle based asset management. Currently, the assets are controlled based on predefined budgets, in a time based system. As a result, it is difficult to determine relationships between assets, and optimize the assets’ performance. By using the Total Cost of Ownership approach, Groningen Seaports aims to harness its benefits in terms of planning, asset and cost control, and optimization. The company is actively pursuing an Asset Lifecycle Management System, and the implementation process and factors identified during this action research case study will aid to achieve that goal. Second, the model of the floating pier will be used to determine its lifecycle cost, and investigate opportunities to optimize the asset. Their expressed interest in the TCO concept is especially important, as it grants management support, and avoids resistance to change. The use of a ‘pilot project’ approach, as suggested by Caniato et al. (2014) and Ellram (1994), further limits the influence of strategic issues, such that the research may focus on identifying the factors influencing practical implementation. What remains is a research setting that can be considered a baseline, void of any biases, other than the willingness to implement the TCO concept.
3.5.1 ‘Drijvende Steiger’
Figure 3.2: Port of Delfzijl with the Floating Pier
The pier was designed and commissioned in 1984, and was moved to its current position in the early 1990’s. The asset can be classified as low-tech, as its load bearing structure is made entirely out of steel. It uses a total of seven poles anchored to the harbor bed to keep it in place and uses these poles as guides to move up and down with the tides. The sides of the asset are lined with wood (called Azobé) to prevent damage to the coating of the pier and the ship. Further, it is equipped with electricity and drinking water.
3.5.2 Investment decision
Over the past years the use of the pier has intensified as a result of the increased activity in the Eemshaven. This trend is expected to continue, and will change the mix of ships using the pier to change from the current mix towards predominantly inland vessels. However, the increased activity has no influence on the business case, as the expectations remain within the boundaries of the original design.
The floating pier is considered End-Of-Life (EOL), and two options are considered: (1) full renovation of the current pier, (2) disposal of the current and acquisition of a new pier.
Either option is considered to be a large investment (upwards of a million euros). This is due to the size of the asset, and the time required to execute either option.
3.5.1 Stakeholders
The stakeholders were identified at the start of the investment project. They are categorized based on the way changes in the current asset affect the respective stakeholder. The stakeholder map is shown in figure 3.3.
Figure 3.3: Stakeholder Map
Figure 3.3 shows the core stakeholders to be Groningen Seaports (owner) and the Water Traffic Control (service provider). Directly affected are the customers of the Water Traffic Control, and the suppliers of the water and electricity. Further, the maintenance department is directly affected as they are responsible for the maintenance coordination.
modification the maintenance department can make to optimize the asset, as the KPI’s related to the maintenance may have an effect on the KPI’s of the service provider.
4 Research Account
As suggested in section 3.4.1, this report will first provide a detailed account of the actions and observations in the research. As the research focusses on the implementation of a TCO-model, the research process is the same as the implementation process discussed in chapter 2.6, and figure 2.3. While figure 2.3 implies a linear process, in reality it is not. In order to gather the required information, the researcher switches back and forth between the three central steps (Identifying and Selecting Cost Drivers, Gathering Data, and Model Building). These iterations are shown in figure 4.1.
Figure 4.1: Research process with iterations
Figure 4.1 also shows which iteration, and which action research phase, answered which sub question. In order to give a more structured and detailed account of the research, the report will condense the iterations into the linear process in figure 2.3.
4.1
Identifying the Need and Purpose
As discussed in section 3.5, the goals of the TCO models are to identify the total cost of ownership of the floating pier, and to identify potential tradeoffs between lifecycles that may reduce the total cost of ownership.
Currently, the company has no asset based ICT system. Instead, files of active projects are stored on a networked server. Once projects are finished they are transferred to an archive folder. While the company does have a software package that allows data to be related to an asset, not all assets are recorded in its database, and the records are not always compete. Finally, there is an accounting system, which is only available to the finance department.
The floating pier did have a record in the database, however, there was no information associated with it, other than a drawing showing the location within the harbor.
An interview revealed there had been an investigation into the purchasing and maintenance costs, and based upon the result of that inquiry it was considered more economical to renovate the existing pier (see also section 4.4.3).
Due to the advanced state of the project, data on the pier had already been gathered. The project folder contained drawings of the pier, the information on which the tender was based, inspection reports on the structural safety, internal and external communication, and some financial data. The folder was structured, although not all folders contained files, and the files were not always located in the proper location. Furthermore, the contents of the files could not always be determined from the file names, and the source of the data was also not always apparent.
Before identifying and selecting the cost drivers, the applicable lifecycles were determined. Given the decision to renovate the pier, the final step of the acquisition phase (see also figure 2.1) is concerned with the execution of the renovation. Because the design of the pier does not change with the renovation the activities in the conceptual design (phase 1) and preliminary design phases are minimal. Therefore, only the final four phases are considered.
4.2
Identifying and Selecting Cost Drivers
Given the decision to renovate the floating pier, the TCO model will focus on this scenario. The identified and selected cost drivers, can be found in the TCO model attached in appendix V (only available in Dutch).
The identification and selection of the cost drivers during the first iteration was done using the information already available in the project folder. The project manager keeps track of the costs associated with the investment process in a budgetary overview. These included the cost drivers related to the design and implementation of the tender and the cost of employees working on the project. The identification of cost drivers for the detailed design phase was relatively easy, due to the limited number of activities needed to support the tendering process.
foundation of the cost drivers, an excerpt from the specifications and conditions is attached in appendix IV (Dutch only).
The costs of quality control were also identified as a cost driver. Quality control is especially important due to the adverse conditions (see also section 3.5.2), and are planned both during and after the renovation.
The second iteration focused on the identification of maintenance costs. The maintenance department identified two types of maintenance, failure based maintenance and time based maintenance. The failure based maintenance strategy applies to all non-critical, low risk, or low value aspects of the asset. The time based maintenance policy applies to the aspects that are governed by laws and regulations or are critical to safety, and includes preventative cleaning of the asset. These decisions are made based on practical experience or were carried over from historical decisions. Based on the maintenance policies, the maintenance costs drivers were separated into planned and unplanned costs. The planned cost elements were identified using the tender detailing the maintenance actions in the port. The unplanned cost elements were identified and categorized based on previous expenses booked on the asset’s ledger account (Dutch: grootboek-rekening). An interview aimed to discuss the ‘Water and Electricity’ costs on the ledger account, revealed an additional cost element, being the weekly inspection and maintenance of the electrical and drinking water installation. This weekly inspection is planned, while the maintenance is based on the results of the inspection. It was categorized as unplanned maintenance, as the duration of the inspection and the results vary from week to week.
The final iteration identified the cost drivers during the operation and utilization phase. These elements included the garbage collection, disposal of chemicals, the water bill, and the electricity bill. Of these elements only the garbage collection and disposal of chemicals were selected, as there is was no data available on the water and electricity bills (see also section 4.3)
Sources of revenue were also investigated. They consist of the rents paid by third parties leasing part of the pier, and rents paid by the users of the pier. Neither sources of revenue were selected as the necessary cost data was unavailable.
4.3
Gathering Cost Data
For the first iteration, the cost data was gathered using two methodologies, historical data, and parametric data. The historical data consists of the expenses already made to support the project, and the quotes given by the company responsible for the quality control. Cost data related to the project management was calculated based on the allocated hours and hourly wage. The cost data related to the renovation is also parametrically estimated by filling out the RAW-specification and conditions. This was done by the engineering company using an estimation of the current market prices.
The data for the cost drivers identified during the second iteration were determined based on the ledger account of the floating pier. The 2013 account ledger is attached in appendix VI as an example. The unplanned maintenance was categorized based on the description, and subsequently associated with the cost drivers. Cost data for the weekly inspection and maintenance was more difficult to find, as these costs not booked on the asset’s ledger. Instead, the maintenance costs are aggregated and filed under the category ‘general maintenance’ in the port of Delfzijl. As a result, there is no link with the asset, which makes it impossible to determine the cost data from the ‘general maintenance ledger’. However, this maintenance is contracted to a single supplier, and it could be possible to identify the costs by reviewing the invoices sent by the company.
In order to determine the costs, and the difficulty associated with identifying the costs, the invoices from this company in the year 2013 were collected and reviewed manually. The inspection costs listed on these invoices are the aggregate of the three floating piers owned by Groningen Seaports. The subsequent repairs were also aggregated, while the materials used for the repairs were listed by their description, but were not related to a specific task.
To still determine a figure, the inspection costs were divided over the three assets, under the assumption that it takes approximately the same time to inspect each floating pier. All other costs that could be positively related to the floating pier in Delfzijl (i.e. through the description of the repairs and the material listed) were also included. The resulting number is considered to be the best estimate of the yearly costs, respective of the assumptions.
The cost data of the planned maintenance is based on the tender issued for this maintenance. During the third iteration, the garbage collection and the disposal of chemicals were identified from the account ledger (appendix VI). The electricity and water bill are unavailable, as this is aggregated across the port of Delfzijl.
stay and the type of ship that was docked to the pier, and calculate the fee based on that data. This calculation was not investigated further, as the (only) person responsible for collecting this data was unavailable.
4.4
Model Building
The model was built using Microsoft Excel. This was to ensure the model could be made available to all employees. First, the structure of the TCO model is discussed, after which an analysis of the model is provided. The full TCO model is attached in appendix V.
4.4.1 Model Structure
The model is structured based on the lifecycle phases presented in figure 2.1. Although not all phases were considered to be applicable to the asset, these phases are still shown, to provide a basis for future models. The phases are shown along the left side of the model, and may cover several cost categories. Figure 4.2 shows the first three phases of the TCO model.
Figure 4.2: TCO-model excerpt
Within each lifecycle phase, the cost categories, drivers, and elements are listed. The cost elements are the smallest unit available in the model, and are the factors that make up the cost drivers. The cost drivers are formatted in Italics. The software gives the user the ability to group the cost drivers associated with the element, which allows the cost element to be collapsed, thereby hiding the details. The inclusion of cost categories allows cost drivers to be grouped based on their overarching activity, such as maintenance or renovation. Non-specific cost drivers are listed in a category that corresponds to the lifecycle phase’ name.
Lifecycle Length 30 jaar
Conceptual Design Details Unit Count Rate Total Cost Group Total
Lifecycle Total 0%
Preliminary Design Details Unit Count Rate Total Cost Group Total
Lifecycle Total 0%
Detail Design and Development Details Unit Count Rate Total Cost Group Total
Extern Advies 0,06% Advies 1 0,01% 0,01% E&W advies 1 0,06% 0,06% Bestek 1,35% Besteks Ontwerp 1 0,76% 0,76% Bestek 1 0,59% 0,59% Manuren 2,98%
Uren Project Management 1 2,98% 2,98%
At the top of the model, the user can define the length of the asset lifecycle, from conception to disposal. Subsequently, the model uses this information to calculate the factors necessary to extend the cost data to the entire lifecycle. For example, the Maintenance category includes unplanned maintenance cost data, gathered over a period of 10 years. To account for a lifespan of 30 years, these costs are multiplied by a factor 3 (i.e. 30/10=3). This is done under the assumptions that (1) the activities within these cost drivers are necessary to maintain the asset, and (2) the costs of these activities is constant throughout the lifecycle. As the company is unable to determine the mean time between failures, the costs are still considered to be unplanned maintenance costs. The same methodology was used for other reoccurring costs.
Wherever the highest level of detail is available (i.e. activity or shape defining characteristics), for example in the renovation phase, these parameters are used to calculate the costs.
4.4.2 Model Analysis
Figure 4.3 shows the lifecycle costs as a percentage of the total cost of ownership. The figure shows that the majority of the costs are concentrated on the renovation category, while the second largest cost category is the maintenance costs. Together, the production and/or construction phase account for 68.95%, while the utilization and support phase accounts for 25.75% of the total cost of ownership.
Figure 4.3: Cost category costs as a percentage of the TCO
4.4.3 Informal Model
Further questioning revealed that, at the time the tender was issued, an informal investigation in the total cost of ownership was made. This investigation, however, was not documented, and limited to the consideration of the investment and maintenance costs. The purpose of the investigation was to determine the amount of maintenance costs that were influenced by aspects of the pier’s design that could be changed.
Both the formal and informal model concluded that the opportunities to reduce the maintenance costs by altering the design of the pier were either non-existent if the pier was to be renovated, or could only be achieved with a complete redesign of the pier, which required the acquisition of a new pier. As the investment costs for the renovation were lower than the costs of purchasing a new pier, the renovation was chosen.
4.5
Model Use
In order to assess the model a small questionnaire was released among the maintenance staff. The results are shown in table 4.1. The questionnaire is attached in appendix I.
Table 4.1: Questionnaire Results (n=4)
Question Score
1) The TCO model is clearly organized 4.5
2) The TCO model is easy to maintain 4
3) The TCO model is complete 4
4) I would use this TCO model 4.5
5) This TCO model can be used as a basis for all assets 4
6) I would make a TCO model for all assets in my portfolio 3
The scores presented in table 4.1 range from 1 to 5, with 1 being completely disagree, 3 being neutral, and 5 being completely agree. Overall, the staff rated the TCO model positively. Their main concern with the TCO-methodology was the effort required to build the model, which is represented in the low score of question 6.
One of the comments related to the model was the inclusion of the operation and utilization cost drivers. This phase contained only the garbage collection cost driver, as the other exploitation costs could not be determined or accurately estimated.
Delfzijl. The procedure is to balance these operation costs with the revenues gained, which is done by a different department, at a lower level of detail.
The tendering procedure ended with only one offer to renovate the pier, which was unexpected. The reason why there was only one applicant remains unclear, however, the offer was still coded into the TCO model for comparison with the estimation. The comparison identified large differences between costs elements used to calculate the cost drivers, which resulted in a difference larger than 15% when compared to the estimate.
Interviews suggested that during the financial crisis most small- and medium enterprises had laid off a large part of their workforce to survive. As the economy has recovered, it has been difficult to attract new employees, which has resulted in more work than the local companies can handle. The increase in demand for work similar to the renovation of the pier, and the limited availability of resources to fulfill the demand (i.e. companies able to do the work), suggests an inflation of the offer. As a result, the market conditions are suggested to have influenced the offer.
5 Analysis
The account provided in chapter 4 is analyzed throughout this chapter. First, the findings are analyzed based on the aspects found in the conceptual model in figure 2.4. Subsequently, the conceptual model is updated to reflect the findings.
5.1
Model type and model execution
The case study shows the purpose of the TCO model to have an influence on both the model type and model execution. In order to determine whether a tradeoff exists between the investment costs and maintenance costs, an informal, unstructured, and undocumented, standard model was used at the start of the investment process. Although the informal nature of the model inhibited the sharing of this information, and was developed separately from the model covered by this research, the identification of a tradeoff required no more than the identification of a link between the design and the maintenance policies, and the associated costs. An informal model was found to be sufficient to identify the tradeoff. In contrast, the development of a unique, formal model to determine the total cost across the lifecycles (the total cost of ownership), also led to the conclusion of the informal model, yet required much more effort. As such, within the case study, the purpose of the model has determined the model type and execution. A comparison between the properties of the two models encountered in the research depending on the purpose is shown in table 5.1.
Table 5.1: Comparison of model properties depending on purpose
Purpose Identifying the total cost of
ownership Identifying a tradeoff between investment costs and maintenance costs.
Model Type Formal Informal
Level of detail High Low
Model Execution Unique Standard
Number of cost drivers 59 2
Documentation Yes, structured No
Advantage Available for future use Ability to share information
Quick
Disadvantage Required effort to gather (accurate) data
Time consuming Possibly complex
Possibly inaccurate due to low level of detail
Excludes pre-investment, disposal, and operation and utilization costs
5.2
ICT systems
The ICT system has an integral role in the accuracy and availability of the cost data. The absence of an asset based system clearly limits the availability of cost data, and the ease with which data is obtained. While the project folder based alternative used by the company provides some structure for the user, the many different projects and cause fragmentation of the data. As a result, it takes much time and effort to gather the data.
5.3
Data accuracy and availability
The availability of information from the financial system varied depending on the action which incurred the costs and in which lifecycle phase these costs are incurred. However, this is not necessarily an ICT related issue. Instead, the availability and accuracy of data is influenced by the standard operating procedures followed within the company. Currently, these procedures require revenue and costs to be charged to general, rather than asset specific cost categories. As a result, it is difficult to identify which data is related to an asset, which means the gathered data may not include all cost elements, or costs may be wrongly associated to the asset. The identification of cost data is further complicated by the invoices sent by the suppliers, who aggregate costs from different assets without a specification of the distribution, and list raw materials used without a reference to the assets.
Table 5.2: Availability, accuracy, and source of data per cost driver or category
Cost driver/category Available Accurate Cost Estimation Method (Data Source)
> Explanation
External advice Yes Yes Engineering procedure (Invoiced)
Specification and Conditions Yes Yes Engineering procedure (Invoiced)
Hours Yes Yes Engineering procedure (Calculated)
Quality Control Yes Yes Engineering procedure (Invoiced)
Renovation Yes Neutral Engineering procedure (Calculated)
> Engineering method of estimation is accurate. Difference between estimate and offer due to difficulty forecasting market conditions (lifecycle uncertainty)
Unplanned maintenance Neutral Neutral Analogy (Historical data)
> Difficult to identify costs due to aggregation and allocation related to operating procedures. Lifecycle accuracy influenced by the extrapolation of initial inaccuracies.
Planned maintenance Yes Yes Engineering procedure (Tender)
Garbage Yes Yes Analogy (Historical data)
As suggested by Fabrycky & Blanchard (1991), the utilized estimation methods in the final phase of the development are the Analogy and Engineering procedure. The case study results show that the estimation method is largely dependent on the source of the information. Indeed, well defined sources (such as invoices or tenders) are automatically best estimated by the engineering method (as shown in figure 2.2. Similarly, cost data that is known, and describes historical actions similar to future actions, are estimated through analogy. This corresponds with the advantages and disadvantages discussed in table 2.1.
The case study, however, does not show a relationship between the cost estimation method, and the accuracy and availability. Indeed, an analogy based on historical data is found to be both accurate and available in the ‘Garbage’ cost driver, but yields opposite results in the ‘Unplanned maintenance’ cost driver.
