Scientific Research and Essays Vol. 6(3), pp. 580-591, 4 February, 2011 Available online at http://www.academicjournals.org/SRE ISSN 1992-2248 ©2011 Academic Journals

Full Length Research Paper

Applying six sigma to manufacturing processes in the food industry to reduce quality cost
Hsiang-Chin Hung* and Ming-Hsien Sung
Department of Industrial Engineering and Management, I-Shou University, Taiwan.
Accepted 13 January, 2011

In recent years, an increasing number of companies have used different types of quality programs in order to increase internal and external customer satisfaction as well as to reduce quality cost. Among all of these programs, Six Sigma is perhaps the most widely-accepted initiative by all a broad range of organizations. The DMAIC (define-measure-analyze-improve-control) approach has been followed here to solve an underlying problem of reducing process variation and the associated high defect rate. This paper explores how a food company in Taiwan can use a systematic and disciplined approach to move towards the goal of Six Sigma quality level. The DMAIC phases are utilized to decrease the defect rate of small custard buns by 70% from the baseline to its entitlement. At the beginning of this project, the defect rate was 0.45% (Baseline), and after the improvement actions were implemented during a six-month period this fell to below 0.141% (goal). The critical successful factors for Six Sigma projects, especially those in the food industry, are discussed at the conclusion of this paper. Key words: Six sigma, food industry, process improvement, DMAIC. INTRODUCTION Since the early 1980s, manufacturing industries worldwide have seen a revolution in the way they operate. Consumers have become more and more demanding, and the key to firm survival is the recognition of the importance of customer satisfaction. Consequently, companies have been forced to enhance the quality of both their processes and products (Efstratiadis et al., 2000). The focus of this study, the food industry, has also become increasingly multifaceted and competitive in recent years (Chong et al., 2001; Knowles et al., 2004; Henchion and McIntyre, 2005; Spiegel et al., 2006; USOCDD, 2007). In this environment, food company managers have to deal with a number of problems. Sales are slowing down and operating costs are increasing, while customers are becoming more demanding and selective (Efstratiadis et al., 2000; Henchion and McIntyre, 2005). Food industry managers must thus consider how to maintain profitability in a shrinking market, while providing increasingly sophisticated customers with high quality products and efficient service. In attempting to achieve this seemingly impossible objective, firms can pursue two strategic avenues. First, they can focus on ways to improve the operational efficiency of the system. Second, they can take actions to enhance its operational quality. In recent years, an increasing number of companies have used different types of quality programs in order to increase internal and external customer satisfaction as well as to reduce quality cost. Process improvement has often been accomplished through an integrated approach, using problem-solving techniques such as total quality management (TQM) and classic statistical analysis (Wiklund and Wiklund, 2002). Among all these programs, Six Sigma is perhaps the most widely-applied. There are many documented case studies of organizational applications of Six Sigma, where large-scale improvements in defect and variability in processes to meet the customer satisfaction. (Antony and Banuelas, 2002; Raisinghani et al., 2005; Chen et al., 2005; Antony, 2008; Chung et al., 2008; Nonthaleerak and Hendry, 2008; Leea et al., 2009). Six Sigma, a statistically-based quality improvement program, helps to improve business processes by

*Corresponding author. E-mail: seanhung@isu.edu.tw. Tel: 886-7-6577711. Fax: 886-7-6578536.

Hung and Sung

581

reducing the waste and costs related to poor quality, and by improving the efficiency and effectiveness of processes (Breyfogle, 1999). Ultimately these measures should lead to improved customer satisfaction and increased profitability (Antony and Banuelas, 2001). Spiegel et al. (2006) indicated that the company shall select and implement specific quality management activities suitable to their situation to increase their production quality. As a business method for eliminating defects, Six Sigma also works well in the food industry. For example, executives at fast-food giant McDonald's started to learn about the management philosophy that General Electric (GE) adopted in the 1990s, and since then, managers have continued to attend GE’s program and initiate related projects (Lee, 2005).
MATERIALS AND METHODS Six Sigma Six Sigma as an improvement program has received considerable attention in the literature during the last few years (Harry, 1998; Hoerl, 1998; Breyfogle, 1999; Bergman and Kroslid, 2000; Hellsten and KlefsjoÈ, 2000; Klefsjo et al., 2001; Chen et al., 2005; Thomas, 2008; Zu et al., 2008; Leea et al., 2009; Kytösaho and Liukkonen, 2009). Motorola launched Six Sigma methods in 1987, and was also the first firm to win the Malcolm Baldrige National Quality Award (MBNQA) in 1988. Today, other companies like Texas Instruments, ABB, AlliedSignal, GE and 3M have been striving to achieve Six Sigma quality, and as a result they have become known as best-in-class companies (Fuller, 2000). Six Sigma is a useful problem-solving methodology and provides a valuable measurement approach. It has a statistical base and with proper utilisation of methodologies can help to improve the quality of both product and process. In addition to providing data-driven statistical methods for improving quality, Six Sigma also focuses on some vital dimension of business processes, reducing the variation around the mean value of the process (Kanji, 2008). At many companies, Six Sigma simply mean a measure of quality that strives for near perfection. It is a disciplined, for eliminating defects in any process, covering manufacturing and transactions, as well as products and services. The fundamental objective of the Six Sigma methodology is the implementation of a measurement-based strategy that focuses on process improvement and variation reduction through the application of specific projects. This is accomplished through the use of two Six Sigma sub-methodologies: DMAIC and DMADV. The DMAIC (define- measure- analyze- improve- control) is an improvement system process for existing processes falling below specification and looking for incremental improvement, and the DMADV (define- measure- analyze- design- verify) apply to the product development and design at Six Sigma quality levels (Linderman et al., 2003). It is a myth that Six Sigma works only in large companies. Six Sigma has evolved into a business strategy in many large companies and its importance in small and mediumsized enterprises (SMEs) is growing everyday (Kumar and Antony, 2008). In fact, the results are quicker and much more visible in smaller companies than in larger corporations (Antony, 2008). The DMAIC approach has been followed here to solve an underlying problem of reducing process variation and the associated high defect rate. There are many case study apply this methodology to solve the company’s underlying problem (Chen et al., 2005; Nonthaleerak and Hendry, 2008; Leea et al., 2009; Kytösaho and Liukkonen, 2009). The DMAIC is a process improvement cycle of Six Sigma program as well as an effective problem solving

methodology. Brewer et al. (2005) indicated that DMAIC is “the primary framework used to guide Six Sigma projects”. Six Sigma projects are based on the DMAIC approach and “the role played by DMAIC in gaining the overall success of Six Sigma is equally critical” (Nilakantasrinivasan and Nair, 2005). In this paper, we are going to adopt the DMAIC quality improvement process. The phase of Six Sigma implementation In order to reduce process variation and the associated high defect rate, Six Sigma focuses on improvement methodology application, then the DMAIC is mentioned most frequently now and a lasting improvement method (Starbird, 2002). The representative's meanings of five English letters are as follows: Define The top management shall identify the problem according to customer feedback, strategy and mission of company, define customer requirements, and set goal. Measure Measurement is a key transitional step on Six Sigma road, one that helps the project team refined the problem and being the search for root causes which will be the objective of Analyze step in DMAIC. Therefore, the project team needs to validate problem/process, refine problem/goal, and measure key steps/input. Analyze In analyze stage, the project team shall use data analysis tools and process analysis techniques to identify and verify root causes of the problem. For the reason, the project team needs to develop causal hypotheses, identify vital few root causes, and validate hypothesis. Improve The goal of the improve stage is to find and implement solutions that will eliminate the causes of problems, reduce the variation in a process, or prevent a problem from recurring. So the project team needs to develop ideas to remove root causes, test solutions, and standardize solution/measure result. Control Once the improvement has been made and results documented, continue to measure the performance of the process routinely, adjusting its operation. It is very important for the project team needs to establish standard measures to maintain performance and correct problems as needs. Without control efforts, the improved process may well revert to its previous state. Case company The case company was founded in the early 1970s, based on exporting frozen prepared eel food products. It then shifted its focus into the domestic wheat flour processed foods and snack market, and successfully launched a range of pork buns. Since 1985, the case company had been certified as one of Taiwan’s Good Manufacturing Practice (GMP) food companies, and has attained

582

Sci. Res. Essays

0.02 0

100 80

0.01 5

Defective rate

60 0.01 0 40 0.00 5

t n e c r e P

20

0.00 0 Prod uct t ype Defect ive r at e P ercent Cum %

Pork b un 0.008533 41. 6 41. 6

Plain b un 0. 008080 39. 4 80. 9

Dum pl ing 0. 0 02679 13.0 94.0

Handmade b un 0.001236 6.0 100.0

0

Figure 1. Pareto chart of product type.

several other international food processed certifications. Although the employees’ general level of education is not high, the case company still decided to implement some Six Sigma projects in order to reduce its operational cost, to improve its financial performance and to better face an increasingly competitive market situation. A Six Sigma committee was established to facilitate the implementation process, including building infrastructure, proposing and selecting projects, tollgate review and decisions related to rewards. As there were many aspects to be improved, the projects were first prioritized systematically and two projects were selected in the first year. Champions, usually the leaders of all the departments and supporting units, were then in charge of monitoring progress and ensuring the success of each selected project. Every year, all the champions had to propose some candidate projects according to their department’s KPI (key performance index) and submit them to the committee to be approved. Several important issues, such as improving customer satisfaction, reducing the A.R. (accounts receivable) collection cycle, lowering product defect rates, reducing recruit cycle time and shortening new product development time were discussed in the committee. Every possible alternative was prioritized by using the C&E (cause and effect) matrix. Criteria being considered to select the most critical projects included the effects on KPI, impact on customers, data accessibility, project hard savings and the time needed to reach the improvement goal. Among all these candidate projects, lowering product defect rates was considered the most critical, as it is highly correlated to KPI and customer impact. During the frozen bun manufacturing process, buns must be steamed thoroughly and then frozen to be stored and transported. The buns are then steamed again by the customer before eating. However, during the re-steaming process, customers encountered some problems with the product such as shrinkage, foreign material, crack, and so on. Consequently, a project to address this issue was eventually approved by the Six Sigma committee. We acted as consultant to coach the project team members to implement project.

RESULTS AND DISCUSSION Define phase After this project was approved, the champion of this project had to select an appropriate Black Belt and form a

Six Sigma team to deal with the improvement. Three tasks must be undertaken during the define phase: Refining feasible project scope, setting up project goals and estimating project hard savings. Because the time period of each Six Sigma project is limited to not longer than six months, a suitable scope for each project is very important if they are to be successfully completed on time. By using a three-layer tree diagram (Figure 4) and three Pareto charts, the project scope was narrowed down. The first layer narrow down was to decide what product defect rates should be improved. Because buns encountered the most serious defects among all the company’s frozen food products, it was chosen to be the project product (Figure 1). Although there were numerous defects with the buns that could have chosen to work on, it is important that any project does not become unwieldy or too complicated. Therefore, the main customer complaint, that the buns had a tendency to shrink, was chosen as the focus of the project in the second layer narrow down (Figure 2). However, the company produces a range of different bun, all which with shrinkage rates, and thus the third layer narrow down decided to put the focus of the project on the firm’s 32 g small custard buns, as it was the most prone to this problem (Figure 3). After the project, scope was specified clearly, various goals needed to be set, namely the project, financial and consequential indices. The project index is also called the primary index, which indicates the measurement and goal of the improvement target based on related time series data. In this case, the primary index was shrinkage defective of 32 g small custard buns, and the goal of this project was to reduce the shrinkage defect rate. Time series data for the shrinkage defects of the bun was collected and is shown in Figure 5. It shows that the defect rate was low in the beginning and then began to rise. The average defect rate was 0.405%, which forms the project baseline. The best situation, also called the entitlement of the project, was 0.028%, which happened

Percent

gv g g d b Bc v b a ` BBXY

dv ` b `v Y b dv q s av b q bv d q bv c ` cv a a cv Y Y a B€a Bx` Bx` B€g Bya Bxg BBwc XBY v d v q v q v q v s v s v d d v Y b c s d BBBpBBBrBBBrBBBbpBBBiBX¦BdhBX¦BeB¦BBa a ` b s b b g b c q d ` b a b a c f g g a f c d Y Y d c

g ‰f g ‘’‘’ ‘’   “‘” •• •• ‰Y g g ˆ` g ˆs g g Bˆd

      

  7  7  7  7  7 

¢¢ ¢¢ §¡ ¢¢ ¢ ¥¥ ¥¥

Shrinkage Foreign Appearance Cracks material

¡¢

leakage

¢4

¢

0 ( ( 0¢ ¥ 0¡ ¢§

" ¢ 0 ) 01 ¡

¢



                

Custo mer

Size

3 2

1%

60

) 5 0¢

¥

61 ¢ ¡

' 60 5 ¥1 3%  ¥ 1 0 ) ( " &§ ¡ ¢ " ' &% ¤ § ' 

§ 5 § 0¡

Frequency

  © © 

$    

       

©      

       

Figure 3. Pareto chart of bun type.

„ ƒ ‚  T u H S A ¦GGDGH t H U S H H I H F E C A 9 XW VT GRQ BPBGBDB@8

g g g g g BBBXf g g g g BBBXY g g g g BBBB` g g g g BBBBs g g g g g BBBBBd g g g g BBBXf¦d

Figure 2. Pareto chart of customer complaint.

       ©

# " ¤ !  ¥ ¢ §¡ ¢  ¨ § ¦¢ ¤ £ ¢ ¡ ¥

Shrinkage 26%

65G sesame

Dumpling 0.2679%

Foreign material 22.8%

32G small custard

Plain buns 0.808%

Product defective rate

65G meshed taro

Appearance 19.5%

Buns 0.8533%

Crack 14.6%

40G taro

Leakage 9.8%

…

–

in January. The objective of this Six Sigma project was to
Figure 4. Project selection tree diagram.
H ‡ T S ‚ I … BX† BT BBBDA t

decrease the defect rate by 70% from the baseline to
Handmade buns 0.1236%
Others

Hung and Sung

Others 7.4%

583

its

Shrinkage defective

Frequency

¢££ ££ ¤¤ ¤¤ ¢¢ ¥¢ ¢ ¨§

Percent

584

Sci. Res. Essays

Shrinkage rate
Ž  Š ŽŽ Ž  ‘‘ ˆ Š‰ ŒŒ ŒŒ ‹‹ ‹‹

Figure 5. Time series data for shrinkage defect rate of 35 g custard bun.

Stuffing Flour

Frozen dough sheet rolling Forming

Inspect Rework

Weighing

Dough stirring

Steaming

Ferment

Screen

Arrangem ent

Gage checking

bottom paper placing

Screen

Pre-freeze

Freeze

Weigh/appear ance screen

Package

Rework Outgoing Frozen Outgoing checking Weigh/package checking

Figure 6. Process flow diagram of custard bun.

entitlement, and the goal was set at 0.141%. After the project goal was set, financial index was calculated accordingly, assuming that the goal could be achieved. Finally, the consequential index considers side effects that may occur when attempting to reach a project’s goal. For example, a longer cycle time is one possible side effect in a defect rate reduction project, and this should thus be monitored. Measure phase The major activity in the measure phase is to understand

the whole situation of the project, including process mapping, defining potential factors (also called Xs) that affect the project index (also called Y), measurement system analysis (MSA) and process capability analysis. Generally, a process flow diagram (PFD) is used to discuss the related process flow of a project, and then a detailed process map to consider inputs and outputs of each process step is prepared for the subsequent cause-and-effect analysis. Figure 6 shows the process flow diagram of the custard bun product. Based on the PFD and detailed flow diagram, the project team members discussed factors that may cause a custard bun to shrink after re-steaming by using

› i u •jl s i š i u &”•eee~ P~ l ž  “ ‚  

— › jl u m™l jw &”•eee„ P~ š ˜ – “ ’ ~ 

s w › je˜

œ &”•7P7€ P~ “ € „ 

d ˜ e™—

i g jhf

jmk n l

q p eho

s p ejr

‡ u ˜ Gi †…v

u p emt

y w zxv

{ en

o

w { xv

}jl |

u jwjr

7P~ P~  €  7P~ P~ ~ €  ee~ P~ ‚ ~  ee~ P~ ƒ ~  ee~ P~ „ ~  ee~ P~  ~  ee~ P~ ~ ~ 

Hung and Sung

585

Figure 7. Fishbone diagram of shrinkage defective.

Table 1. Attribute gage agreement analysis plan.

Method Sampling Three inspectors looks at 30 sample buns, with two inspections for each sample. Randomly sampled three from 12 inspectors Buns were inspected to see if they shrank or not 15 shrinkages and 15 standard buns

brainstorming, and summarized all the factors in the cause-and-effect diagram (also called a Fishbone diagram, because of its shape). Factors were further categorized into three kinds, major, medium and minor. Figure 7 shows the Fishbone diagram of this project. A very important characteristic of a Six Sigma project is that improvements are made based on data analysis, usually statistical analysis, so the reliability of the data collection system is very important. If the data does not precisely represent the true situation, then the results of the analysis would be useless. There are two kinds of measurement system analysis, the use of which depends on the type of collected data, gage repeatability and reproducibility (also called GR&R) analysis for metric data, and attribute gage agreement analysis for attribute data. In this project, the shrinkage defective was judged by an inspector and the outcome is binary data, so attribute gage agreement analysis was performed to verify the reliability of the measurement system. To perform the agreement analysis, three inspectors were selected

randomly among the complete group of 15. Thirty small custard buns, 15 good and 15 defective were arranged randomly to be judged by the inspectors twice; Table 1 shows the agreement analysis plan and Table 2 shows the results. The final screening effective score vs. attribute was 83.33%, as shown in Table 2. The score shows the probability that all three inspectors made the same correct assessment of a bun, giving an identical attribute value. This score is higher than the recognized standard for an attribute gage agreement analysis of 80%, so the conclusion is that inspection for shrinkage defects is suitable for further analysis. The next job in the measure phase is to explore potential influential factors for bun shrinkage. From the process flow diagram and the cause-and-effect analysis result, numerous factors from the related process were scrutinized using the cause-and-effect matrix (also called an X-Y matrix), which weighs each factor by three indexes: Data collection difficulty, impact on defect and controllability of factors. Table 3 shows the X-Y matrix.

586

Sci. Res. Essays

Table 2. Attribute gage agreement analysis result.

All operators agree within and between each other Sampling Operator #1 Sample Attribute Trv#1 Trv#2 1 y y y 2 y y y 3 n n n ˙ ˙ ˙ ˙ ˙ ˙ ˙ ˙ 29 n n n 30 v v v (1) Appraiser score (%) 100.00 (2) score v.s attribute (%) 100.00 Operator #2 Trv#1 Trv#2 y y y y n n ˙ ˙ ˙ ˙ n n v v 100.00 100.00 Operator #3 Trv#1 Trv#2 y y n n v v ˙ ˙ ˙ ˙ y y v v 100.00 83.33
(3)

All operators agree with standard Y/N agree Y N Y ˙ ˙ N Y

Y/N agree Y N Y ˙ ˙ N Y

Screen effective score (%) (4) Screen effective score vs. attribute

(%)

83.33 83.33

Table 3. X-Y matrix.

No 1 2 3 4 34

Process steps Weighing

Priority factor

Data collection Impact on defective Controllability 5 1 1 5 3 5 5 1 3 5 5 5 7 1 1 3 3 1

Total 17 27 71 61 57

Weight Quantity Dough stirring Color difference Dough sheet rolling Thickness Outgoing Shift

Nine potential factors were selected, including: color of the dough, number of times the dough sheet is rolled, temperature of stuffing, ferment time, steaming pressure, volume of ice water input, steaming time, type of steaming box and production shift. Analyze phase In the analyze phase, two types of analysis were performed. One was the analysis of variables for which data can be readily collected, and then statistical analysis was used to test whether these factors (Xs) had an influence on the project’s index (Y) or not. The other analysis was for variables that data are hard or even impossible to collect, and for which the failure mode and effect analysis could be performed. For the statistical analysis, the small frozen custard buns were sampled one from each containing plate, which contained 49 buns, 30 times during a two-week data collecting period. Each bun was re-steamed to evaluate whether or not it would shrink. Figure 8 sows the main effects plot for each factor with regard to shrinkage. Among these, type of steaming box

and steaming pressure were collected at only one level in that period. For all the other factors, Color difference of dough, number of times for dough sheet rolling, ferment time, volume of input ice water, steaming time, and shift seemed to be influential to the defect rate. The same data were analyzed by statistical method ANOVA (Analysis of Variances). Table 4 shows the tested result. Except for X4 and X5, 5 factors were concluded significant. Failure mode and effect analysis (FMEA) is the task of finding possible faults in a system and evaluating the consequence of the fault on the operational status of the system. It’s a method of reliability analysis intended to identify failures which have consequences affecting the functioning of a system. In other words, FMEA can be explained as a group of activities intended to recognize and evaluate the potential failure of a product or process and its effects, to identify actions that could eliminate or reduce the chance of the potential failure occurring, and to document the process. In a FMEA, manufacturing steps are listed by project members and all potential failure modes are identified in each step. All possible failure modes are then discussed with their respective effects and the severity (SEV) regarding every effect, the cause

Hung and Sung

587

Mean of shrinkage rate

Figure 8. Main effects plot for each factor to shrinkage.

Table 4. ANOVA tests for each factor.

Variable Variable name X1 X2 X3 X4 X5 X6 X7 X8 X9

Conclusion

Stuffing temperature P=0.043