A RAPID ASSESSMENT TOOL TO ASSESS FACTORY SUSTAINABILITY

Over the last decades, there has been an increased interest in sustainability and it has become an important issue in production and manufacturing research. To use the traditional definition provided by the Brundtland Commission, sustainable development is “development that meets the needs of the present without compromising the ability of future generations to meet their own needs” (Brundtland p.43). This concept of sustainability might be understood intuitively, but to express and assess specific goals poses an important challenge. As a result sustainability assessment is becoming a rapidly developing area with a growing number of frameworks and tools. However, most of the sustainability assessment tools focus on a national, regional or community level. At this point, the company level has not been considered sufficiently and those tools that are actually used within industry focus mainly on a product level within the organization (Labuschagne et al. 2005). Furthermore, the existing tools require a lot of effort and insight data in order to be completed. This study presents a tool that overcomes this issues and aims to fill the gap of a missing factory assessment tool. Based on existing integrated sustainability assessment tools a set of indicators is compiled and integrated into a framework that calculates an overall composite index. The developed tool distinguishes itself from other tools, because it is constructed as a user-friendly software that allows the assessment of a factory’s overall sustainability with a minimal time effort. It can be used from an external as well as from an internal perspective and considers the differences between industries. Furthermore, it provides the possibility to compare different alternatives and to assess a factory’s development over time.


CHAPTER 1 -INTRODUCTION
This thesis will develop an integrated assessment tool to measure the sustainability of factory related operations. Therefore, the first section of this chapter presents the background of sustainability in manufacturing and the motivation of the thesis. The second part of this chapter describes the objectives of the study and the procedure by which they can be achieved in more detail.

Background and Motivation
For the last two centuries, industry and economy has evolved on the premise that the earth is an unlimited 'store of resources' and a stable ecosystem (Graedel, Allenby 2010). However, as the population exceeds seven billion and the standards of living improve enormously, the interest and awareness towards the limited natural resources increases as well. The goal is to use the resources consciously in order to satisfy human demand (Davidson et al. 2010). One approach to this challenge can be found in the key concept of sustainability. By regarding the three dimensions: social, environment and economy, it aims for our society to meet present as well as future needs worldwide. Obviously, manufacturing is a major factor in this approach towards a more sustainable society (Despeisse et al. 2012).
Against this background many manufacturing companies have already started to reconsider the idea of being "green" and how to deal with sustainability. However, this change of attitude was of course supported by even more factors. Local environmental regulations have a significant global impact, especially if they are supported by political decisions. Therefore many global manufacturers feared to be locked out of the market, if they do not change their policies towards the concept of sustainability (Srinivasan 2011). Furthermore, investors are also interested in the sustainability performance of companies and some of them integrated it into their portfolio decisions. They are one of the target groups that use indexes and tools to evaluate companies. This trend towards socially responsible investing is another important factor that forced companies to adapt their strategy (DJSI 2013).
Although rethinking has begun, it is important not to limit the scope of sustainability to the product itself, but to consider the production process as well. There has been a lot of work on researching sustainability on different levels, but sustainability assessment at factory level is still lagging behind (Labuschagne et al. 2005).

Objectives and Procedure
Against this background the larger goal of the thesis is to focus on sustainability at factory level and to describe the relationship between factories and sustainability dimensions in a basic concept and to develop an integrated assessment tool based on that relationship. In order to achieve this goal, several sub-goals will be pursued during the study. These objectives are summarized below:  Giving an insight into the history and development of sustainability.
 Reviewing the current state of sustainability assessment tools and categorizing them.
 Examining the impact of factories on their environment and classifying industrial sectors.
 Developing a framework to assess the sustainability performance of factories and to calculate an overall composite index.  Implementing the model into a computer-based tool by using Visual Basics for Applications.
 Testing the tool by applying an exemplary case study and developing a usability questionnaire.
The procedure which will be performed in this study in order to achieve the set goals is illustrated in the following figure.

CHAPTER 2 -SUSTAINABILITY ASSESSMENT
In order to develop a new sustainability assessment tool, it is necessary to begin with understanding the background and concept of sustainability and to analyze the state of the art in this field. Therefore, the first section of this chapter will present the basic ideas behind sustainability and its development. In the second section, a comprehensive literature review will categorize existing sustainability frameworks and will identify their characteristics and field of application.

Basics of Sustainability
Becoming "sustainable" has become central to many aspects of everyday life. Not only does this relate to environmental decisions, but many products, services, production systems and developments now claim to be sustainable. However, in most cases when the term sustainability is used, the definition and the meaning of it are not clear. Sustainability has become a buzzword in the media, and is widely used in a diverse range of contexts with disparate meanings.

Background of Sustainability
Sustainability is derived from two Latin words, sus which means up and tenere, which means to hold (Theis, Tomkin 2012). After all, the term sustainability is comparatively modern and was hardly used until the 1980s. The timeline in Figure 2.1 illustrates the development.

Figure 2.1: Timeline of Sustainability
The first milestone in the history of sustainability was initiated by the Club of Rome and a group of young scientists from the Massachusetts Institute of Technology (MIT).
In 1972 they published the controversial report The Limits to Growth, which reported that "the limits to growth on this planet will be reached sometime within the next hundred years". (Donella Meadows, III 1972)  Apart from the CSD, other organizations such as the Global Reporting Initiative (GRI) were founded over the past two decades and they have developed other indicators and matrices to assess sustainability on different levels (more in the following section of the chapter).

Sustainability and Sustainable Development
Acoording to the Brundtland definition of sustainable development, sustainability is a state that will be achieved through sustainable development. Therefore, the literature supports the thesis that both terms can be described and measured as the same and even Agenda 21 uses them interchangeably. (Dresner 2002, p.65) However, this is also the reason why other authors criticize the Brundtland definition. Tim O'Riordan expressed his concerns about the meaninglessness of the term 1988 in his essay The Politics of Sustainability. He complains that the formulation is too vague and it allows people to claim everything as being part of the sustainable development (O'Riordan 1988). Nevertheless, keeping with the common practice, both terms will be used interchangeably in this study.
Besides the definition of sustainable development the Brundtland report contains also two key concepts: "the concept of 'needs', in particular the essential needs of the world's poor, to which overriding priority should be given; and the idea of limitations imposed by the state of technology and social organization on the environment's ability to meet present and future needs." (Brundtland,p.43) Thus, the report implies that sustainability has three dimensions that it seeks to integrate: economic, environmental and social. Today the common understanding in literature illustrates the three dimensions as overlapping circles as represented in Figure 2.2.

Figure 2.2: Three dimensions of sustainability
This illustration implies that there is an interaction between the different dimensions of sustainability, and progress can be achieved only by considering them simultaneously (Seliger 2007).

Social Sustainability
Environmental Sustainability Economic Sustainability

Sustainable Manufacturing
Sustainable manufacturing can be considered to be a part of the larger concept, sustainable development. Although it focuses only on one specific aspect, it is still based on the same problems and aims for the same goals.
The most quoted definition is given by the U.S. Department of Commerce. They define sustainable manufacturing as "the creation of manufactured products that use processes that minimize negative environmental impacts, conserve energy and natural resources, are safe for employees, communities, and consumers and are economically sound" (U.S Department of Commerce 2007). This definition demonstrates again the need to consider all three dimensions -economic, social and environmental.
Furthermore it also states that sustainable manufacturing includes both the manufacturing of sustainable products as well as the sustainable manufacturing of all products (NACFAM 2009). Therefore, it has to take the entire life-cycle with the stages pre-manufacturing, manufacturing, use and post-use into consideration. However, with regard to the goal of this study, sustainable manufacturing will be limited to the stage "manufacturing" within the life-cycle and it will focus only on the second part of the statement: sustainable manufacturing of all products.

Categorization of Sustainability Assessment Tools and Indicators
As mentioned in the previous section, there have been different organizations over the last years that have developed tools and defined frameworks to assess sustainability. The CSD and GRI referred to above are named as two examples. In the literature several authors categorized these tools and frameworks based on numerous factors and dimensions. For example Ness et al. conducted an overview of tools by considering the focus of the tool (i.e. product level or policy), the temporal characteristics and the degree to which it integrates environmental, social and/or economic aspects (Ness et al. 2007). Feng et al. on the other hand categorized sustainable assessment tools into a hierarchy of global, country, sector, corporation, process, and product levels (Shaw C. Feng et al. 2010). Moreover, Labuschagne et al. conducted an overview of tools that include a set of indicators, integrate all three dimensions of sustainability, have a wide focus and are independent (Labuschagne et al. 2005). This study categorizes tools by considering the following three factors:  Integration of all three dimensions of sustainability, i.e. if the tool considers environmental, social and economic aspects.
 The hierarchy/focus, i.e. if the focus is at the global, country, sector, corporation or product level.

 Developed by a company or by an organization
The developed categorization and overview of sustainability assessment tools is illustrated in Figure 2.3. It consists of two main branches; the non-integrated and the integrated indicators. The non-integrated indicators include indicators that do not consider all three dimensions of sustainability simultaneously. Therefore, they are further broken down into development, economy based and eco-system based indices.
The second branch on the other hand covers all integrated tools and divides them first into macro and micro tools and subsequently into a hierarchy of global, country, sector, corporate and product level. While the macro tools are developed by superordinate organizations, the micro tools are developed by a company. This separation is based on the main issue of macro frameworks and tools. Their focus is mainly "on the external reporting for stakeholders, rather than on internal information need to decision-making and re-design or optimization for actual eco-innovation." (Shaw C. Feng 2009, p.2).
The tools developed by a company (micro tools) on the other hand give the manufacturers the possibility to evaluate and track their sustainability performance within the environment they are in. But the issue with those tools can be seen in the fact that they are designed mainly for the specific environment of a company or supply chain. Therefore it is important to include both in the overview.

Review of Current Non-Integrated Indicators
Non-integrated indicators include all indices that do not consider the three traditional dimensions simultaneously. Therefore, they are divided into three different levels (see Figure 2.3). The development indices focus mainly on the social dimension,

Intra-Company
Walmart Scorecards while the economy based indices focus on the economy and the eco-system based indices on the environmental dimension. In the following the most quoted indicator at each level will be described briefly.

Development Indices
The best known indicator at this level is the Human Development Index

Economy based Indices
At this level, Life-Cycle Costing (LCC) is one of the most important methods. It is an economic approach to get the total cost of goods by examining all the parts of the cost over its lifetime. This includes costs for research and development, production, maintenance and disposal. Thereby, Life-Cycle Costing is not associated with environmental costs, but with costs in general. Overall, it is an important tool to support decision making. (Gluch, Baumann 2004)

Eco-System based Indices
At the eco-system based level, the Ecological Footprint (EF) is one of the most quoted indicators. The Ecological Footprint developed by Wackernagel and Rees (Wackernagel, Rees 1996) is defined as the quantitative land area on earth that is required to sustain the given living standard until infinity. This includes also areas, which are needed to produce food and clothes or to supply energy. Moreover, it takes also the waste assimilation requirements in terms of a corresponding land area into account.
Finally, the result is expressed per hectare per person and year. In other words, "EF analysis is an accounting tool that enables us to estimate the resource consumption and waste assimilation requirements of a defined human population or economy in terms of a corresponding productive land area" (Wackernagel, Rees 1996, p.9).

Review of Current Integrated Tools
In contrast to the non-integrated indicators, the integrated tools are characterized by the fact that they consider the three traditional dimensions of sustainability at the same time. Generally, all of these tools follow the same structure, which is illustrated in the following figure. After a thorough literature analyses the most quoted and relevant integrated sustainability assessment tools will be described in the following, including the first three levels of their frameworks: tool, dimensions and themes (see Figure 2.4).

Global level
At the global level, the Core Environmental Indicators (CEI) developed by the Organization for Economic Co-operation and Development (OECD) are considered to be the most relevant indicators. They can be used to measure environmental performance, to report on the progress towards sustainable development and also to monitor the integration of economic and environmental decision making as well as society's response (OECD 2001(OECD , 2003. The core set contains about 50 indicators with a strong focus on environmental issues, but it integrates also society and economic aspects (OECD 2001). The hierarchical structure is shown in Figure 2.5.

Corporate level
With the regard to the goal of the study, to assess factory sustainability, the corporate level is considered to be the most important hierarchical level. It also includes factories as an aspect. Therefore, it is important not only to focus on one tool, but to describe this level extensively.
One of the most quoted tools on this level is the Global Reporting Initiative (GRI).
The GRI was launched in 1997 by the United Nations Environment Program (UNEP) together with the US non-profit organization the Coalition for Environmentally Responsible Economics (CERES). It is designed to be used by organizations of any size, sector or location and to report on sustainability of the entire organization. Additional items measuring the environmental impact over its entire lifecycle. It measures at least eleven environmental impacts in six main themes (see Figure 2.10). The results are then aggregated using weighting schemes for each category. Another aspect of the Seebalance concerns the full economic impact of all alternatives, in order to determine an overall total cost of ownership. All identified costs are summed, normalized and combined in appropriate units, without weighting them. Finally, the socio-eco-efficiency analysis assesses also the social fingerprint. Therefore, it takes five themes into account and weights them. The themes are shown in Figure 2.10. Overall, this tool allows it to quantify sustainability for different alternatives and to compare them. Therefore, it is useful for supporting strategic decision-making, marketing and also for prioritizing R&D activities (Saling et al. 2005).

Product level
Like the BASF Seebalance, the Ford Product Sustainability Index (FPSI) is also considered to be a micro-tool. It is directly used by Ford's engineers to improve the sustainability performance of the products and not to report to a superordinate organization. The tool looks at eight different indicators, reflecting key impacts of automotive products. The dimensions and themes are illustrated in Figure 2.11. Since the tool focuses on only few key elements with available data, the effort to complete the tool is rather easy and it can be done in approximately 10 -15 hours for the whole product development process. The tool has been applied the first time for the vehicles Ford Galaxy and S-MAX and resulted in a significant improvement of the sustainability performance (Schmidt 2006).

Figure 2.11: Hierarchical structure of the Ford PSI
Another framework at this level was created by the company Wal-Mart. In contrast to Ford's approach, Wal-Mart designed their tool to be used not only within the company, but mainly from their suppliers. For this reason it is considered to be at the inter-company level. As one of the worldwide leading retailers, Wal-Mart was accused of unfavorable business practices with a significant ecological impact and high carbon footprints. Therefore, they changed their mission towards a sustainable development and implemented an environmental initiative. (Nandagopal, Sankar 2009)  with the intention of helping suppliers to improve packaging sustainability and to conserve resources. This scorecard is a measurement tool that allows the suppliers to evaluate themselves relative to other suppliers. The evaluation is performed using a specific metric which is based on the "7 R's of Packaging": Remove, Reduce, Reuse,

CHAPTER 3 -FACTORIES AND INDUSTRIAL SECTORS
Since the thesis develops a tool to assess factory sustainability, it is important not only to look at the sustainability aspect, but also at the factory aspect. Therefore, this chapter presents a basic description of factories and illustrates their importance in terms of sustainability, based on energy use and CO2 emissions. Besides the influence of general factories, the industry specific influence by sector is also considered. Finally, the chapter concludes with explaining the need for a factory specific sustainability assessment tool that will be developed in the next chapters.

Basics of Factories
The term factory is derived from the Latin word facrica, which means workshop.  Wiendahl considers primarily the space that will be needed by the resources (Wiendahl et al. 2007).
Moreover, it is important to classify factories also in an overall system. For this purpose, different descriptive models have been developed. In terms of sustainability the life cycle assessment appears to be the most relevant approach, where factories are considered to be a stage in the product life cycle. This approach attempts to evaluate the environmental impact of products throughout the entire life cycle of a product from raw material extraction, manufacturing, and use to ultimate disposal (see Furthermore, it takes the ecological backpack of input products into account and also the environmental impact of the output products (Müller et al. 2009).
This section of the chapter gives a rough impression about the complexity and significance of factories, based on the hierarchical order and the entire product life cycle. However, in the following course of the study it is necessary to reduce the complexity and to limit the scope in order to create a rapid assessment tool. Therefore the factory will be considered in its entirety and the pre-and post-stages of the factory will not be taken into account.

Impact of Factories on their Environment
The previous section of the chapter has already indicated that the manufacturing industry produces adverse environmental impacts such as waste generation and consumption of natural resources. The significance of factories becomes particularly   Besides the industrial energy use and industrial CO2 emissions, the significance of factories is also shown by regarding the general pollution. According to the European Environment Agency (EEA) "manufacturing contributes 22% of European global warming potential as well as 14% of acidification potential, and 21% of tropospheric ozone potential" (OECD Sustainable Development Studies, p. 65).
So far, only air pollution and energy consumption have been considered for describing the relationship between manufacturing and sustainability. Nevertheless, there are even more issues, which indicate that sustainable manufacturing will become one of the major objectives within industry in the twenty-first century. Not only Other 28% improvements in efficiency and reductions on pollution have to be made but also traditional paradigms for doing business have to be changed.

Classification of Industrial Sectors
As the section above has already indicated, the different industrial sectors need to Based on the explained significance of industrial sectors it is necessary to divide the manufacturing industry into sectors within this study as well. The indicators need to be weighted sector-specifically in order to receive a meaningful sustainability score.
The figure below presents the classification of different sectors.

Figure 3.5: Classification of industrial sectors
The left side of Figure 3.5 lists the 19 supersectors that are used by the DJSI.
However, their classification is not suitable for this study. In order to reduce the complexity it is necessary to combine some of the supersectors. Hence the right side of the figure presents the results of the combination. Moreover, only factory related sectors are relevant for this work. Since the sectors Financials and Consumer Services do not operate with factories they will not be considered any further. As a result six main sectors: Basic Materials, Industrials, Consumer Goods, Health Care, Utilities and Technology remain.

Needs of an Individual Tool at Factory-Level
This chapter demonstrates that the manufacturing industry is a main consumer of natural resources and a main producer of adverse environmental impacts. It signifies that there is a high responsibility of factories towards their environment. For this reason it is important to design a tool for the sustainability evaluation of factories. Although the purpose of such a tool is primarily to assess the sustainability performance of the factory, it can also guide factory managers to think and act in the right direction and to discover possible improvements in order to increase the sustainability metrics related to factory operations.
Even though the literature review from chapter 2 indicates that several frameworks and tools are already available to assess sustainability, it also demonstrates that they vary depending on the subject of investigation and it needs a lot of insight knowledge and effort to use those tools. Moreover, current tools focus primarily on regional, national and global levels. At this point, the company level has not been considered sufficiently and those tools that are actually used within industry focus mainly on a product level within the organization (Labuschagne et al. 2005).

CHAPTER 4 -DEVELOPMENT OF A FRAMEWORK
In order to fill the gap of a missing factory assessment tool it is necessary to develop effective sustainability indicators and a reasonable framework. Therefore, it is important to specify the purpose of the tool in the beginning. This chapter presents general criteria for indicators and also specific criteria for each dimension of

Purpose of the Tool
Generally, the study aims to develop a tool that assesses a factory's sustainability performance. Furthermore, it is the goal to ensure a rapid and integrated assessment for all industries. Besides these general characteristics, the tool also has to meet specific criteria listed below, which distinguish it from other tools.
 It should be possible to use the tool as an external user without internal information. That means the data for the indicators should be available through published sustainability reports, webpages etc.
 It should be possible to evaluate a single factory.
 It should be possible to evaluate two or more factories and assess them as alternatives against each other.
 It should be possible to evaluate one factory over time, in order to observe its sustainable development.
Based on the integration of these criteria the tool's purpose is intended for external investors as main users who integrate sustainability consideration into their portfolio.
The tool provides a quick and general overview of the sustainability performance and supports the comparison of different alternatives. At the same time internal factory managers may also use the tool to compare themselves to other companies or to identify possible improvements or deteriorations in terms of sustainability.

Criteria of Sustainability Performance Indicators
Indicators are simple measures; most often quantitative that represent a state or condition of something. An example of an indicator is a thermostat displaying 32 degrees. In this sense, indicators typically provide key information about a physical, social, or economic system and also allow analyzing trends and relationships. Thus, indicators are usually a step beyond primary data (Veleva, Ellenbecker 2001). They vary depending on the type of system they monitor. In terms of this study, sustainability indicators can be defined as "information used to measure and motivate progress toward sustainable goals" (Ranganathan 1988, p.2 The characteristics listed above help to distinguish indicators from primary data, goals, parameters, or issues. The following example demonstrates the importance. "Using renewable energy" is often labeled as a sustainability indicator by the media, even though it is not. In fact, it is a goal. In order to define an indicator it is necessary to consider all the mentioned characteristics. In terms of renewable energy use a possible indicator would be "percent of energy from renewables, measured at a facility over a period of one year" (Veleva, Ellenbecker 2001).

Identifying and Grouping of Sustainability Performance Indicators
In order to identify and group indicators it is necessary to define a hierarchical structure for the framework.  Therefore, the framework adopts this view and contains the same three dimensions.
In contrast to the dimensions, the themes and indicators require more effort as each sustainability tool in the literature focuses on different aspects. Therefore, it is important to analyze and compare the main sustainability assessment tools that have already been identified in chapter 2 further. Table 4.1 organizes the most important sustainability tools from chapter 2 by focus level, dimension, themes, and subthemes.
Based on these information it is possible to derive dimension specific themes in the following section of this chapter.

Themes for Environmental Sustainability
The environmental dimension traditionally gains most of the attention in terms of sustainability, and it is the dimension discussed in most detail in the literature.

Themes for Social Sustainability
Recently, the public and especially stakeholders shifted the focus from environmental-related to social-related issues. Therefore, businesses pay increasingly more attention to the social dimension of sustainability, although the work on this topic is still insufficient (Labuschagne et al. 2005). It is striking that the more modern tools like EICC, 2004 and BASF Seebalance, 2012 contain significantly more social aspects then the older tools, such as DJSI or OECD-CEI.
In contrast to the environmental dimension, most of the tools considering the social dimension have an internal view instead of an external view. Since the tool developed within this thesis is aimed at assessing the social sustainability at the factory level, the focus is also internal. The following themes are derived from

Themes for Economic Sustainability
In terms of economic sustainability the review of current integrated frameworks from chapter 2 shows that there are two different understandings of economic sustainability. Since OECD and UN-CSD are located at the global and national level, it is obvious that they take impacts from the economic system at the national and global levels into account. However, GRI assesses sustainability at a company level and considers "organization's impacts on the economic circumstances of its stakeholders and on economic systems at the local, national, and global levels" (Global Reporting Initiative 2011). All three frameworks focus on the general economic performance and development (see Table 4.1). Thus, there are two approaches that can be taken: one approach takes the external impacts on the entire economic systems into consideration, while the other focuses on the internal economic impacts of a business.
The DJSI as well as the EICC consider economic performance in terms of the internal management, whereas the BASF and FPSI frameworks attempt to minimize their costs (see Table 4.1). Consequently, it is necessary to choose between the two different approaches.
With regards to the statement that the first goal of businesses towards sustainability is to stay in business, the focus within this study is internal. Activities at the factory level contribute to the overall profitability of the company and only subsequently contribute to the economic system on a broader, national level (Labuschagne et al. 2005). Therefore the following themes are derived based on the

Sustainability Performance Indicators
After defining themes for each dimension and the general criteria for indicators, it is now required to define and constraint the concept to a number of key performance indicators that meet all the criteria and can be measured, monitored, and recorded on a regular basis. A wide range of possible sustainability performance indicators can be found in the literature (see chapter 2). However, every indicator is not relevant to the industry and can be evaluated from an external perspective. Therefore, suitable key indicators have to be identified. To accomplish this, existing tools have to be compared and the most common key indicators have to be identified. Again, the main sustainability tools from chapter 2 are used for this analysis. A detailed overview of each tool can be found in the digital appendix. Additional sets of indicators found in the literature that focus on sustainable manufacturing are included as well: Krajnc and Glavic (2003), Velena et al. (2001) and Veleva and Ellenbecker (2001). Finally, the indicators are also tested and compared with sustainability reports published by different companies to ensure the data availability for the external use of the tool. To achieve this, the BMW Group Sustainable Value Report 2012, the BASF Report 2012 and the AkzoNobel Report 2012 are analyzed. These reports are published annually by the companies to report their figures and goals in terms of sustainability.
Generally, the study aims for using only quantitative indicators, as these are more objective and less biased than qualitative ones. It should also be possible to express each indicator in relative terms and not only in absolute terms, as different factories have to be compared on a meaningful level. Social indicators for example should expressed relative to the size of the workforce and environmental indicators relative to an appropriate measure of production such as produced units of product or an indication of produced weight.

Where
, is the quantity of resource summed up over a period ; represents the weighting factor of that resource based on the total estimated world reserves (see Table 4.3). Renewable materials are calculated with a weighting factor of 0 and are therefore not taken into account. Dividing the total materials used by an appropriate measure of production presents the material intensity.  Access to fresh water should be a universal and human right, but limited resources and a growing population are increasing its economic value.
Therefore, a goal of sustainable manufacturing is to reduce consumption of freshwater.
 Goal: Reduce the freshwater consumption.
 Calculation: Where is the quantity of freshwater summed up over a period . To gather this data, a factory's water utility bills can be used. Dividing the total amount of freshwater by an appropriate measure of production presents the water intensity. Where is the amount of solid waste generated over a period .
Dividing the total amount of solid waste by an appropriate measure of production presents the relative waste generation per unit of product/service.  Goal: Reduce greenhouse gas emissions.
 Calculation: Where , is the quantity of emission summed up over a period ; represents the equivalent factor of that emission relative to carbon dioxide (see Table 4.4). Dividing the total GWP by an appropriate measure of production presents the relative intensity. Indicator 6: Acidification potential (AP) -total and adjusted per unit of product  Significance: Acidification potential is a measure of how much a particular gas contributes to the acidification, by comparing each gas at a relative scale with sulfur dioxide. As presented in Table 4.4, sulfur dioxide has been assigned an AP of 1. Upon release, plants and soils can absorb acidic gases, leading to decreased biomass and poor soil quality. Additionally, surface waters and other water bodies may be acidified, resulting in poor water quality, thus, endangering ecosystem health.  Where , is the quantity of emission summed up over a period ; represents the equivalent factor of that emission relative to carbon dioxide (see Table 4.4). Dividing the total AP by an appropriate measure of production presents the relative intensity.   Therefore, every factory's goal should be to be as safe as possible.
 Goal: Achieve zero working accidents.
 Calculation: Where is the number of workplace related accidents over a period . Dividing the total amount of working accidents by the total number of employees presents the relative number of working accidents per employee. should not be exposed to such materials.
 Goal: Reduce the amount of hazardous materials.
 Calculation: Where is the amount of hazardous waste generated over a period . Dividing the total amount of hazardous waste by an appropriate measure of production presents the relative generation per unit of product. Since companies usually do not publish the use of hazardous materials, the indicator considers the generation of hazardous waste instead. Where is the quantity of training hours summed up over a period .
Dividing the total number of training hours by the total number of employees presents the relative number of training hours per employee.  (4.14) Where is the total number of women divided by the total number of managers .

Minimum-Maximum
This method normalizes indicators with a positive impact on sustainability by the equation: Where , ℎ is benchmark for indicator from the group of indicators . In this case, it is possible that the normalized value is higher than 1, which indicates that the performance of the factory is better than benchmark. (Zhou et al. 2012; OECD

Percentage over annual differences
Finally, the method "Percentage over annual differences" is the third main normalization approach discussed in this chapter. It focuses on the development of the indicators over time. Therefore, each indicator is transformed using the following formula: The normalized indicator is dimensionless. Nevertheless, the disadvantage of this method concerns the case = . In that case, the indicators cannot be normalized by the given equation and the data would be lost during the analysis. (Zhou et al. 2012)

Evaluation of Normalizing Method for Factory Sustainability
After reviewing the main normalizing methods, it is necessary to evaluate which method fits best. This evaluation is not as straightforward as evaluating the weight of the model. All of the described methods require a database or a set of reference data in order to transform the indicators. However, since there is no database available for the indicators, normalization is not possible for one data set of indicators. Nevertheless, the tool created within this thesis does not only attempt to assess a single factory, but also to compare different factories with each other and to evaluate the development over time. These three different cases lead to the following conclusion:  Assessing a single factory → No normalization possible  Comparing different factories → "Distance to a reference"  Development of a factory → "Percentage over annual differences"

Implementation within the Framework
As mentioned above, the selection of the normalizing method depends on the purpose of use. Regarding the comparison of factories the best method to use is the "distance to a reference". However, some aspects of this method have to be slightly modified in order to meet the requirements within this thesis. As there is no large database available, there is also no external benchmark value available. Although, it is not possible to normalize one factory with a reference value, it is still possible to compare two or more factories, by assigning the value 1 to the inferior factory for each indicator and therefore making it a reference factory. The remaining factories are evaluated relatively to that factory with a value between 0 and 1. The closer the value is to 0, the better the performance of the factory according to that indicator.
The second case, where it is necessary to normalize the indicators, regards the development of a factory over time. Here, the method "Percentage over annual differences" provides the best comparison. In contrast to the other method, this method can be implemented as described in the previous section, without modifying any aspects. Nevertheless, if only few data sets are available it is recommended to use the same normalization method as explained above, because the data for = would be lost during the analysis.

Weighting of Sustainability Performance Indicators
The next step towards developing a sustainability assessment framework focuses on weighting the indicators. Not only is it important to weight the indicators against each other, but also with regard to the different industries (see chapter 2), and their impact on factory performance. The main reason for weighting indicators is to determine the individual importance of these indicators towards an overall goal. Although, this purpose is understood intuitively, it remains difficult to determine weights with sufficient accuracy (Krajnc, Glavič 2005). "The relative importance of the indicators is a source of contention" (OECD 2008, p33). Therefore, a number of different weighting techniques exist in the literature. Some are derived from statistical methods such as the Data Envelopment Analysis and others from participatory methods like the Analytic Hierarchy Process. Additionally, there is also a method that avoids the relative importance of the indicators and weights them equally. In the next section, the theory behind those weighting techniques is discussed and how they meet the criteria for weighting factory assessment indicators.

Budget Allocation Process (BAP)
This weighting procedure determines the indicator weights based on expert opinion. In general, the BAP has four different phases: first, experts in the field have to be selected for the assessment. It is essential that the experts represent a wide spectrum of knowledge and experience. Second, the selected experts have to allocate a "budget" of one hundred points to the indicator set, based on their personal judgment of the relative importance. In a third step weights are calculated as average budgets.
As an optional fourth step the procedure could be iterated until convergence is reached. (Hermans et al. 2008;OECD 2008) The main advantages of BAP are its transparent and simple application as well as its short duration. On the other hand, it also contains several disadvantages: the weights are fairly subjective and could reflect specific conditions that are not transferable from one factory to another. (Zhou et al. 2012)

Analytic Hierarchy Process (AHP)
The Analytic Hierarchy Process is another participatory method similar to the Budget Allocation Process. However, this method is far more complex and consists of a mathematical approach. The AHP was developed by Saaty in the early 1970s and is a widely accepted technique for multi-attribute decision making. Singh et al. used this method to develop a composite sustainability performance index for the steel industry (Singh et al. 2007), Krajnc et al. applied it to a case study on the sustainability performance of the oil and gas industry (Krajnc, Glavič 2005) and Hermans et al.
implemented it to a limited extend in the road safety research (Hermans et al. 2008).
As a first step of this method it is necessary to translate a complex problem into a hierarchy. The top element of the hierarchy is the overall goal of the decision model and the criteria and indicators contributing to the decision are represented at the lower levels. The second step requires a pair-wise comparison between each pair of indicators. Experts have to judge "how important is indicator j relative to indicator i ?".
Values on a scale from 1 to 9 are assigned to show the intensity of preference. The larger the number, the greater the importance. (Saaty 1980) In the next step the results are presented in a matrix to obtain the relative weights of each indicator. For a matrix Q x Q, only Q-1 comparisons are necessary to find weights for Q indicators (OECD 2008). Finally, it is required to find the eigenvector with the largest eigenvalue from the matrix. The eigenvector presents the weights and the eigenvalue measures the consistency of each judgment. Inconsistency within this method can always occur, because it is based on people's beliefs and it is human nature that they may be inconsistent. However, a consistency ratio of 0 indicates a perfectly consistent matrix, while a ratio equal to 1 indicates meaningless or random judgments. A suggested ruleof-thumb says that a ratio of less than 0.1 does not drastically affect the consistency of the weights. (Saaty 1980;Singh et al. 2007) Aside from the problem of possible inconsistency, the subjectivity of judgment is another negative characteristic of the method. Each expert judges the indicators based on his or her own knowledge and experiences. With that, the possible inconsistency is also related to subjectivity (Hermans et al. 2008 (Yang, Kuo 2003).

Benefit-of-the-doubt (BOD)
The benefit-of-the-doubt presents another application of the DEA in the field of composite indicators. In contrast to the original DEA model, BOD evaluates the relative performance of the factories and not the efficiency (Cherchye et al. 2004). However, it is based on the same model and follows the same process. The composite index in this case is calculated as the ratio between the actual performance of the factory and the external benchmark: Where is the sub-index for the group of indicators , while ℎ is their benchmarks and the corresponding weight (Zhou et al. 2012). Whereas, the subindicator is calculated by: Since BOD can be seen as a specialized version of the original DEA model, the DEA's advantages and disadvantages also apply for this method. However, this method has already been used for a number of indeces. It was originally proposed in the context of a macroeconomic performance assessment by Melyn and Moesen in 1991 and later adapted by Cherye and Kuosmann for a cross-country assessment of human development and sustainable development performance (Cherchye et al. 2004).

Equal Weighting (EW)
As its name already indicates, the same weight is assigned to each indicator. This implies that all indicators have the same importance and that no statistical or participatory approach is used to determine the weights. The value of the weights is simply calculated by 1 where is the number of all indicators and 1 represents the sum of all weights (Zhou et al. 2012;Hermans et al. 2008).
Although this method appears too simple from a scientific point of view, several composite indicators like the Environmental Sustainability Index or the European Innovation Scoreboard are constructed by equal weighting (Hermans et al. 2008). The main disadvantage is the fact that is does not offer any insights on indicator importance and it does not reflect reality. However, this method can be considered as a solution in case no other weighting method presents valid results.

Evaluation of Weighting Method for Factory Sustainability
In order to analyze which weighting method is best fitted and suitable for a framework to assess factory sustainability it is required to develop specific criteria that has to be fulfilled.
 Quantitative and qualitative data: Since the set of indicators that are used for this framework may be extended by quantitative indicators it is necessary that the weighting method can handle both types of data.
 Objectivity: Indicators should be weighted without bias in order to be meaningful and to decrease personal preferences.
 Insights into indicator importance: The overall goal of the tool developed within this study is to assess the sustainability performance of factories. Therefore, it is important that indicators reflect their individual importance towards factory sustainability.
 Transferability: The developed tool is supposed to allow the user to compare factories with each other. In order to do so, it is required that the indicator specific weights are always valid and transferable from factory to factory.
 No need for a database: Due to the fact that there is no large accessible database for each indicator, the weighting has to be possible without including a lot of data.
In order to identify a weighting method the methods introduced previously are presented in a structured way in Table 4.6, where each method is assessed towards the fulfillment of the before derived criteria:

Implementation within the Framework
The first step towards implementing the AHP requires the formulation of an AHP model, which synthesizes the composite sustainability performance index into a systematic hierarchal structure. The overall goal of the problem, to develop a composite sustainability performance index, is represented at the top level of the hierarchy as shown in Figure 4.3. The three dimensions of sustainability, which are identified to achieve the goal form the second level. The third level consists of the various key performance indicators, which are grouped with respect to the three dimensions and shall be weighted specifically to the industry that is being evaluated.

Level 3
Indicator are weighted industry specific In the next step, all three levels have to be assessed using the AHP approach of pair-wise comparisons according to their impact on the next level. A group of four sustainability experts is asked to judge the indicators by estimating a preference factor of each indicator relative to another. The preference factors follow a scale from 1 to 9, where 1 indicates equality between the two indicators and 6 for example means that one indicator is six times more relevant than the other. However, the evaluation has to be carried out in an industry specific manner for each one of the six main industries that have been identified in chapter 3. For this reason, the assessment team that is asked to carry out the evaluation is composed of experts from different sustainability leading companies, like BMW, Cisco, and AkzoNobel, for example. This ensures that the evaluation is practically oriented and comprehensive. The exact questionnaire that was distributed to the experts is shown in the appendix A1.
The process of pair-wise comparisons and relative weight evaluation is presented in the following based on an example considering the environmental dimension within the basic materials/resources industry. The pair-wise comparison matrix for this example is shown below. The first column of the matrix includes the indicators and is provided to the expert.
The second column is filled in by judging indicator 2, 3,..n with respect to indicator 1.
Then the process of comparison is repeated for all other columns of the matrix.
The next step requires a normalization of the weights. Therefore each column of the matrix in Table 4.7 is normalized by dividing each indicator weight by the sum of all relative weights in the column and then averaging them. The results are presented in Table 4.8. After calculating the weights, it is required to check the consistency of the judgment. Inconsistency is likely to occur when the expert exaggerates or makes errors during the pair-wise comparison. For example, if material use is preferred over energy use and material use is not as important compared to waste, consequently waste should be more preferred over energy use. In case this logical chain is not followed, inconsistencies will occur. As stated above, the consistency index ranging from 0-1 can be applied in this scenario to test for discrepancies in the evaluation and weighting of the indicators. To check for consistency it is necessary to find a vector by multiplying the pair-wise comparison matrix with the weight vector. In the next step the resulting vector has to be divided by the weights vector.
( 1.73 /0,27 0.97 /0.15 0.72 /0.11 0.75 /0.12 1.49 /0.24 0.69 /0.11 ) = ( 6.40 6.49 6.40 6.23 6.32 6.27 ) (4.33) Then, the consistency index has to be calculated by inserting the overall average of the final vector is = 6.35 into the following formula: Finally, the consistency ratio can be calculated using the following formula: Where is divided by a random matrix consistency index, , providing a normalized value (Deturck 1987). With regard to the value of 0.056, it can be concluded that the judgment is consistent. The consistency ratio has to be calculated for each judgment and also for the overall weights combining each judgment. However, the procedure is always similar to the example shown above.
The results of the entire assessment procedure for each industry are summarized in the appendix A2.

Calculating the Sub-Indices
After weighting and normalizing each indicator, the next step requires to group these basic indicators into the sustainability sub-index for each group of sustainability indicators. In the context of this thesis there are three groups of indicators; environmental, economic, and social and therefore also three sub-indices, respectively.
Sub-indices can be derived as shown in the following equation: = ∑ * (4.36) . . ∑ = 1, ≥ 0 (4.37) Where is the sustainability sub-index for each group of indicators . Since the framework uses the AHP weighting method, the first constraint restricts the sum of all weights of indicator for the group of sustainability indicators to be equal to 1.

Combining the Sub-Indices into the Composite Index
As a final step it is required to combine all three sub-indices into one overall composite sustainable performance index. This index can be calculated as shown in the following equation: Where is the overall sustainability composite index for the factory that has been assessed.

Final Framework of the Factory Sustainability Assessment Tool
Reviewing the concept of the framework that is used to assess factory sustainability in this thesis shows that three different cases can be evaluated using this assessment tool.

Figure 4.4: Scheme of the final framework
First, it is possible to consider only a single factory. The collected data for this case can be judged according to the results of section 4.4, but since there is no reference data available for each indicator, normalization cannot be performed. This process ends here with the presentation of the results. The other case regards the comparison of two or more factories. After collecting the relevant data for each indicator, they can be normalized by using the method "distance to a reference", where the value 1 is assigned to the worst factory for each indicator. In a next step it is possible to weight the data according to their importance towards an overall goal. Afterwards the normalized and weighted indicators can be combined to a sub-index and then to an overall composite index. The third case considers the development of a factory over time. This case is similar to the case "comparison of factories". The only difference is that the normalization step uses the method "percentage over annual differences". This is the best suited method, because only one factory is considered over time.

CHAPTER 5 -IMPLEMENTATION OF THE TOOL
After developing the framework of the tool in the previous chapter, it is now required to convert it into a computer-based tool. For a user friendly digital assessment and data processing it is decided to implement the tool in MS Excel by using VBA.
Therefore, the first section of the chapter presents the programming environment Visual Basic for Application and its characteristics, in order to show that it satisfies all needs.
In the second section, the implementation and structure of the tool will be presented in more detail.

Visual Basic for Applications (VBA) is the programming environment for Microsoft
Office and its associate applications. It allows object-oriented programming by using a modern language that resembles most of the popular programming languages such as Pascal or C. VBA is used for the same reasons macros are used, but it offers a finer degree of control and more possibilities than macros alone (Microsoft 2013). Moreover, it was decided to implement the tool in VBA because of the following characteristics:  Stepwise processing: The input and processing of the data is complex and should be performed in several steps. VBA simplifies the coordination and the process.
For example: Each dimension has its own input mask.
 Error prevention: The data has to be entered in a specific way and the tool should prevent the user from making mistakes. VBA can ensure to accept only a certain format.
For example: The input mask cannot be closed before the user has entered all data.
 Error messages: When the user makes an error, the error and the possible remedy should be displayed with individual information and instructions for the user. This ensures a minimal rate of errors.
For example: A message box will be displayed to show the user which information is missing in order to match the input mask.
 Automatic analyses: The tool calculates sub-indices and a composite sustainability index. Moreover, the results should be presented graphically based on different charts. This requires an iterative calculation process, as well as a process to create and format the charts. These tasks can be automated by using VBA to write explicit instructions for Excel.
For example: Any company from the database can be evaluated at the touch of a button.
These characteristics indicate that by implementing the tool in VBA, it can easily be used on any computer with an MS Excel installation without the need of sophisticated programming experience. Furthermore, MS Excel is usually available at any company.

Implementation of the Tool
The start page of the tool is shown in the figure below. It is simply structured with clear symbols and colors. It is divided into the left side with buttons for data entry and the right side with higher level functions and buttons for the data processing. These buttons and functions are described in more detail in the next sections of this chapter.

Implementation of the Data Input
The data input is divided into four different steps. As mentioned above, the start page of the tool in Figure 5.1 shows all four buttons for the input at the left side. One for the general factory information and one for each dimension.
The input mask for the general information is presented in Figure 5.3. Besides the start and end date of the evaluation period, the user has to insert general information about the company (name, size of workforce and number of produced units).
Additionally, it is asked in which sector and industry the company is operating. Here the user can choose between the sectors identified in chapter 3. This decision influences the weighting in the next steps.

Figure 5.3: General factory information input mask
After entering all information, the user has to confirm the input and close the mask by pushing the button "OK" and all information are automatically stored in a general database. The following figure presents the program code for this process.

Figure 5.4: VBA code to store information in database
In a next step the user has to enter data into the input mask for the environmental dimension (see Figure 5.5). Starting with the first indicator to the last one. This mask is similar to the input mask described above. After confirming the input, a code comparable to the code in Figure 5.4 will store the information in the database at the correct position.

Figure 5.5: Environmental dimension input mask
Additionally, the user has to complete the data entry for the social and economic dimensions accordingly.
Generally, the tool provides assistance to avoid entry errors. Some entry fields allow only integers or strings to prevent errors during calculation. Moreover, it is not possible to close the input mask before the user entered all information.

Implementation of the Data Processing
Besides the aspect of data entry, the tool considers also the aspect of data analysis and evaluation. By pushing the button "Results" on the start page another user form opens and the user can select one out of three different cases:

Case 1: Assessing one factory
The first case assesses only a single factory. The user form provides checkboxes to select a factory from the database. The selection is made based on the name of the company and also on the date of the evaluation period. Since the same company might be listed more than once, it is necessary to search the database for two variables in order to make a clear identification. The following program code shows the iterative forloop, used to combine both variables.

Figure 5.6: VBA code to combine two variables
After selecting the desired company, the respective values have to be analyzed and presented. In order to transfer the correct data it is required to search the database for both variables, company name and date. The code in Figure 5.7 presents the code for this process. At first the two search variables company name and date are declared (search, searchc). Next, the database is searched iteratively with a do-while loop for an entry with the correct company name in one column and with the correct date in the next column (offset) which matches the second search variable. If an entry is found the respective database entry is returned otherwise a massage box with an error is displayed.

Figure 5.7: VBA code to search for two variables
Finally, the results will be presented in a structured and clear overview, as shown in Figure 5.8. The button "Print Results" will automatically format the page and print it out on the standard printer.

Case 2: Comparing different factories
In this case the user has the possibility to assess two or more factories at the same time and to evaluate them as alternatives against each other. As in case 1, the user can select the companies from the database by entering the name and the date.
Therefore, this case uses similar codes as shown in Figure    Finally, it combines the results to one overall index. Based on these numbers the user is able to determine which factory has the best sustainability performance in comparison to the other factories.

Case 3: Assessing a factory over time
The third case concerns the assessment of one factory over a period of time. The user can select different evaluation periods for the same factory and analyses the results. Generally, the data entry as well as the data processing is very similar to the second case. Therefore it is not presented any further.

CHAPTER 6 -CASE STUDY: AUTOMOTIVE INDUSTRY
After implementing the framework into a computer based tool, it has been applied to a practical case study in order to demonstrate its application and effectiveness. The case study is divided into two parts. Case 2 and/or Case 3 are performed. The entire case study is carried out by using the developed sustainability assessment tool and data based on public records.

Comparison of different Factories
In this section of the case study the sustainability performance of two different factories from the automotive industry are compared with each other. One factory that is being assessed is the BMW plant in Dingolfing. This plant belongs to the BMW Group since 1967, employs a workforce of around 18,500 and produces about 1,500 cars per day. The other plant is located in Sindelfingen and was founded by the automotive producer Daimler in 1915. The annual production of this site is estimated to be 424,609 and it employs a workforce of around 25,947.
For this case study the data of the calendar year from 1 January to 31 December 2011 are considered. All data and information are gathered from public records e.g. environmental declarations, webpages and sustainability reports. The data entry and data processing is then carried out by using the developed assessment tools.
According to the data processing process described in chapter 5, the first step of this analysis results in a structured overview including all indicators for each production site. Figure 6.1 and Figure 6.2 illustrate the results.  In order to make a general statement about the sustainability performance of both production sites, the results of each indicator have to be weighted and combined into a composite index. The results of this process are presented in the following table. It can be seen that the BMW plant Dingolfing performs better in all areas. However, the overall index of 0.81 compared to the index of 0.93 shows that the differences are not as significant and both production plants are in the same range. The results are also visualized in Figure 6.6.

Figure 6.6: Presentation of assessment results by social indicators
Finally, it can be concluded that the comparison of different factories provides clear results and presents them in a comprehensible form. Since BMW is the sustainable leader in the automotive industry, it was expected for the BMW plant to perform better.
However, there are also indicators where the Daimler plant achieved better results.
This might be interesting for a factory manager of either one of the two companies in order to improve the performance in the future. On the other hand these results are also interesting for any investor with a focus on sustainable investing. In this case the analysis clearly supports the decision to invest in the BMW plant in Sindelfingen.

Assessment of a Factory over Time
In this section, the sustainability performance of the company BMW is assessed from 2010 to 2012. It is a leading company in terms of sustainability and has been awarded with several prizes. The BMW Group was named for example best automotive producer in the Dow Jones Sustainability Group Index several times in a row and it is ranked at GRI level A+, which means that BMW meets the maximum requirements detailed by the GRI guidelines.
For this case study the evaluation period for all three periods is again the calendar year from 1 January to 31 December. For the most part the data were taken from the BMW Sustainability Report 2012, because all data in this report were audited and verified by a third party and it is therefore a reliable source. In this case the data includes the 17 main production sites e.g. Landshut, Leipzig and Munich and presents an average over all of them.
After gathering and entering the data for each evaluation period into the computer based tool, the data processing is performed. According to chapter 5 the first step of the data processing results in a structured overview. The following three figures illustrate the results and provide detailed information for each evaluation period.  In a next step of this case study it is important to analyze the weighted and combined sub-indices. The indices are presented in Table 6.2. Recalling that the closer the index is to 0 the better the sustainability performance, it can be determined that the environmental and social indices have improved significantly over time. However, the economic dimension shows a minor difference.
Here the index for the year 2011 is slightly better than the index for the year 2012.
Nevertheless, the overall composite sustainability index indicates again the continuous improvement of the sustainability performance over the last three years. The results are also visualized by the sustainability assessment tool (see Figure 6.13).

Figure 6.13: Presentation of assessment results by dimensions for BMW from 2010 to 2012
As a conclusion of this case study it can be summarized that the general development of the sustainability performance of BMW demonstrates a continuous improvement over the last three years. However, it shows also that in the future production managers at BMW should focus more on economic indicators and also on the employee attrition rate.

Results of the Case Study
The case study has been carried out without any significant complications. The results of the study are clearly visualized and provide detailed information that might be used by factory managers as well as investors to support decisions and to guide future activities.
However, the data collection based on public records required more time than expected. The general sustainability report, a report about the main production sites of a company, offered all data needed for this study. Therefore, the second part of the case study was completed within 90 minutes, including the data collection, data entry and data processing. The first part of the case study on the other hand was more time consuming. It has been proven more difficult to gather data for an individual production site than for all sites combined. However, the environmental declarations are very useful and provide data for each environmental indicator and also for some social and economic indicators. According to the Eco Management and Audit Scheme (EMAS), a new environmental policy instrument developed by the United Nations, these environmental declarations are required for each production site in order to be certified by EMAS. Therefore, the use and popularity of such environmental declarations is increasing steadily. Other indicators had to be sought in press releases or webpages.
Nevertheless, there were still some indicators such as the number of working accidents that had to be derived from the general sustainability report, by calculating the share of the total number of working accidents for the specific production site based on its size.
However, this information is not completely accurate.
Overall it can be concluded that data collection for large companies with a focus on sustainability is significantly easier than for small and individual production sites.
However, since sustainability is attracting more and more attention it becomes also more important for companies to be certified by environmental audits such as EMAS and therefore they have to publish more figures and data in the future. For now it cannot be guaranteed to find all data on public records, but there is always the option to get in touch with the sustainability contact person in order to gather more information about a specific production site.

CHAPTER 7 -VERIFICATION OF THE USABILITY
Besides testing the developed factory assessment tool on its functional capability and effectiveness, it needs to be tested on its usability as well. In order to get feedback for initial improvements, the tool is tested under real conditions in collaboration with different test users. Thus the first section of the chapter describes the development of a suitable questionnaire. In the second part, the test persons use the questionnaire to evaluate the software based tool. The results are then analyzed and discussed in order to optimize the assessment tool.

Developing a Usability Questionnaire
Questionnaires are the most frequently used tool for the evaluation of software usability. The goal of the evaluation is to detect weaknesses in the tool and to develop suggestions for improvement. In order to achieve this, the test users have to answer all questions based on their personal experience with the tool.

Usability and ISO Norm 9241
The term usability has been defined by many researchers in many different ways.
However, the International Organization for Standardization (ISO) established an official standard on usability. ISO 9241 defines usability as "the extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency, and satisfaction in a specified context of use" (p.2). Additionally, ISO 9241 part 10 formulates seven principles regarding the description, design and evaluation of software:  Suitability for the task from ---(very negative) to +++ (very positive) (Prümper, Michael 1993).

Structure and Layout of the Questionnaire
Based on the ISONORM 9241/10, a questionnaire for the evaluation of the rapid factory assessment tool is being developed. It consists of the same seven principles and uses the same rating scale. However, since the original ISONORM 9241/10 is usually used for more complex and larger software, certain bipolar items have to be adjusted or simply neglected in order to meet the demands of the rapid factory assessment tool and to decrease the complexity. Thus, the final questionnaire contains only fourteen items instead of thirty five; two for each principle.
The basic layout of the questionnaire is shown in the figure below. Each block contains the name of the principle, the general topic, the bipolar items and a rating scale. The complete questionnaire can be found in the appendix A3.

Error tolerance
Does the tool ensure a minimal rate of errors?
The tool… Agreement to the statement The tool… does not prevent the user from making errors.
prevents the user from making errors.
provides error messages which are difficult to understand.
provides error messages which are easy to understand.

Testing the Factory Assessment Tool
The tool has to be tested under real conditions in order to obtain meaningful results. As the purpose of the tool specifies that any external user should be able to work with the tool, it is not necessary to test the tool in collaboration with a real factory.
Therefore, the evaluation of the tool is carried out based on the case study from the previous chapter. Each test user is given the task to enter the data of the company BMW for the year 2012 based on its sustainability report. Furthermore, they have to compare the results with the results from the years 2010 and 2011 from the database (see section 6.2). It is assumed that the test user has never seen the tool before and works without further help.
Once the task is completed, the test user has to answer the usability questionnaire.
Every question has to be answered based on the personal experience during the case study. For a comprehensive evaluation all questions need to be answered.

Outcome of the Test
The test has been carried out without any significant complications. Each test user was able to enter the data for each indicator and to analyze them as instructed. It was noticeable that it was fairly time consuming to convert the values from the sustainability report into the right format. Nevertheless, each test user was able to complete the tasks within 45 minutes, including the data entry and data processing. Thus, it is proved that the goal of a rapid assessment tool is achieved.
After completing the tasks each test user answered the usability questionnaire.
The results of this survey are illustrated in Figure 7.2.

Figure 7.2: Presentation of testing results
The evaluation of the questionnaire demonstrates that the results for six out of seven principles are more than satisfactory. The principles Suitability for Learning,

Conformity with User Expectations, Self-descriptiveness, Controllability, Error
Tolerance and Suitability for the Task are already in the very positive area. In contrast to these six principles, the results for the principle Suitability for Individualization are rather negative. Since the test users were no VBA-experts they stated that it is very complicated to expand the tool for new tasks or to adapt it to the individual working style.
A subsequent feedback discussion with the test users revealed further suggestions for improvement such as introducing a tabindex and an overview of the different types of industries. In summary, testing the practical case study in collaboration with different test users provided new insights and also possibilities to improve the usability of the tool.

CHAPTER 8 -SUMMARY AND CONCLUSIONS
In the beginning of the study, the basic concept of sustainability as well as the history of its development was presented. Additionally, a comprehensive literature review in the field of sustainability assessment categorized existing sustainability frameworks. It can be concluded that integrated sustainability assessment tools are available at different levels e.g. global, national, company and they are either developed by a company or by a superordinate organization. However, sustainability assessment at factory level is still lagging behind and is not considered sufficiently.
Besides the aspects of sustainability, the study looked also at the factory aspect.
In the next phase, the study presented a basic description of factories and illustrated their importance in terms of sustainability, based on energy use and CO2 emissions. It was pointed out that the manufacturing industry is a main consumer of natural resources and a main producer of adverse environmental impacts. Based on the high responsibility of factories towards their environment, the need for a factory specific sustainability assessment tool was explained. Moreover, it was pointed out that the influence on the environment varies depending on the specific type of industry. In this context, different industrial sectors have been classified as well within this phase.
Based on this situation a framework for a tool at factory level was developed in this research. The framework has a hierarchical structure with the three dimensions of sustainability at the highest level, followed by themes and indicators for each dimension. It was demonstrated that the traditional view in the literature considers three dimensions: social, environmental and economic. This view was adapted for the framework. However, the definition of suitable themes and indicators required more effort. The main sustainability assessment tools that have been identified in the literature review had to be analyzed and compared in order to derive suitable themes and indicators for the framework.
Furthermore, the framework includes a model to calculate an overall composite index. The development of this index followed various steps. First, the indicators had to be judged whether they support or harm a company's sustainability. Then, they had to be normalized in order to avoid adding up incompatible data sets that can lead to inaccuracies. Therefore different normalization methods were analyzed and selected.
It was concluded that a single set of data for one factory cannot be normalized, because currently there is no standardized scale for the assessment values available. However, the method "Distance to a reference" was selected for the case when different companies are compared to each other and the method "Percentage over annual differences" was selected for the case when one factory is being assessed over time.
In the next step it was necessary to weight each indicator based on the type of industry in order to obtain a meaningful evaluation of the sustainability performance of factories within each industry. After analyzing and evaluating different weighting methods based on the fulfillment of criterias is was decided to implement the AHP-method. This method provides insights into indicator importance, handles quantitative and qualitative data, is transferable from factory to factory and does not require a large database in order to be calculated. However, since the weighting is based on experts judgment this method is not as objective as methods that are derived from statistical methods. In a next step a formula had to be defined to calculate a sub-index for each sustainability dimension from this model. Finally, all three sub-indices were combined into one overall composite sustainable performance index. In summary, the framework considers three different cases. In the first case a single factory is being assessed; in the second case two or more factories are being compared and in the third case the development of one factory is being assessed over time. The tool can be used from the external perspective for all three cases and the assessment can be completed rapidly with a minimal time effort.
Based on the given framework, a computer-based tool was developed. Therefore it was necessary to implement the framework into MS Excel by using Visual Basics for Applications. It was pointed out that based on its characteristics VBA is the best fitted solution for the tool.
After implementing the computer based tool, it was then applied to a practical case study in order to demonstrate its application and to test its effectiveness. The first part of this case study focused on the comparison of two production sites. The BMW site in Dingolfing, Germany and the Daimler site in Sindelfingen, Germany. The second part considered the assessment of the BMW Group from 2010 to 2012. The data collection for both cases was based on public records and used the developed sustainability assessment tool for data entering and data processing. It can be concluded that the tool provides clear results and presents them in a comprehensible form. However, the data collection from public records has revealed some difficulties. Smaller, individual production sites without a strong focus on sustainability have not yet published extensive data or figures on this topic.
Finally, the computer based tool was tested on its usability. In order to get feedback for initial improvements, the tool was tested under real conditions in collaboration with different test users. Test users were asked to perform the same case study as mentioned above, by assessing the sustainability performance for BMW in the year 2012 and to compare the results to the years 2010 and 2011. Once the task was completed, the test user had to answer a usability questionnaire. The questionnaire was developed based on the ISO Norm 9241 and modified to meet the needs of the evaluation. As a result of the evaluation, it can be concluded that the tool meets the goal of a rapid assessment tool. All test users completed the task within 45 minutes.
Furthermore, the results of the questionnaire indicated that the tool prevents the user from making errors, is easy to learn, self-descriptive, suitable for the task and easy to control. However, it was also pointed out that it is rather complicated to expand the tool for new tasks or to adapt it to the individual working style.
Finally, it can be concluded that the research objectives of the study were all achieved. It gives an insight into the history of sustainability, reviews and categorizes the current state of sustainability assessment tools, analyses the impact of factories on their environment and classifies industrial sectors. Furthermore, it develops a framework and implements it into a computer-based tool. Finally, it also tests the tool in collaboration with different test users.
However, the larger goal of the study was to fill the gap of a missing sustainability assessment tool at factory level. Theoretically the tool is verified to achieve the goal, but this needs to be confirmed in practice. The tool enables external user such as investors as well as internal users such as factory managers to compare the sustainability performance of different companies or to evaluate the development of a company in terms of sustainability performance. On the one hand this tool supports the investors decision on sustainable investing and on the other hand it may also guide factory managers to think and act in the right direction and to discover possible improvements in order to increase the sustainability metrics related to factory operations. However, there is still potential for future research on this topic, especially when it comes to data collection of small and medium sized factories.

CHAPTER 9 -RECOMMENDATIONS FOR FUTURE RESEARCH
Although ideas for improvement and extensions have been mentioned throughout this study, they shall be summarized at this point. Therefore, the section is divided into two parts. First, the ideas to improve the usability of the developed tool are presented and then the ideas to further improve the assessment framework.

Usability of the Tool
As the results of the questionnaire pointed out, it is important to improve the Suitability for individualization. Since it cannot be assumed that any user is an Excelexpert, a detailed user manual shall be developed. This manual has to instruct the user on how to adapt the tool to the individual working style, to change options and to expand the tool for new tasks.
Minor suggestions on the improvement of the tool such as a tabindex or an overview for the different industrial sectors were already implemented in the latest version of the tool.

Assessment Framework of the Tool
The model of the tool provided in this work does not offer the calculation of an index for the assessment of a single factory. This is due to the fact that no database or standardized scale for the assessment values is available and the values cannot be normalized. This issue offers potential for further development. One possibility would be to collect data for the leading company in each industrial sector. In case a single factory is being assessed the results can be normalized relative to the benchmark company of the specific industry. The leading sustainability companies could be identified based on the DJSI. Also, the case study proved that data collection especially for factories with a focus on sustainability is rather easy to accomplish.
Another step of further improvement considers the weighting of indicators by using the AHP method. Since it is a participatory method, the results will be more sophisticated the more experts participate. Therefore the questionnaire should be placed on a webpage where experts have the possibility to evaluate the indicators continuously. The results of the evaluation have to be stored automatically through an interface into a database.
In the future, it might also be possible to place the entire tool on a public webpage.
Therefore different users have access to it and they would be able to share a database.
The variety of factories in the database would increase and the users have the opportunity to compare the results of different factories with a minimal amount of effort.
This may also solve the problem concerning the complicated and time consuming data collection for small and midsize factories based on public records.

Environmental Dimension
Please do a pair-wise comparison between each pair of indicators, by judging "how important is indicator j relative to indicator i ?". Values are given on a scale from 1 to 9 to show the intensity of preference (see

Economic Dimension
Please do a pair-wise comparison between each pair of indicators, by judging "how important is indicator j relative to indicator i ?". Values are given on a scale from 1 to 9 to show the intensity of preference (see table below). The larger the number, the greater the importance.
Factor of preference Importance Please fill out the next tables for each type of industry, by moving the orange dot (an example can be found at page "environmental dimension"):

Basic Materials/Resources (Oil & Gas, Chemicals, Basic Resources)
Indicator

Social Dimension
Please do a pair-wise comparison between each pair of indicators, by judging "how important is indicator j relative to indicator i ?". Values are given on a scale from 1 to 9 to show the intensity of preference (see

A3: Usability Questionnaire
Please evaluate the factory assessment tool by judging each statement of the questionnaire. Values are given on a scale from ---(very negative) to +++ (very positive).
The goal of the evaluation is to detect weaknesses in the tool and to develop suggestions for improvement. In order to achieve this, please answer every question based on your personal experience.

Suitability for the task
Does the tool support to realize the tasks more effectively and efficiently?
The tool… Agreement to the statement The tool… is complicated to use. is not complicated to use.
requires unnecessary input. does not require unnecessary input.

Self-descriptiveness
Is every step understandable in an intuitive way?
The tool… Agreement to the statement The tool… uses terms, definitions and/or symbols that are difficult to understand.
uses terms, definitions and/or symbols that are not difficult to understand.
does not offer contextsensitive explanation, which are concretely helpful.
does offer context-sensitive explanation, which are concretely helpful.

Conformity with user expectations
Is the tool consistent with common expectations and habits?
The tool… Agreement to the statement The tool… complicates orientation due to an inconsistent design.
facilitates orientation due to a consistent design.
provides insufficient insight regarding its current status.
provides sufficient insight regarding its current status.

Suitability for learning
Is the effort for learning the tool as low as possible?
The tool… Agreement to the statement The tool… requires a lot of time to learn. requires little time to learn.
cannot be used without previous knowledge or training.
can be used without previous knowledge or training.

Controllability
Is the user able to start the sequence and influence its direction?
The tool… Agreement to the statement The tool… forces the user to follow an unnecessarily rigid sequence of steps.
does not force the user to follow an unnecessarily rigid sequence of steps.
does not support easy switching between individual menus or masks.
supports easy switching between individual menus or masks.

Error tolerance
Does the tool ensure a minimal rate of errors?
The tool… Agreement to the statement The tool… does not prevent the user from making errors.
prevents the user from making errors.
provides error messages which are difficult to understand.
provides error messages which are easy to understand.

Suitability for individualization
Does the tool allow customizing according to the task and individual preferences?

The tool…
Agreement to the statement The tool… is difficult to expand for new tasks.
is easy to expand for new tasks.
is difficult to adapt to the individual working style.
is easily adaptable to the individual working style.