Making the Case for Quality

December 2012

Thai Brewery Deploys QFD Tools

to Tap Into Consumer Motivation
by Theera Vongpatanasin and Glenn H. Mazur

As Thailand has become an increasingly popular tourist destination, affection for the country’s food
At a Glance . . . and beverages has been shared by returning travelers, resulting in three ­phenomena: a growing number
of Thai restaurants worldwide, experimentation with Thai cooking at home, and a growing demand for
• Boon Rawd Brewery, the Thai ingredients in stores.
maker of Singha beer, saw
an opportunity to better Consumer demand for Thai products has been a big boon for Singha, which has led to a global network
understand its American, of distributors. To ensure continuing success, Boon Rawd Brewery, the maker of Singha, and Japan
British, and Australian
Business Consultants, Ltd. used quality function deployment (QFD) tools to research the company’s
customers as demand
for Thai food products in non-Thai customers to better understand their interests, motivations, self-image, and other character-
these market segments istics. Their findings were used to tailor marketing messages for Boon Rawd’s American, British, and
continued to boom. Australian customers.
• The company partnered
with Japan Business The researchers used the QFD-related tool kansei, a Japanese-created approach to identify product
Consultants, Ltd. and ­sensory attributes that affect consumer emotions and image, and analytic hierarchy process (AHP),
deployed QFD tools using a U.S.-created method to determine priorities based on human decision-making modes.
a five-step approach to
uncover the thoughts,
beliefs, feelings, and About Boon Rawd Brewery
perceptions influencing the
consumption of its products. Boon Rawd Brewery, owner of the Singha beer brand, is a
• The team’s findings were
privately held, family-owned business that was founded in
used to craft marketing 1933. It is Thailand’s first brewery and in 1939 it became
messages that tap the only Thai beer producer granted the country’s Royal
into consumers’ inner Permission. The Royal Warrant and the designation “By
psyches and began to Royal Permission” still grace the neck of every bottle of
appear in print ads
Singha beer. The brewery also produces non-alcoholic
beginning in 2009.
soft drinks.
Singha beer is available in over
50 countries.
Methodology

While traditional QFD tools have enjoyed longtime use among product developers1, the modern tools
in Blitz QFD®2 add deeper analytic insight into the unspoken needs of the customer. In this project,
tools such as the house of quality were unnecessary and were omitted in favor of adaptations of some
of the Blitz QFD® tools.

© QFD Institute. Published with There is no “one size fits all” way to deploy QFD, since each company, its management of the product
permission from QFD Institute. development process, its customers, technology, and other factors present unique challenges. QFD

ASQ www.asq.org Page 1 of 6

 Figure 1—Custom-tailored QFD process for Singha beer

Analytic
1 2 3 Hierarchy hierarchy
diagram process
Gemba visit Customer voice Affinity
4 5
table table diagram Structured High-value
Clarified Customer customer customer

Priorities
items image images images Marketing
Go to gemba Items Image
collateral

Gather the “voices of Analyze the Structure the Analyze Prioritize

your customers”: customer customer images customer image customer
what they say and do self-image structure images

should be tailored to fit the company. The unique approach used often added in custom tailoring. In this case, the purpose was to
in the Singha beer project is depicted in Figure 1. note the best time and place to conduct the gembas. Unknowns
were also identified and explored in detail during this step.
Step 1. The Gemba
The table is also valuable for identifying key customers to visit,
The project began in the U.K. with the help of local Singha but this was already decided prior to starting the project.
agents. They arranged for the Singha team to visit the largest Thai
restaurant chain in London, as well as some pub chains and super- The gemba visits included 26 open-ended interviews with 57
markets for the gemba. The gemba team included: the authors; drinkers, observation studies of shoppers, and some demographic
Pichet Changkasem, distribution development manager; Jeerachot questions to qualify the respondents. QFD research has shown
Sathirakul, marketing manager; Boonchai Asavalitiporm, consul- that to obtain qualitative data, sample sizes of 20 will usually
tant; and Kasamapong Suttangkakoon, consultant. yield about 75 percent of the data, with additional samples
tending to repeat rather than add new data3. Interviews were
Gemba is a Japanese term that means the place where the truth can audiotaped so that the interviewer could focus on the customer
be learned, and it is commonly used in process quality improve- without worrying about writing large amounts of data.
ment to help find the true causes of deficiencies or variability. In
marketing and new product development, gemba now refers to the The audio data (over 271 MB) and interviewer notes are what
place where customers use the product and it is where we can learn QFD practitioners call raw voice of the customer data. In this
the truth about their needs, and in this project, self-image. Since the unanalyzed state, the raw data is complex, fast-flowing, and
best results occur when we can observe the customer in the act of includes many pieces of information that must be thoroughly
buying or consuming the product, timing is very important. understood. Modern QFD teaches several techniques to help
avoid these common cognitive errors:
A customer segments table, shown in Figure 2, was used to
ensure the correct identification of gemba opportunities. The 1. Affective errors—we only see what we wish to see.
table is built on the journalistic questions of who, what, when, 2. Attribution errors—we see only stereotypes and this bias
where, why, and how a product is used. Additional columns are closes our minds to differences.

Figure 2—Customer segments table for the U.K.

Who uses product? What is product Where is When is Why is product used? How is product used?
used for? product used? product used?
Business people at lunch Go with meal Restaurant Lunch Lunch beverage, take edge off Glass
bad day or end of week
Business people relaxing Quench thirst Pub, indoors After work Relax after work, socialize with friends, Pint or bottle
after work at pub or outdoors while waiting for friends to arrive
Business people having Go with meal Restaurant Night Dinner beverage, relax in evening Glass
dinner with friends or family
Shoppers at market Drink later Home, BBQ ? Always buy beer ?

ASQ www.asq.org Page 2 of 6

 Figure 3—Excerpt of gemba visit

Observations Verbatims Documents Notes Clarified Items

Eight men, one woman Like Thai food because less salty Need to go back to work Fast (I have one hour for lunch,
having buffet. than Chinese or Indian food, so it and be in appropriate and I walk 10 minutes from office)
Four Coronas, two Singhas, is milder and easier to eat at lunch. mood for work (not drunk). Easily available
three soft drinks. Usually have a beer at lunch = Good deal
prefers Corona because it is less Fragrant
“gassy.” Like authentic flavors and Mild
fragrance. Cooks Thai style at home
With work team
by adding coconut milk and green
Break for the usual
curry which are easily available,
for a special meal that is unusual Light
and romantic (and it works!). Authentic
Special moments
Romantic

Figure 4—Customer voice table

Clarified Items # Customer scene Benefit Product

Need/Image Problem
Fast (I have one hour for lunch, 1 Lunchtime in London’s central I want to have fun with my friends. Good deal
and I walk 10 minutes from office) business district on a warm, sunny I want something convenient. Fragrant food
Easily available Monday. Mostly office workers I enjoy the fragrance. Mild spices
Good deal in business or casual clothes, I want to feel refreshed afterwards. Light meal
Fragrant younger ones are single. Warm I like to be adventurous.
sunny day, in upstairs a la carte
Mild I want to spend my time the way I want.
menu dining. Those in a hurry
With work team I’m in control.
enjoying downstairs buffet (6GBP),
Break for the usual Authentic.
those with more time (business
Light lunch, colleague birthday party) Romantic.
Authentic in upstairs a la carte menu dining. Special moments.
Special moments Most having soft drinks as it is too A break from the ordinary.
Romantic early in week for alcohol at lunch I’m a leader.
– Monday (first day of week). I know what I like.

3. Confirmation bias—we cherry pick those facts that agree with Clarification is one of the more difficult steps in modern QFD
our preconceptions and rationalize away contradicting data. because it requires the team to discover the unknown unknowns.
4. Search satisfaction—we stop searching once we find In Figure 3, the clarified items are extractions of several of the
something we think is significant.4 verbatims collected, such as “mild” or “authentic.”

The team’s next task was to simplify the complex, sort the com- A little deeper analysis about a preference for preferring Corona
ingled, and analyze the unspoken. One tool for this is the gemba (a Mexican beer) because it is less gassy at lunch, as well as the
visit table. Figure 3 is an excerpt from one of the gemba visits. need to get back to the office without being drunk, formed the
clarified item of “work team.” Notice that the clarified items
There are different kinds of data inputs to collect for the table,
refer to Thai food, lunchtime, entertainment, beer and beverages,
namely observations (things we see), verbatims (things we hear),
self-image, and other matters. The purpose of this table is to cap-
documents (written data provided by the customer), notes (writ-
ten data and inferences produced by the gemba visit team), and ture these details and to trace the evolution of the data.
evidence (physical specimens, samples, failed or destroyed pieces).
For traceability, it is helpful to segregate these in the table. Step 2. Voice of the Customer

Another section of the table that described the interviewee char- The clarified items in Figure 3 contain data that is both directly
acteristics and the circumstances of the gemba was omitted from relevant to this study and some that must be translated into
this case study, but it is crucial to the step of simplifying, sorting, relevant data. A customer voice table, depicted in Figure 4,
and analyzing. This is called clarification, and it produces the was used to facilitate this translation process. The table should
clarified items shown in Figure 3. include necessary information to retain fidelity to the original

ASQ www.asq.org Page 3 of 6

voice of the customer, as well as circum- Step 4. Customer Kansei Image Hierarchy
stantial data to shed light on the outcomes
desired by the customer. The second stage of the prioritization process was to combine the affinity diagrams of
the various focus groups into a unified hierarchy diagram. Combining several affinity
In Figure 4, the shaded column titled need/ diagrams into a unified hierarchy diagram is best done by an experienced QFD facili-
image shows the results of this analysis tator and confirmed with customers.
for a customer. For example, the statement
“break from the usual” was translated into The hierarchy diagram was analyzed to make sure that levels of abstraction are not
“I like to be adventurous.” Though the accu- comingled among the groupings. This is critical for the AHP, which relies on forced
racy of the translation was unknown, the comparisons. For example, if we were to ask which was preferred—an apple or an
QFD team itself had ventured outside the orange—we could get a discreet answer. However, if we asked which was preferred—
obvious statement to infer that there may an apple or a fruit—the respondent would be unable to reply since an apple is a
be additional meaning. type of fruit. This is an abstraction problem that must be corrected prior to the AHP.
Figure 6 is an excerpt from the unified hierarchy diagram.
They tested statements later in the process
when customers were asked to prioritize the
images. If the translation were correct, it Figure 6—Hierarchy diagram of kansei image
would receive a high priority and be con-
sidered for inclusion in marketing collateral. I feel relaxed
Since it was unspoken by the customer, the Feel refreshed
company may able to take advantage of dis-
covering something first.
Private life

Step 3. Customer Kansei Image Things are convenient

Structure
Romantic
The kansei image needs captured in Figure I feel special
4 were presented to another group of cus- Fragrant
tomers in the same market segment to
determine the preferred images. In other Refined

words, how did they wish to see themselves

or be seen by others?
Step 5. Customer Kansei Image Prioritization With AHP
The kansei image needs were written on
cards and presented to three small focus The third stage of the prioritization process was to apply the AHP5 to learn which
groups. Participants were asked to cluster images are most preferred by the customers. Prioritization in multi-criteria decision
these into groups based on common themes making was advanced by the research of Dr. Thomas Saaty in the 1970s at the U.S.
and then to assign a name to the group. Department of Defense and later at the Wharton School of Business at the University
Figure 5 is an excerpt from one of the of Pennsylvania.
groupings.
Saaty found that decision makers facing a multitude of elements in a complex situ-
ation innately organized them into groups sharing common properties (an affinity
diagram), and then organized those groups into higher level groups, and so on until
Figure 5—Affinity diagram of
a top element or goal was identified (a hierarchy diagram). When making informed
kansei image needs
judgments to estimate importance, preference, or likelihood, both tangible and
intangible factors may be included and measured. The AHP was created to manage
Free time this process in a manner that captures the intuitive understanding of the participants
and also yield mathematically stable results expressed in a numerical, ratio scale. A
Spend time the Fun with numerical, ratio scale is preferred for the following reasons:
way I want friends
• Numerical priorities can be applied to later analyses to derive downstream priorities.
Break from
• Ratio scale priorities show precisely how much more important one issue is
Feel refreshed than another. Ordinal scales only indicate rank order, but not the magnitude
ordinary
of importance.

ASQ www.asq.org Page 4 of 6

 Figure 7—AHP of one tertiary branch

Things are
I feel special convenient Romantic Fragrant Refined sum row avg
Things are convenient 1 5 9 1 0.433 0.543 0.375 0.409 1.760 0.440

Romantic ⁄5
1
1 5 1
⁄3 0.087 0.109 0.208 0.136 0.540 0.135

Fragrant ⁄9
1 1
⁄5 1 1
⁄9 0.048 0.022 0.042 0.045 0.157 0.039

Refined 1 3 9 1 0.433 0.326 0.375 0.409 1.543 0.386

2.311 9.200 24.000 2.444 1.000 1.000 1.000 1.000 4.000 1.000

Inconsistency 0.04

Figure 8—AHP priorities applied to the kansei image needs hierarchy diagram

I feel relaxed

50.0% local Feel refreshed 12.0% 5.0%

41.7% global

Private life

Things are convenient 44.0% 18.3%

Romantic 13.5% 5.6%

I feel special
Fragrant 3.9% 1.6%
50.0% local
41.7% global Refined 38.6% 16.1%

• Numerical scales can be tested for judgment inconsistency, “spend time the way I like,” “I am adventurous,” “break from
sensitivity, and other useful properties. the ordinary,” and “exotic.”

Starting with the highest level of abstraction in the hierarchy, Marketing Collateral
three more small focus groups were asked to express their pref-
erence for the kansei image needs. In addition to the kansei image needs, the Singha marketing team
also considered their brand truths of “authenticity,” as the first beer
The words were presented to them in pairs and they were asked in Thailand, “prestige,” as the only beer in Thailand brewed under
to identify which word was preferred and by how much using a
royal permission of HM King Prajadhipok Rama VII, as well as
verbal scale of somewhat, more, very much more, or extremely
the brand asset of Thailand as a travel destination. Thus, the
more preferred. In the AHP method, these verbal scales are
elements for defining the new brand essence were inferred from
entered into a numerical matrix and the calculated eigenvalues
the following formula: image needs + brand truth + Thailand.
(row average) quantify the ratio scale preference of the words.
Judgment inconsistency was also calculated and was within The kansei image needs hierarchy was generalized to a meta-
acceptable limits. Figure 7 shows the results of one of the hierar- metaphor of the journey of “drinking Singha from the first drop
chy branches of one of the focus groups. to the last drop.”

Figure 8 shows the AHP results applied to the hierarchy. The advertising agency Publicis was asked to create market-
Checked items are kansei image needs that would be given to ing collateral such as advertisements and promotions using the
an advertising agency to build marketing collateral, as they team’s findings. With “journeys, adventures, discoveries, experi-
indicate which self-image the market segments ultimately ences, fun, and authenticity—all captured in a bottle” in mind,
­prefer. Other highly ranked kansei image needs included the agency distilled a new brand essence: “Singha Transports

ASQ www.asq.org Page 5 of 6

You With Every Drop.” It also created the tag line, “For the Glenn H. Mazur has been active in QFD since its introduction
Road Less Travelled.” to North America in the 1980s, and has worked extensively with
the founders of QFD on their teaching and consulting visits from
These messages helped generate advertising ideas such as: Japan. He is a leader in the application of QFD to service indus-
tries and consumer products, conducts advanced QFD research,
• Singha beer takes you on a journey of discovery, taking you
and is the conference chair for the annual North American
to places that are less traveled.
Symposium on Quality Function Deployment. Mazur is the
• Singha beer is a journey of discovery for the world’s
executive director of the QFD Institute and International Council
adventurers.
for QFD, an adjunct lecturer on total quality management at the
University of Michigan–College of Engineering (ret.). He is the
Print ads were generated to appear in magazines beginning in 2009.
president of Japan Business Consultants Ltd. and is a senior mem-
ber of ASQ and the Japanese Society for Quality Control (JSQC).
Conclusion He is a certified QFD Red Belt® (the highest level) and is one of
While quality methods like QFD, kansei, and AHP have long two in North America. He is a certified QFD-Architekt #A21907
been used to create new products and services, Singha beer has by QFD Institut Deutschland. He is the convenor of the ISO work-
extended its utility to better understand why its customers like ing group to write the new ISO 16355 standard for QFD.
the product from an emotional perspective and generate adver-
tisements that speak to the consumer’s inner wishes. Singha References
believes that by crafting advertisements to kindle these wishes,
1. Akao, Yoji, ed. 1990. Quality Function Deployment:
consumer interest will continue to grow.
Integrating Customer Requirements into Product Design.
Translated by Glenn Mazur. Cambridge, MA: Productivity
For More Information Press. ISBN 0-915299-41-0
• Contact Glenn Mazur at [email protected] or Theera 2. Hepler, Carey and Glenn Mazur. 2006. “Finding Customer
Vongpatanasin at [email protected]. Delights Using QFD.” Proceedings of Quality Institute
• Visit the QFD Institute website at http://www.qfdi.org/. for Healthcare Annual Conference. American Society
• Visit Singha Beer’s website at for Quality. Section 1. http://www.mazur.net/works/
http://www.singhabeer.com/. Hepler_Mazur_2006_Finding_Customer_Delights_at_Blue_
• Learn more about QFD in the ASQ Knowledge Center at Cross_Blue_Shield_Florida.pdf
asq.org/learn-about-quality/qfd-quality-function-deployment/ 3. Pouliot, Fred. 1992. “Capturing the Voice of the Customer.”
overview/overview.html. Proceedings from the GOAL/QPC Ninth Annual Conference,
Advanced QFD Session. Methuen, MA: GOAL/QPC.
www.goal-qpc.com
About the Authors
4. Groopman, Jerome, M.D. 2007. How Doctors Think.
Theera Vongpatanasin is the managing director of Boon Rawd Marinerbooks.co, ISBN 978-0-547-05364-6
Trading International Co., Ltd. He is a QFD Green Belt® and 5. Hepler, Carey and Glenn Mazur. 2007. “The Analytic
QFD Black Belt®. He has worked as a strategy consultant with Hierarchy Process: Methodologies and Application with
Coopers & Lybrand (now PriceWaterhouseCoopers) from 1993– Customers and Management at Blue Cross Blue Shield of
1997 and Auriga, a Bangkok consult firm, from 1997–1999. He Florida.” Transactions of the 19th U.S. and 13th International
joined Accenture as a senior manager in the strategy and busi- Symposia on QFD. QFD Institute. ISBN 1-889477-19-2
ness architecture group from 1999–2003. From 2003–2007 he http://www.mazur.net/works/Hepler_Mazur_2007_AHP_
was the managing director of Asia Books Co., Ltd. with_Customers_and_Management.pdf

© QFD Institute. Published with permission from QFD Institute.

ASQ wants you to share your organization’s quality improvement success story with its worldwide audience.
No time? No problem. Authors are available to do the writing for you. Contact [email protected] for details.

ASQ www.asq.org Page 6 of 6