international journal of

production
e c o n o mi c s
E L S E VI E R Int. J. Production Economics 46 47 (19961 575 585
What and how about quality function deployment (QFD)
C. P . M. Go v e r s *
Eindhoven Universi O, o/" Technology,, Graduate School of Industrial Engineering and Management Science, Frits Philips blstitute./or Quali(v
Management, P.O. Box 513, 5600 MB Eindhoven, The ,Netherlands"
Abstract
QFD is more a process than just a tool for product as well as production process development based on the concept of
Company Wide Quality Control. Essential characteristics are: customer orientation, team approach and a way of
concisely structuring communications and linking together information. The methodology is described to discuss
experiences and some implementation problems.
Ahhough first used by the Japanese, experiences from "Western" companies support the results of better products and
production planning. Key factor for success is the Cross Functional Management approach.
Kevwords: Quality function deployment methodology: Practice in the Netherlands; Implementation aspects: Dutch
quality award
1. I ntr oducti on
Qual i t y funct i on depl oyment ( QFD) is a cus-
t omer - or i ent ed appr oach to pr oduct i nnovat i on. It
gui des pr oduct manager s and desi gn t eams
t hr ough the concept ual i zat i on, cr eat i on and realiz-
at i on process of new pr oduct s. QF D suppor t s de-
sign t eams to devel op pr oduct s on a st r uct ur ed way
t hat relates mar ket de ma nd via engi neeri ng speci-
fications to par t s speci fi cat i ons and t o pr oduct i on
process var i abl es and t hus to pr oduct i on oper a-
t i ons pl anni ng.
To discuss possi bl e i mpr ovement s of devel op-
ment processes by QF D we need t o under s t and the
* Tel.: I 4 31/40) 472285/47 21 70. Fax: ( + 31/40) 45 1275.
phi l osophy and concept s t hat are the r oot s of this
met hod. Appr oaches t o qual i t y based on the con-
cept of Tot al Qual i t y Cont r ol ( TQC) as i nt r oduced
by Fei genbaum [1] is f undament al l y different from
the Japanese TQC concept . In this vision TQC is
" Co mp a n y Wi de Qual i t y Cont r ol ". It is mor e
compr ehensi ve and char act er i zed by depl oyi ng cus-
t omer desires hor i zont al l y and vertically t hr ough-
out the or gani zat i on ( Japan Indust ri al St andar d
Z8101 - 1981).
The ori gi ns of the Japanese CWQC are the same
concept s of statistical qual i t y cont r ol (SQC) and
TQC as t hey were br ought over f r om the U.S. aft er
Wor l d War I I but t hey are depl oyed as means of
securi ng qual i t y account abi l i t y at each level in t he
or gani zat i on [2] and combi ned with mar ket ori -
ent at i on. The "voi ce of the cus t omer " drives all
activities.
0925-5273/96/$15.00 Copyright ,.~'~ 1996 Elsevier Science B.V. All rights reserved
SSDI 0925- 5273( 95) 001 13-1
576 C.P.M. Govers/lnt. J. Production Economi cs 4 6- 4 7 (1996) 575- 585
Often in many "Western" companies the execu-
tive's or engineer's voice dominates because there is
still a strong influence of the ideas of scientific
management. The separation of development and
preparation from implementation and doing, as
advocated by Taylor, brought us to organizations
subdivided into more or less isolated functional
departments, staffed with specialists for quality,
cost and delivery (QCD-aspects), paying attention
to output characteristics with separated perfor-
mance objectives.
Because of that orientation the TQC concept is
too often exclusively directed to the quality of
product and service in a proper balance with costs
and usually identified with manufacturing and as-
sembly activities. US and European companies put
a greater emphasis on problem solving and effici-
ency improvement during the implementation and
production stage. In the CWQC approach more
effort is put into designing quality at the develop-
ment stage and the QCD aspects are managed by
interrelation. Roughly, the differences concerning
product development can be illustrated as depicted
in Fig. 1.
The CWQC philosophy is characterized by cus-
tomer orientation, cross functional management
and process rather than product orientation. It
refers to quality of management and the quality of
work being done. Within that concept QFD pro-
vides a means of translating customer requirements
for each stage of product development and produc-
tion (i.e. marketing strategies, design, planning,
process development, production control). It is
a mechanism that serves as an "operational defini-
Europe Japan
market research
product definition
design/cOStengmeermg, . engineering 1
purchase neg otiation
cos t evaMation
I I
production cos t reduction
~
- ~ orrectiv e actions "7
efficiency) improvements ~
design
Fig. 1.
market research
product definition
establishment of
QCD-targets
I
engineering purchasing
p o l i c y
i 1
production
ontinuous improvements[
tion" of CWQC [3]. It means that although quality
professionals are important participants to facilit-
ate the QFD process, the marketing, development
and manufacturing professionals play an even more
vital role.
An organized QFD approach follows all the
rules for project management, which means project
definition, team selection and is not restricted to
a single action within just one department. Teams
should be cross-functional, expertise oriented and
consisting of six to eight members of comparable
peer levels.
QFD is a process that can help companies to
make the key trade-offs between what the customer
wants and what the company can afford to build. In
essence, QFD encompasses same activities that
people did before but it replaces erratic, intuitive
decision making processes with a structured meth-
odology that establishes all relevant information
and experiences that are available throughout the
organization. As such, QFD lays a basis for organ-
izational learning.
In general, the product development process
from customer-needs to manufacturing process op-
erations, can be outlined by a step by step approach
marking the points at which the requirements for
intermediate results are established and go-or-no-
go decisions can be made. Usually we can discern
four phases:
Strategy
and concept ~ Product Process Manufacturing
Definition design ~desi gn ~ operations
In the strategy phase, product policy and deter-
mination of the customer will be established and
the customer needs are translated into a product
concept. The design requirements (WHAT's) serve
as input to establish the component characteristics
(HOW's) of the product design which on their turn
serve to define the process plans and next the
manufacturing process operations. Because of the
complex relationships between the inputs and out-
puts, these relationships are mapped into matrices.
The basic structure is depicted in the relationship
matrix of Fig. 2.
The flexibility of the method of approach allows
for adding any other information which may be
C.P.M. Govers/lnt. J, Production Economi cs 46 47 (1996) 575- 585 577
rel at i onsh i pmat ri x
Fig. 2. Basic st ruct ure QFD.
P H A S E I
Fig. 3. Cascade of QFD charts.
useful to the deci si on maki ng. When viewing
a QFD- c h a r t the first time, l ook for t he What ' s
How' s rel at i onshi ps. Each How will be appr ai sed
to set t arget goal s or values, t he How- Muc h' s we
want to achieve. These How- Muc h' s shoul d be
meas ur abl e as much as possible.
Measur abl e i t ems pr ovi de mor e oppor t uni t y for
anal ysi s and opt i mi zat i on. Usi ng QF D char t s the
out l i ned devel opment process can be depi ct ed in
four char t s (Fig. 3) al t hough in act ual use as ma ny
levels of chart s as necessary ma y be used.
Of all the steps in the t ot al pr oduct i on devel op-
ment process, none deserves mor e and receives less
at t ent i on t han the defi ni t i on of the right pr oduct
for the right cust omer .
Thi s first st ep is the mos t critical par t of the
process and it usual l y is the mos t difficult because it
requi res obt ai ni ng and expressi ng what the cus-
t omer t rul y want s and not what we think he or she
expects. The great est gains of QF D will be realized
when the "voi ce of t he cus t omer " gets to be de-
pl oyed to the mos t det ai l ed level of manuf act ur i ng
oper at i ons. Thi s means depl oyi ng all phases al-
t hough it is possi bl e to achi eve subst ant i al benefits
by i mpl ement i ng QF D onl y in the first phase.
Ther e are several useful ext ensi ons to the basic
QFD- c ha r t s which great l y assist in the t r ade- of f
pr ocedur e to est abl i sh the val ues of the How-
Much' s. Deci si ons will be based on all the i nf or ma-
t i on nor mal l y available: busi ness and engi neeri ng
j udgement as well as var i ous anal ysi s techniques.
Once the first char t is compl et ed, the downs t r eam
stages will be det er mi ned mor e and mor e by
specific t echnol ogi cal charact eri st i cs of a par t i cul ar
or gani zat i on. So the most el abor at ed i mpor t ance-
rat i ng syst ems and assessment t ool s arc t ai l ored to
the first stage.
Here, in c o mmo n with ot her general i zed i nt ro-
duct i ons to QFD, we will discuss some of the most
i mpor t ant and well known expedi ent s by bui l di ng
a ~' House of Qual i t y" as it was named by Haus er
and Cl ausi ng [4].
2. QF D met hodol ogy: " The Hous e of Qual i t y"
St art i ng a QFD- pr oj ect , t eam member s shoul d
reach agr eement on issues as:
whi ch pr oduct or pr oduct charact eri st i c are we
goi ng t o focus on
- who do we consi der as our cust omer s
- which compet i ng pr oduct s will be used as a refer-
ence for pr oduct eval uat i on
how does the QF D a ppr oa c h fit i nt o pr oduct
and process pl anni ng.
In the initial phase the scope of the proj ect has to
be est abl i shed and shoul d be communi cat ed to and
agreed upon by management . Management sup-
por t is al ways very i mpor t ant because all avai l abl e
expert i se as well as mar ket i nf or mat i on will be
requi red. In or der to t urn a pilot proj ect i nt o a suc-
cess it is critical to select an appr opr i at e pr oduct to
be empl oyed (step 1: see Fig. 4). Tr y to find a pr o-
j ect wi t h br oad appeal t hat may pi que interest f r om
several areas of the company. A first proj ect shoul d
be simple, but not trivial, and present a real oppor -
t uni t y for i mpr ovement . Do not t ry to t ackl e your
t oughest pr obl em in your first QFD- ef f or t .
578 C.P.M. Govers/Int. ,Z Product i on Economi cs 4 6- 4 7 (1996) 575- 585
~ Wh ich product ~ ................... , ( ~
Wh[ch customer / / ~ 0 ~ ' ~ " - ' " ' ~
(~) Customer R e q u | r m/ H O ' ~ s { ~ i
I @'
i-2 1, 1 '
@ L
Correlatlonmatrlx
Design Parameters
Competetlve
Bench marklna
i....... S C O R E S
.... ~ ) Relationsh ipmatrlx
~ . : ~ Target~
Bench marklng
Fig. 4. The "House of Quality" showing the "rooms" of the
various steps in the QFD process.
c u st o me r
sa t i sf a c t i o n
E X C I T E M E N T
M A N C E
- - - - ~ ~ / degree of
~-~BAS~C achieVem"
/
"k
St ep 2 deals wi t h t he ki nd of cus t omer t o focus
on. Especi al l y for cons umer pr oduct s a cl ear cus-
t omer profi l e is needed. A good descr i pt i on in-
cludes t he end-users but coul d al so include profiles
of per sons or i nt er est - gr oups who influence the
pur chase decisions, e.g., retailers, cons umer s ' asso-
ci at i ons or publ i c aut hor i t i es ( envi r onment al regu-
lations!). To collect i nf or mat i on about cus t omer
r equi r ement s (step 3) var i ous dat a col l ect i on
met hods are avai l abl e.
For professi onal pr oduct s it will be r at her easy to
ascer t ai n requi rement s. I mpr ovi ng a cur r ent (con-
sumer) pr oduct , t han we al r eady know a lot about
the cust omer s ( mar ket surveys, service calls, etc.). I f
we t ake a new pr oduct this will be mor e difficult. In
t hat case a cl ear cus t omer profi l e can hel p to esti-
mat e what is i mpor t ant , less i mpor t a nt or not i m-
por t ant . Somet i mes a compar i s on bet ween different
t arget gr oups ma y help.
Aski ng consumer s about t hei r requi rement s,
a di st i nct i on shoul d be made bet ween expressed
r equi r ement s and i mpl i ci t requi rement s. The Ka no
model [2] rel at es cus t omer sat i sfact i on to t he de-
gree to whi ch pr oduct feat ures (or requi rement s)
are achi eved (see Fig. 5).
The st rai ght line represent s performance features.
We will be mor e satisfied if the per f or mance ex-
ceeds our expect at i ons and dissatisfied if t hey fall
short . Gener al l y, onl y these (most l y one- di men-
sional) r equi r ement s will be expressed by the cus-
t omer when we ask for.
- ma r k e t r e se a r c h
" k . o n e d i me n si o n
Fig. 5. The Kano model.
The i mpl i ci t feat ures fall i nt o t wo groups: basic
and exci t ement features. The basic feat ures are
expected. These include the f undament al funct i ons
which must be present al ong wi t h safet y and relia-
bility consi derat i ons. Even if all of the basic feat ures
are i mpl ement ed perfect l y we woul d not achi eve
real cus t omer sat i sfact i on - we woul d onl y elimin-
ate dissatisfaction.
The t op cur ve represent s the so-cal l ed excitement
features. Somet i mes these are seemi ngl y mi nor
i t ems whi ch the cust omer s percei ve as super i or
value. Focus on the exci t ement feat ures as sales
points can l ead t o a maj or compet i t i ve oppor t uni t y.
Requi r ement s shoul d be expressed in c ommon
parl ance. So in t he case of cons umer pr oduct s it is
i mpor t ant to use expressi ons like: easy to carry,
moder n l ook, nat ur al sound in st ead of: xx kg; yy
mm or zz Wat t s I-5].
The "voi ce of t he cus t omer " needs to be wor ked
out in or der to gai n a collective underst andi ng.
Usi ng a funct i on tree a r ough r equi r ement can be
det ai l ed i nt o t wo or mor e levels (Fig. 6) to descri be
the " WHAT' s " for the first QF D - chart .
C.P.M. Govers/lnt. J. Production Economi cs 46- 47 (1996) 575 -585 579
level 1 level 2
I//{, Easyto find information
Easy to operate ~, Cont r ol s have logic positions
' ~ Controls are easy accessible
Fig. 6. Funct i on tree with t wo levels.
( ~ \, , ~ ) I mp o r t a n c e Rating
r ~ - - 4 ~ (low) 1 - 5 (high)
I ~z*l I ........ t--~--".;L~ (for 2 di f f erent groups)
Easy t o use 3 5
" ( ~ D i mensi ons 2 5
Ener gyConsump. 3 3
Natural Sound 5 2
Fig. 7. What ' s list and i mpor t ance rating.
The list of " WHAT' s " shoul d be sufficiently de-
tailed t o make j udgement s about t he i mpor t ance of
each item to t he cust omers on whom we focus.
Dependi ng on the t arget gr oup the relative i mpor t -
ance of t he vari ous requi rement s can be rated. Thi s
can be done using a scale f r om 1 (not very i mpor -
tant) to 5 (very i mpor t ant ) (step 4 see Fig. 7). In the
case of obvi ous rat i ng differences bet ween the t ar-
get groups, it is necessary to consi der to devel op
t ai l ored pr oduct types. (in Fig. 7, for example, a de-
vice for "sound freaks" or for "ent er t ai nment "l .
Cust omer s choose bet ween pr oduct s of different
brands. Therefore, it is of strategic i mpor t ance to
know how the pr oduct s of our most i mpor t ant
compet i t or s mat ch up to the cust omer requi re-
ment s compar ed with our own product . Compet i -
tive benchmar ki ng (step 5) answers the quest i on
" WHY" we shoul d focus on which requi rement s
and will allow a pl an to be deri ved for i mprove-
ment. Thi s compar i son is shown in Fig. 8.
The heart of the QF D met hodol ogy in the first
phase is t he gener at i on of the design paramet ers:
t he "' HOW' s' " list. The design requi rement s result
from t he t r ansl at i on of cust omer wishes into tech-
nical specifications (step 6). This list must be in
bal ance with the available expertise and the given
time and cost frames of the project. To depict the
st rengt h of t he rel at i onshi p bet ween the What ' s and
How' s, symbol s and/ or an i mpor t ance rat i ng can
be used for pri ori t i zi ng efforts and maki ng t rade-
off-decisions. Some commonl y used symbol s and
weighing factors are shown in Fig. 9. The 9-3-1
weighing often achieves a good spread between
i mpor t ant and less i mpor t ant items, al t hough any
weighing system which makes sense, may be used.
Scientifically it will be always possible to i mprove
the list but you shoul d ask yoursel f whet her addi-
t i onal i nf or mat i on gat heri ng will pay off in the
project.
The design par amet er s refer to concret e observ-
able characteristics and met hods of measur ement
(see Fig. 10). Fr om an organi zat i onal poi nt of view
somet i mes it will be helpful if we ar r ange the char-
acteristics under headings like: mechani cal , electri-
cal, software, etc.
The design par amet er s must reflect a valid
measur ement of cust omer requi rement s. As al ready
said, t here u'ill be no one- t o- one rel at i onshi p and
the i nt eract i ons vary in intensity. The weighing
and compl et i on of the rel at i on mat ri x (step 7) trans-
late the proj ect objectives into a technical pri ori t y-
list.
This also permi t s us to cross-check our thinking.
Blank rows or bl ank col umns indicate places where
our t ransl at i on of What ' s into How' s has been
i nadequat e!
The operat i onal i sat i ons of How' s are the HOW
MUCH' s. The How Much' s shoul d be measurabl e
as much as possible. If How Much' s are not
measurabl e or nondescri pt i ve, then we have not
been detailed enough in our definition of the How' s
(anot her cross check of our thinking! 1.
We want HOW MUCH' s for the following
reasons:
To det ermi ne priorities and di rect i ons for im-
pr ovement s of the How' s (Sometimes an ext ra
row to i ndi cat e this is added to the How- Much
mappingt.
To pr ovi de an objective means of assuring t hat
requi rement s have been met.
580 C.P.M. Govers/Int. J. Production Economics 46- 47 (1996) 575- 585
Easy t o u se 3 5
D i me n si o n s 2 ' 5
EnergyConsump. 3 3
Na t u r a l Sound 5 2
SCORES
\
\
Competitive Bench marking
poor
1 2
v ery g ood
3 4 5
.... Ul~ ......... : O i ow n product
O : ' ' ~ l l l , & competitor A
i 1 ~ - ~ ............ I . ~ competitor B
Fig. 8. Competetive benchmarking.
- To provide targets for furt her detailed develop-
ment (step 8).
The targets are enumerat ed in the bot t om part of
the house.
To set the targets, it is quite common to perform
a compet i t or' s analysis on technical data. The
benchmarki ng on technical performance (step 9)
reveals our technical position with respect to our
relationsh ip s y mb o l w ei g h i ng factor
WEAK Triangle 1
MEDIUM Circle 3
ST R O NG D ot 9
Fig. 9.
competitors. This ki nd of benchmarki ng provides
a check for consistency of the relation mat ri x (step
7) and the competitive benchmarki ng dat a (step 5).
For instance, a high score for cust omer require-
ment X shoul d be reflected in high scores for design
paramet ers which are strongly related with t hat
requirement.
Most l y you will find i nt erdependency between
the design parameters. In the attic of the house
support i ng and conflicting design paramet ers are
identified by a correl at i on mat ri x (step 10). Differ-
ent degrees of i nt eract i on will again be represented
by symbols. (see Fig. 11).
The assignment of positive or negative correla-
tions are based on the influence of How' s on achiev-
ing ot her How' s regardless of the direction in which
the How Much values move. Positive correlations
are those in which one How supports another. The
ot her way, negative correlations are those in which
one How adversely affects the achievement of an-
C.P.M. Govers/'hlt. J. Production Economics 46 47 (1996) 575 585 581
\ .
ES~ tO USe 3 ] 5 \
D i me nsi ons 2 ~ 5
Ener g y Consump. 313
Natural Sound 2
SCORES
D esig n P arameters
R elationsh ip

strong
medi um 3 9
w e a k 1
o >
0 e 3
) Targ ets
B ench mark ing
Fig. 10. Est abl i shment of desi gn paramet ers.
ot her Flow. These conflicts are ext r emel y i mpor t ant
as t hey represent condi t i ons to di rect trade-offs.
Ther e are several useful ext ensi ons t o the basic
QF D char t s whi ch great l y enhance t hei r usefulness
[2, 6]. These pr ovi de some addi t i onal met hodol ogy
to assist in the deci si on process. For i nst ance, in the
figures we made use of symbol s. A popul a r met hod
is the use of importance rating. For each cell, a rela-
tive wei ght is cal cul at ed by mul t i pl yi ng the rat i ngs
of the Wha t ' s and the assi gned weights to each
r el at i onshi p mat r i x symbol . Summi ng the wei ght s
for each col umn provi des a rel at i ve i mpor t ance of
each How in achi evi ng t he collective What ' s (see
Fig. 12).
However , it is i mpor t ant t hat we are not blindly
dri ven by these number s. These val ues as such have
no di rect meani ng but r at her mus t be i nt er pr et ed
by compar i ng the magni t udes to each ot her. We
shoul d quest i on the rel at i ve val ues of the number s
in light of our j udgement . Is it r easonabl e t hat the
How val ued at 90 is about ten t i mes as i mpor t ant
as the How val ued at 9? And is it r easonabl e t hat
Howl s with si mi l ar rat i ngs are nearl y equal in
i mport ance' ?
If our j udgement is vi ol at ed we shoul d review the
char t for possi bl e errors. I f the i mpor t ance rat i ng
can be accept ed and a t r ade- of f decision is neces-
sary bet ween the How' s wi t h the 90 and 9 ratings,
great er emphasi s shoul d be pl aced on the How with
the 90 rating.
When the first phase has been compl et ed, we
have got a compi l at i on of:
cust omer r equi r ement s and t hei r i mpor t ance
a compet i t i ve assessment of our pr oduct
the rel at i onshi ps bet ween cus t omer r equi r ement
and design par amet er s
pri ori t i es for i mpr ovement based on a cross func-
t i onal a ppr oa c h
a means t o facilitate communi cat i on ensuri ng
t hat the obj ect i ve val ues and t r ade- of f decisions
are not "l ost " and s uppor t the c ompa ny' s l earn-
ing process.
582 C.P.M( Govers/lnt. J. Product i on Economi cs 46 47 (1996) 575- 585
( ~ "':_
C orrelationmatrix
/ C orrelation
. .
st r ong pos! t , v e
w e a k po st t , v e
w e a k n e g a t , v e
st r ong n e g a t i v e
Fig. 11. Correl at i on matrix.
As already mentioned, in the next phase, the
design requirements (HOW's) are carried on as
WHAT' s to the next chart to establish product or
part characteristics. This is continued to define the
process characteristics and subsequently manufac-
turing operations.
3. Pr act i ce
QFD, as a formalized approach started in 1972
when Yogi Akao introduced his "Quality tables" at
the Kobe Shipyards. A survey of QFD usage
conducted in 1986 among the larger member
companies of JUSE showed that QFD had grown
significantly. About 50% of the respondents re-
ported that they were using QFD [11]. Frequently,
Japanese success stories were attributed to a cul-
tural difference. However, one must keep in mind
that not all Japanese companies are equally succes-
sfull! Mainly the best ones demonstrate achieve-
ments with respect to their philosophies and
methods. But what is more some of them demon-
strate also to be successfull with their production
plants in foreign countries.
In 1984 Clausing introduced the QFD approach
in the United States to the Ford Motor Corpora-
tion. As a result of the article "The House of Qual-
ity" by Hauser and Clausing [4] the first case
studies outside of Japan became known. The use of
QFD in the United States is becomming quite
popular. Each year various conferences bring to-
gether speakers to describe various analysis tech-
niques and successful applications of QFD for
a wide variety of industries.
Companies which used QFD reported the fol-
lowing benefits
decreased start up problems
competitive analysis became possible (improved
market research)
C.P.M. Govers/lnt. J. Production Economics 46 47 (1996) 575- 585 583
sO@
3 o A
s- Ao A
1 O A @
s 0 0
2 @ O @
2 @ A @
Fig. 12. Importance rating.
A =1
O =3
@ =9
cont r ol poi nt s clarified (reduced devel opment
time; bet t er planning)
effective communi cat i on bet ween divisions (de-
part ment s)
design i nt ent is carri ed t hr ough t o manuf act ur i ng
(Qual i t y is built in "upst ream").
However it is difficult t o obt ai n specific case
mat eri al to witness the i mprovement s. Gener al l y
compani es are very rel uct ant t o br oadcast their
results because the results of a QF D process are
highly confi dent i al and of strategic val ue for the
company.
One of the first appl i cat i ons of QF D in The
Net her l ands was within the Philips Cor por at i on.
Philips concent r at ed at t ent i on on QF D since 1986.
The first successful appl i cat i on within the frame-
work of the Qual i t y I mpr ovement Pr ogr am was
achieved in the Chungl i moni t or s fact ory in Tai wan
[5, 7]. Aft erward (early 1989) it was i nt r oduced at
the Ei ndhoven research and devel opment depar t -
ment of hi gh-end TV-t ubes. This pr oduct is an
ext remel y compl ex mi xt ure of many i nt erdepen-
dent t echnol ogi es which resulted in a present
House of Qual i t y of appr oxi mat el y 150 times 120
positions. Filling in all the t echnol ogi es (How' s)
and technical know- how in t he chart pr ovi ded
a growi ng insight i nt o the i nt eract i ons bet ween the
weighed What ' s on one hand and the How' s on the
other. This knowl edge facilitated deci si on-maki ng
and decreased dependence on "good feeling" for
specification setting. This was especially useful for
the experts. Each of whom who mast er a par t of the
t echnol ogi es coul d not formerl y make use of the
"hi dden" knowl edge of his colleagues.
The use of QFD revealed that:
- the knowl edge about cust omer requi rement s was
insufficient
- dat a on compet i t or s was handl ed i ncorrect l y
i nt erdependences bet ween t echnol ogi es were
onl y part l y known.
The out comes of the appr oach (first phase) are
summar i zed/ t r ansl at ed i nt o one or mor e
"scenari o' s" to suppor t management decisions
concerni ng strategies. The i nf or mat i on is widely
spread among specialists of the depar t ment so
ever ybody can cont r i but e to the process by his
observat i ons. It is difficult to express benefits in
exact figures with respect to di mi ni shed lead times
or costs. However it has become clear t hat t op
management get mor e insight in expect ed conse-
quences (e.g., t he need for adapt at i ons or devel op-
ment s of pr oduct i on technologies) or not yet solved
probl ems. Compar ed to the past t hey have become
the "real " deci si onmakers. Before somet i mes solu-
tions had to be selected on ad- hoc basis by l ower
management levels with restricted insight i nt o pos-
sible consequences of t hei r decisions.
Nowadays it is common policy to i ncor por at e
QF D i nt o all devel opment processes of this par-
t i cul ar depar t ment . Most proj ect s requi re pr oduct
expertise from many different sections of the de-
part ment . Ther ef or e teams are cross funct i onal
groups of individuals represent i ng appr opr i at e dis-
ciplines like pr oduct planning, market i ng, engineer-
ing and manufact uri ng. Sometimes, in the case of
less compl i cat ed projects, t he faci l i t at or himself fills
in the QFD- char t s by consul t i ng the specialists.
The most i mpor t ant cont r i but i on of the facilitator
is concent r at i ng on the pr epar at i on of scenario' s to
communi cat e the QF D out comes to peopl e who
are less familiar with the met hodol ogy.
An appr oach within ~Van Door ne' s Tr ans-
missie" (VDT) was qui t e out of the ordi nary. Here
t he driving forces t o appl y QF D by some i nt erest ed
groups were very compl ex probl ems of process con-
trol. Because the pr oduct (a met al push belt fl)r the
aut omat i c transmission} al ready existed they
st art ed to anal yse the i mpor t ance of and relations
584 C.P.M. Govers/lnt. J. Production Economi cs 46- 47 (1996) 575- 585
bet ween process r equi r ement s and design variables
(phase three). The results of this proj ect per suaded
qui t e a number of designers i nt o t he appl i cat i on of
QFD in ot her areas of the devel opment process
and t hey will st art to discuss the want s of the
aut omot i ve i ndust ri es with respect t o t he devel op-
ment of new generat i ons of cont i nuous vari abl e
transmissions.
The concept s of QF D are not restricted to t he
devel opment of pr oduct s or services. Sullivan [8-]
describes an appr oach for Pol i cy Management in
concept ual t erms but he does not offer pract i cal
reference. Also in t hat field clear communi cat i on
and cust omer or i ent at i on are f undament al for suc-
cess. In an MSc- gr aduat i on proj ect his ideas were
wor ked out for the devel opment and management
of an annual pol i cy decl ar at i on for a pr oduct i on
pl ant of an i nt ernat i onal I C- manuf act ur er [9].
As far as I had cont act s with Dut ch QF D facili-
t at or s most of t hem r epor t ed t echnol ogi cal im-
provement s. Initial proj ect s of QFD do not yield all
expect ed benefits in one go. Ear l y appl i cat i ons re-
qui re mor e time and addi t i onal effort but results as
knowl edge transfer, bet t er pr oduct s and bet t er
under st andi ng of the cust omer expect at i ons are
di rect l y exploitable. Once a t eam has gai ned experi-
ence lead time and cost reduct i on as well as furt her
r educt i on of pr oduct and pr oduct i on pr obl ems will
be accumul at ed by fostering a bet t er under st andi ng
of cust omer s' requi rement s and what is needed t o
meet these requi rement s.
Most pr obl ems t hey had to unt angl e were rel at ed
to or gani zat i onal ci rcumst ances like proj ect defini-
t i on and proj ect management as well as t eam selec-
t i on and t eam building.
A critical f act or concer ni ng proj ect definition
is the "Voi ce of t he Cus t omer " (see Ka no model )
and what are the Critical Qual i t y Charact eri st i cs
[10].
Wi t h respect to proj ect management and t eam
selection, it can be ment i oned t hat focus on the
expertise is requi red but t ake care t hat the ranks of
the t eam member s shoul d be about equal in or der
to avoi d decisions being mani pul at ed by ranking.
Keep in mi nd t hat suppor t f r om (top) management
will be needed. It woul d be t he best to have recep-
tive, open- mi nded member s on the t eam who are
willing to challenge established practice.
4. Implementation
QF D is not a panacea for solving design pr ob-
lems nor for devel opi ng "perfect " product s. The
ai m is i mpr ovi ng the pl anni ng and cont r ol of the
development process. This implies t hat the ot her
processes of pr oduct i on are mor e or less under
cont rol . A company t hat still struggles with the
qual i t y per f or mance at the expect ed and the speci-
fied level, has t o stress basic qual i t y t echni ques first
and to change the cul t ure t owar ds mor e Tot al
Qual i t y Management . Wi t h respect to this the
Eur opean Qual i t y Award assessment model can be
seen as a reference to derive criteria for pr oduct i on
and devel opment funct i ons but also for the mana-
gerial functions. The Dut ch Qual i t y Award is
copi ed from t he Eur opean Award but for the inter-
pr et at i on of the established qual i t y level, the pat h
t owards TQM is split up into five phases. Each level
r oughl y cor r esponds with a score for the award:
Phase I Activity ori ent at i on: Pr obl ems are sol-
ved, but the process is not receiving
at t ent i on. (0-200 pts)
Phase II Process ori ent at i on: Based upon pro-
cess cont rol , probl ems are solved in
a syst emat i c way. (200-400 pts)
Phase III System ori ent at i on: All funct i ons of
the organi sat i on are cont rol l ed.
Cust omer or i ent at i on is achi eved in-
cl udi ng i nt ernal cust omers. At t ent i on is
pai d to pr event i on of probl ems.
(400-600 pts)
Phase IV Chai n or i ent at i on with suppliers and
customers: Opt i mal use of knowl edge
and capabilities for cust omer satisfac-
tion. Cooper at i on is sought in or der to
minimise costs. ( 60~800 pts)
Phase V Tot al Qual i t y Management : Philos-
ophy and st rat egy are based upon a
sense of responsibility within the so-
ciety. (800-1000 pts)
The maj or i t y of "good" compani es is at the level of
appr oxi mat el y 400- 500pt s which cor r esponds
with the obt ai ni ng of an ISO certificate. Ar ound
700 pts we find serious candi dat es for wi nni ng the
award.
To i mpl ement QF D successfully a company
shoul d have reached r oughl y the upper level of
C.P.M. Govers/Int. J. Production Economi cs 46 47 (1996) 575 585 585
phase It. This also can be an expl anat i on t hat at
present most l y the bigger firms benefit by t he QF D
met hod. Not onl y because t hey have the resources
but pr obabl y mor e because t hey have al ready es-
tablished a system (chain) ori ent at i on.
When an organi sat i on is r eady for QFD, t here is
t he need for a good faci l i t at or who knows the
met hod very well and has also t he social skills to
build and to manage a team. Usual l y the first pro-
ject will be mor e time consumi ng and appeal to
open mi nded discussions. So l ook for peopl e who
has a positive at t i t ude t owar ds new appr oaches
(early i nnovat ors).
5. Conclusion
It can be said t hat QF D is a synthesis of numer-
ous met hodol ogi es ori gi nat i ng from t he USA (e.g.,
Value Engi neeri ng and mar ket research) but t hey
are perfected and i nt egrat ed by t he Japanese. Suc-
cess stories from Japan are fairly known but do not
forget t hat t hey have al ready a l ongst andi ng experi-
ence. Pr obl ems encount er ed by facilitators in t he
"West " concent r at e on the real i zat i on of Cross
Functional Management.
Ther e is a wide vari et y of i mpr ovement t ool s t hat
will enabl e compani es to achieve high quality.
Tool s al one however cannot pr ovi de results on
themselves. They must be devel oped t o reflect the
compani es' cul t ure and management vision. To im-
pl ement QF D successfully a company has t o be
system ori ent ed. QF D provi des activities t hat bri ng
t oget her all requi red disciplines t o wor k and pl an
the devel opment efforts in a highly disciplined,
communi cat i ve and effective manner. QF D as such
is not a high t echnol ogy r at her it is a t echnol ogy
devel oped by users based on common sense and
effective i nf or mat i on transfer. Many Japanese and
Ameri can compani es have experi enced QF D to be
wort h the effort.
When under t aki ng a proj ect it is critical to t ake
the time to plan and organize your efforts. An ap-
pr opr i at e proj ect needs to be selected with respect
to its scope and objectives. Key fact ors for initial
success are management suppor t and the const i t u-
tion of the team. The t eam members need t o be
given t he time to establish its rules of oper at i on and
t rai ni ng requirements.
References
[1] Feigenbaum, A.V., 1961, 1983, 1986. Total Quality Con-
trol. McGraw-Hill, New York.
[2] Akao, Y., 1990. Quality Funct i on Deployment, Integrating
Customer Requirements into Product Design. Productiv-
ity Press. Cambridge, MA.
[31 Sullivan, E.P., 1986. Quality function deployment, Quality
Progress. June, 39 50.
[4] Hauser, JR. and Clausing, D., 1988. The House of Quality,
Havard Bus. Rev. 63 73.
[5] Nijdam, L.M., 1990. Quality function deployment, in:
Quality Matters, a publication of the CQB of Philips
International, no. 21.
[61 King, R., 1989. Better Designs in Half the Time. Implemen-
ting Quality Funct i on in America. 3rd ed. Goal QPC,
Methuen. MA.
[7] Rijsingen. P.H. and de Vries, R.F., 1991. Quality ['unction
deployment in picture tube production, in: Quality
Matters, a publication of the CBQ of Philips International,
no. 28.
[8] Sullivan. L.P.. 1988. Policy management t hrough quality
function deployment, Quality Progress. June, 18 20.
[9] Philips, M., Sander, P, and Govers, C., 1994. Policy formu-
lation b~ use of QFD techniques: a case study. Int. J.
Quality Reliability Management, 11(5): 46.58.
[10] Govers, C.P.M., 1993. Quality of services "applicable to
production'?", Int. J. Prod. Econom., 30-31:385 397.
[11] Goal / QPC, 1990. Quality Funct i on Deployment: A pro-
cess for Iranslating customers' needs into a better product
and profit, Research Report no. 89-10-02. Goal / QPC
Methuen, MA.
[12] Americala Supplier Institute, 1987. Quality Funct i on
Deployment, Instruction Manual ASI, Dearborn, MI.