Quality Function Deploymnet

Quality Function Deployment

A typical statement made in a quality environment is that “... we must meet the needs
of the customer”. A simple statement but one that is very important. The major
difficulty in having such statements is how do you establish the needs of the
customer? In many ways this is very impracticable. What one customer wants is not
always what others require.

This may be most highlighted in the automotive industry; drivers of cars have many
different wants, some will be the same while others will be completely different.
Quality function Deployment, QFD is a quantitative reliability tool to work with
customer requirements, both potential and actual.

Like most techniques it is useful if used in isolation, but if used as part of an overall
strategy it will probably be of greater use. Nowadays, many companies are working
within the context of a total quality environment, usually TQM. In the figure below
the relationship is shown between the common quality techniques.

Taguchi
SPC Mission

FMEA

Kaizen
Tools Values
QFD
Total Quality
Management
Technical Policy
Development
Expertise Ethos
Quality Corporate
Management Systems Attitude
5750/Q1/QS9000
Quality
Costs

Relationship of TQM & QFD

1
 Quality Function Deploymnet

QFD is a new forward planning tool which is pro-active in preventing quality

concerns. As quality systems have developed over the past 40 years they have gone
from being mainly reactive to mainly pro-active. This approach has an advantage in
that it becomes difficult to move more than one stage at a time because the previous
stages are pre-requisites for the next. In the following table we can see how these
stages overlap.

Ins SPC Design of Expts & QFD

Inspect SPC D of E & QFD

Inspection SPC D of Expts

Time
Inspection SPC

Quality by Inspection SPC

0 100%

Developing Quality Activities

The Japanese have followed this approach and their success is clear for all to see. In
the 1960‟s the phrase “Made in Japan”, was synonymous with poor product quality.
Using the developments from the above they have changed this view. Nowadays it is
more probable that people look for “Made in Japan” as a quality safeguard.

An advantage (by-product) of a good quality design is that it seldom needs changing.

Of course, on large and/or complex products there will always be some design change.
Any change incurs cost and hence increases the overall budget. In two automotive
cases, one in Japan and the other in America, the number of design changes were
recorded, see figure overleaf. The difference in number of changes is considerable,
but even more important, after production starts the number of changes is
exceptionally low. This reduces the otherwise high costs associated with post-
production changes.

2
 Quality Function Deploymnet

QFD
Company US
No of
Changes

Job 1 Time

Number of Engineering Changes

Good Design

Customers are not always capable of determining what is a good design. In the
modern world new designs are sometimes so complicated that only technical experts
are equipped with the skills to evaluate a design. A customer can only judge on their
knowledge, or what perceived experts say: this is especially true in the automotive
industry. However, customers can evaluate a design in terms of how they find the
product is in use. More importantly if you exceed what a customer wants the question
of what is a good design becomes secondary.

Exceeding customer expectations may be achieved in many ways, but as a guideline

we may say:

 improving performance level, i.e., better than accepted standards

 assuring minimum variations of performance

 reduce cost to company and consumer benefit

 offer performance features that are not expected.

3
 Quality Function Deploymnet

QFD, if directed towards any of the above can assist in finding solutions to all of the
expectations.
The House of Quality

A QFD programme is essentially a detailed planning chart that is a live document of

the features being planned for a product. Customer focus is very strong, providing the
rationale behind the relative significance and inter-relationship of each feature and an
assumed target value for each.

The QFD contains many items of data and may be used to concentrate on specific
customers. Unlike other techniques it takes note of competitive designs in a formal
and logical manner.

1 4 2

7
8
9

House of Quality

1. Customer Needs. A difficult task at the best of times and even harder when it
is these points that will be built upon. Consider a mouse-trap. First, who is the
customer? It is not always easy to define. Is it the mouse or the person who
wishes to catch the mouse? Depending upon your answer will be the customer
wants. A mouse would have a need to be killed instantly with no pain, a quick
and clean death. However, is this a want of the mouse catchers? It is more
likely that their wants would be simply catch and kill, the effectiveness of the
kill is probably of secondary importance.

4
 Quality Function Deploymnet

What is apparent is that the design could reflect these different wants if necessary. It
may be achieved by having various types of springs with different tensions. Material
properties, or the cost of heat treatment will result in a different cost.

Although this may seem a simplistic argument it can be followed through to many
designs. Within the market for off-road vehicles, the customers will have a wide
range of wants. Some, as is fashionable in certain sectors, will use the vehicle for
taking the children to their private schools and then to Harrod‟s for shopping. Many
owners will be driving over ploughed up fields. What is clear is that each will have
different wants. Therefore, under this section it is likely the wants will be lengthy. It
would be noted that only in exceptional circumstances should more than 10 be used
otherwise it is too cumbersome and direction will be directed away from the task in
hand.

Methods of Understanding Customer Needs

There is no guaranteed way of ensuring this, many people spend their whole working
life trying to achieve this.

Nevertheless, companies often have a wealth of useful information that is readily

available. The following list, although not exhaustive, is a common approach to
gather what is accepted to be a very comprehensive breakdown of the customer voice.

 Market surveys
 Dealer information
 Trade shows
 Test marketing
 Product Reports - trade magazines and newspapers
 Customer questionnaires
 Direct customer contact
 Direct consumer contact
 Why do consumers buy competitive products
 Warranty returns
 Customer complaints
 Consumer association reports
 Product to market share trend information

5
 Quality Function Deploymnet

The last five points are probably filling someone‟s in-tray in an office. Likewise the
other points are often covered by large companies during periodic or cyclic planning
activities.

Structured Questions

Professor Kano (1982) proposed an approach at the 12th JSQC Conference entitled:
“Attractive Quality vs Must be Quality”. In this presentation he demonstrated that
careful questions could identify different customer needs. He divided these into three
distinct groups: first, second and third levels.

First Level of Quality - Basic Functional Needs (MUSTBE)

A customer will rarely express these needs, they are expected to be there. A knife will
cut, a car jack will lift the car, a letter will arrive at the address stated on the envelope.
If it lacks in this there is total dissatisfaction.

Second Level of Quality - Performance Related

A customer will be more than satisfied, but pleased with this facet. The knife cuts
tough meat, a car jack lifts a car with nominal effort or a letter is delivered within a
few hours. When no standard performance is recorded a customer is totally
dissatisfied.

Third Level of Quality - The Excitement Features

When these features are present and unexpected the customer is satisfied. If not
present they are not dissatisfied. Today‟s Excitement features are tomorrow‟s
Performance features.

This is, perhaps, clarified by Kano‟s Model.

6
 Quality Function Deploymnet

Example: Oil filter

1st Level

Q: What is your reaction if you don‟t have to change the oil filter every 3000
miles?

A: I wouldn‟t expect to!

Q: What is your reaction if you have to change the oil filter every 3000 miles?

A: I would be very annoyed!

2. Degree of Importance of Each Need. To further assist in determining the

customer needs they should be weighed with respect to each other. Thereby,
the effort of complexity of a need will be valued. Comparing these with
competitors and their ratings, the necessity for assumed or planned
improvement levels are readily determined.

Competitive Comparison and Planned Improvements

In this section the problem of prioritisation and decision of critical design

requirements and, ultimately, the key critical characteristics are introduced.

Step 1: Degree of Importance. This may be from customer surveys,

together with team knowledge and is the key for grading each
customer need on a scale of 1 to 5, with 5 being the most
important.

Step 2: Our Company‟s Rating. This will list an objective view of each
customer on a scale of 1 to 5, where 5 is very good.

Step 3 & 4: Competitors Rating. Again using a scale of 1 to 5 in the same

manner as above.

Step 5: Planned Level. Listed here is the companies desired rating as

opposed to the actual rating in Step 2 .

7
 Quality Function Deploymnet

Step 6: Improvement Ratio. The planned level divided by our

company‟s rating. Step 5 divided by Step 2.

Step 7: Sales Point. A maximum of 1.5 for a strong marketing feature

down to 1.0 for expected features. Excitement qualities enter
into consideration. Reference should be made to the product or
sub-assembly life cycles, and fewer points are preferred.

Step 8: Importance Weight. The result of multiplying the degree of

importance by the Improvement Ratio by the Sales Point.

Step 1 x Step 6 x Step 7 = Importance Weight.

Step 9: Relative Weight. This is a percentage of the Importance

Weight.

3. Design Requirements. For each need there will be design requirements. If we

use for an example a bottle of carbonated mineral water, a customer need
would be that the water must be full of bubbles when opened. To achieve this
the seat (cap) must ensure the elevated pressure in the bottle is maintained.
How this is done will depend on the style, price and facilities available.

4. The Relative Relationship Between Customers’ Needs and the Design

Requirements. This is the central relationship diagram. In the relationship
table a key is suggested, but this is dependent upon the cause. If another one is
used ensure others can see this change. It is in effect a comprehensive Pareto
Chart. Thus, there will not always be a relationship. No relationship may
suggest the feature is engineer driven and not required. There will naturally be
a subjective element and it is suggested that it is tackled by a multi-disciplined
task force.

5. Relative Weights of each Design Requirement. This is a simple mathematical

calculation that indicates the strength of each design requirement in respect of
the others. The relative weight of each customer need is used with the
relationship of the Whats vs Hows.

If in the first design feature we have a strong relationship, 9, and the relative
weight is 12, we have:

(12 x 9) = 108

If we have a strong relationship, 9, and relative weight of 14, plus a weak

relationship, 1, and its relative weight is 6, then:

(9 x 14) + (1 x 6) = 132

8
 Quality Function Deploymnet

To establish the new relative weight the Importance Weights have to be

converted into percentages and then ranked in order of their percentages,
(round up or down to suit).

6. Design Feature Interactions. All design features need relating to each other if
possible. This is recorded in the roof of the House of Quality.

Positive Relationships give opportunities for trading off against negative ones.
Any negative relationships require attention. They are likely to be potential
reasons for dissatisfied customers.

When these negative relationships are established it does not require the
complete process to be started again. The whole point of QFD is to tackle
designs before they are undertaken. They must be used to set targets, see next
section.

7. Target Values. This may be looked upon as a customer benchmark, every

design feature must be addressed and a target set. This will be of great use to
the designer. It should be highlighted that these are not mandatory
specifications, it may be physically impossible to be met. Although the final
detailed design should be related back to these, shortcomings will be addressed
accordingly.

8. Technical Comparisons. These may be done in comparison to the company‟s

other products and competitors. It is best if these are reviewed by an engineer
addressing product breakdowns, testing and field data.

It is usual to compare using „S‟ for same, + for better and - for worse. Again,
at this stage if the competitors is found to be better a review of the target value
may be necessary. This is yet another chance to identify a potential
shortcoming in the design.

9. Special Requirements. This block is reserved for special regulatory situations,

e.g., government and national standards. The design can be viewed to see if
any design feature fails to meet these minimum specifications.

Developing QFD Teams

Being a practising QFD expert is not an easy position to reach. Qualitative techniques
are developed by people at different speeds. When a company first embarks on such a
task it may be a slow learning curve. It is usual to develop a core team, whom, when
they are familiar with the technique, can integrate into other teams, thereby, assisting
them to be fluent at a faster rate, although overcoming barriers and obstacles from
others could be the hardest part. It is difficult to manage change and this must be
addressed by top management.

9
 Quality Function Deploymnet

QFD‟s in Further Analysis

Recognising, and acting upon customer needs provides a powerful quality technique.
That is to say that all customer needs are found, but those which are established and
ranked will be of great importance in producing a good design. Many companies
consider the QFD process complete at this stage. All that is outstanding is to cross
reference the design specifications against target value; a relative simple process to be
performed by an engineer.

QFD may be further developed to include other aspects of producing a quality product.
Although this is not often undertaken in the West, the Japanese usually do develop the
work.

Design features are useful but not comprehensive for all the future work. Likewise,
we ultimately wish to arrive at operating systems to produce the parts. The figure
overleaf shows the stages involved and how they link together to arrive at our end
condition.

In the stages 1 to 6, the customer needs are related to design features. These in turn
are related to:

Functional features (Stage 2)

Part features (Stage 3)

Part characteristics (Stage 4)

Part characteristics (Stage 5)

Operating systems (Stage 6)

Stages 5 & 6 will produce, in principal, the process definition. Compare this to a
conventional programme. Designs are given to manufacturing who ultimately pass
them onto production. Each department works in isolation. Simultaneous
Engineering has made these barriers smaller, but each acts independently. Some needs
and experience may pass on, nevertheless, this is on an ad hoc basis (not a way to
handle £millions of investment).

Following the above argument, how does QFD tackle this in a different manner?
First, decisions are arrived at through collective team work of a multi-disciplined task
force.

The questions that are raised will be focused towards what is required to meet the
customers needs. The voice of the customer is driving all decisions and the process is
a direct result of this action.

10
 Quality Function Deploymnet

It must not be forgotten that a QFD requires a considerable amount of effort. Many
man hours are required to achieve design features; for an operating system it will be
considerably more. Nevertheless, the benefits may be considerable as can be seen in
diagram comparing the number of changes when QFD is not used.

Quality Function Deployment (QFD) Tutorial, Student Bag - D Pearce

Suggested Tutorial Procedure

Read the whole of the instructions once fairly quickly to gain an overview of the
process. Then, as a group, work through the student bag exercise using the drawing
board example as a step-by-step guide. Don't worry too much about developing a
definitive result, its more important that you complete and understand the process. It
may be useful to use a voting process to decide customer requirements and relative
importance and nominate a secretary to record results and a nominal chairperson to
have a casting vote and maintain progress.

Example - Drawing Board

Product Identity: Free standing, desk top drawing board for student home use.
Product Function: To provide a flat surface for drawing and reference.
Guide for developing the QFD matrix (see also figure 1):

Identify the user requirements next to the identification letters A to H at figure 1, i.e. A
= Board to suit Al paper size. Re-appraise the requirements to ensure that they do not
assume a solution; for example requirement B might have specified a carrying handle
rather than "Board to be easily lifted", this would have restricted the eventual design to
incorporate a carrying handle that might not be the best solution.

2. Rate each customer requirement between 1 (least important) and 10 (most important).
This part is both important and difficult to do; ratings will have a dramatic effect upon
the output and subsequently the design solution. Effective interpretation of customer
wants and needs is essential to this step. In practice companies are likely to use a
combination of techniques to establish comparative importance, one such 1c method is
the Customer Evaluation Matrix shown at figure 2. Follow the instructions at figure 2
to input customer ratings on the QFD matrix. Note that in the example the functional
requirements have outweighed the aesthetic considerations, this is not always the case.

3. Now identify on the vertical axis the measurables of product performance that will
address the customer requirements. Note that at this stage the vertical output should
not identify design solutions. If you do this you will limit the freedom of the design
evaluation team at too early a stage. However, finding the most effective measures is

11
 Quality Function Deploymnet

not always easy and in order for you to make progress in the tutorial it may be
necessary for you to include some design decisions. In the example "Angle horizontal
to 45 degrees in 5 degree steps" has been specified. These outputs form the basis of the
physical attributes of the Product Design Specification (PDS) and ultimately form the
basis, with external standards, of the product test programme.

4. Next rate how well the vertical outputs satisfy the customer requirements and rate
them 1 (low match), 2 (medium match) or 3 (high match). Since you have established
the outcomes to match the customer requirements it will be no surprise that a diagonal
of high match is established, however, subsidiary correlation is also probable. For
example "Frame folds flat to board" is intended to address the need for "F = Compact
for storage" but also impacts "B = Board to be easily lifted". Note that you should only
identify positive matches not negatives, see item 5.

5. Some of the design criteria/outputs may be in conflict; these conflicts can be identified
in the, so called, roof of the matrix as "X" marks. For example as the size increases the
weight of the board is likely to increase hence the X at the intersection of these two
outputs: The way in which the design team address such conflicts is likely to reflect
the relative importance to the customer. In the example the customer has rated "A =
Board to suit up to Al paper size" as a 9 and "B = Board to be easily lifted" as a 4,
hence it is unlikely that the board size will be compromised against weight. .

12
 Quality Function Deploymnet

13
 Quality Function Deploymnet

14
 Quality Function Deploymnet

15
 Quality Function Deploymnet

16