How to avoid car recalls: A case study at Honda of the UK Manufacturing

How to avoid car recalls: A case study at Honda of

the UK Manufacturing

Master Thesis

By

Niels-Karsten Pluym – s1629948

University of Groningen

1

Foreword

This research is part of my Master degree in Technology Management. The Master Thesis is the ultimate examination of my student career and involves skills, experiences and knowledge from several courses.

It was an honour to work with the Honda of the United Kingdom Ltd. What makes my stay even more special is the fact that I am the first Master student who finishes his Master degree in one of the most famous automotive manufactories. The novelty of a graduate intern gave me opportunities to develop my practical and academic skills. Honda showed me what ´the real-World’ expects and, on the other hand, gave me room for experiments. Therefore I would like to thank all my colleagues of the Quality Team, AF managers and general associates.

I would like to thank my manager and mentor Luke Hughes. Luke supported me to think out-of-the-box and corrected me when needed. Moreover, he helped me enjoying the7 months at HUM and the UK.

2

3

Content

Foreword ... 1

Management Summary ... 2

1. Honda Worldwide ... 5

2. Honda of the UK Manufacturing ... 7

3. Project Introduction ... 9

3.1. Assembly frame customisation ... 9

3.2. Outflow ... 10

3.2. Recalls ... 12

4. Problem definition ... 13

4.1. Quality ... 14

4.2. Hard and soft aspects ... 15

4.3. Causal relationships ... 17

4.4. Detection ... 18

4.5. Mistake Prevention ... 19

4.6. How to avoid outflow on the long term? ... 20

4.7. Summary research problem ... 20

5. Methodology ... 22

5.1. Common metrics ... 22

5.2. Stepwise problem solving ... 22

5.3. Analysis with the tools ... 24

5.4. Application of methodology ... 24

6. Literature framework ... 25

6.1 SQ1: What are the critical items in AF2? ... 25

6.2 SQ2 & SQ3: Mistake detection ... 27

6.3 SQ4: How can we reduce complexity ... 28

6.4 SQ5: How can we use the quality improvement projects and avoid outflow on the long term? ... 30

7. Data protocol ... 32

7.1 QFD process ... 32

7.2 Study worksheet and the use of it ... 33

7.3 Database ... 35

8. Results ... 37

4

8.2 SQ4; Complexity ... 37

8.3 SQ2 & SQ3: Detection ... 38

8.4 QFD results ... 39

9. Analysis ... 40

9.1 Analysis per sub-question ... 40

9.2 Answering the short term sub-questions: ... 47

10. How to use the improvement activities and avoid outflow in the future? ... 48

10.1 HUM’s Continuous Improvement Infrastructure ... 49

10.2 Role of Knowledge ... 52

10.3 Role of knowledge in the continuous improvement infrastructure ... 53

11. Benefits ... 56

11.1 Direct cost savings ... 56

11.2 Expenses for improving critical items... 56

11.3 Direct savings AF strategic level ... 57

11.4 Indirect cost savings ... 59

11.5 Total benefits ... 60

12. Conclusion ... 62

Literature implications ... 65

Further Research Honda ... 65

References ... 67

5

1. Honda Worldwide

Honda is a Japanese car, motorcycle and power product manufacturer founded in 1948 by ‘the dreamer’ Soichiro Honda. Nowadays, Honda creates innovative products that enhance human mobility and benefit society. Honda’s slogan ‘the power of dreams´ is based on the visionary principles of Soichiro Honda and encourage all associates to pursue their dreams. This pursue in combination with a global thinking ensures to create and produce products of high quality at a reasonable price for customer satisfaction. ‘The power of dreams’ will lead to a continuous improvement in the field of mobility, which brings the society forward.

The public awareness of environmental issues and major structural shift in economy makes the customers buy decision more careful. To satisfy the customers demand Honda strives for advanced technologies that fulfil the safety and environmental requirements while enriching product appeal. Further cost reductions and a strengthened manufacturing structure should lead to the reasonable price for the whole society.

Honda determined that ‘Providing good products to customers worldwide with speed, affordability and low CO2 emissions’ will direct the company for the next eight years. Honda’s innovative hybrid and hydrogen-fuelled engines are examples of Honda mission to fulfil the environmental requirements. As a result of the innovative engines and high fuel price, Honda was able to cope with the financial crisis, which lead to net sales of 12,002,834 million Yen (about 96,649 million GBP, 20th of june 2012) and net income of ¥600,039 million. Competing car manufacturers introduced hybrid engines as well, resulting in decreased net sales of ¥8,936,867 million in 2011. However, due to efficient cost-cutting, Honda maintained its profits with ¥534,088 million net profit in 2011. Figure 1 shows the net income for the different segments over the last 5 years (Honda Motor Co., 2012)

Figure 1. Net sales by business segments

6

Figure 2. Overseas production plants (Honda World, 2012)

7

2. Honda of the UK Manufacturing

Honda of the UK Manufacturing (HUM) was established in 1985 and is the Honda Car manufacturer of the Civic, Jazz and CRV for the UK, Europe, Middle and Near East and Africa. In 1989, the Engine Plant began producing engines and had a capability of 1000 engines per day, enough to satisfy HUM’s demand (Honda of the UK Manufacturing, 2012). The first car plant was opened in 1992 and has a production capacity of 150,000 cars per year and a size of 24,000m². The production line comprises Weld, Paint, Assembly Frame (AF) department, Material Logistics and Vehicle quality department (final quality check). Figure 4 provides an overview of HUM’s production facilities. In the current situation the Honda Jazz and Honda CRV are produced in the plant one.

In July 2001, Honda started the production of the Honda Civic in a new completed 15,000m² second car plant. With the addition of a second plant increased the capacity from 150,000 to 250,000 units. The current daily production rate for the year 2012-2013 is 792 units /day, equals 180,000 cars for the year 2012-2013. In total are 3,500 associates employed at HUM. The focus of this study is mainly on the assembly activities in AF of the Honda Civic in car plant 2. The next chapter provides more detailed information about AF2 and the production processes.-

Engine plant Weld 1

Paint 1 Weld 2 Paint 2 Assembly Frame 2 Assembly Frame 1 Plant 1 Plant 2

Honda of the

UK

Manufacturing

Figure 4. Overview production facilities Honda of the UK Manufacturing

HUM’s ethics and values play a major role in the car production and are clearly visible in the different departments. The ethics and values are founded on Honda’s worldwide philosophy. Since the founding of Honda in the 50’s forms the Honda philosophy the basis of associates’ actions on Honda’s worldwide operations and thus at HUM too. The fundamental beliefs are based on two pillars (Honda of the UK Manufacturing, 2012):

1. Respect for the individual:

1. Initiative: Encourage creativity and freedom of each individual. This provides the environment to create the best products.

2. Equality: Create equal opportunities for every associate.

8 2. The three joys:

1. The joy of buying: Make the owner of Honda products and services proud by exceeding customer’s expectation

2. The joy of selling: Be proud to represent Honda and take satisfaction in making customers happy

3. The joy of creating: Providing a quality product to the customers every time

The philosophy encourages ‘living’ the three core values and making HUM a successful chapter of ‘Honda worldwide’. The three core values are:

1. Make a difference today: ‘Recognise the impact you make on the business and on your team. Seek out ways that we can improve what we do, get involved and make tomorrow better’. 2. Do the right thing: ‘Be honest and open in your dealing with everyone around you. Believe in

each other and be prepared to step out of your comfort zone. Learn as much as you can to be the best as you can be’

3. Make time for each other: ‘Be approachable, listen to others’ viewpoint and consider different opinions. Be open minded, care about what happens and help us create a great workplace.

Figure 5. Quality improvement processes

The philosophy encourages empowerment of all associates. Different Kaizen programmes stimulate continuous improvement at all levels. The aim for continuous improvement must lead to the good quality car, with speed and affordability. Honda has a worldwide quality approach to achieve consistently high levels of quality and is working towards the ultimate goal of zero-defects. For Honda it is unacceptable that even 1% of the customers receive a defective product. The customers who become the owner of the 1% defect product consider the car as a 100% defective. That is why Honda and so HUM aims for 100% product quality. Honda has created a quality cycle that continuously improves quality at every stage in every facility: Design, Development, Production, Sales and After-Sales service. HUM is involved in the development and production stage. Figure 5 shows the processes that create new levels of quality for HUM and Honda Worldwide.

9

3. Project Introduction

This chapter describes what the consequences of producing defects are. Firstly, the development and relation between car customisation andAF2 processes is explained. The high car customisation leads to a high variety in cars, what makes the process more complicated for the Line Associates. Complicated processes result in increased chance of quality defects. A defect car becomes outflow if the car leaves the AF department. The second sub-chapter defines outflow and shows the need for a quality improving research. Lastly, car recalls might happen if the outflow concerns a critical item. Car recalls cause besides the direct ‘outflow’ costs high indirect costs. The last sub-chapter defines why critical items cause car recalls and the consequences. Diagram 1 provides an overview the relations between defect, outflow and car recalls.

Defect Outflow Recall Direct + Indirect costs Critical Item? No Recall Direct costs Yes No

Diagram 1. Defect Flow

3.1. Assembly frame customisation

HUM’s AF2 is the final process department for the Honda Civic. The department fit 831 items on the frame to deliver the final product, a 100% quality car. Diagram 1 provides an overview of the car’s process flow. This research focuses at the quality issues in the AF department. Thus, Quality issues concerning Weld, Paint or Engine are out of scope.

Die cast engine Model Weld Frame Type Paint frame Colour

AF2

HUM

Diagram 2. Process flow HUM and AF2

10

attractive for customers. To fulfil the customers’ demand of high quality customised cars, HUM produces according to Just-in-Time (JIT) principles. This sophisticated pull-production system is subject of Continuous Improvement (CI) and enables production on demand. HUM works with batch sizes from 1-40 cars, dependent on the customers demand and cluster opportunities. The planning of production is out of scope and will not be discussed in this research. Every car is supplied with a unique Vehicle Identification Number (VIN) at the first production process; Die casting the engine in the engine plant. The VIN represents the cars Model, Type and Option. First customisation step is the engine choice. For the Honda Civic, the customer can choose between 1.4 petrol, 1.8 petrol, 1.6 diesel or 2.2 diesel engine, representing the four models TV0, TV1, TA9 and TV2.

The weld department customises the frame for the demanded model and type. The type represents all possibilities that have effect on the driving ability of the car. For example a G stands for a Left Hand Drive (LHD) and E for Right Hand Drive (RHD). Other examples that determine the type are the shifting (manual/automatic), country; Russian legislation differs from South-African legislation, etc. In total the Civic 109 has different types to satisfy the customers worldwide. The model and type specified frame goes to the paint department, where the frame gets painted in one of the nine Civic colours. A conveyor chain transports the painted car to the AF2 department, where the engine and about 837 items are installed to complete the customised car. Besides the model, type and colour, the customer might add options to the car. Options are added in the AF department and improve the convenience for the driver. Examples are, sound system, leather seats, navigation, etc. In total AF has to cope with 9 single options, which can form 17 different option combinations.

The Model, Type and Options (MTO) are not affecting all the 838 items, but create different processes for most of the AF processes. By providing the cars with a buildsheet HUM tries to simplify the part selection process. The buildsheet contains an item specific identity code, what tells the associate which items suits to the MTO. However, this is not possible for all parts and even with a buildsheet is the diversity of different items, and so processes, significant. The high customisation makes the AF processes more complicated, what affect the desired zero-defect quality.

3.2. Outflow

Honda’s total outflow represents the amount of defect cars that leave the AF department. A defect and thus outflow can be caused by several causes. The defect cars are from an AF point of view detected at three different stages after the AF department (Figure 6):

1. PIPS - Market Quality issue 'PIPS' (Priority Item Improvement Requests) to AF when issues are identified on the Market and the defect is AF’s responsibility.

2. QIR - Vehicle Quality Department issue QIR (Quality Improvement Requests) are AF’s responsibility when vehicle quality department associates complete their inspection and conclude the defect is made by the AF.

3. VCS – Vehicle Stock Checks are cases where Honda has identified an issue and had to inspect vehicles in the stock yard just in case of other affected vehicles.

11

shipped to the car dealers in GB or to the main distribution centre in Ghent, Belgium. This is the last opportunity to detect outflow before it goes to the car dealer and become PIPS.

Figure 6. Product flow AF - Customer

As depicted in the figure 7, AF experienced an increasing trend in the quantity of outflows/quality reports over the last three years (2009-2011). It is impossible to identify one root cause for the entire outflow. Hence, we quantify and qualify the yearly outflow per cause. Figure 8 provides an overview of the causal factors per stage for the total outflow of 411 cars in the year 2011 (88Ki).

The two most significant causes in the PIPS, QIR and VQ stage are miss assembly and wrong part specification installation, which are responsible for respectively 34% and 27% of 2011’s total outflow. According to Honda’s aim for zero-defect quality product, every PIPS is one PIPS too much. Every PIPS will be considered as a 100% defective car and cause a car recall. On the other hand shows the QIR and Vehicle Stock checks a more representative overview of the defects. In special wrong spec installations are hard to detect for the customer. The customer won’t feel the slightest difference in the items, but

it can have disastrous consequences for the driving performance and safety. Hence, it is not possible to identify which causal factor has the biggest influence on a model’s total outflow. Nevertheless, Figure 7 and 8 show the relevance of the quality problem in the AF department.

Figure 7 . Outflow 2009-2011

12

3.2. Recalls

Moreover, Honda wants to avoid a car recall at any time. Recalls occur when the product fails to meet or exceed the safety standard and cause a risk of serious injury or death, or fails to meet a voluntary standard set by the specific market and is often referred to as a “Product-Harm Crises” or “Brand misconduct” (Mullan, 2004; Dawar & Pillutla, 2000, Huber et al., 2009). Huber et al. (2009) distinguish four types of product-harm crises:

1. Product quality differs from expectation: The brand is not able to fulfil customer’s expectations of functionality. The relationship between customer and brand is harmed by security risks arises while using the car.

2. Lack of service orientation: Bad experience or attention in customer service influence purchase decisions.

3. Symbolic-psychological harm crises: Social and psychological values harms the product brand, i.e. links with child labour, or use of hazardous waste.

4. Socially debatable actions: Consumers base the purchase decision on norms and values, which may be violated by the brand.

AF outflow can be categorised as the first type of product-harm crises; product quality differs from expectations. Not all outflow causes a risk of serious injury or death. For example, a defect audio installation causes no serious risk; while a defect brake-system might has disastrous consequences. HUM defines items that exceed the safety standards as critical items. Hence, HUM’s critical items are responsible for possible car recalls.

Berman (1999) argues that recalls are more likely to occur due to the increase of product complexity, increasing demand and closer monitoring by government and the specific firms. Honda has to deal with the trend of increasing product complexity since the customer demands customised cars with continuously innovating high-tech equipment on board.

13

4. Problem definition

HUM’s goal is reducing the amount of outflow and so recalls, avoiding direct and indirect costs, which tarnish the customers’ goodwill and company’s reputation. HUM’s AF quality team is responsible for solving the AF quality issues. They asked me to research the quality improvement opportunities for the critical items, by developing an easy applicable methodology. Therefore, the AF quality team supports me with gathering the required data. HUM’s direction appointed it as a priority project and demands solutions for the critical items before Christmas 2012. During my stay at HUM, I noticed that HUM invents sophisticated methods to solve quality issues. However, the methods and knowledge are not used after solving the specific quality issue. The managers re-invent the wheel again and start from scratch instead of adopting previous used methods. HUM needs an approach that supports the managers to re-use and customise the gathered knowledge and improvement methodologies. This need has a long term strategic character and ensures outflow avoidance for non-critical items. The combination of short and long term needs result in the following objective: Elimination of outflow in HUM’s Assembly Frame departments.

To gather the correct data the following research question is formulated:

The gathered data must entail into deliverables to achieve the objective. The distinction between short and long term leads to two deliverables:

1. Elimination of outflow for all critical items in Assembly Frame 2

2. Widely applicable scientific approach to eliminate outflow for the non-critical items

The first deliverable focuses on short term solutions for AF2 critical item quality defects. An understanding of why the outflow occurs is needed to develop elaborated solutions. This chapter explores the literature and HUM’s knowledge about the different outflow causes. On the basis of this knowledge are the sub-questions defined. The sub-questions determine which data is needed to achieve elimination of outflow for all critical items in AF2. The second deliverable satisfies the need for a long term solution and requires a more holistic approach.

The distinction in short and long term is a distinction between critical and non-critical items. As described in the previous chapter are critical items responsible for car recalls. However, what defines HUM as critical items? And how many critical items are assembled in AF2 on the Honda Civic? Hence, the first sub-question is:

SQ1: What are critical items in AF2?

As shown in figure 8is PIPS outflow caused by 7 causal factors: 1. Miss Assembly, 2. Wrong Spec, 3. Poor repair, 4. Damage, 5. Missing information, 6. Torque and 7. Complex causes. HUM already solved the torque and complex causes. The missing information outflow has miscellaneous causes. Because of time limitation is identification of these causes out of scope. Four outflow causes remain after excluding three of the causes (Diagram 3).

2. Wrong spec installation

+ 1. Miss assembly 3. Poor repair 4. Damage + + + Ouflow

Diagram 3. Outflow causes

14

To find out why the specific outflow causes happen is a better understanding of the literature required. The next two sub-chapters explain: (1) ‘what is quality and which quality defect causes are defined in the literature?’ And (2) ‘What are the Total Quality Management practices and how can they be distinguished?’ Comparison between the literature and HUM’s outflow issues results in a cause-relationship model. This model enables formulating the sub-questions for the short term. The formulation of the long term sub-question completes this chapter, what makes the chapter a comprehensive overview of the research problem and requirements.

4.1. Quality

Rupp (2004) argues that the potential recall-related goodwill loses for Japanese car firms are higher than for American car manufactories. Japanese manufactories, like Honda, gained high quality reputation through a sophisticated and scientific approach of quality and manufacturing. Honda’s philosophy as explained in chapter 1 is based on the Just-In-Time (JIT) approach. The JIT approach focuses on improving the process and resources efficiency to achieve high standard quality what meets or exceeds the customer expectation

(Wedgwood, 2006). A cornerstone in the JIT philosophy is elimination of waste (Nicholas, 1998). Waste is defined as every non value adding activity and element of a system. Taiichi Ohno (1978) defined 7 types of wastes:

1. Waste from producing defects 2. Waste in transportation 3. Waste from inventory 4. Waste from overproduction 5. Waste of waiting time 6. Waste in processing 7. Waste of motion.

According to the Taiichi Ohno (1978), the AF quality issues are defined as ‘Waste from producing defects’. Defect production accrues additional material expanses and labour costs related to the rework and required disassembly (Nicholas, 1998). The waste consists of all scraped items and the related labour, material and resource expenses. Furthermore, the production holds up through the defects, what increases the production lead time. The challenge is to eliminate the waste by doing it right the first time.

Despites Honda’s strive for zero defects, it is impossible to ensure a 100% quality; there will always be a machine, worker, or process which produces a defect. Two causes of defects can be defined; (1) variation and (2) mistakes (Stewart and Grout, 2001).

1. Variation: Honda’s manufacturing processes produce variable outputs and might turn into consequential quality affecting variations. Variation is distinguished in two causes:

I. Common causes: Are small variations that act continuously on a specific process (Nicholas, 1988). Continuously small variations in the operation environment, equipment, materials and so on, become predictable.

II. Special causes: Are abnormal and unpredictable difference variation of the output (Nicholas, 1988). Special causes are often referred to as assignable causes, because caused variation can be traced back to them.

15

items and determination of the specific cause, action can be taken to eliminate the variation and decrease the number of defects.

2. Mistakes: SPC’s major drawback is the reliance on sampling. Given the random nature of errors, the production of a defect probably occurs during a non-sampling time (Nicholas, 1988). The critical items have major impact on direct and indirect costs, thus a zero defect quality is required. Hence, the short term emphasis is on 100% inspection instead of sampling inspection.

Nikkan Kogyo Shimbun (1988) authors related 10 kinds of errors to different defect sources. All the errors are an outcome of human mistakes. Understanding of human mistakes helps to fully understand the importance of the mistake subject. Most human error studies are grounded on the human’s cognitive science. The cognitive scientists view the human brain as an information processor. The processor has to make the right decisions and actions with a limited resource (Reason, 1990). Mistakes occur when the processor is overtaxed, or when known routines and rules are used in inappropriate situations. Norman (1981) identifies two sources of mistakes; (1) Faulty activation of schemas, unconscious routines as a result of commonly performed actions, and (2) faulty triggering of schemas. Stewart and Grout (2001) summarise the cognitive mistakes mechanisms in relation with quality control problems (Diagram 4). All relevant information about the task is input of the cognitive resource ‘black box’. Besides information, the selection of the appropriate rules and routines in the ‘black box’ is influenced through other environmental factors. The output of the black box is an action or decision, which can be manifested as an outcome to the environment. The required task is successfully fulfilled if the outcome is a positive manifestation. A negative manifestation leads to failure if the manifestation is important to complete the task.

- Required task - Environmental cues and noises

Cognitive ‘’Black Box’’ Negative manifestation Positive manifestation Failure Succes No mistake Outcome important Outcome unimportant Mistake

4.2. Hard and soft aspects

HUM’s quality approach is guided by the principles of Total Quality Management (TQM). TQM links the different core concepts that embody the way HUM operates to achieve targets and enhance quality (Hackman and Wageman, 1995). Bou-Llusar et al. (2008) summarise the two main points of successful TQM operations:

1. TQM practices are classified into two categories, the social/soft TQM and technical/hard TQM. The social issues emphasise the importance of good human resource management. Aspects as leadership, associate participation and training facilitate HUM’s quality mission. The hard technical issues orientate on improvement of production methods. The continuous seek for process improvement results in enhanced and well-defined processes and

16

procedures. These points are reflected in the European Foundation for Quality Management (EFQM) excellence model, what makes it a valid embodiment of TQM (Eskildsen, 1998; Westlund, 2001). As shown in Diagram 5, the enabler side of the model is the infrastructure of the TQM, which enables achievement of results, and so Honda’s mission statement. The five enablers are classified into soft and hard TQM issues. The social soft aspects of TQM includes ‘Leadership’ and ‘People Management’, while ‘Partnership and Resources’ and ‘Processes’ represents the technical hard aspects (Brown, 2002). The fifth enabler ‘Policy and Strategy’ is related to both social and technical aspects (Black and Porter, 1995).

Mission statement Vission Key succes factors

Leadership (Soft) People management (Soft) Policy and strategy (Soft &

Gard) Partnerships and resources (Hard) Processes (Hard) People results Customer results Society results Key performance results Enablers Results

Learning, Creativity andImprovement Diagram 5. EFQM Model of Excellence

2. TQM initiatives are distinguished by its holistic character. This holistic character makes it impossible to isolate the social or technical aspects. Successful TQM initiative demands interrelationship between the two issues.

HUM’s elimination of outflow mission fits perfectly into the EFQM Excellence model. Therefore, the 5 enablers facilitate this project and thus the elimination of outflow in the short and long term. Honda contracted an independent consultancy organization to improve the social aspects of quality management and CI. This organisation launched the ‘Engage Change’ project, what contains ‘social’ improvement throughout the whole organisation. Because of the Engage Change project, HUM asked me to focus on the hard aspects of quality management for the critical item improvement. However, the soft aspects must be included in the development of a strategic long term solution. Hence, the soft TQM practices are only excluded for the critical item research.

17

4.3. Causal relationships

The literature intermezzo distinct two causes for quality defects: ‘Variation’ and ‘Human mistakes’. Furthermore, it introduces the 5 TQM enablers, which are classified in soft/social and hard/technical practices. For the short term, the emphasis is on human mistake defects and the hard ‘processes’ enabler. This sub-chapter combines the literature with the causes as detected by HUM detected. Hence, it is possible to define the sub-questions and gather the data for the short term deliverable.

Wrong spec installation Human Mistakes + Outflow + Miss assembly Poor repair Damage + + + Variation + Special Causes Common Causes + + ₊ Poor transport + + Lack of skills associate + Training -+

Diagram 6. Cause-relation Diagram

Diagram 6 shows the cause-relation diagram for outflow. The causes are defined with preliminary research by the Quality Team, interviews with the Quality Team and literature. The ‘Damage’ outflow is caused by poor transport. After the final AF item installation covers the car a long route to the customer. Sometimes the car gets damaged during the product transport from AF to the customer. This research is limited to HUM’s AF department. Because of this boundary setting is the ‘Damage’ outflow out of scope. The 8 ‘Poor repair’ outflows are caused by a lack of training. with the new model introduction of the Civic and the CRV HUM employed a new associates, who needs training. Training is part of the soft ´People management´ practice and out of scope for the short term research.

18

The last identified cause is miss assembly. Miss assembly represents more than 50% of 2011’s PIPS. Previous research indicated that 24 of the 40 miss assembly frames has miscellaneous causes, which are caused by human mistakes, variations and unqualified associates. The remaining 16 PIPS are caused by mistakes in the product launch and already solved. Unlike the wrong spec installation are most of the miss assembly outflows easier to detect for the customer, what decreases the exposure. Therefore, the chance of hazardous car accidents and so recalls is less likely, thus the indirect costs will be lower. However, there is still a chance that human mistakes cause miss assembly outflow. Hence, the miss assembly outflow will be included in the short term research.

HUM’s quality team started several projects to decline the number of outflow caused by variations. These projects bear fruit already. However, human mistake outflow is under exposed, since the causes are not as obvious as the variation outflow causes. Hence, the next sub-chapter focuses on the human mistake affecting variables. Determination of the variables enables formulation of the sub-questions, which are necessary to gather the required data.

4.4. Detection

This sub-chapter connects the cognitive science to operations and mistake detection, the first aspect of 100% inspection. The aim of the study is elimination of all defects. However, inadvertent errors will always occur (Nicholas, 1998). Hence it is important to detect the defect items before they leave the manufactory (Ghinato, 1998). Informative inspection detects mistakes and provides feedback (Hirano, 1988). The feedback ensures that the assigned associate takes the required actions and prevents outflow. Mistake detection is grouped in two kinds of inspection (1) Self-inspection and (2) successive-inspection.

Self-inspection is performed by the associate who is responsible for the installation of the spec. This is the most effective and thus preferred method of informative inspection (Ghinato, 1988). The key of an effective self-inspection approach is instantaneous feedback and entails in quickly applied corrective actions, what leads to a decrease of outflow. At the moment, Honda applies self-inspection, although it is unclear which processes have self-inspection and how it is used. An overview of the currently executed self-inspection reveals which processes use self-inspection methods and the efficiency of the methods.

SQ2: Which processes use self-inspection and can we create an improved self-inspection method? Self-inspection is the most effective way of informative inspection, although it can be liable to subjectivity. A method to deal with the subjectivity is using successive-inspection (Ghinato, 1998). With successive-inspection the succeeding operations check the predecessors work (Shingo, 1986). A disadvantage of inspection is the longer feedback loops. The location of the successive-inspection determines the length of the feedback loop. Longer feedback loops need more time before action can be taken. Hence the location of successive checks is an important topic. The third sub question includes the topic of successive-inspection and the location of them.

SQ3: Which processes needs an objective successive-inspection and how can we improve the feedback loops of the current successive checks?

19 Detection - Required task

- Environmental cues and noises

Cognitive ‘’Black Box’’ Negative manifestation Positive manifestation Failure Succes No mistake Outcome important Outcome unimportant Mistake Feedback & Action - Self-check -Successive-check

Diagram 7 Mistake detection

4.5. Mistake Prevention

Detection of mistakes will serve to remedy defects, but none of them eliminate the waste. Shingo refers to source inspection (1986) in case of finding the sources of the causes and eliminate them. The first step in mistake prevention is identification of the source. The source can be monitored and eventually eliminated when it is identified (Nicholas, 1998). Prevention of mistakes relies on an understanding of the cognitive mechanism. Prevention is possible when we are able to prevent black box overcapacity. An adjustment of the environment highlights correct cues for particular cognitive response (Stewart and Grout, 2001). In other words, reduce the complexity of the associate’s environment. Several variables influence the associate complexity. The different variables which influence the items are defined in Chapter 6.3. The fourth sub-question reflects the challenge for mistake prevention by reducing the complexity for each specific critical item:

SQ4: How can we reduce the complexity for each critical item and prevent mistakes?

Diagram 8 shows how complexity reduction avoids overtax of the cognitive black box.

Detection

- Required task - Environmental cues and noises

Cognitive ‘’Black Box’’ Negative manifestation Positive manifestation Failure Succes No mistake Outcome important Outcome unimportant Mistake Feedback & Action - Self-check -Successive-check - Reduce complexity Prevention

20

4.6. How to avoid outflow on the long term?

Mistake detection and mistake prevention avoids human mistakes for the critical items. However, the outflow solution on the long term needs to take the other outflow variables into account. Hence, HUM needs a widely applicable scientific approach that enables application of the developed tools. SQ5: How can we use the quality improvement projects and avoid outflow on the long term?

4.7. Summary research problem

Avoidance of outflow is essential for Honda due to the high direct cost and recall related indirect costs. Hence, the objective of the research is ‘Elimination of outflow in HUM’s Assembly Frame departments’. To gather the correct data the following research question is formulated: ‘How can we eliminate quality defects in the Assembly Frame and reduce the amount of outflow for the short and long term?’ The short term deliverable is ‘Elimination of outflow for all critical items in Assembly Frame 2’ and long term deliverable is a ‘Widely applicable scientific approach to eliminate outflow for the non-critical items.’ The emphasis for the short term is on human mistake related outflow, what includes wrong spec installation and miss assembly. Four sub-questions are formulated to guide the research and gather the required data for the short term. The data is needed to avoid outflow for critical items caused by wrong spec installations and miss assembly answers. The fifth sub-question enables outflow avoidance for the non-critical items. The long term part of the research has a more holistic and strategic character. The sub-questions are:

The relations between research question, variables and sub-questions are shown in cause-relation diagram (Diagram 9).

Short term

2. Which processes use self-inspection and can we create an improved self-inspection method. 3. Which processes need successive inspection and can we improve the feedback loops of the

current successive checks?

4. How can we reduce the complexity for each critical item and prevent mistakes?

Long term

21 Wrong spec installation Human Mistakes Mistake Detection Mistake Prevention Reduce Complexity -- + + Source inspection Informative inspection Successive inspection Self inspection + + + + Outflow + Miss assembly Poor repair Damage + + + Variation + Special Causes Common Causes + + Pokayoke + + Poor transport + + Lack of skills associate + Training -+

Diagram 9. Cause-relation diagram

Out of scope:

The limitations and boundary settings determine the focus for the short term. On the other hand, not all the variables that might influence Honda’s performance are taken into account. The three main limitations and boundaries are: distinction in Hard/Soft Issues, time and physical boundaries.

The complementary ‘Engage Change’ project deals with all the soft TQM issues. Car recalls caused by poor repairs are caused by soft associate issues. Other variables which have to do with ‘Leadership’, ‘People Management’ are not included in the short term data protocol. Combination of time and Soft issues restrict the short term research in ‘Policy & Strategy’.

Due to the urgent need of short term activities it is not possible to change the design of the items. Design changes are only possible with new model introduction, what will not occur in the next 4 years for the CIVIC and CRV. Hence, design improvements are out of scope for the short term. However, it must be taken into account on a more holistic and strategic level.

The physical boundary is the department AF 2. All researched items are processed in AF2. All other departments, as showed in figure 4 are out of scope. Hence, it is impossible to solve transport issues and eliminate damage outflow. A combination of time and physical boundary settings put supplier issues out of action.

22

5. Methodology

The use of structured methods represents an important item in problem solving quality programs (Imai, 1986; Ishikawa, 1985). The methodology makes problem solving a manageable program, which use the similarities in the patterns of problem solving. The cognitive processes as described in the previous chapter can be represented as ‘a sequence of states’ and can be effective in facilitating problem solving (Choo et al., 2007). The methodology consists of a consistent language, defined sequence of steps and a set of tools for quality improvement. Choo et al. (2007) defined three methodological elements which are required for a successful project; (1) employing common metrics, (2) adhering the stepwise problem solving approach and (3) analysis with the tools. This chapter describes which methodology and how it will be applied.

5.1. Common metrics

Introduction of common metrics helps coordinating improvement activities by aligning the problem solving steps across the whole manufactory (Orlikowski, 2002). They enable a shared language which facilitates communication about issues and mistakes in quality (MacDuffie, 1997). Choo et al. (2007) argues that implementation of common metrics ensures a consistent use of language for executing technical quality. The common metrics are especially helpful in coordinating the acquired knowledge across the company.

HUM created its own jargon in 25 years of manufacturing. This jargon is applied in the research and definition of terminology are provided inline the text. Appendix I; Definitions and common metrics’ list up the jargon and abbreviations. Furthermore provides the research common metrics for the coordination and acquisition of knowledge throughout the different chapters.

5.2. Stepwise problem solving

While the metric ensures clear communication, the stepwise approach is often used for problem analysis. Honda’s aim for continuous quality improvement suits with thePlan-Do-Check-Act (PDCA) approach (Shewhart, 1939). The PDCA approach provides a systematic analysis due the capacity to include multiple perspectives of the problem and the formulation of solutions (Gibbons, 2000). The PDCA, also known as the PDSA (Plan-Do-Study-Act), involves the next steps (Nicholas, 1998; Moen & Nolan, 2010: Nicolay et al., 2012):

 Plan: Define the problem and hypothesise possible causes and solutions

1. Collecting Data: Before thinking in solutions, existence of the problem and the root cause should be identified.

2. Defining the problem: Once the source and causes of dissatisfaction are known, the problem can be defined and grounded.

3. Stating the goal: Continuous improvement goals are only provisional. Goals are

23

4. Solving the problem: define all the variables and the multiple ways of achieving the goal.

5. Strategy formulation: The end of the plan stage consists of a plan which describes all the specific steps to be taken.

 Do: Implementation of the strategic plan. Monitor all changes and modify the plan to specific circumstances.  Check/Study: Collect and analyse the data and see to

what extent the goals are achieved.

 Act: Return to the plan step if the results are unsatisfactory, or standardize the solution if the results are satisfactory

Figure 9 provides an overview of the different steps. The study is sequenced according to the PDCA steps. Table 1 shows how the different chapters are related to the Phases of the PDCA approach

Step Chapter

Plan – Collecting data 3. Project Introduction Plan - Defining the problem 4. Problem Definition Plan - Stating the goal 4. Problem Definition Plan - Solving the problem 6. Literature framework Plan - Strategy formulation 7. Data Protocol

Do 8. Results

Check 9. Analysis

Act

10. How to use the improvement activities and avoid outflow in the future?

11. Benefits

Table 1. Chapters connected to PDCA

The project introduction shows the different causes of outflow and provides the data for defining the problem. The problem definition chapter defines what is researched and makes the distinction between the short and long term. The sub-questions reflect the needs for the short and long term, to achieve the goal of outflow avoidance. A literature research required is to solve the quality problem caused by human mistakes. The literature describes for the short term sub-questions which variables affect the probability of human mistakes. These variables must be included in the critical item studies. The chapter ‘Data Protocol’ describes how the study worksheet assure objective and comprehensive studies of the in the literature section defined variables. Furthermore recaps the chapter how the study worksheet must be used at HUM to get elaborated results. Hence, the ‘Data protocol’ chapter forms the strategy formulation to research critical items for the short term issues. Besides the critical item researches the literature framework how HUM can avoid outflow in the future. The long term sub-question demands a different approach and is not included in the data protocol. Instead of variables introduces the literature framework a strategic infrastructure.

24

By studying the appointed processes with the study worksheet, the required data is gathered in a database. The ‘Do’ phase is execution of the studies and categorisation of the items with the ‘Data Protocol’. An in-depth analysis is necessary to answer the short term sub-question. The ‘Analysis’ chapter checks why the items are lacking and if it is possible to apply the findings on a strategic level. The Act phase shows how to improve the critical items with implementation of the improvement activities and answering the long term sub-question; how can we avoid outflow in the future? What results in significant benefits, as described in the benefits chapter.

5.3. Analysis with the tools

Tools ensure a rational analysis for the different critical items (Choo et al., 2007). Proper use of tools provides well-grounded evidence about the efficiency and aids in analysing the possible improvement themes (Victor et al., 2000). The evidence helps to convince the different management layers of the urgency and furthermore helps achieve the improvement targets in and efficient manner.

To gather the data we use two of the seven so-called ‘improvement tools to uncover the root causes and achieve improvements’ (Nicholas, 1998). The first tool is the process flowchart; a tool used process analyses by sequentially showing all relevant steps and relation between the steps (Nicholas, 1998). Value-added and non-value-added activities are included and show where waste occurs. The process flowchart provides an overview of the process per critical item.

Cause-and-effect analysis is the second tool. The diagram is used to identify all causes for the wrong spec installation. Causes of each critical item will be defined with a fish-bone diagram. Short interviews with line-associates provide information about the different possible causes and opportunities to avoid wrong spec installation. The final tool is the usage of pictures. This helps to win buy-in from the management. Pictures of each cause shows what is going on and why it does not contribute to the Honda’s zero defect quality mission. Hence, all root-causes per critical items need to be displayed and documented to gain the trust of the management.

5.4. Application of methodology

25

6.

Literature framework

The problem definition chapter defined the problem and stated the sub-questions. This chapter creates a better understanding of what critical items are, why human mistakes are made and how we can avoid outflow in the future. The literature framework shows which variables influence the sub-questions and determine which variables will be included in the research for critical items. Modification of the variables with HUM’s existing approaches ensures suitability of the literature variables in the practice. Sub-questions 2 and 3 are both related to detection of human mistakes literature. Hence, these sub-questions are clustered together and named SQ2 & 3: Mistake detection

6.1 SQ1: What are the critical items in AF2?

Car recall researches indicated several potentially important attributes as nationality of the manufactory, defect severity and vehicle age. The different attributes affects the owners urgency to solve the specific recall case (Hoffer et al., 1994). In addition, Rupp (2004) researched which items are especially costly for the car companies and cause shareholder losses. This research proved the significant differences between the different item groups and assign products that might cause serious injuries or fail to meet the standards (Mullan, 2004).

Further knowledge of HUM’s recall policies is needed for identification of critical items. An interview with the Marketing Quality (MQ) manager about the current recall policy provides information to tailor the groups and criteria. HUM’s MQ department defined standards to ensure communication and progress of the recall investigation. The MQ information enables analysis of the critical items and implementation of the necessary countermeasures.

Detailed information of MQ’s responsibilities and Honda’s market quality information flow is provided inAppendix II; Examples for Grade Judgement of Market Quality Information. This research investigates not the quality of the MQ’s policies. Our aim is to avoid any car recall. This sub-chapter describes how to define the critical items for car recalls, which need to be researched before Christmas 2012.

6.1.1 Item grading

Part of the standards is the grading guide. The grading guide distinguishes A, B and C grades, which are applicable on all items. An item is A graded when the product could cause serious effects in term of:

1. Safety

2. Regulatory requirements 3. Environment

4. Future sales of products

Examples of A grades are similar to Rupp’s (2004) definition of critical items and results in: - Personal injury, accident or fire explosions

- Unexpected acceleration or sudden engine stop - Insufficient braking or uncontrollable steering - Unable to exit the vehicle safely

- Violation of certain government regulations as air pollution and safety

26

B grades affect the functional performance of the car and are easily identifiable by most customers. Criteria under normal usage could be:

- Care driveability is reduced but an accident is avoidable

- An issue which not result in an A grade symptom, but does affect an functional part of the cars performance

- Cosmetic defect which is easily detectable and would be obvious to most customers

Any issue with a product which does not belong to grade A or B and is judged to be either a cosmetic concern, a customer wish or opinion which is outside of the original design intent for the vehicle is defined as a C grade. C grades in general reflect a cosmetic concern which is judged to be only detectable by a very small percentage of customers.

Only A grades cause car recalls and are critical. Hence, A grades threatens directly Honda’s reputation and might result in large indirect costs. Therefore, grading the items will help to study the parts which are directly involved in car recalls.

The different items are grouped into different categories to ensure correct grading. MQ uses eleven different item groups, which groups are necessary to determine the possibility of wrong spec installation and miss assembly for the AF department. The groups and corresponding definitions of a grades are given in Appendix II; Examples for Grade Judgement of Market Quality Information.

6.1.2 Product features

After allocating the grades, it is possible to filter the items on the product features. The product’s feature determines if it is possible to fit the item wrong. How many specs have the item? How easy is it to fit the item wrong? Hence, the easiness of wrong installation depends on the ‘product feature complexity’.

The devices used for mistake proofing the product features are called ‘Pokayoke’ (Nicholas, 1998). Pokayoke has two broad functions in quality management. The first kind of pokayokes is ‘regulatory pokayokes’. ‘Warning pokayoke’ gives the associate a signal when it detects an abnormality, where the ‘control pokayoke’ shuts down the operation in case of wrong spec or miss assembly. ‘Setting pokayokes’ are devices that guarantee correct setting of the process. Stewart and Grout (2001) consider pokayokes as a method to reduce ‘Product Feature complexity’. Pokayokes can substitute the selection process of the spec and/or correct it automatically it wrong installation occurs.

6.1.3 The variables at HUM

27

6.2 SQ2 & SQ3: Mistake detection

The problem definition chapter shows the importance of detection to avoid human mistakes. However, which variables influence the effectiveness of detection in this research? As stated by Nicholas (1998), it is not possible to prevent all mistakes. Informative inspection is a form of mistake detection and created to reduce the amount of defects. Mistake detection alarms the associate in self-inspection and successive-inspection. As mentioned in the previous chapter, self-inspection is liable to subjectivity, emphasising the need of an objective successive-inspection which provides a more rigorous inspections. Self-inspection ensures quality control at the source (Anand, 1999). Both forms of inspection involves an associate who is in control can be held responsible if: (1)Has the correct information – knows what to do, (2) Receives feedback – knowledge of what he/she is doing and (3) Make correction whenever the process goes wrong

Advantages of well implemented successive inspection are (Harmon, 1992): (1) Successive inspections performed by following associates are free and automatic. No additional resources are required. (2) Rate of quality increases if a 100% inspection is applied. The associate detect and correct the mistakes which are caused by a lack of concentration of the first associate. (3) Usually defects in the previous process interferes the task of the next associate. The next associate cannot execute his tasks until the defect is corrected, what ensures a compulsory appraisal. (4)It is more reliable because of the objective approach

One of the disadvantages of successive inspection is the long(er) feedback loop. In some cases, the actual circumstances of wrong spec installation are vanished when the feedback comes to the responsible associate (Anand, 1999). Short feedback loops activates quick action. To determine the appropriateness of a successive inspection is location research required.

Another distinction is sensory and physical inspection (Shingo, 1986). Sensory inspections are performed by the human sense. Sensory inspection is liable to subjectivity, since the associate judges the product with his own sensors and own tolerances. Physical inspection is more reliable inspection and relies on measurement devices. Measurement uses numerical values provided through the measuring devices. Items are accepted if the values are equal to the pre-set tolerance rate.

Self and successive-inspections are desired to guarantee a 100% human mistake detection (Anand 1999). Important is the positioning of the feedback and the action points. Where are they situated and how long is the feedback loop in the current situation?

Another subject which should be incorporated in the data protocol is process interference. Are succeeding associates able to fulfil their tasks according to the operation standards? If not, where in the line is the process interfered? Is it impossible to fulfil the task or is it still possible to execute the task with some extra effort? Is the interfered associate assigned as successive check and what are standard activities if he/she detects a defect?

6.2.1 The variables at HUM

28

6.3 SQ4: How can we reduce complexity

As described in the problem definition chapter ensures complexity reduction mistake prevention. Mistake prevention tools support the associate in making the right decisions. However, what is the complexity for each critical item and how can we reduce it. Clear definition of complexity and complexity variables enables identification of the critical item’s complexity concerns.

6.3.1 Complexity

Honda’s high degree of diversity ensures a high variety of components, production processes, and other requirements. The variation is the considered main causes of the product complexity. Processing products with high complexity increase yhe set-up costs, need for materials, WIP and moreover difficulty in balancing the assembly line process (Orfi et al., 2011).

Huang and Inman (2010) argue why product complexity is just a part of the total ‘plant build complexity’. The second driver is Marketplace complexity (e.g. Russian cars are supplied with an extra coolant label in the glove box). Supply chain complexity is the third drive and comprehends supply chain reliability. The fourth and last driver deals with the number and similarity of products. As Diagram 10 shows, this research focuses on complexity caused by product complexity and number & similarity of products in the plant.

Diagram 10. Drivers plant build complexity

The four drivers influence the three main elements of plant build complexity: 1. Task complexity, e.g. associates work instructions

2. Equipment and Facilities complexity, e.g. appropriate battery tool 3. Management coordination complexity, e.g. balanced workload

29

fundamental work elements - direct relation

fundamental work elements - indirect relation

2 Search 1 Transport empty

3 Select 4 Grasp

7 Position 5 Transport loaded

8 Assemble 6 Pre-position

9 Disassemble 10 Release load

13 Inspect 11 Use

16 Plan 12 Hold

14 Avoidable delay 15 unavoidable delay 17 Rest to overcome fatigue Table 2. Relation fundamental work elements - quality

6.3.2 Variables and influence on complexity

The work elements and their relation to quality are determined. However, the relation between work elements and complexity is required to prevent human mistakes. One of the complexity related variables is process visibility (Norman, 1988).

Usually, information availability is not the problem. However, the communication of information is not always as effective as it should be (Bilalis et al., 2002). Every associate must be able to understand the different processes at any time. Moreover enables a transparent process immediate feedback of the current status (Womack & Jones, 1996). Thus the first variable is the accessibility of information and process.

Variable two and three are related to simplification (Nicholas, 1988). Nicholas (1988) defines simplification as: ‘means accomplishing the same ends but in a less complex, more basic way or with fewer systems. Simplification also means cutting out or cutting down on features which do not add value.’ Simplification can be distinguished in three groups of simplification; (1) Product, (2) Process and (3) Procedure. The short term research is on ‘Process’ and ‘Procedure simplification’. ‘Product simplification’ is out of scope due to the limited time and budget for a design change. Process simplification objective is to create a process flow with the least number of processing step per associate or zone (El-Halwagi et al., 2012). Procedures are defined as all used elements alongside the production line, except for the installed parts. Simplification of the procedures results in minimising the complexities of the physical constraints (White et al., 2010)

Fourth complexity reducing variable is Process balancing (Nicholas, 1998). The assembly process is paced by a chain or roller conveyor, which transports the frame through all assembly station. The processes have a constant takt time of 111s. Despite the constant chain pace, the task might require less or more time than the takt time. Honda is wasting resources if the associate’s capacity is not fully utilised. On the other hand, if the process activities take more time than the takt time, the associate get. Rushed associates make more mistakes (Huang and Iman, 2010). Hence, it is necessary to streamline the process and remove the noises in the process flow.

30

the associates are connected. This rule prevents grey zones in decision making. It creates a supplier-customer relationship between each individual. Linking the five variables to the three main elements of complexity provides a short summary of the complexity sub-chapter so far (Table 3)

Element of complexity Complexity reducing variable

Equipment & facilities A. Information and process accessibility B. Simplification of procedure

Task C. Simplification of process Management coordination D. Process balancing

E. Standardisation

Table 3. Relation element of complexity – complexity reducing variable

6.3.3 The variables at HUM

The five defined complexity reducing variables are the core for critical item research. However, the variables are too abstract and not applicable for the research. Thus, we translate the abstract variables into HUM critical variable. This is done after process observations and conversations with HUM’s general associates, quality team and comparisons with the literature. The five complexity reducing variables are dissected in 20 criteria. Definitions of the criteria are provided in Appendix III; Definition of critical variables. Table 4 shows an overview of all critical variables in relation to elements of complexity and the complexity reducing variables

ELEMENT OF COMPLEXITY COMPLEXITY REDUCING

VARIABLE

CRITICAL VARIABLES

I. Equipment & facilities

A. Information and process accessibility 1. Visual aid 2. Area Layout 3. Spec sheet 4. Build Sheet 5. Racking Ident

6. Part packaging (Ident) 7. Barcode on part B. Simplification of procedure

8. MTO change points 9. Racking layout

10. Part packaging (Material) II. Task C. Simplification of Process

11. Part decanting 12. Process flow 13. Walk path

III. Management coordination

D. Process balancing

14. Process time 15. MTO change point 16. Density

E. Standardisation

17. Part segregation 18. Stillage movement 19. Part decanting

20. Associate Checks spec Table 4. Overview variable criteria Honda

6.4 SQ5: How can we use the quality improvement projects and

avoid outflow on the long term?

31

major pitfall is ignoring the need for a CI infrastructure that provides organisational learning and dynamic capability (Zollo and Winter, 2002). Dynamic capability is: ‘a learned and stable pattern of collective activity through which the organisation systematically generates and modifies its operating routines in pursuit of improved effectiveness.’ (Anand et al., 2009). The CI infrastructure ensures sustaining and coordination of the organisation’s learning efforts towards systematically improving processes. Figure 10 shows the coordination and execution of projects to get the sustained CI.

Figure 10. Sustained CI initiative

6.4.1 Continuous Improvement Infrastructure

The project involves different kind of associates, who all contribute to outflow reductions. These associates need to be encouraged to stay in touch with the project and continuously improve the methodology. The most comprehensive kind of structures are ‘purpose-process-people’ management frameworks, facilitating proactive improvement on the shop floor and middle-management, while securing strategic congruence (Teece et al., 1997). As showed in Figure 11 are the purpose, process and people issues separated in eight topics. The answer of this sub-question defines the eight topics for HUM and shows how HUM can use it.

32

7. Data protocol

This chapter describes the data protocol for gathering data of the critical items. Firstly, a Quality Function Deployment process is defined. This process shows how the sub-questions are captured in the data protocol. The study worksheet forms the backbone of data gathering and includes the in the literature framework defined variables. The required data is gathered through a structured interview, interaction with line associates and observations. The data is input for the database, which translates the data into qualitative data with the audit judgement matrix.

7.1 QFD process

The data protocol ensures an objective and comprehensive study approach for the critical items. The literature framework defines which variables affect the quality of the process. With the Quality Function Deployment (QFD) process are the 4 short term sub-questions and its variables captured. The QFD process is a sophisticated process, revealing all critical attributes through matrices at different levels. The first matrix and core of the QFD (Nicholas, 1998) is the ‘House of Quality’ (HoQ). The HoQ translate the customer requirements into the engineering goals (Liu, 2010). The customer requirements are converted into MQ requirements to fit the HoQ into this research. Thus we assume MQ plays a role as extension piece for the customer. The engineering goals are translated to the 838 Civic’s items. Which items are needed to build the car and how are they scoring according to the MQ policies? Hence, Execution of the HoQ results in the critical items and answers the first part of SQ1: Which items are critical?

The Succeeding ‘Features matrix’ ranks the critical items by comparing them with the item’s product features (Nicholas, 1998.) As part of SQ1, the design matrix shows which critical items are uncontrolled based on the item’s features. The uncontrolled items require immediate improvement and so adequate resources. The result chapter is defines when the item is uncontrolled on its features

All the critical items must be controlled on complexity. The ‘Complexity matrix’ contains all complexity affecting variables and shows which items score an insufficient on complexity. The measured scores highlight attributes that needs reduction of complexity and provides the data to solve SQ4; How can we reduce the complexity for each critical item and prevent mistakes?

The last matrix is the ‘Detection matrix’. The matrix measures the relation between the critical items and detection of mistakes. Items with high scores need extra inspection

safety D ri ve s h af t R ad io M od u la to r La b el s H ea te r Steering Engine Etc.

House of Quality

Can fit wrong