i STOCKHOLM, SWEDEN 2020
Supply Chain Traceability
A framework for a future traceability system in
the electrification industry
WILLIAM BERGSTRÖM, FLORIAN HAAS
KTH ROYAL INSTITUTE OF TECHNOLOGY
SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT
Production Engineering and Management
Supply Chain Traceability
A framework for a future traceability system in the
electrification industry
William Bergström, Florian Haas
We hereby confirm that we have written the accompanying thesis by ourselves, without contributions from any sources other than those cited in the text and acknowledgements. This applies also to all graphics, drawings, tables and images included in the thesis.
Stockholm, Sweden
______________________________
Signature
ii
Society's demand for supply chain traceability is increasing. The need for traceability has been prevalent in food and pharmaceuticals for quite some time. However, the demand for traceability is now spreading to other industries, for instance, the electrification industry. This thesis aims to establish a traceability system framework for an electrification company that strives to have sustainable products of a specific origin. The framework must satisfy both the company's mission and existing legislation. Furthermore, both sustainability and having a supply chain of a specific origin is something that the food and pharmaceutical industries have used traceability systems to verify. The electrification industry could implement traceability systems similar to those in other industries. Therefore, the suggested traceability system framework is based on a literature review from other industries. The framework is also based on stakeholder analysis and a value tree analysis. The result is a framework presented through data flow and entity-relationship models. The suggested traceability system tracks data from customer RFQ and the most upstream sub-supplier to the end-of-life of the product. The result is a system that can verify the company's sustainability, quality, and the specific origin of the supply chain while complying with existing legislation.
iii
Samhällets efterfrågan av spårbarhet inom värdekedjor ökar. Behovet för spårbarhet har varit stort inom livsmedels- och läkemedelsindustrin under en längre tid. Nu börjar spårbarhet även bli relevant för andra industrier, till exempel, elektrifieringsindustrin. Uppsatsen syftar till att etablera ett ramverk för ett spårbarhetssystem för ett elektrifieringsbolag som strävar mot att leverera hållbara europeiska produkter. Ramverket måste tillfredsställa både bolagets affärsmodell och den lagstiftning som finns för spårbarhet som angår bolaget. Ytterligare, både hållbarhet och att ha en värdekedja med aktörer från ett visst område, är något som livsmedels- och läkemedelsindustrin har använt spårbarhet för att verifiera. Därför bör elektrifieringsindustrin anamma de lärdomar inom spårbarhet som finns i andra industrier.
Således är spårbarhetssystemets ramverk baserat på litteraturstudier från andra industrier.
Ramverket är också ett resultat av en intressentanalys och en multi-kriteriers beslutsmetod.
Resultatet är ett ramverk som presenteras genom en dataflödesmodell och en enhetsrelationsmodell. Det föreslagna spårbarhetssystemet spårar data från kundens offertfråga och den första underleverantören i kedjan tills slutet av produktens liv. Detta resulterar i ett system som kan försäkra bolagets hållbarhet, kvalitet, och dess värdekedjas specifika ursprung, samt följa existerande lagstiftning.
iv
First of all, we would like to thank the company; nothing is nobler for an engineer than contributing to powering the world sustainably. An opportunity which you gave us, and we have approached with utmost dignity. Similarly, you have treated us with respect and cared about our project considerably. We are especially thankful for our supervisor's brilliant input about all things related to traceability. We hope that our work will contribute significantly to the company and help us all to a greener future.
Secondly, we would like to thank Per Johansson, our supervisor at KTH, for his great support in this project. More importantly, Per has been a crucial part of ensuring that these two years have been great. Per's care for the well-being and success of the people around him is truly admirable and inspiring. We are grateful to have called you our supervisor and continue to be grateful to call you our friend.
Lastly, we would like to thank our classmates, friends, siblings, cousins, parents, and grandparents. This Covid-19 period would have been dreadful without the love and care of those who love us the most. You contributed to our clarity and peacefulness, which shaped this work, for that, and so much else, we are eternally grateful.
v
List of abbreviations ... vi
1. Introduction ... 1
1.1 Background ... 1
1.2 Description of the company ... 3
1.3 Problem statement ... 3
1.4 Scope and delimitations ... 3
2. Methodology ... 4
2.1 Value tree analysis ... 4
2.2 Data Flow ... 6
2.3 Entity-Relationship model ... 7
2.4 Model verification ... 9
3. Traceability in industries and literature ... 11
3.1 Traceability definition ... 11
3.2 Traceability in various industries ... 11
3.3 Applicability of traceability in other industries and literature ... 20
4. Application of the methodology: specification for the company’s traceability system .. 23
4.1 Stakeholder analysis ... 23
4.2 Value tree analysis: Specification of objectives ... 24
4.3 Data flow model for the traceability system – from customer RFQ to dispensed product ... 24
4.4 ER model – a framework to the traceability system ... 29
5. Model verification ... 31
6. Discussion: Applicability of results ... 35
6.1 Brand image ... 35
6.2 Financial traceability ... 36
6.3 Legal control ... 37
6.4 Product functionality and quality ... 37
7. Conclusion ... 40
References ... 41
Appendix ... 45
Appendix I: Value tree: Brand image ... 45
Appendix II: Value tree: Product functionality ... 46
Appendix III: Value tree: Financial traceability ... 47
Appendix IV: Value tree: Legal control ... 48
Appendix V: Pivot table resulting from the value tree analysis ... 49
Appendix VI: Data flow for the company’s traceability system ... 50
Appendix VII: Upstream supply chain data flow ... 51
Appendix VIII: Simplified upstream supply chain data flow ... 52
Appendix IX: Logical ER model of the traceability system ... 53
vi
List of abbreviations
AQL Acceptable quality limit
AS After-sales
BOM Bill of material B2B Business to Business
CO Customer order
DF Data flow
DFD Design flow diagram
EOL End-of-Line
ER Entity relationship
NDA Non-disclosure agreement
OT Off the shelf
PCC Production capacity check PLM Product Lifecycle Management
PO Purchasing order
RFQ Request for quotation
SC Supply chain
V&V Verification and validation VAT Value added tax
VTA Value tree analysis
1
1. Introduction
1.1 Background
Consumers increasingly want to know what they are buying. While previously, it was enough for a product to be safe and of high-quality, today's consumers are concerned about more than that. Consumers have realized that possibly the most significant effect that they have on the world is rooted in their purchasing power. Therefore, each consumer’s core values become relevant for their consumption patterns and thus for producers. As a result, today's successful products need to be safe, high-quality, and sustainable (Marucheck et al., 2011; Papetti et al., 2019).
However, most environmental impacts of products are not observable for consumers (Styles et al., 2012). Additionally, today's global supply chains are so complex that companies struggle to keep their supply chains free of criminal activity and worker exploitation (Venkatesh et al., 2020). Companies struggle to keep their supply chains sustainable, and they are often unaware of sustainability issues in their supply chains (Venkatesh et al., 2020). If companies cannot identify sustainability issues, then they cannot convey them to the consumer. As a result, it is challenging for a consumer to make an educated decision on what to buy.
Furthermore, today's long and complicated supply chains create a significant risk for product safety as a larger number of actors are in contact with the product before it reaches the customer. Not only are safety issues caused by manufacturing errors, but some arise due to improper handling or storage of the product (Marucheck et al., 2011). Product safety issues can lead to incredibly expensive lawsuits and recalls for a manufacturer, even if the issue itself stems from one of their sub-suppliers. For instance, Toyota had to do recalls for 9 million cars worldwide in 2009 (Marucheck et al., 2011). Another example is Johnson & Johnson's Tylenol recall in 1982, which involved 31 million bottles of Tylenol and set them back 260 million in today's U.S. dollars (Kost, 2019). Therefore, it is in every manufacturing company's interest to minimize the risk and cost of recalls.
Both sustainability and product safety issues have a crucial thing in common: a need for tracing what happens to products during production, which is referred to as traceability. Moreover, the need spans across the entire supply chain, even sub-suppliers' processes and sustainability need to be monitored (Jansen-Vullers et al., 2003; Papetti et al., 2019). The tracking needs to be detailed; complete traceability as to what components go into the product is necessary to verify its safety and ease a potential recall process (Jansen-Vullers et al., 2003). Furthermore, the actual production operation's parameters and the capacity unit in which they are performed should be tracked (Jansen-Vullers et al., 2003). All of this data needs to be saved and accessible so the company can fully benefit from traceability in terms of product safety, cheaper recalls, and quality improvements (Jansen-Vullers et al., 2003). The system responsible for gathering, storing, and distributing such data is called a traceability system.
Furthermore, a traceability system can also be used to track sustainability aspects in a supply chain. Tracing sustainability entails, for instance, relating production to its environmental impact and tracking the working conditions at every level of the chain (Papetti et al., 2019;
2
Venkatesh et al., 2020). Since a product is the result of actions taken during its supply chain, it is essential to track activities throughout the supply chain (Germani et al., 2015). Only by tracking the entire supply chain can the product be verified to be sustainable. However, it is in actors' interest to portray their products as more sustainable than they are. Furthermore, for components with high brand-value, there will be an incentive to supply the downstream with counterfeits (Marucheck et al., 2011). Thus, it is a challenge to maintain authenticity in the traceability system.
Despite the challenges, some traceability systems that have been implemented have resulted in resounding success. Currently, traceability systems have a profound role in ensuring food safety (Alfaro and Rábade, 2009; Marucheck et al., 2011; Saak, 2016). Traceability systems are used in the food industry to track the storage and handling of the products since food can quickly become spoilt or contaminated. Moreover, traceability systems are utilized to facilitate recalls in case of a discovered contamination.
Other industries are a bit behind in traceability compared to the food industry. Mainly since the food industry has experienced legislative requirements on food to be traceable in the European Union (Saak, 2016); however, other industries are catching up to the food industry.
Some industries are starting to implement traceability systems because they view traceability as a means for differentiation (Alfaro and Rábade, 2009; Marucheck et al., 2011). Additionally, some industries must adapt to certificate standards and upcoming legislation, for example, about traceability for tin, tantalum, tungsten, and gold (3TG).
The electrification industry is one of many industries concerned with 3TG traceability requirements. Additionally, since the electrification industry is rather new, it has much to prove in quality compared to older fossil-fuel types of power storage. Thus, continuous quality improvements are also of utmost importance. Furthermore, many actors are switching to battery solutions for sustainability reasons; thus, the batteries' sustainability must be verified.
Therefore, traceability systems have an essential role in the electrification industry. Both to allow the electrification industry to adhere to legislation, but also to differentiate battery solutions from older solutions.
However, there is not much research on traceability in the electrification industry. Instead, it is up to electrification actors to identify their industry's most effective traceability system design. This thesis means to formulate the framework for a traceability system in an electrification company ("The company"), using existing theory from other industries and interviews of stakeholders within an electrification company. The specification will lay the foundation of a traceability system that will allow the company to comply with legislation and its legal obligations. Additionally, the traceability system will allow the company to verify that what they are marketing is precisely what the customer is buying.
3
1.2 Description of the company
The company is on a mission to build sustainable electrification solutions. Additionally, the company intends to have a clean manufacturing process, setting a benchmark for not just the electrification industry but also manufacturing as a whole. Furthermore, the company aims to increase its sustainability by powering its manufacturing entirely on renewable energy, resource-efficient process design, embracing circular economy, and sustainable handling of products at their end-of-life. Finally, in addition to environmental sustainability, the company aims to have a socially sustainable supply chain through its due diligence and traceability strategy.
In conclusion, the company has made promises about sustainability in their marketing.
Additionally, they mention a traceability strategy to track their social sustainability. As a result, a traceability system is of importance for the company to be confident that they can deliver as advertised. Without a traceability system to verify the company's sustainability, they would be exposed to bad public relations and legal liability.
The company's supply chain consists of a large number of suppliers, and sub-suppliers, supplying components and sub-assemblies for manufacturing. Additionally, the company has a significant after-sales department and is focused on business to business (B2B) sales.
1.3 Problem statement
The project aims to define the specification for a traceability system which aptly traces the company's products and process from customer RFQ to the products end-of-life handling.
Moreover, the traceability system should trace the sustainability of the company's entire supply chain, including its sub-suppliers. The company must also be able to continuously improve its own, its suppliers, and its sub-suppliers manufacturing process through the traceability data. Furthermore, the system should track the cost-of-life and all repairs and replacements on a serial number level of the products.
1.4 Scope and delimitations
The framework for the traceability system only handles what data should be collected, where the data should be collected, and how the data is linked together throughout the chain.
Meaning, that the thesis does not handle the technology for gathering data, nor does it handle the system architecture on which the system should be built.
4
2. Methodology
Traceability systems are complex systems involving stakeholders across the whole supply chain and large sets of data. The report proposes several models that can be later used as a framework for the design and implementation of the traceability system. Different methods and modeling techniques that support the elaboration of these models will be covered. First, the concept of value tree analysis will be explained. Although value trees are decision-making methods, they will be particularly helpful in determining the requirements the traceability system needs to fulfill. Based on this problem-structuring, two main modeling techniques will be handled: Data Flows and Entity-Relationship models. Data flows (DF) depict the flow of information within a system. Entity-Relationship (ER) model is a common database-design method that demonstrates how information is linked together. Finally, the aspect of model verification will be explained.
2.1 Value tree analysis
A value tree analysis (VTA) is a multi-criteria decision-making tool that enables the evaluation of several design alternatives (Vaughn-Cooke and Kremer, 2011). The value tree represents the core of VTA and is a hierarchical structure built of objectives and attributes. Objectives are split into sub-objectives, creating a tree-like shape structure. The attributes are a measure of the fulfillment of the objectives (Varde et al., 2019). A decision is then made based on what alternative best fulfills all objectives. For the traceability system problematic, a slightly different approach is taken. The aim is not to find a best solution from different alternatives, therefore the use of attributes is not taken into consideration. Instead, through the value tree, it shall be determined what objectives and requirements should be considered.
A VTA analysis consists of different steps, as seen in the diagram below (P. Hämäläinen, 2002):
Figure 1: Phases of value tree analysis according to Hämäläinen
5
The phases of the problem structuring are detailed to give an overview of their purpose and how they relate to the framework of the traceability system.
2.1.1 Problem structuring
The first part of the VTA consists of addressing the problem and clearly defining the objectives to achieve through rounds of interviews (P. Hämäläinen, 2002). In this part of the analysis, the actual value tree is built. Structuring the problem represents the most significant part of the VTA, as it builds up the foundation for the DF and ER model later on. Preference elicitation, recommended decision, and sensitivity analysis are intentionally left out of the methodology.
Their aim relies on finding the best solution to a problem (P. Hämäläinen, 2002). The framework for the traceability system as such is not considered as an issue, where alternatives are identified and the best solution is chosen, but rather as a problem, which shall incorporate all necessary requirements to fulfill objectives.
2.1.1.1 Decision context
In the first stage of structuring the problem, the nature of the decision and the defining context are determined. In the case of the traceability system, the factors by which the design of the system is framed (e.g. economical, sustainable, political) are analyzed. As a result, the decision context helps in determining the main purpose of the traceability system and provides an overview of possible stakeholders. (P. Hämäläinen, 2002)
2.1.1.2 Identifying and generating the objectives
Once the decision context has been defined and the stakeholders identified, the objectives of the traceability system are gathered. Requirements for the traceability system, in terms of data to track, are discussed with stakeholders through an extensive round of interviews.
Stakeholders are invited separately to avoid conflicting requirements in group discussions, where the aim is to achieve the same level of understanding of traceability and generate a discussion on what is important for the system. The same structured approach is followed for each interview to gain as much input as possible from the stakeholders. Firstly, a definition of traceability is presented alongside the extent within the supply chain to which it could be applicable to the company. Secondly, the list of stakeholders is shown to give an overview of who is involved in such a system. Finally, some of the benefits discussed in previous interviews are presented and possible advantages and suggestions of relevant information for the stakeholder proposed. Based on that, a discussion is generated, and objectives and requirements written down. The list is then sent to the stakeholder and signed off.
6
2.1.1.3 Generating and identifying decision alternatives
At this stage of the problem-structuring, decision alternatives are supposed to be discussed together with the stakeholders (P. Hämäläinen, 2002). Since the round of interviews focusses on requirements, such decision alternatives are not presented. Instead the stakeholder’s requirements lists are signed-off, to agree on the data that should be tracked. However, the value tree in the next section will determine which of the requirements expressed by the stakeholders bring value to the traceability system objectives, and which can be excluded.
2.1.1.4 Creating a hierarchical model of the objectives
The requirements expressed by the stakeholders leads generally to a preliminary list of data to track. The information is unorganized and may not be complete at this point. Stakeholders express their requirements and objectives of the traceability system based on experience and without consideration of other stakeholders, which leads to the potential problem of not necessarily considering objectives on a broader scope. The aim now relies on creating a deeper understanding of the requirements of the traceability system and which objectives they fulfill (P. Hämäläinen, 2002). The actual value tree is built here by determining fundamental and lower-level objectives in a hierarchical model. The fundamental objective of “traceability in the supply chain” is completed with sub-objectives based on literature research. Several approaches exist to create such hierarchical structures. A distinction between top-down and bottom-up structure is made (P. Hämäläinen, 2002). In the top-down approach, a tree is built from the fundamental objective of achieving traceability in the supply chain, down to the lower-level objectives. The bottom-up approach follows the inverse construct. A top-down approach is chosen for the value tree, where the lowest layer represents the information to track. This method proves to be more flexible to identify the most necessary requirements for the given objectives.
2.1.1.5 Specifying the attributes
As previously mentioned, the attributes are a measure of how the objectives are fulfilled. In the case of the traceability system, no quantitative attributes can be given. Rather, the DF and ER model verify if the requirements specified guarantee traceability and therefore achieve the objectives. Instead of attributes, the data to track represents the lowest layer of the value tree.
2.2 Data Flow
The VTA generates an extensive list of requirements and data to track for the traceability system. However, it does not provide information on what stage of the supply chain the required information is gathered. Data flows are useful at this stage, as they show how the information flows through the traceability system. The various stages of the supply chain are depicted together with the data that is added each time. Data flow diagrams (DFD) are commonly used for the design of computer systems as they model the logical structure of the program (Sharp, 1992). However, they are common practice in any other application, from depicting production flow on the shop floor to activities within a business organization.
Standardized notations exist to model data flow models. They mainly consist of four elements:
flow arcs, processes, data stores and terminators (Sharp, 1992). The flow arcs represent the direction of the flow, while also providing information on what input goes into each process
7
and what output is produced (Dennis et al., 2019). Input can be distinguished in information going from one block to another or the movement of physical parts. Dotted arcs represent the flow of information, while full lines physical parts. The process transforms the input into an output. A data store represents the entity that actually stores the information (What is a Data Flow Diagram, no date). The terminator is an external entity that shows the source or destination of the information (Sharp, 1992). For practical reasons, it has been decided to slightly deviate from the presented standard notation. The traceability system does not process information but instead adds it together at each stage. Furthermore, the data store blocks are intentionally left out, since the purpose of the DFD is not to show where the information is gathered. This aspect will be explained later in the ER model. The following notation is introduced for the DFD:
Figure 2: Block notation for the data flow
Each block is similar to the terminator described previously and represents an activity or even a document. As seen in the image above, a terminator lists the information added, as well as the information it can be linked to in the previous blocks. In the picture above, the offer sent by the company lists data such as the quote number, customer order ID, and the offer lines.
The customer order ID is the attribute, linking it with the customer PO.
2.3 Entity-Relationship model
Traceability systems involve the gathering and structuring of large sets of data, where the key relies on maintaining an unbroken data chain so that all types of information can be retrieved.
This implies the creation of a connected structure of entities and links.
The backbone of such a structure can be determined by means of a conceptual model, which is commonly used for database modeling. These models aim at describing the semantics of software application at a high level of abstraction, and to create a common understanding and a way of communication between analysts and programmers (Pichler and Eder, 2011). The Entity-Relationship (ER) model, proposed by P. Chen, is a technique and foundation of conceptual models describing the real world in so-called “entities” and “relations” (Chen, 2011). A major advantage of ER-models relies on the creation of a unified view of the data (Pichler and Eder, 2011). A unified view is helpful since the architecture of the traceability
8
system and how it is programmed is yet to be determined. As a result, the ER model proposes a structure that defines the semantics of the traceability system by showing how the data is interconnected.
2.3.1 Characteristics of ER-models
ER models represent a system in entities and relationships as previously mentioned. Entities are parts/objects, such as products, persons or organizations (Chen, 2011). When talking about an instance, it is referred to as a particular and identifiable object of an entity. Relationships describe how the entities are linked to one another. Additional characteristics of ER models are attributes and cardinalities (Chen, 2011), where attributes define the properties of the entities. Relative to our previous example of a “person” being an entity, possible attributes could be age or height. Attributes take values in form of different data types such as numbers, dates, strings, lists or text. Considering our example, the age attribute would be given a number as a value. In some cases, attributes are given unique identifiers as a value. A unique identifier, is an identifier, which is ideally only used in one instance of an entity and is stable, meaning that it is not reassigned to another instance (McMurry et al., 2017). These characteristics make such instances of entities clearly differentiable. Unique identifiers within the traceability system are for instance serial numbers, PO numbers or customer IDs.
Cardinalities set the boundaries for the relationship between two entities by determining how many times an instance of an entity can be found within the other entity. This report uses mainly the Martin notation, also called Crow’s foot notation, which is commonly used in practice (Teorey et al., 2006) and the (Min, Max) notation for clarity purposes. The diagram below characterizes these notations:
Figure 3: Martin and (Min,Max) cardinality notation for ER models
The following example illustrates the characteristics of a conventional ER model with the crow’s foot notation of cardinalities:
Figure 4: Example of an ER model notation
9
The two entities here are “Department” and “Employee”. The relationship between them shows that an employee in an organization is assigned to a department. The cardinalities specify that an employee works in a “one and only one” department. On the other hand, a department employs “one to many” employees.
2.3.2 Conceptual, Logical and Physical data models
In the conception of the ER model for the traceability system, a two-phase approach has been adopted. First, a generic model (conceptual model) has been determined where the relationship between the entities of the system are represented. This enables an overview of the entities and their connection with cardinalities that need to be considered in the traceability system. This model has a descriptive purpose and thus assists in the discussion of data concepts with the stakeholders of the system (Simsion and Witt, 2005). Based on this first model, a second model (logical model) has been created, where the attributes of the entities are represented in tables. Here, a reduction of abstraction and a higher granularity is characteristic, and the logical structure of the conceptual model is mapped (Simsion and Witt, 2005). The physical model offers an extension to the logical model and describes the actual implementation of the database (Simsion and Witt, 2005). Here, for instance, the data type of the attributes is specified. Since the scope of the report does not include the architecture of the traceability system, the physical model is not proposed.
Figure 5: Abstraction levels of ER models
2.4 Model verification
The ER model depicts the real world through entities and relationships. It suggests a framework for the traceability system by connecting the items that are involved in the system. To determine how well the model represents the desired supply chain traceability system, model verification and validation (V&V) is necessary. It is important to mention the distinction between a ‘model’ and ‘system’ V&V. The following sections mainly describe a model V&V since the actual traceability system is not built.
Model verification is the process of confirming whether the specified requirements are satisfied by the model. Model validation is the process of validating whether the model truly represents the real system (Carson, 2002). The traceability model is therefore verified if all requirements specified in the value tree and data flow are covered. For the validation, it shall be proven that the traceability system depicts the information from the real world. This report focuses only on model verification.
10
Verification is an iterative process, where the re-design of the model is regularly required to fit the requirements. Hence, errors should be identified and corrected during this process.
Modeling errors can be categorized as follows (Carson, 2002):
• Project Management errors: Communication errors between stakeholders and analysts (interviewers).
• Data & Data Model errors: Errors in input data.
• Logical Model errors: Errors in the made assumptions and specified requirements and implementation in the model. These errors are closely linked to project management miscommunications.
• Experimental errors: Errors due to poor experimental design.
For the verification of the traceability model, mostly logical errors will be examined. Data model errors refer to wrong input data, which would be for instance the wrong naming of an attribute or entity. It was our intention, however, not to reveal the true wording of every instance to maintain confidentiality. The verification will also not focus on project management errors since a round of interviews with all the relevant stakeholders of the traceability system was carried out and signoffs of their requirements gathered. The verification will mostly cover logical model errors and ensure that the requirements are fulfilled through a well-defined logical structure of the ER-model.
A PLM software is used as a support tool to identify and correct the ER model on logical errors, and hence verifying it in regard to the requirements. The software enables to build the ER model as a database with instances of entities and relationships. Features of the PLM software can then show if an unbroken data chain is guaranteed. The implementation of the ER model in the software should not be mistaken with the actual implementation of the traceability system. Instead, it should only be considered as a tool for verification.
11
3. Traceability in industries and literature
3.1 Traceability definition
Traceability is defined in the ISO 9000 as the "ability to trace the history, application or location of an object" (2015). Additionally, the ISO 9000 notes that traceability, with regard to products or services, can involve the "the origin of materials and parts; the processing history; the distribution and location of the product or service after delivery." Traceability has also been defined as accurate accounting in manufacturing. Furthermore, traceability is defined as "The registering and tracking of parts, processes, and materials used in production, by lot or serial number" by the American Production and Inventory Control Society.
This thesis defines traceability as the registering and tracking of the history, application and location of orders, materials, components, sub-assemblies, and products by article, batch, or unique identification numbers. As a result, the traceability system should completely cover the company's operations from customer RFQ to the disposal of the product at its end-of-life. As such, individual orders need to be linked to instances of products. Moreover, the products need to be linked to deliveries and their ingoing material. Tracking ingoing material means keeping track of article numbers, batch numbers or serial numbers, depending on the criticality of the material, for all material in the product. Additionally, the organizations that produce or order goods are viewed as a vital part of the orders or goods' history.
3.2 Traceability in various industries
Currently, traceability systems are most prevalent in food and pharmaceutical supply chains.
Since pharmaceuticals and food got a head start in the use of traceability systems, there is a massive amount of knowledge to be adopted from the two industries. Additionally, traceability is increasingly being used in the manufacturing industry. Therefore, studying food, pharmaceutical, and existing manufacturing industries is relevant to the company’s manufacturing supply chain.
3.2.1 Traceability in food supply chains
Food is a unique product category as it is one of the most fundamental necessities for people, and it is consumed in the most literal sense. Additionally, food is something that can spoil or become contaminated and is then useless or harmful to the consumer. Ensuring that food is safe when it reaches the customer has become increasingly difficult, due to the increased complexity of food supply chains (Piramuthu et al., 2013). The increased complexity has led to some high-profile product safety issues that have severely influenced the public’s confidence in companies’ and government’s ability to assure food safety (Marucheck et al., 2011). For example, the salmonella outbreak in peanut butter paste in 2009, which resulted in the most massive food recall in US history, ended up with the company’s owner going to prison for 28 years, and the company becoming defunct (Marucheck et al., 2011; Basu, 2015). In order to ensure food safety, traceability systems are frequently used (Alfaro and Rábade, 2009; Feng et al., 2020).
12
For European food producers, traceability is not a choice. Instead, it is a requirement, since 2002, to maintain traceability in food supply chains to ensure food safety (European Parliament and Council, 2002). Additionally, food safety incidents carry a considerable cost for its producers in terms of penalizing fines and brand damage (Marucheck et al., 2011; Rueda et al., 2017). It can easily be understood that people are less likely to eat peanut butter from a brand whose products have resulted in other people’s death. Therefore, there is a clear incentive for manufacturers to minimize the impact of food safety incidents (Marucheck et al., 2011). Due to legislation for traceability and obvious risks of not having it, the food industry was early in its implementation of traceability systems. Thus, the food industry is at the forefront of traceability.
Marucheck et al. identified three core vulnerabilities of food supply chains (2011). First, most food products are perishable and become spoilt with time and poor handling. Secondly, non- local food supply chains can be long and complex, meaning that they are exposed to significant risks. Third, food is very susceptible to adulteration and can even be a target for terrorists. All of the three core vulnerabilities are, in some way, protected by a solid traceability system.
To keep food safe, one should trace its movement throughout the supply chain (Marucheck et al., 2011). One distinctive trait of food is that it is more sensitive to quality degradation over time than most products (Piramuthu et al., 2013). Because food is perishable, but customers still expect the same quality, it is essential to continually track supply chain information to identify degradation at any level (Gautam et al., 2017). By tracking how much time, in what packaging, and in what condition, food is stored at different levels of the chain, one can see where food might perish before reaching the consumer (Piramuthu et al., 2013; Gautam et al., 2017). Also, by identifying the perished products, and having traceability data, it is possible to see the key factors that cause the product to perish (Gautam et al., 2017). Thus, the supply chain can optimize costs based on the critical factors for keeping the product safe (Alfaro and Rábade, 2009). The traceability data allows the supply chain to evaluate investment decisions based on quantified data (Alfaro and Rábade, 2009).
Additionally, it is critical to record what goes into each product because of the complexity of food supply chains. Moreover, tracking the lots of food that go into each product eases the recall process if one lot is identified as contaminated (Jansen-Vullers et al., 2003; Piramuthu et al., 2013). However, one major problem with food supply chains is that there is typically much mixing of batches (Piramuthu et al., 2013). If a contaminated batch is mixed with other batches, then those batches can quickly become contaminated. Thus, one key aspect of food traceability is the process in which batches are handled. For instance, all bins must be thoroughly cleaned out before handling a new batch (Piramuthu et al., 2013). If and only if food suppliers can keep batches separate, they will be able to efficiently trace goods in the supply chain.
Furthermore, an agro-food supplier’s raw material can be dispersed across many products (Piramuthu et al., 2013). Even if only small amounts of the raw material are dispersed into a product, that small amount may still be enough to contaminate the entire product.
Additionally, some products contain many ingredients, all of which have the potential to contaminate the product. Thus, it is difficult to identify the source of contamination (Piramuthu et al., 2013). Typically, to be on the safe side, without an identified source of contamination,
13
the supply chain has to be temporarily shut down (Piramuthu et al., 2013). Therefore, it is of utmost importance that contamination sources can be found quickly and accurately.
Furthermore, identifying the source of contamination has enormous implications for the actor liable for damages and recalls (Piramuthu et al., 2013). When it is unclear where contamination comes from, the entire chain shares the cost. However, when it is possible to pinpoint the contamination to one actor, that actor is responsible for the cost (Piramuthu et al., 2013).
Therefore, precise traceability will not be as sought-after by every level of the supply chain.
Actors who have higher costs for quality improvements and traceability than recalls and brand damage spread across the chain are less likely to push for traceability implementation.
However, downstream actors have a great incentive to push for traceability throughout the supply chain (Piramuthu et al., 2013). For example, Hudson Foods, a meat supplier for Burger King, had an E. coli contamination in 1997 but could not determine its origin. As a result, Hudson foods were held accountable for the entire outbreak and went bankrupt (Piramuthu et al., 2013). If Hudson Foods would have been able to identify a sub-supplier as the culprit, they would perhaps still be in business. Several technologies have been utilized in order to trace food products throughout their supply chain. Radiofrequency identification (RFID) tags are proven to be an effective way of tracking products (Piramuthu et al., 2013; Choi et al., 2015; Gautam et al., 2017). While traditionally, bar codes have consistently been used in the food industry, RFID is increasingly used since it is faster, more accurate, and can store more data (Gautam et al., 2017). However, due to the cost of RFID chips, they can, in some cases, only be economically sane if the supply chain traces at a lower level of granularity (Piramuthu et al., 2013). For example, since bar codes are practically a sticker, it can cheaply be attached to each unit of fruit, but for RFIDs, it might be more feasible to attach it to each box of fruit.
Box marking can be seen as the equivalent of RFID tracking on a batch level instead of bar code tracking at a serial level. However, while Piramuthu et al. argue for the use of RFID scanning, they also argue that going from tracing batch to serial has a higher return on investment than going from tracing article to batch. Alfaro and Rábade implemented a traceability system in a small canned vegetable company (FF) that recovered its initial investment in less than two years (2009). The traceability system consisted of bar codes and scanners, tracking product on a serial level. By tracing their products, FF was able to evaluate suppliers based on a price- quality ratio, optimize their inventory, and increase customers’ trust in the company (Alfaro and Rábade, 2009).
Blockchain is another technology that is seeing increased use in food supply chains. Essentially, blockchain offers a way to distribute immutable records of data (Song et al., 2019; Feng et al., 2020). The public has concerns regarding the transparency and integrity of traceability data in the food industry (Feng et al., 2020). One promising way to ease the concern is by implementing blockchain. Additionally, blockchain shows promise in complex supply chains, such as food supply chains, since it is managed by a group of computers rather than a centralized system (Song et al., 2019). As a result, blockchain-based traceability systems do not place the burden of management on one entity but instead the entire chain. The fact that blockchain systems are not managed by one actor also results in the customer deeming them as more trustworthy (Feng et al., 2020). Therefore, blockchain is a technology that could allow the food industry to redeem the trust in its product safety that it has lost (Marucheck et al., 2011; Feng et al., 2020).
14
The lack of trust for food supply chains is not solely in terms of product safety; the sustainability of the food industry is also under scrutiny. Most consumers are unable to track products’ point of origin and their journey from upstream to downstream (Feng et al., 2020). As a result, the consumer cannot be sure how their food has been produced. Additionally, consumers remain unaware of the environmental damage of their consumption (Styles et al., 2012). Consumer demand is one of the critical drivers of supply chain sustainability; however, consumers must know the sustainable impact of products to consume sustainably (Styles et al., 2012). As a result, if a product’s sustainability can be proven, that is an enormous competitive advantage for the product (Papetti et al., 2019). To prove a product’s sustainability, it is necessary to trace it. Environmental sustainability has been successfully tracked using “traditional” traceability system combined with resource consumption data (Papetti et al., 2019). Although sustainability also includes social sustainability, which has become challenging to maintain in today’s global supply chains (Venkatesh et al., 2020). Social sustainability information is something consumers increasingly expect sellers to disclose (Venkatesh et al., 2020). In agri- food products, it is crucial to trace back food to a farm level to verify sustainability (Feng et al., 2020). Once the traceability system identifies the farm, third-parties can determine its sustainability.
In conclusion, traceability is both a legislative requirement and a vital competitive advantage in the food industry. Extensive traceability is something consumers are willing to pay for because it ensures food safety, quality, and sustainability (Alfaro and Rábade, 2009; Piramuthu et al., 2013; Feng et al., 2020). In other words, traceability can verify that food, our most fundamental necessity, is safe, and it can also differentiate and verify the other demanded
“luxury” features of food.
3.2.2 Traceability in pharmaceutical supply chains
GS1 defines traceability in the healthcare industry as something that “enables you to see the movement of prescription drugs or medical devices across the supply chain” (Healthcare traceability and GS1 standards, no date). Additionally, by tracking products in the healthcare supply chain, it is possible to expose counterfeits, increase product safety, and make sure that the product is compliant with the burdensome regulations in the pharmaceutical industry.
Similarly, to food, pharmaceuticals are also consumed. Thus safety is of the essence. Moreover, pharmaceuticals are typically consumed by frail individuals, meaning that safety is imperative.
Contrary to the food industry, pharmaceuticals have a more significant counterfeit problem (Khezr et al., 2019; Hastig and Sodhi, 2020). The World Health Organization (WHO) estimates that “1 in 10 medical products in low- and middle-income countries is substandard or falsified”
(Substandard and falsified medical products, 2018). However, WHO also claims that drug counterfeiting is a global problem (Substandard and falsified medical products, 2018).
Additionally, WHO states that both generic and innovator medicines are subjects for counterfeiting, meaning both expensive cancer medicine and inexpensive pain relievers are counterfeited. For example, the United States had counterfeits of a formidable cancer medicine reach patients (Hastig and Sodhi, 2020). The counterfeits were a mix of wheat and a component of nail-varnish remover and were manufactured in Africa, moved through legit supply chains in Europe, and finally reached the United States (Hastig and Sodhi, 2020).
15
Additionally, counterfeits in healthcare are not limited to medicines; instruments also share the same issue. A compromised instrument can be perfectly safe and still be dangerous due to decreased efficiency. For example, a medical instrument used to diagnose a patient that fails to give an accurate diagnosis can result in the patient getting the wrong care (Marucheck et al., 2011). An incorrect diagnosis can both result in delayed proper medical treatment and unsuitable treatment. Meaning that even if the instrument itself does not hurt the patient, the patient’s care suffers (Marucheck et al., 2011).
Counterfeits hurt the healthcare industry, not just due to missed out profits, but also due to the impact it has on the industry’s reputation, due to the serious consequences they can have for patients (Khezr et al., 2019; Hastig and Sodhi, 2020). Not only is there an issue of final products being counterfeits, but there is also an issue with ingredients used in manufacturing being counterfeit (Hastig and Sodhi, 2020). Additionally, for medicines, it is difficult to track individual pills; instead, the bottle is typically tracked. Thus, counterfeits pills can enter the legitimate supply chains by infiltrating the tracked containers (Hastig and Sodhi, 2020).
Moreover, counterfeits that are proficient at copying the labels of legitimate goods can infiltrate legitimate distribution (Hastig and Sodhi, 2020). Thus, obtaining complete traceability of the healthcare supply chain is crucial.
GS1 has a standard called EPCIS for maintaining supply chain integrity that is meant to be applied to the healthcare industry. The standard requires supply chain participants to share information about four types of events: What, when, where, and why (Healthcare traceability and GS1 standards, no date). The standard tracks: the products impacted, timestamp of event, where the product is and where it has been, and why the event was observed (Healthcare traceability and GS1 standards, no date). The goal behind tracking these events is to be able to trace the product backward and track it forward. Tracing backward means being able to identify the transaction history and location of the product, from manufacturing to the consumer (Healthcare traceability and GS1 standards, no date). Additionally, tracking forwards means that one should be able to see the route that each product will take to the consumer (Healthcare traceability and GS1 standards, no date).
However, what is troublesome with the GS1’s standard is that it does not handle the input of raw material and ingredients during the manufacturing process. Thus, EPCIS is not sufficient for revealing counterfeit inputs in production, which was indicated as a key issue by Hastig &
Sodhi in 2019. However, tracing according to EPCIS seems to be sufficient to adhere to current legislation. Both the U.S.’s Drug Quality and Security Act and the EU’s Falsified Medicines Directive only demand traceability for the distribution of medicines (H.R.3204 - Drug Quality and Security Act, no date; Falsified medicines: overview, no date). Maintaining traceability of the distribution is enough to facilitate recalls and verify that the product is from the right manufacturer. Nevertheless, distribution traceability does not track the inputs of the manufacturing process. However, it should be noted that the production of pharmaceuticals and active pharmaceutical ingredients (APIs) is heavily regulated. In the U.S, the FDA has established good manufacturing practices (GMP) for APIs, and similar legislation exists in the EU, Canada, and many other countries (Guidance for Industry, Q7A Good Manufacturing Practice Guidance for Active Pharmaceutical Ingredients, 2001; Marucheck et al., 2011). GMP includes rigid quality inspections of raw materials (Guidance for Industry, Q7A Good Manufacturing Practice Guidance for Active Pharmaceutical Ingredients, 2001). Due to the life-
16
or-death seriousness of pharmaceuticals, it might be better to rely on incoming quality inspections rather than traceability to ensure medicines safety. Even the most rigid traceability system runs the risk of becoming infiltrated by malicious actors, while proper testing should allow the manufacturer to verify the quality of the raw materials at hand sufficiently.
Additionally, even if traceability would be a cheaper solution for quality verification, one can expect the pharmaceutical industry to be slow to adapt due to its rigorous regulation (Hastig and Sodhi, 2020). As a result, one might expect traceability to primarily act as a means of maintaining the distribution’s integrity in the healthcare industry.
In 2019, 60 % of pharmaceutical companies were evaluating blockchain as a means to increase their traceability (Khezr et al., 2019). If successful, the implementation of blockchain in healthcare companies’ distribution could guard the public against medical counterfeits. The result would be increased public confidence in healthcare, more accurate diagnosis, and the correct treatment reaching more patients. Successful traceability systems in healthcare would save lives.
3.2.3 Traceability in manufacturing supply chains
Manufacturing industries share food and healthcare’s urgent need for traceability (Wang, 2014). However, traceability has historically not been a legislative requirement in manufacturing, to the same extent as in the industries mentioned earlier. Nevertheless, traceability is necessary for manufacturing to ensure compliance with increasing government regulations, ensuring sustainability, protect against counterfeits, and making right and timely supply chain improvements (Wang, 2014; Choi et al., 2015; Abeyratne and Monfared, 2016;
Song et al., 2019). Additionally, traceability is also a means of certifying quality (Jansen-Vullers et al., 2003). Here certified is defined as some level of guarantee that the product meets its requirements (Jansen-Vullers et al., 2003).
The quality certification can be made after rigorous tracking of production processes and production tools. The tracked production data can be linked to the quality attributes of the output (Jansen-Vullers et al., 2003). Production data needs to be linked to an individual product or a batch, where the definition of a batch is a homogeneous group of products, in terms of production specification (Jansen-Vullers et al., 2003; Papetti et al., 2019). Moreover, being able to link production data to products enables new kinds of analytics for operations, but also for risk and sustainability (Abeyratne and Monfared, 2016). For instance, the traceability data can be combined with AI to improve processes and decision-making (Wang, 2014). Thus, traceability data and AI could advance the continuous improvement concept in manufacturing.
To continuously improve products and certify quality based on traceability data, all products, components, and processes in manufacturing must be tracked (Jansen-Vullers et al., 2003;
Wang, 2014). The result of such tracking is an as-assembled bill of materials (BOM) for each product instance. An as-assembled BOM needs to relate the batch numbers of the components to the parent’s identification number (Jansen-Vullers et al., 2003). The links between components and the parent form the product instance’s composition. Jansen-Vullers et al.
identified four elements of traceability that must be fulfilled, which has also been adopted by Marucheck et al. (2003; 2011). The four elements are physical batch integrity, data collection, product identification, and process linking, and reporting (Jansen-Vullers et al., 2003).
17
First, physical batch integrity refers to how well separateness between batches is maintained.
A shared batch number can identify batches, and each item of a batch should be processed using the same process parameters (Jansen-Vullers et al., 2003). The batch number of each product, or component, and its process parameters must be determined, registered, and locked at the manufacturer (Choi et al., 2015). Otherwise, the rest of the chain cannot trust the initial traceability information. Furthermore, establishing batch and serial numbers for products and components in production is crucial for manufacturing traceability. The American Production and Inventory Control Society (APICS) defined traceability as “the registering and tracking of parts, processes, and materials used in production, by lot or serial number” (Alfaro and Rábade, 2009).
Furthermore, maintaining physical batch integrity also includes determining the size of the batch and sticking to it. For critical components, manufacturers might have a “batch” size of one, which results in tracking on a serial number level. Nevertheless, components with serial numbers might have a batch number, to allow the producer to identify components with common characteristics. The batch number does not have to include every item with the same process parameters. If a manufacturer wants to maintain a level of secrecy around their batches, they might split the batch into several batch numbers. However, most customers of the manufacturer will want to be able to say that they do not want any items with the same process parameters as a batch that they have determined to be deficient. The manufacturer must then be able to link parts of the split batch numbers to the actual main batch. However, the manufacturer is not free to include any number of items in a batch, since all items identified by a batch number must have been produced through a homogeneous process (Jansen-Vullers et al., 2003). If the batch is not produced through a homogenous process, then the purpose of the batch is lost, and physical batch integrity is not maintained.
Second, traceability requires data collection of movements and transactions within the supply chain and the process data associated with each product and component (Jansen-Vullers et al., 2003). This data must be linked with each batch or serial number to form continuous data links throughout the supply chain. Data collection is possibly the costliest part of traceability as it takes valuable time to perform on the factory floor and has considerable technology costs (Venkatesh et al., 2020). Typically, traceability data collection in manufacturing relies on some tag technology, like RFID, barcodes, NFC, or QR-codes (Feng et al., 2020). These technologies allow data to be collected in real-time, but there are pros and cons to each. For instance, barcodes are cheap but can be hard to read, and data cannot be added later (Tu et al., 2018).
Meanwhile, RFID is easier to scan and can accept more data, but it is a more expensive alternative (Miaji et al., 2013; Tu et al., 2018). Regardless of the technology used, traceability data must be collected at the appropriate time to create a history for each product (Jansen- Vullers et al., 2003; Chongwatpol and Sharda, 2013).
Third, full manufacturing traceability can only be achieved by registering what operation happens in which capacity unit to each product (Jansen-Vullers et al., 2003). Furthermore, it is crucial to keep track of actual input variables rather than the process plan’s planned inputs for each operation (Jansen-Vullers et al., 2003). Only by tracking actual input parameters can a product’s traceability data be accurate to its actual history. In order to track the process parameters, there needs to be a recorded link between the products and components and the
18
performed processes. Creating a valid traceable link requires rigid batch integrity and accurate data collection for the process data (Jansen-Vullers et al., 2003). The continuous link of traceability data throughout the supply chain will allow the actors to certify the products’
quality and safety, but also certify its sustainability, protect it against counterfeits, and facilitate recalls (Jansen-Vullers et al., 2003; Marucheck et al., 2011; Choi et al., 2015; Abeyratne and Monfared, 2016; Papetti et al., 2019).
Fourth, the three paragraphs below addressing sustainability, counterfeits, and recalls, illustrates some examples of data that can be reported from a traceability system. These examples are significant in terms of profitability, competitiveness, legislation, and brand reputation.
Complete traceability of the supply chain allows the actors to effectively monitor the supply chain’s sustainability (Papetti et al., 2019; Venkatesh et al., 2020). It is not sufficient to consider internal manufacturing activities; external activities must also be considered to certify sustainability (Papetti et al., 2019). Being able to verify one’s supply chain’s sustainability is crucial to meet increasing market pressure and to gain a competitive advantage (Abeyratne and Monfared, 2016; Papetti et al., 2019; Venkatesh et al., 2020). Currently, finding an accurate history of where, how, and when for each product is difficult due to globalized production (Abeyratne and Monfared, 2016; Papetti et al., 2019). Thus, it is difficult for consumers to make informed and responsible decisions (Papetti et al., 2019; Venkatesh et al., 2020). A product that provides the information necessary to make an educated, sustainable choice will be considerably more attractive for the increasing number of consumers that want to consume responsibly (Papetti et al., 2019; Venkatesh et al., 2020).
Similarly, to certifying quality, certifying sustainability requires complete tracking of the products, components, and sub-assemblies in the supply chain (Papetti et al., 2019).
Additionally, tracking of the external and internal manufacturing processes, transportation within the chain, and the scrap during production is required (Papetti et al., 2019). In some cases, sustainability data will not be directly related to the product but instead an “overhead”- factor. For instance, the energy consumption for lighting in the factory is not directly caused by a single item. However, it is still possible to allocate the lighting’s energy consumption to each item that was produced in the factory during that day. To get a fair understanding of a product’s sustainability contribution, “overhead”-factors should be allocated to each item (Papetti et al., 2019). Nevertheless, direct data is to be preferred over overhead allocation, when possible. The sustainability of the supply chain can be certified by maintaining continuous links of products and components in the supply chain and linking sustainability data to each item. The result is similar to an as-assembled BOM of environmental and social resource consumption for each product.
Protecting supply chains against counterfeits is a difficult task due to the modern supply chain’s complexity (Choi et al., 2015; Miehle et al., 2019). However, track-and-trace has been proven as an exemplary mean for protecting against counterfeits (Choi et al., 2015). Mainly, track-and- trace involves tagging products and components and tracking them throughout the supply chain. Tracking and tracing include recording all events that a part is exposed to; additionally, each record must contain the current time and location of the part (Miehle et al., 2019). The
19
result of track-and-trace is an as-assembled BOM where each component and its history is linked to a parent.
Furthermore, track-and-trace also tracks the product from production until it reaches the customer. Thus, it creates a continuous link from the sourcing of components to customer delivery that verifies that the goods are from the right supply chain actors. Moreover, the tag on the final product can remain accessible to the consumer so that they can verify the product (Choi et al., 2015). For example, Gucci has RFID chips in their handbags that can be read by the consumer to verify that the product is authentic (Li, 2013). These RFID tags are also used within the supply chain to track-and-trace Gucci’s goods (Li, 2013).
For high-value goods, like Gucci handbags, the costs of track-and-trace solutions are relatively low and, therefore, justifiable (Choi et al., 2015). There is a massive amount of money in protecting supply chains against counterfeits. Currently, OECD estimates that 3.3 % of world trade is in fake goods (Trade in fake goods is now 3.3% of world trade and rising, 2019).
Additionally, in the European Union, 6.8% of imports from non-European countries are counterfeits (Trade in fake goods is now 3.3% of world trade and rising, 2019). These counterfeits are dangerous for companies’ brands, and potentially the well-being of their consumers (Li, 2013; Choi et al., 2015).
Moreover, both counterfeit components in products and other quality deficiencies can lead to product safety issues. The result of product safety issues might be a product recall. A recall means retrieving products with a particular type of deficiency from the supply chain and customers (Jansen-Vullers et al., 2003). Recalls are incredibly costly to facilitate but will continue to be a necessity in manufacturing; therefore, there is an incentive to reduce recall cost (Wang, 2014). For example, General Motors (GM) had to do a recall of 2.6 million vehicles in 2014, due to faulty ignition switches, the recall and repairs cost them 700 million USD (Lawrence, 2019).
The faulty ignition switches resulted in 124 deaths before the vehicles were taken off the road (Lawrence, 2019). Recalls must be quick and efficient to make sure that products are recalled before people are hurt. A key aspect of recalls is identifying the products that are impacted by the issue. However, it is challenging to identify which other products are affected when a quality defect is detected (Jansen-Vullers et al., 2003). The traceability of assembly processes makes it possible to link quality defects in one product to other products. For example, the faulty products might share a quality deficient component from the same batch as other products. Then, the other products with a component from the same batch as the faulty products might be subject to recall (Jansen-Vullers et al., 2003). As a result, the granularity in the traceability system is highly relevant. If good and bad products and components share the same ID, then it is impossible to differentiate between the two (Dai et al., 2015). If it is not possible to identify the other products which might share the same deficiency, then all questionable products have to be recalled, and the company might have to rely on expensive advertisement to announce their recall (Jansen-Vullers et al., 2003; Dai et al., 2015). Thus, full traceability of the as-assembled BOM is vital for product traceability (Jansen-Vullers et al., 2003).
