Logistics Decision Analysis Methods

1
Logistics Decision Analysis Methods

• Quality Function Deployment Part IV
• Presented by Tsan-hwan Lin
• E-mail percy_at_ccms.nkfust.edu.tw

2
Construction of the HOQ

• The first section of the HOQ to be constructed
  will almost always be the Customer Needs/Benefits
  section.
• Sections are also referred to as rooms.
• The Planning Matrix (also, Preplanning Matrix) is
  often the second section to e constructed.
• The third section of the HOQ to complete is the
  Technical Response (also, Corporate Expectations)
  section.
• The fourth step is to complete the Relationship
  section of the HOQ.
• The fifth and sixth steps in completing the HOQ
  are Competitive Benchmarking and Target Setting.
• The seventh and usually final step in completing
  the HOQ is to fill in the Technical Correlations
  Matrix.
• This part is also referred to as roof.

3
Q A
4
Technical Correlation - Introduction

• QFD is a key to concurrent engineering because it
  facilitates team members communicating with each
  other.
• The Technical Correlations section will show us
  for which technical areas close communication and
  collaboration are important, and for which it is
  not.
• It will also show us where design bottlenecks may
  occur, and therefore where design breakthroughs
  are necessary.
• The section is probably the most underexploited
  part of the House of Quality. Few QFD
  applications use it, yet its potential benefits
  are great.

5
Technical Correlation Meaning (1)

• The Technical Correlations section maps
  interrelationships and interdependencies between
  Substitute Quality Characteristics.
• The section consists of that half of a matrix
  that lies above the matrixs diagonal.
• Very often, especially after a technical concept
  has been decided upon and is somewhat understood,
  the developers will be able to see that as SQCx
  is moved in the direction of goodness, SQCy will
  be influenced, either in its direction of
  goodness or in the opposite direction.
• The degree and direction of influence can have a
  serious impact on the development effort.
• Example For an automobile, increased BTU rating
  of an automobile air conditioner (SQCx, more is
  better) may have a negative impact on automobile
  weight (SQCy, less is better).
• Notice how the SQCs are somewhat solution
  dependent (higher BTU rating gt heavier
  equipment).
• Had this incompatibility not been discovered
  during the product planning phase, dollars will
  be wasted in preliminary development work.

6
Technical Correlation Meaning (2)

• In QFD we usually identify five degrees of
  technical impact.
• These symbols carry no directional connotation.
• It is more constructive to indicate a direction
  of impact, since a developer can often make a
  strong argument for impact of SQCx upon SQCy, but
  not impact of SQCy upon SQCx.

?? Strong positive impact
? Moderate positive impact
ltblankgt No impact
X Moderate negative impact
XX Strong negative impact
?? Strong positive impact, left to right
? Moderate positive impact, right to left
ltblankgt No impact
X Moderate negative impact, right to left
XX Strong negative impact, left to right
7
Roof of the House of Quality

• Example Moving SQC 1 in the direction of
  goodness has a moderate negative impact on SQC
  5s direction of goodness.

• Direction of goodness
• ? More is better
• ? Less is better
• ? Target is best

8
Responsibility and Communication

• One of the most important benefits of the
  Technical Correlations is to indicate which teams
  or individuals must communicate with each other
  during the development process.
• One method for making this information more
  explicit is to construct a Responsibility Matrix
  (in addition to using roof directly) .
• This matrix would display the SQCs along the
  left, and the possible responsible teams along
  the top. A cell in the matrix would indicate the
  relationship of the team to the SQC.
• A good management practice is to assign
  responsibility for an objective to a single
  individual or a single organization. Therefore,
  the responsibility matrix would have a single ?
  in each row.

? Primary responsibility
? Supporting role
? Should be informed
9
Correlations Network

• An alternative but equivalent representation of
  the correlations in the roof is the Relationship
  Network Diagram (or, Relationship Digraph).
• In this diagram, the SQCs are represented by the
  circles, and the SQC affected is shown by the
  arrows connecting the circles.
• The degree and direction of influence is shown by
  the and indications written alongside the arrows.
• SQC with arrows emanating from it only is called
  the driver on the sense that it influences other
  SQCs but is not in turn influenced by any SQCs.
• SQC with incoming arrows only is called
  indicator. It is usually not worthwhile to
  invest resources in it.

10
Technical Benchmarks Introduction - 1

• No organization would invest in the development
  of a product or service without knowing enough
  about the competition to be sure that their
  design is competitive.
• Only the most important SQCs (i.e., with highest
  priorities) will be benchmarked and target-set.
• A critical question is how should the targets be
  set? How aggressive do they need to be?
• To a great extent, development teams can be
  guided by the competitions performance as well
  as their own performance (on the most important
  SQCs) (to make crucial strategic decisions to
  match, exceed, or concede).

11
Introduction - 2

• In general, competitive benchmarking is the
  process of examining the competitions product or
  service according to specified standards, and
  comparing it to ones own product or service,
  with the objective of deciding how to improve
  ones own product or service.
• The QFD process provides the basis for strategic
  competitive benchmarking (i.e., examining only
  those highest-ranking SQCs).
• In the process, the language of Substitute
  Quality Characteristics and the definition of
  direction of goodness become important
  determiners of the work.
• Only two types of SQCs, performance measures and
  product functions, will be examined in the
  benchmarking process here.

12
Target Setting - 1

• Setting targets is of course a matter of greatest
  interest to product and service developers.
• Obviously, setting SQC targets will drive all
  subsequent development activity.
• Development teams set targets for themselves
  whether or not they use QFD to plan their
  project.
• It (i.e., setting targets) takes up the QFD
  process after the development team has determined
  the most important SQCs and has benchmarked the
  competition.

13
Target Setting - 2

• Some of the linkages in the Relationships section
  might not be linear, because the associated SQCs
  may be Dissatisfiers or Delighters.
• The target-setting stage is a good time to deal
  with these Kano classifications.
• For those SQCs classified as potential
  Delighters, the team must decide how aggressively
  it can afford to be in target setting.
• There is relatively little downside risk in
  setting a conservative goal customers will not
  notice the absence of a Delighter. However, the
  potential gaining of setting a goal that beats
  the competition is high.
• For those SQCs classified as Dissatisfiers, the
  team cannot afford not to be aggressive.
• For those SQCs classified as Satisfiers, the team
  can expect that the better they perform on the
  SQCs, the greater the customer satisfaction
  performance will be for the linked customer
  needs. (Focus of the following discussion)

14
Target Setting - 3

• We will look (1) at setting numeric targets for
  SQCs that have been expressed as performance
  measures, and (2) at setting function or feature
  (nonnumeric) targets for SQCs that have been
  expressed as features.
• Numeric Targets
• Comparison with Competitions
• Mathematical Modeling
• Nonnumeric Targets

15
Q A
16
Matrix above the Diagonal

• The SQCs are arrayed along the top and side.
• The matrix is then rotated 45 degrees, and since
  the SQCs are already available along the top (of
  the HOQ), they double as the labels for both the
  rows and the columns (of the roof), making the
  row and column labels unnecessary.

17
Responsibility Matrix
Organization A Organization B Organization C Organization D Organization E Organization F Organization G
SQC 1 ? ? ?
SQC 2 ? ?
SQC 3 ? ? ? ?
SQC 4 ? ?
SQC 5 ? ?

• Because changes in SQC 1 strongly affect SQC 4,
  the organizations responsible for SQC 4
  (Organization E) must be informed of progress on
  SQC 1 (by Organization A).

18
Relationship Network Diagram
X
? ?
XX
?
? ?
19
Benchmarking Performance Measures

• If the SQCs were defined as performance measures,
  the benchmarking process becomes one of measuring
  the competitions performance and ones own
  performance in terms of these measures.
• To the extent that the performance measures were
  defined independently of the design of the
  product or service, the benchmarking process
  provides ideal apple-to-apple comparative data
  between the two.
• The results of measuring the two products or
  services can be laid down side by side (one above
  the other in the HOQ) and evaluated at a glance.

20
Benchmarking Functionality

• If the SQCs were defined in a more
  solution-specific manner, with product or service
  functions explicitly defined, the comparisons
  must be much more subjective.
• One way to deal with differences in functionality
  (designed in each product/service) is to
  decompose the high-ranking SQCs into sub-SQCs
  (i.e., lower levels of the Function Tree or the
  Affinity Diagram) and compare these sub-SQCs to
  the competitions.
• The number or percentage of subfunctions that
  correspond (to the competitions subfunctions)
  provides valuable numerical information (such as,
  where the competition provides more
  functionality, or how the competitions design
  solves the same problem differently from the
  development teams design).

21
Comparison with Competitive Benchmarks
22
QFD and Target Setting

• With QFD, the targets have a context(????)
• They (i.e., targets) are related to customer
  needs, to the competitions performance, and to
  the organizations current performance. (section
  A and B)
• The rank ordering of the targets is based on the
  systematic analysis done in the Relationships
  section (and all the prior QFD analysis).
  (section C and D)
• The rank ordering process is traceable, because
  all the decisions affecting the rank ordering are
  recorded in the QFD matrix.
• The QFD process itself provides no cookbook
  approach for setting targets for SQCs.
• The most vital information not explicit visible
  in the HOQ is the business know-how and technical
  expertise of the development team. However, the
  HOQ provides much of the strategic information
  needed, laid out in a compact form.

23
Comparison with Competition - 1

• One approach to setting targets is similar to the
  process of setting customer satisfaction
  performance goals in the Planning Matrix.
• The primary inputs to target value setting of
  SQCs are
• Rank order of substitute Quality Characteristics
  (Priorities)
• Competitions technical performance (Competitive
  Benchmarks)
• The development teams products technical
  performance (Own Performance)
• The primary inputs to goal setting for customer
  satisfaction performance in the Planning Matrix
  are
• Importance of customer attribute to customer
• Our current satisfaction performance rating
• Competitions satisfaction performance rating.

24
Comparison with Competition - 2

• The line of reasoning for setting targets is also
  similar (to that used in setting goals in the
  Planning Matrix).
• Starting with the highest ranking SQC, determine
  the strength of the development teams position
  relative to that of the competition.
• Based on the teams knowledge of the difficulty
  of performing well on the SQC, the team can
  decide whether to aim to do better than the
  competition, to match the competition, or to
  concede technical leadership to the competition.
• As a general rule, the goal should be set for
  technical performance that exceeds the best in
  the world for those SQCs that matter the most to
  overall customer satisfaction.

25
Mathematical Modeling - 1

• While QFD is certainly not a precise mathematical
  model of the relationship between technical
  performance and customer satisfaction
  performance, a little bit of simple mathematics
  could serve as a guide to the development team in
  setting targets.
• In the case of a SQC for which Less Is Better
• Equation
• A similar but slightly more complex relationship
  can be modeled in the case of Target Best (TB)
• Equation

26
Mathematical Modeling - 2

• Mathematical model provides some insight for
  setting the target for a SQC, but care must be
  exercised in its use.
• There is no guarantee that the relationship
  between any SQC and the corresponding customer
  attribute satisfaction performance is precisely
  linear or precisely quadratic.
• Customer satisfaction performance for any
  attribute is usually not the function of a single
  SQC, but of several SQCs.
• Since there are no reliable multivariate models
  for setting target values in QFD, a reasonable
  approach could be the following
• Treat each SQC as if it were the only SQC
  contributing to customer satisfaction performance
  of an attribute.
• Use the simple models to create a first estimate
  of an appropriate target value.
• Repeat this analysis for all the customer
  satisfaction attributes that this SQC is linked
  to. This creates multiple target values for the
  same SQC.
• Choose the most aggressive of these target values.

27
Satisfaction Vs. Technical Performances - LTB

• denotes customer satisfaction performance on a
  customer need as a function of a SQC p of the
  type Less The Better
• denotes customer satisfaction performance with
  the best product in the market (by market
  research)
• denotes customer satisfaction performance with
  the development teams current product (by market
  research)
• denotes technical performance of a SQC
• denotes technical performance of a SQC with the
  best product in the market (by competitive
  benchmarking)
• denotes technical performance of a SQC with the
  development teams current product (based on
  laboratory measurement)

sLTB(p) sworld-class s0 p pworld-class p0
28
Satisfaction Vs. Technical Performances - TB
29
Relationship Section
SQC U SQC V SQC W SQC X SQC Y SQC Z
Attribute A ? ? ?
Attribute B ? ?
Attribute C ? ? ?
Attribute D ?
Attribute E ? ? ? ?
30
Nonnumerical Targets - 1

• Setting targets for SQCs defined as features or
  processes is obviously more difficult than
  dealing with numbers.
• A number is one-dimensional, but features and
  processes are multidimensional and multifaceted.
• There are two helpful ways of thinking about
  targets for nonnumerical SQCs the continuum
  model and the subfeature model.

31
Continuum Model

• In the continuum model, we may imagine the SQCs
  to be on a continuum.
• This continuum could have as its endpoints
  stripped down and deluxe.
• The development team could judge where on the
  continuum their current offering lies, and where
  the best in the world lies.
• To clarify these judgments, they would do well to
  make their subjective judgments as objective as
  possible by documenting
• The differences between Best in world and
  Deluxe
• The differences between Development teams and
  Best in world
• The differences between Development teams and
  Target

32
Continuum Model - Figure
Development teams
Best in world
Target
Deluxe
Stripped down
33
Subfeature Model

• By using the subfeature model, the development
  team can explode each feature to be targeted into
  its component subfeatures.
• Each subfeature could be evaluated according to
  the continuum model, or could be exploded into
  lower-level subfeatures.
• Targets could be set by continuum, where Best in
  world and Development teams subfeatures line
  up, and by identifying subfeatures to be added,
  where the features dont line up.

Development teams Best in world Target
Subfeature a 7 7 8
Subfeature b 5 4
Subfeature c 6 8 8
Subfeature d 3 3
Subfeature f 8
Subfeature g 6 6
Outperform
Sustain
Concede