MSc Business Administration
Track Digital Business
Master Thesis
The Implementation of Blockchain in the Food Industry
by
Mireille Smeets
11644494
June, 2018
15 ECTS
Research conducted from January 2018 to June 2018
Supervisor/Examiner:
i
Statement of Originality
This document is written by Mireille Smeets who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
ii
Preface
This master thesis has been written as a part of fulfilling the final requirements for completing the Master Business Administration Program at the University of Amsterdam. There are a number of people I would specifically like to thank, since without them, I would not have been able to write this thesis.
First, I would like to express my gratitude to my thesis supervisor Prof. em. dr. ir. H. J. Oppelland for his continuous support, feedback, and guidance during the thesis writing process.
Secondly, I would also like to express my gratitude towards all the respondents, who were willing to give some of their time. Without their insights and expertise, this research would not have been conducted successfully.
Last but not least, I would like to thank my family, friends and boyfriend for their incredible and best possible support from the beginning of the thesis writing journey onwards.
iii
Abstract
The objective of this research is to identify the factors which determine the success (e.g. success factors) of blockchain technology implementation in the food industry. Prior research literature has mainly focused on how the technology has been accepted in various contexts. However, limited research has been done on the acceptance of blockchain technology in the food industry. From the review of earlier research literature, eight propositions were derived which were tested via semi-structured interviews with management level employees from different food companies. A total sample of ten respondents were interviewed, and six out of eight propositions were (partly) accepted as true. These results build further on the call for generalizability and applicability of the technology acceptance model (TAM). In this research, the strongest relationship was found between the perceived usefulness explaining the intention to adopt new technology. Due to the industry specific nature of this case-study, future research is needed to test the generalizability of the theory in other industries and a larger sample would be preferred to be able to do statistical tests.
Keywords: blockchain, food supply chain, performance, technology acceptance, transparency, traceability
iv
Acronyms
ERP Enterprise Resource Planning FSCM Food Supply Chain Management IoT Internet of Things
IT Information Technology
IS Information System
NGO Non-Governmental Organization RFID Radio Frequency Identification TAM Technology Acceptance Model
UTAUT Unified Theory of Acceptance and Use of Technology VIoT Value Internet of Things
v Table of Contents Statement of Originality ... i Preface ... ii Abstract ... iii Acronyms ... iv 1 Introduction ... 1 1.1 Research Problem ... 2 1.2 Research Objective. ... 3 1.3 Research Methods ... 3
1.4 Structure of the Thesis ... 4
2. Literature Review ... 5
2.1. Introduction of Blockchain ... 5
2.2. Technology in the Food Industry ... 6
2.3 Blockchain in the Food Industry ... 8
2.4 Technology Adoption Research ... 10
2.4.1 UTAUT ... 11 2.4.2 TAM... 12 2.4.3 Perceived Trust ... 13 2.4.4 Perceived Risk ... 14 3. Research Design ... 15 3.1 Conceptual Model ... 15 3.2 Propositions ... 16 3.2.1. Perceived Trust ... 16
3.2.2. Perceived Ease of Use ... 17
3.2.3. Perceived Usefulness ... 18 3.2.4. Perceived Risk ... 18 4. Method Selection ... 19 4.1 Research Design ... 19 4.2 Sample ... 19 4.3. Data Collection ... 20 4.4 Level of Measurement ... 21 4.5 Data Analysis... 22 5. Research Results ... 25 5.1 Perceived Trust ... 27
5.2 Perceived Ease of Use ... 30
5.3 Perceived Usefulness ... 31
5.4 Perceived Risk ... 32
6. Discussion and Conclusion... 34
7. Limitation and Future Research ... 36
Bibliography ... 38
Appendix ... 42
1
1 Introduction
Lately there has been a growing interest in blockchain and specifically in cryptocurrencies. Since the introduction of Bitcoin in 2009, several researchers have studied the underlying concept of Bitcoin, the blockchain technology, and discovered the great potential for the possible implementations in other industries and services (Underwood, 2016). Foroglou and Tsilidou (2015) call it the next generation technology. In their research, Zhao et al. (2016) compared Google search volumes for the keywords “blockchain” and “Bitcoin” from 2011 until 2016. In the first two years, it seems that there was no great interest in these keywords, however, the search volume for Bitcoin reached its peak in 2013. The search volume for “blockchain” started to rise by the end of 2014 and has been increasing ever since. Yli-Huumo et al. (2016) collected 41 relevant research papers on blockchain technology and found that 80% of the papers were focused on Bitcoin and less than 20% consider other applications of the blockchain technology. Today, however, companies in financial services are greatly investing in blockchain. Nonetheless, the adoption of blockchain applications in the business world is still fairly unexplored (Godsiff, 2015).
There are also positive examples of blockchain implementation in the food industry. A group of leading retailers and food companies including Walmart, Unilever, and Nestlé announced a blockchain collaboration with IBM in order to strengthen consumer confidence. They envisioned a fully transparent food system. The food supply chain is complex. It includes farmers, processors, manufacturers and distributors. In today’s traceability system each party along the supply chain is responsible for their own procedure to administrate (either on paper or not), process and distribute their own product or service. However, the systems used do not always correspond or communicate properly with other systems that regulate a product. As a consequence, it is not always possible to fully see what happens in the food system. By implementing blockchain technology in the food industry, the three above mentioned
2
companies hope to bring all involved stakeholders together and collaborate, in order to quicken and optimize the food tracking process.
1.1 Research Problem
Food concerns everyone and the last couple of decades the trust and confidence of the customer in the food industry has been tremendously damaged due to food scandals and safety risks (Tian, 2016). In society some concerns have been raised on food safety, and there is an increase in demand for information, safety and transparency regarding food (Ge et al., 2017). The diminished trust and confidence causes the customer to question whom to trust. Blockchain technology, however, promises to solve many problems related to the lack of trust (Ge et al., 2017). Attention from academic and industrial areas has been drawn due to increasing concerns of customers regarding food safety and the quality of food. However, in order to implement such a technology, every part of the supply chain should be on board. Since so far little research has been done on the possibilities of blockchain implementation in the food industry and the acceptance of companies and managers towards this new technology. This research will focus on measuring the attitude of managers towards this new technology, determining which factors are needed to successfully implement the blockchain technology in the food industry. There is existing literature on the technologies currently used in the food industry, though it has not yet incorporated the technology of blockchain. Academia has paid little attention to the use the technology of blockchain in the food industry. To bridge the existing gap in the literature, this thesis will study technology adoption literature and analyze how the blockchain technology can be implemented in the food industry.
3
1.2 Research Objective.
In recent years, there have been various examples of information technology (IT) failures, which resulted in huge financial losses. Low adoption of IT has been indicated to be one of the key causes for these failures (Venkatesh & Bala, 2008). The objective of this research is to identify the factors which determine the success (e.g. success factors) of blockchain technology implementation in the food industry. In this research, the implementation of blockchain technology is successfully accomplished when it tracks food correctly, at acceptable cost, and when it increases trust and traceability. The main research question in this research will be:
What are the factors that determine the success (e.g. success factors) of blockchain technology implementation in the food industry?
To achieve the research objective, this research will analyze the effectiveness of adoption of the blockchain technology in the food industry from different organizations’ perspectives. The adoption literature will be explored to establish a theoretical foundation and a conceptual model. From this theoretical framework propositions will be used and tested accordingly.
In the end, this research will provide both theoretical and practical knowledge. Practical knowledge can be used for managers interested in implementing blockchain in the food industry as the adoption research will investigate the applicability of adoption theories in a fairly unexplored field, namely blockchain in the food industry.
1.3 Research Methods
For this research, a single case study done within the food industry will be analyzed. A qualitative research approach, with interviews as a mono data collection method, will take place in order to collect the data. A purposive sampling strategy is used to select a sample of food professionals. These respondents are selected based on their knowledge and position within a food supply chain. The sample includes farmers, manufacturers, distributors and retailers in order to ensure an informed overview of the industry.
4
1.4 Structure of the Thesis
The remainder of this thesis is divided into six chapters. Chapter two provides a brief overview of the literature on blockchain technology, the current technology in the food industry and the driving force behind the increasing concerns on food safety. TAM by Davis (1989) will be taken into account to investigate to what extent blockchain could be accepted by organizations. Chapter three presents the research design, including the conceptual model and the propositions following from the conceptual model. Chapter four introduces the used methodology, including an explanation of the selected sample and the data collection method. In chapter five the findings and results of this research are found. Chapter six contains the discussion and conclusion, which summarizes the main findings. The last chapter addresses the limitations of this research and addresses some suggestions for future research.
5
2. Literature Review
The aim of this literature review is to provide a thorough overview of the literature on blockchain technology, food supply change management and the current technology used in the food industry. To investigate whether blockchain would be accepted by organizations, several technology acceptance models will be studied.
2.1. Introduction of Blockchain
The blockchain technology was introduced in 2008, in the white paper published by Nakamoto (2008). It introduced the Bitcoin cryptocurrency. Blockchain is the technology behind Bitcoin and has the potential to disrupt existing business models in several industries. In its generic form, a blockchain enables a digital transfer between two unknown parties without a trusted middleman based on a peer-to-peer basis (Kewall et al., 2017). Blockchain eliminates the need for intermediaries and can be considered to be a chronological database of transactions documented by a network of computers, also known as a ledger (Peters & Panayi, 2016). No single party owns or controls the ledger and all parties in the network can view the recorded transaction (Kewall et al., 2017). Instead of storing the transaction on a single database, transactions are distributed on the network of computers of the system (Kewall et al., 2017). These transactions are being stored and recorded in blocks while the chain of these blocks becomes the approved history of transactions (Peters & Panayi, 2016). To make sure that only correct transactions are stored and recorded into a blockchain, the network has to confirm that the new transaction is valid. The transaction only becomes valid once it is included in a block and published to the network. A new block of data will only be added to the end of a blockchain after the computers on the network have reached consensus on the validity of all the transactions that constitute it (Peters & Panayi, 2016). This process is called mining in which users of the network get an incentive to verify the correctness of the transaction (Yli-Huumo et al., 2016). Through this process, blockchain enables parties to coordinate individual
6
transactions without needing a third party to verify and clear all transactions (Peters & Panayi, 2016). The main objective of this technology is to achieve database consistency and increase the integrity in a distributed decentralized database (Ge et al., 2017). Experts in the field consider this blockchain technology to be a disruptive and transformative innovation with revolutionary power (McKinsey, 2016).
2.2. Technology in the Food Industry
Due to globalization, the distance that our food travels from “farmer to fork” has increased significantly (Tian, 2016). Therefore, the challenge to keep our food safe has increased simultaneously, as has the challenge of guaranteeing the quality of the food along the food supply chain (Aung & Chang, 2013).
Nearly all companies in the food supply chain run computerizes enterprise resource planning (ERP) systems and supply chain management software. About twenty years ago, the US food industry developed and implemented traceability systems to improve the food supply chain management (FSCM) and to facilitate the tracking ability for food safety reasons (Golan et al., 2004a; Zhong et al., 2017). Some of these systems track food in a precise way, from the retailer back to the source back to the minute of production, and others focus mainly on cursory information in the supply chain like large geographical areas of production (Dickinson & Bailey; Zhong et al., 2017).
Information asymmetry between actors in the supply chain could withhold customers from making ethically-based decisions when purchasing a product (Giddens et al., 2016). To reduce information asymmetry, greater interaction between the actors is needed; and better information sharing is critical when coordinating the supply chain (Zong et al., 2017). To increase the available information of one’s environmental and social impact third-party certifications and labels are used, such as Fairtrade (Sayogo et al., 2015). The Ecolabel Index is the largest global directory of ecolabels and currently tracks 464 ecolabels (Ecolabelindex,
7
2018). However, limited information available on the meaning of each certification makes it very complex for the end user to assess the credibility of the labels (Sayogo et al., 2015). The objective of information disclosure is, in fact, to help consumers make better choices. Therefore, it is of high importance that data is relevant, accurate, reliable, available and traceable in an appropriate quantity (Sayogo et al., 2015). Moreover, to earn the trust of shareholders, it is crucial to provide relevant and reliable information.
Food traceability refers to a trail of data which follows the physical trail of food (Smith et al., 2005; Zhong et al., 2017). Traceability is increasingly used as a method to connect the producers with the consumers (Regattieri et al., 2007). A product traceability system requires the identification of all the physical entities of the product in the supply chain process, including where it is processed, packaged, stored, and stocked (Regattieri et al., 2007). In fact, good traceability systems help to keep the production and distribution of poor quality and unsafe products to a minimum (Aung and Chang, 2013). Also, they make it possible to have a prospective product recall for safety reasons and the possibilities for efficient investigation about the cause when problems occur (Regattieri et al., 2007). The current labeling system in the food industry cannot ensure that the food is of good quality, safe, and authentic. Therefore, traceability systems have been designed as a tool to guarantee the quality of the food, safety, and to build consumer trust (Aung and Chang, 2013).
In their research, Zhong et al. (2017) considered traceability for FSCM when implementing IT-based solutions. Radio Frequency Identification (RFID) is a wireless sensor technology, similar to the bar coding concept (Roberts, 2006). Most of the RFID-based supply chain traceability systems make use of a centralized system in which a third-party organization needs to provide information transparency along the supply chain (Tian, 2016). However, these kinds of centralized systems are obscure and make it impossible for the user to receive more detailed information about the product or transaction. This non-transparency could result in
8
information fraud and exertion for parties in the supply chain (Tian, 2016). WaltonChain (2018) tries to connect the blockchain technology with RFID hardware with the Internet of Things (IoT). This is also known as the Value Internet of Things (VIoT) (WaltonChain, 2018). Their RFID chips add all the collected data immediately to the blockchain without intervention from a trusted third party.
As just mentioned, with the use of blockchain, there will be no need for a trusted middleman or centralized organization. Blockchain as a new decentralized traceability system could become a disruptive innovation in the food supply chain and increase the transparency, traceability, information credibility and safety assurance along the supply chain (Tian, 2016). With the use of RFID and blockchain the information of the entire food supply chain will be available for the end consumer. When the consumer uses the RFID reader he can immediately obtain basic information on the product. Due to blockchain technology, the information along the supply chain is fully auditable when the final product is scanned. Tian (2016) states that this technology will lead to an increase in consumers’ trust in the product and the confidence in the food market as a whole. Aung and Chang (2014) conclude that food traceability from farm to fork will become a reality as future government regulation in combination with consumer demand will force the supply chain to take the idea of visibility to the next level.
2.3 Blockchain in the Food Industry
Several pilot tests have been conducted to implement blockchain in the food industry. According to Galvin (2017), the blockchain technology could be a game changer in the food industry. He argues it could save time by reducing the transaction time from days to nearly instantaneous. Secondly, it removes costs, including overheads and the cost of intermediaries. In addition, it reduces risks, by tampering fraud and cyber-crime and lastly it increases the levels of trust through shared processes and recordkeeping.
9
Provenance in the tuna industry. Provenance is a company that helps brands to build consumer trust through transparency. In its pilot test, Provenance used the blockchain technology with smart tagging and mobile use in order to track tuna fish, while taking into account the verified social sustainability claims (Provenance, 2015). The aim of this pilot was to aid proof of compliance to demonstrate that all parties complied with relevant regulations and legislations towards standards at origin, to counter double spending of certificates, and explore how blockchain technology could be the basis for more transparency and traceability for the consumer. Provenance (2015), in fact, used this peer-to-peer technology to track tuna caught in Maluku, Indonesia, which is the largest tuna producing country (Sunoko & Huang, 2014). The application links identity, location, material attributes, certification and audit information with a specific item or batch ID. Thereafter, the data is stored in an immutable and decentralized format which is globally auditable and protects identities by default, which then ensures secure data verification (Provenance, 2015).
The new application is designed to work through a simple smartphone interface through which local fisherman send a text message in order to register their catch. By sending this message, a new asset on the blockchain will be issued (Provenance, 2015). Then, the tuna with its new permanent and unique ID, will be transferred from the fisherman to the supplier through both the physical transaction and the digital transaction on the blockchain. The identities and further additional information of the fishermen is saved and stored permanently in a list of previous owners of that blockchain (Provenance, 2015). The social (for instance ethical labor, fair pay) and environmental conditions for the fishermen are verified through trusted local Non-Governmental Organizations (NGO), whose audit systems confirm the fishermen’s compliance to an external standard. Using a blockchain explorer allows an individual or organization to verify the raw content of the digital asset that illustrates an item on the blockchain (Provenance, 2015).
10
The next phase of this pilot test is to link blockchain to existing systems. Supply chain management systems already exist these days. However, these systems run mostly on internal hardware or private cloud environments with data silos that rarely take into account the full supply chain. Blockchain, however, provides a neutral open platform where there is no need for a third party to authorize transactions. Provenance (2015) developed a community-owned open standard through which a unique ID in the system takes the form an address on the blockchain. The blockchain provides an audit layer on top of an existing data management system which then makes it irrelevant which specific platform or system each company uses, as long as the party commits its transaction to the blockchain. By collecting the data stored at the address on the blockchain, other parties can access more adequate information and details about that particular item (Provenance, 2015). The final phase is about consumer experience and building an interface for trust, in which they explored how information from on its origin and the entire supply chain can be reached and trusted by the end consumer. With the use of QR codes and the users’ smartphone, customers can scan their products and are able to get access to the full and complete history of the product before they purchase it. Through this method, the customer is able to fully track the product in a unique way, from farm to fork.
2.4 Technology Adoption Research
Although blockchain is receiving an increasing amount of public attention, this does not automatically mean it will become widespread adopted in every industry. The final decision whether an individual will make use of technology is based on the acceptance of somebody towards a technology (Yi et al., 2006). The rising interest in technology acceptance studies has led to the existence of many competing models on technology acceptance, in which each has a different set of acceptance determinants (Venkatesh et al., 2003). In this research, two models will be compared, the technology acceptance model by Davis (1989) and the Unified Theory of Acceptance and Use of Technology (UTAUT) by Venkatesh et al. (2003). These models
11
have both been proven to be robust models in predicting a consumer’s intention whether or not to use a new technology.
2.4.1 UTAUT
UTAUT was developed in 2003 by Venkatesh et al. to measure consumer acceptance and technology usage. The model was developed and integrates elements of eight prominent models, namely the technology acceptance model, innovation diffusion theory, model of PC utilization, social cognitive theory, motivational model, theory of planned behavior, theory of reasoned action, and a model combining the technology acceptance model and the theory of planned behavior. According to Venkatesh et al. (2003), UTAUT measures the behavioral intentions for the use of the technology based on four factors. First, it measures the performance expectancy, which is the degree to which one believes that using the system will enhance job performance. Secondly, it measures the effort expectancy, which is the degree of ease associated with the use of the system. The third factor has to do with the facilitative conditions; it takes into account to which degree it is believed whether an organizational and technical infrastructure exists to support the use of the system. The fourth and last factor is social influence, which is the degree that somebody believes that other people think that the new system should be used. According to the authors, UTAUT is considered to be of high value for managers who want to measure the likelihood of success, when introducing a new technology. Also, it helps to more fully understand what causes the acceptance so that strategies can be designed in order to target users that may be less likely to adopt and use new systems in the first place. In 2012, Venkatesh et al. complement the UTAUT model by integrating hedonic motivation, price value and habit to the initial model. This new model is called UTAUT2. Compared to UTAUT, the extension proposed in UTAUT 2 produced a considerable improvement in the variance explained in technology use and behavioral intention (Venkatesh et al., 2012).
12 2.4.2 TAM
The second model taken into consideration in this research is TAM by Davis (1989). TAM is an information systems (IS) theory that models whether a user may or may not adapt and implement a new technology (Venkatesh & Davis, 2000). TAM theorizes that the behavioral intention of an individual whether to use a system is set by two beliefs: the perceived usefulness and perceived ease of use (Davis, 1989). Venkatesh and Davis (2000) refer to this model as a robust and powerful model for predicting the user acceptance of a new technology. TAM is later extended to TAM2, which then has included social influence processes and cognitive processes. In 2008, Venkatesh and Bala extended the model to TAM 3. In this latest extension, the authors have tried to bridge the gap of research in the IT implementation literature on the role of interventions that have influenced managerial decision making. From an organizational point of view, this is of significant importance as they try to measure how managers make informed decision regarding these interventions and how this then can lead to greater acceptance and utilization of IT (Venkatesh & Bala, 2008).
2.4.2.1 Perceived Usefulness
Perceived usefulness is defined by Davis (1989) as: “the degree to which a person believes that using a particular system would enhance his or her job performance” (p. 320). The author states that people tend to use a new technology to the degree that they believe it will increase job performance. The author refers to this as the first variable perceived usefulness.
2.4.2.2 Perceived Ease of Use
Perceived ease of use refers to “the degree to which a person believes that using a particular system would be free of effort” (Davis, 1989, p. 320). The author states that it could be the case that a potential user believes that a given technology is useful, however, also believes that it is too complicated to use. Then the performance benefits of usage are outweighed by the effort necessary to use the application.
13 2.4.3 Perceived Trust
The main benefit of blockchain technology is the absence of a trusted third party (Nakamoto, 2008). The Economist (2015) described blockchain technology as “the trust machine” indicating that blockchain technology runs without trust concerns, which makes a transaction between parties “trust-free” once it is documented in the blockchain. Blockchain technology has become very popular in the finance industry where trust, transparency, and security in transactions between two or more parties is crucial (Economist, 2015). Banks are investing heavily in the development of this technology and its applications. Applications, in fact, also do not require any trusted intermediary, such as a central bank to guarantee and ensure trust and security (Kewell et al., 2017).
Up until now, little research has been done in the field of trust and blockchain; however, the concept of trust has been studied in a variety of disciplines in the academic field. Gefen (2003) conceptualized trust as the willingness to depend based on beliefs in ability and integrity. McKnight et al. (2002) studied trust measures for e-commerce. They state that trust is a multidimensional concept and plays a crucial role in helping consumers to overcome perceptions of risk and insecurity. Bhattacherjee (2002) found in his research that the lack of trust in web shops can, in fact, discourage the adoption of e-commerce.
Trust is a critical factor to cherish commitment between supply chain partners. It can overcome perceptions of risk and uncertainty, and foster trust-related behaviors (McKnight et al., 2002). They also found that perceived trust indicates whether a person perceives a new technology as secure and trustworthy. Koka and Prescott (2002) state that trust is a significant predictor of positive performance within inter-organizational relationships. The absence of trust between partners along the supply chain often results in ineffective and inefficient performance (Kwon & Suh, 2004). According to Hoffman (2018), the only way to make market transformation work is through trust. To gain trust, greater transparency is needed. However, with the growing demand for data, the idea of transparency has taken a new level, which can
14
create both opportunities and new awareness’s of risks within the supply chain (Hoffman, 2018).
2.4.4 Perceived Risk
Perceived risk is mostly defined as feeling uncertainty because of potential negative consequences when using a product or system (Featherman & Pavlou, 2003). In 1974, Ostland argued that risk functions can be seen as an additional measure when considering IT adoption and the study by Nicolaou and McKnight (2006) found evidence that perceived risk negatively influences the intention to use new technology.
A new technology is generally not implemented quickly. In the adoption process, it is critical to consider that the end user is influenced by the risk they perceive from a new technology or innovation, apart from whether there is an actual risk or not. Also, perceived privacy violation regarding the blockchain technology can cause the potential user to postpone the decision to adopt such a technology. This is therefore of significant importance to consider when the adoption process of new technology is studied.
15
3. Research Design
TAM is an important analytical tool in the field of research on the social mechanism of technology adoption (Folkinshteyn & Lennon 2016). It is referred to as being most influential and commonly used in technology acceptance studies (Benbasat & Barki, 2007). The use of TAM is, to a wide extent, studied and analyzed on web-related technologies including extensions of trust, privacy, risk and social awareness (Venkatesh & Davis, 2000; Folkinshteyn & Lennon, 2016). Given that TAM has formerly proved to be a robust and powerful method to analyze rather disruptive and new technologies (Folkinshteyn & Lennon, 2016), this model will be used as the theoretical framework for this part of the research, in order to analyze the potential acceptance of the blockchain technology in the food industry.
3.1 Conceptual Model
This research makes use of a conceptual model based on TAM by Davis (1989), see figure 1. Based on the literature review, the conceptual TAM model is applied in this research and visualized in table 1. Perceived risk and perceived trust are added to the model and specific factors are applied for testing the blockchain technology. This conceptual model provides theoretical guidance, in order to set research propositions for the adoption of blockchain in the food industry.
16
Table 1. TAM factors applied to blockchain technology Perceived risk Perceived
ease of use Perceived usefulness Perceived trust Risk of business failure Performance risk Privacy risk Common language Ease of learning Record integrity Performance Secure Trustworthy 3.2 Propositions
Based on the theory in section two and the researchers’ own interpretation, certain assumptions have been made, which can be found below. These will be empirically tested. It might be the case that other additional motivators are found during the research. For that reason, this conceptual model will be re-evaluated in the discussion section. The following propositions explore the relationship between the variables that influence the intention to use a new technology.
3.2.1. Perceived Trust
In this research, the variable trust is added to charter the effect of trust in organizations’ adoption decisions. Trust is a critical factor to foster commitment between supply chain partners. It can overcome the perceptions of risk and uncertainty, and engage in trust related behaviors (McKnight et al., 2002). Higher levels of trust could reduce the level of uncertainty and anxiety. It therefore can be assumed to have a negative impact on perceived risk. Gefen et al. (2003) argued that both understanding of a piece of technology, and trust issues are important in determining the behavioral intention to actually use a new technology. The study by McKnight et al. (2002) found that perceived trust indicates whether a person perceives a new technology as secure and trustworthy. Trust is stated to increase certain aspects of the perceived ease of use as well as perceived usefulness (Gefen, 2003).
17 Therefore, the first four propositions emerged:
Proposition 0. Higher perceived trust in blockchain technology results in higher intention to use the new technology.
Proposition 1. Higher perceived trust in blockchain technology results in higher perceived ease of use.
Proposition 2. Higher perceived trust in blockchain technology results in higher perceived usefulness.
Proposition 3. Higher perceived trust in blockchain technology results in lower perceived risk.
3.2.2. Perceived Ease of Use
Perceived ease of use is defined by Davis (1989) as “the degree to which a person believes that using the system will be free of effort” (p. 320). In his study Davis (1989) argued that it could be the case that a potential user believes that a given technology is useful, however, at the same time be convinced that the system is too complicated to use. Therefore, when using the product, the performance benefits are outweighed by the effort it takes to use the application. Blockchain provides an audit layer on top of an existing data management system. By storing the data in the blockchain, other parties can access more adequate information and item specific details (Provenance, 2015). This could increase the common language within the supply chain. Most blockchain applications are still in the phase of development and their perceived ease of use differs significantly. The perceived ease of use is then based on whether using a particular technology will be effortless and the degree of ease of learning. This led to the following propositions:
Proposition 4. Higher perceived ease of use results in higher perceived usefulness
18 3.2.3. Perceived Usefulness
Perceived usefulness is defined by Davis (1989) as “the extent to which a person believes that using a particular technology will enhance her or his job performance” (p.320). Therefore, the following proposition will be tested:
Proposition 6. Higher perceived usefulness results in higher intention to use a new technology
3.2.4. Perceived Risk
Perceived risk is mostly defined as the feeling of uncertainty towards potential negative consequences when making use of a product or system (Featherman & Pavlou., 2003). Cunningham (1967) typified perceived risk as having six dimensions, including performance risk and privacy risk. Featherman and Pavlou (2003), found for both performance risk and privacy risk important causes for concern. They argued this could lead to reduces system adoption. Thus, perceived risk is incorporated in this model as previous research has argued that perceived risk is of significant relevance to adoption research. This has led to the last proposition:
19
4. Method Selection
In this section, the chosen research methods will be justified. The method used will answer the research question: What are the factors which determine the success (e.g. success factors) of
blockchain technology implementation in the food industry?
4.1 Research Design
Since the study of blockchain technology is a rather a new area of research, a qualitative research approach qualifies as a useful method (Blumberg et al., 2008). In this thesis, a rather new phenomenon is studied through an exploratory research. Exploratory research can provide the researcher with new insights about a phenomenon and help to understand the problem better (Yin, 1994). This research is a combination of both an inductive and a deductive approach, as the limited available literature will be used deductively to provide theory that can be used as a basis for this research. By conducting interviews, the researcher will develop new theory, while also testing the existing theory. This research is cross-sectional in nature and allows the researcher to compare different variables at the same time (Saunders, 2011). The research strategy of this thesis entails investigating a single case study and according to Blumberg et al. (2008), case studies are an appropriate research strategy when there is a lack of specific and appropriate literature written on a certain topic or phenomenon.
4.2 Sample
The unit of analysis in this research is the food industry. Within this industry, ten managers from different food supply chains will be interviewed as a mono data collection method. According to Miles and Huberman (1994), the use of a purposive sampling strategy, in order to select your sample, presents outstanding opportunities to develop theory. The sample members are selected based on their knowledge, relationships and expertise. Respondents were selected from companies which have not yet implemented blockchain technology in their food supply chain. Also, the respondents are all part of a (different) food supply chain. The sample
20
includes farmers, suppliers, manufacturers, and retailers from different supply chains. The aim of the researcher is to develop a sample that reflects several parties of different supply chains to acquire a thorough overview. See table 2 for an overview of the sample. The focus of these particular characteristics enables to answer the research question and meet the objective for this research (Saunders, 2011). Contact via phone or email was used to communicate with the respondents and set the interviews. When appropriate, the private network of the researcher was used for arranging the interviews.
Table 2: Sample overview Interview
number Company Sector
1 Chef’s Culinar Wholesale distributor for food and non-food products
2 VosVis Fish Wholesale
3 Fresh Valley Tomato farmer
4 Arla Foods Dairy product producer
5 Gebrs. Fuite BV Animal feed supplier
6 MARS Manufacturer of confectionery, pet food and other food
products
7 Jumbo Supermarket Retailer, supermarket
8 Mooren’s farm Dairy farmer
9 Jumbo Headquarters Retailer, Headquarters
10 Marel Poultry Supplier of poultry processing technology
4.3. Data Collection
The data was collected via semi-structured interviews held between May and June 2018 and followed a specific interview protocol. The interview protocol is based on the propositions from the literature review. Before the interviews took place, a pilot test was run in order to check the wording of the interviewer, the duration, format and flow necessary. Each interview had an approximate duration of one hour. With permission from the interviewees, the interviews were recorded and transcribed after the interview took place. Then, the data of the interview was systematically analyzed with Nvivo 12, a qualitative data analysis software program in order to identify emerging patterns or common themes (Strauss & Corbin, 1990).
21
According to Eisenhardt (1989), information-processing bias can be prevented through the use of software and coding since it enables more systematic qualitative data analysis.
To develop a rigor case study design, reliability and validity are taken into account as they are the cornerstones of scientific methods (Yin, 2009). In order to ensure reliability, a case study database was set up, with all the collected data in order to ensure that the same insights would be derived from the data if the research was conducted at a later moment in time. Validity has three dimensions: internal-, external- and construct validity. Construct validity checks if the research actually measures what it claims to measure. Internal validity refers to whether the researcher provides plausible and logical arguments that logically lead to the conclusions made (Yin, 2009); this is also known as logical validity (Yin, 1994). In this research construct validity is pursued through the use of a clear research framework and through considering and incorporating rival explanations based on earlier written literature. According to Eisenhardt (1989) including triangulation in your research is of major importance. By implementing triangulation from collected data and existing theory the researcher ensures construct validity (Yin, 2009). As blockchain is a rather new technology, one of the main limitation of this research is that there is only little research in this field. This makes it hard to establish a chain of evidence. However, construct validity is ensured through the view of the draft case study report by one of the interviewees. Another limitation of this research is that the findings will be hard to generalize to other sectors and supply chains since a single case study often offers a poor basis for generalization (Gibbert & Ruigrok, 2010).
4.4 Level of Measurement
In 1932, Rensis Likert, developed the Likert scale, which is most commonly used to measure the attitude of a respondent towards a given statement (Jamieson, 2004). Since the attitude of the respondents towards blockchain technology is measured, the use of the Likert scale seems to be an appropriate choice for this research. Still, there are some contradicting arguments in
22
the academic field on which level of measurement is appropriate when using the Likert scale. According to Stevens (as cited in Jamieson, 2014), there are four types of data namely, nominal ordinal, interval, and ratio scale. With nominal data, the order of the values is arbitrary, and it is therefore never correct when numeric values are assigned to the response of a participant. According to Jamieson (2004), the Likert scale can be seen as an ordinal level of measurement. He states that the response categories have rank orders, i.e. strongly disagree (1), and strongly agree (5), however, the interval between these values cannot be presumed to be equal. Nevertheless, in this research, the researcher is willing to assume that the values on the 5-point Likert scale are approximately evenly spaced and assign sequential numeric values to the answers. Due to the small sample size, the researcher has not consulted statistical software such as SPSS to statistically analyze the data. Instead the data will be analyzed manually.
4.5 Data Analysis
The data analysis process using the software Nvivo consists out of three steps. The first step was to make a word cloud, a tool of Nvivo to identify keywords in the data set, see figure 2. Figure 2. Word cloud
23
Secondly, the data was divided into the main codes. According to Yin (2009), there are four general strategies that guide a researcher when making an argument. In this research, the strategy of relying on theoretical propositions was used; the propositions set in chapter three have shaped the data collection plan (Yin, 2009). In this analysis, a four-step coding process was conducted based on literature from Miles and Huberman (1994). First, the researcher revealed relevant statements and assigned codes to the data (open coding). Secondly, the researcher searched for quotes and statements that fit into the relevant categories. New codes were developed, and themes were identified in this stage, also known as axial coding. Afterwards, possible patterns and explanations were considered. Codes were grouped together under a more general code and positioned in a more logical sequence. Pattern-matching is a technique that compares empirically based patterns with those that can be predicted (Trochim, 1989; Yin, 2009) and according to the author, this is the most desirable technique used in case studies. Saldana (2009) states that patterns can be characterized by their similarity, difference, frequency, sequence, correspondence and causation. In the last step, the researcher chose a category as the core category (selective coding). She related all other categories to this core category and looked at evidence that was contradictive or confirmative.
A combination of both a deductive and inductive approach was used for the coding process. The researcher made use of starter codes, while also taking into consideration the possibility of new codes and themes during the process. While coding the researcher, in fact, added new codes and merged and separated other codes. New coding structures were developed, and new categories were created.
A within-case analysis was held to gain familiarity with the data (Eisenhardt, 1989). As there is little available expertise in this field of research, conducting a content analysis served as a suitable analysis method in order to report common issues mentioned in the data (Green & Thorogood, 2004; Vaismoradim et al., 2013). Content analysis includes both the coding and
24
classifying of the data (Neuendorf, 2002). The purpose of content analysis is to determine the characteristics of a documents’ content and to highlight any important messages, features and/or findings (Neuendorf, 2002).
25
5. Research Results
In this chapter the results will be explained based on the codes, themes and patterns that have emerged from the data analysis process. Solely, the results that are of influence in answering the research question will be covered in this particular section. However, the full transcribed interviews are stored in a secure database and are available upon request. The findings have been categorized in four predefined categories, namely: perceived usefulness, perceived ease of use, perceived risk and perceived trust. As discussed in the methodological section of this research, the ordinal data will be treated as if the values on the scale are approximately evenly spaced and sequential numeric values are assigned to the answers given by the respondents. The answers of the respondents have been categorized in a 5-point Likert scale, ranging from ‘very likely’ to ‘very unlikely’. ‘Very likely’ and ‘likely’ are categorized as positive and the number 1 is assigned as a numeric value. ‘Unlikely’ and ‘very unlikely’ are categorized as negative and the number -1 is assigned to those attitudes. When the attitude was ‘neutral’, the numeric value of 0 is assigned to the given answers. See table 3 for a visualization of the attitudes of the respondents and the assigned numeric value. With the use of content analysis, the researcher can measure the frequency of different themes and categories and analyze the data qualitatively and simultaneously quantify the data (Gbrich, 2007; Vasimoradim et al., 2013).
26 Table 3: Data results
27
As the sample consisted specifically of companies which did not already work with blockchain technology, the prior level of knowledge of blockchain technology was limited. Most respondents were only familiar with cryptocurrencies; others were more advanced in their knowledge and were familiar with terms as ledgers, decentralized database and smart contracts. In fact, two of the participants (R5, R9) had recently been contacted by consultancy firms who would like to introduce their own blockchain technology platform. However, R9 experienced the existence of various forms of blockchain platforms used by various consultancy firms as a proliferation of the technology. He argued:
“I strongly believe this technology has great potential, and I can imagine many firms see it as a new business model. However, we now experience proliferation on different platforms from different firms of the same technology. This results in even more insecurity and ambiguity for the end user”.
5.1 Perceived Trust
The purpose of the first four propositions is to test the effect of perceived trust on the variables intention to use new technology, perceived ease of use, perceived usefulness and perceived risk.
Proposition 0. Higher perceived trust in blockchain technology results in higher intention to
use new technology.
Overall, the respondents experience the technology as secure and trustworthy; they believe that this technology could truly change the food supply chain. It seems that there is a slightly positive relation between perceived trust and the intention to use the technology, though it is not a strong relationship. More than once, respondents mentioned that too little understanding and obscurity of the technology are factors that withhold them from the intention to use this new technology. R5 believes in the revolutionary power of this new technology, however, he states:
28
“I believe this technology could revolutionize our industry, however, to me it is still pretty vague what the impact will be and how this could be implemented in our industry”.
R2 and R8 both fall in the category ‘negative intention’, to use this new technology. However, they both do perceive trust in the new technology and think of it as secure and trustworthy. An explanation could be that R2 is currently in a take-over and therefore biased in their intention to use a new technology. However, there is no strong evidence to either reject or accept this proposition since no strong relationship between the two variables was found.
Proposition 1. Higher perceived trust in blockchain technology results in higher perceived
ease of use.
There is a mixed result in the variable perceived ease of use. Within the case data there is a positive relationship between perceived trust and common language. However, there is a negative relation between perceived trust and ease of learning. R10 also had a negative relation for both factors. For R5, R8 and R9 there was a positive relationship between the two variables. Interestingly, these three respondents already work with pretty advanced software packages in their business, which could have an effect of the level of perceived ease of use towards the blockchain technology. Proposition 1 could only be partly accepted; higher perceived trust in the blockchain technology results in higher perceived ease of use regarding common language. However, there seems to be no impact of the level of trust in the technology and the level of ease of learning.
Proposition 2. Higher perceived trust in blockchain technology results in higher perceived
usefulness.
From the case data, one could say there is enough support to accept proposition 2. The level of trust, indicated with six and eight unit points, both increased with two to eight points ten for
29
perceived usefulness. Only R10 was not convinced it could improve their own performance in the supply chain. He argued:
“As a machine producer, we manufacture machines for the food supply chain. For us implementing blockchain technology would not have that much of an impact. However, I do believe it could enhance the performance of our slaughterhouses. As a supplier of machines, an operating software could provide additional and valuable incoming and outgoing information. With blockchain technology, we might better help our customers’ product and eventually enhance their performance within the food supply chain”.
Although R10 is not convinced it would enhance their own performance, from the additional data it could be assumed that the respondent believes the overall performance of the food supply chain would increase.
In addition, R9 mainly emphasized product performance instead of supply chain performance and indicates that due to the trust in this technology the performance of the product would increase simultaneously. He argued the following:
“With this technology we can ensure our customers that the organic meat they buy is truly organic and not simply a more expensive version of the same product”.
Therefore proposition 2 will be accepted as true.
Proposition 3. Higher perceived trust in blockchain technology results in lower perceived risk. From the data in table 3, we can assume that there is a negative relationship between perceived trust in blockchain and risk of business failure and performance risk. However, no evidence was found that higher perceived trust results in lower privacy risks. As already was emphasized in the previous propositions, respondents’ do have a positive attitude towards trust in this technology, Nonetheless, some do experience a high level of insecurity when it comes to privacy and sharing their data. R4 argues:
30
“If there is a malicious person somehow has access to your data in the blockchain and is able to unpack and decode that data, it could become a very unpleasant situation. I still need to get convinced that this part will be well covered as hackers are always smart enough to get around some safety issues”.
This proposition will be partly accepted and perceived as true; trust in blockchain technology results in lower levels of perceived risk of business failure and performance risk. However, it is rejected and perceived as false for the factor privacy risk.
5.2 Perceived Ease of Use
The purpose of the next two propositions is to test the effects of perceived ease of use on the variables perceived usefulness and intention to use new technology.
Proposition 4. Higher perceived ease of use results in higher perceived usefulness.
As discussed in proposition 1, there is no ambiguous result within the variable perceived ease of use. However, when the factors are considered separately, the results indicate that there is a positive relationship between common language and perceived usefulness. R4 believes that this technology could increase the common language within the supply chain and argues it has a positive effect on the performance and record integrity. He argues:
“We work with many international partners, which all have different currencies. Because of this variety, our systems are not able to communicate with each other and do not communicate in a common language. This sometimes results in exchange rate differences on invoice of the customers since the systems did not have the same currency exchange rate on date X and time Y”.
R4 thinks that with the use of blockchain technology, a lot of time would be saved, which would increase their overall performance. In this case, the proposition will be partly accepted, as higher perceived ease of use regarding common language results in higher perceived usefulness. Unfortunately, there is not enough strong evidence to accept that ease of learning
31
will result in higher perceived usefulness; therefore, it will not be accepted as true.
Proposition 5. Higher perceived ease of use influences the intention to use a new technology Most respondents do not receive a high feeling of ease of learning, the underlying concept of the technology was found complicated and hard to understand. Due to this belief, there is a negative relation between the perceived ease of use (ease of learning) and the intention to use the new technology. Therefore, this proposition cannot be accepted; however, from the case data, one could assume that if the level of ease of learning would increase, the intention to use this new technology would increase simultaneously. R1 argues:
“I think it will cost a lot of time and effort before it would work sufficiently and everyone in the supply chain would implement it in their business. There are plenty of systems that store data. From past experiences, we learned that it simply takes a lot of time and effort before new systems are integrated and work as they are supposed to work”.
In addition, she argued:
“Eventually, when this technology has reached a higher level of maturity and more research has been done, I would say that the intention to implement such a technology would increase simultaneously”.
5.3 Perceived Usefulness
Proposition 6. Higher perceived usefulness results in higher intention to use new technology. This research found a positive attitude from all respondents towards perceived usefulness. When analyzing the data, the researcher found a positive relationship towards the intention to use a new technology. Overall, all respondents were convinced that blockchain technology could enhance their supply chain performance and improve the record integrity. Since performance is a rather broad concept, a variation of sub codes were used among the interviewees. The main themes that emerged from the data were reduction of paperwork, data
32
sharing, honesty, traceability, transparency, efficiency and trust. Each respondent envisioned different possibilities for their supply chain; they had different beliefs on ways it could increase their performance. They had different reasons for why they wanted to implement the technology. Therefore, the proposition will be accepted as true.
5.4 Perceived Risk
Proposition 7. Lower perceived risk results in higher intention to use new technology.
As discussed earlier, a negative impact is found for two out of three factors within perceived risk. Unfortunately, there is no clear relationship found within the factor privacy risk, and there is no direct impact on the intention to use new technology. However, when business failure and performance risk is considered, a relationship is found between lower perceived risk and higher intention to use new technology. Most respondents do not believe that this will lead to a business failure. However, R4 argues:
“I think arranging the investment could be a hurdle. Almost all new kinds of technologies get stuck on the part that it hasn’t proved its value yet, and moreover, it’s hard to measure the payback period of this new technology”.
However, seven out of ten respondents were convinced that this will not lead to business failure in the end; and three out of these seven believe that the customer will demand the use of the technology or that the technology will be imposed by the government or NGO’s in the future. R3 stated:
“In the end, I think it will be imposed by influential parties like “Wakkerdier” (a Dutch NGO fighting for animal welfare). Wakkerdier has the power to say: we don’t want those broiler chickens anymore, and if you keep selling them, we will publicly attack you. I think this will also happen when implementing blockchain technology. Companies will be forced to implement the technology and otherwise it will be a name and shame kind of approach”.
33
He also argued that besides NGO’s, the society will demand it and it will be necessary to obey and respond towards those demands.
R5 argues:
“We see an increasing trend in what consumers demand. They desire more transparency, better controllability and verifiability. We think that these demands will only increase. There is the demand for a better solution and more efficiency in this process, and with the great potentials of the blockchain technology, I consider the chance of a business failure relatively small.”
34
6. Discussion and Conclusion
This research has taken the first steps towards understanding the use of blockchain technology in the food industry. The goal of this research was to determine the success factors of blockchain implementation in the food industry. The former TAM model of Davis (1989) was used to study theoretical technology acceptance factors applied to the adopt blockchain technology. In addition, perceived risk and perceived trust were added to the model, based on the characteristics of the blockchain technology. The findings of this research support some of the propositions defined to measure the attitude of food supply chain managers towards the intention to use blockchain technology.
The expected improvement in performance and record integrity seem to have a positive impact on the intention to use blockchain technology. It is perceived as an application that could reduce the paperwork and increase levels of efficiency, traceability and transparency. These findings support prior research that the perceived usefulness determinant is a primary determinant of technology adoption (Davis, 1989).
For the other primary determinant, the perceived ease of use, there was no consistent evidence that higher levels of perceived ease of use will lead to higher intention to adopt this technology. Perceived ease of use was divided in ease of learning and common language. The attitude towards common language was fairly positive. However, the majority was not convinced that this technology is easy to implement and argued that it is still in its infancy phase. The major constraints respondents perceived was the complication to get the whole supply chain on board, which is critical to make the technology implementation a success.
One of the variables added in this research was the level of perceived risk. Perceived risk was divided in risk of business failure, performance risk and privacy risk. Although, blockchain ensures better provision of privacy risk control (Yue et al., 2016), respondent argued to feel high levels of privacy risk. They argue that this technology is still in its infancy phase and hackers often seem to be able to outsmart the most sophisticated technologies and
35
hack them. The results of this research support previous findings by Featherman and Pavlou (2003) that higher levels of privacy risk can be seen as a cause for concern and reduce the level of system adoption. However, overall, respondents perceived low levels of uncertainty regarding risk of business failure and performance risk.
Perceived trust was the second variable added to this research. Overall, the respondents feel a positive attitude towards trust in the technology. The believe the system has already proven its value and it is perceived as secure and trustworthy. However, little knowledge and obscurity of the technology seem to be factors that could withhold them from the intention to use the technology.
To conclude, overall there seems to be a positive attitude towards the intention to use blockchain technology. The respondents perceive high levels of usefulness; however, they also perceive a high level of uncertainty regarding privacy risk and ease of learning, which has a negative impact on the final intention to use the blockchain technology. A determinant which was not included in the model but was mentioned several times by the respondents is social influence and government regulation. Even though it could be the case that the industry has not yet reached a certain point of development and implementation, some of the respondents argued they believe that in the future the consumer will simply demand these kinds of technologies. This thinking is in line with the argumentation of Aung and Chang (2014). They believe that based on consumer demand, in the future there will be government regulations which force the food supply chain to improve the traceability and visibility of products. Maybe the industry is not yet ready to adopt this technology, since it is still in its infancy phase and there is much insecurity revolving this relatively new technology. However, in the future there might be no other possibilities then to use the technology.
36
7. Limitation and Future Research
When critically considering the results of this research, there were various limitations that cannot be ignored. In this section also, suggestions for future research will be addressed and discussed, in consideration of the limitations of this research.
To a certain extent, this study was limited due to a lack of earlier research done on the topic. Making use of other research papers and citing them is, of course, a process which helps researchers to first of all build a foundation to understand the research problem. As blockchain technology is a rather new phenomenon, there was little prior research done that could be used to develop the theoretical framework, which as a result, make it more difficult to establish a chain of evidence.
Moreover, given the small sample size, caution must be applied when interpreting the results of the data that was obtained. Ten interviews were conducted, which made it difficult to identify significant relationships from the data. Future research could gather a larger sample size of data which could then be statistically tested in order to further test how representative hypotheses are for a larger group and/or industry. In addition, qualitative data was collected through interviews, and evidence to support or reject a proposition was based on an individual’s attitude towards certain variables.
Another limitation of this research is that the researcher plays an important role in the research done and this could have biased the findings.This limitation had been reduced to a certain extent through the use of the literature review to support the analyses and also, through the use of additional quotes from the interview transcripts.
To conclude, blockchain technology has high potential for many industries and there are great possibilities in the future enabled because of its introduction to the market. This research focused merely on the food industry, which made the results hard to generalize for other industries. Another possibility for future research could be to consider the next phase of the food supply chain and the end-consumer; it could focus on testing the attitude of the
37
consumer towards the technology. It could question whether the consumer ultimately wants the ability to track the full history of the product he or she bought.
