So far, we have discussed the voice of the customer and related performance measures.

We have discussed many factors that impact the value of a product or service.

Different customers in different markets may have

different preferences with respect to performances,

to ability, reliability, aesthetic, et cetera.

The next step is to develop a model that provides a single performance measure for value.

A performance measure that represents the value

for a variety of customers in different markets.

In the following discussions,

we will see some models designed to measure value,

and we will discuss the pros and cons of these models.

A very simple model is based on the quality parameters discussed earlier.

Each quality parameter is viewed as a criterion that is

measured on a specific scale using an appropriate performance measure.

The relative importance of each criterion is judged by

the customer and aggregated over all the customers by the model.

The end result is a summary table in which each row correspond to one of the criteria.

Each of the criteria is assigned a weight or relative importance.

And each alternative receives a score,

representing its performances relative to each of the criteria.

By multiplying all these scores in a specific alternative,

by the weight of the corresponding criteria,

and summing up over all the criteria,

a weighted average score is calculated for each alternative.

This weighted average score is based on the scores given by

the customers as well as the weight given to each of the criteria by the customers.

And therefore, it presents

the relative value of the alternatives as judged by the customers.

The main problem with this approach is how to estimate

the weights given by the customers to each of the criteria.

And in some cases, such as aesthetics,

how to estimate the score of each alternative with respect to the criteria.

In this example, an air traffic control radar system for airport is considered.

The first performance measure is the radar range.

The minimum acceptable range is 10 miles,

while the required range is 12 miles.

In addition, required values of quality and reliability are also specified.

Each criteria has a relative weight of presenting the voice

of the customer and its importance to the stakeholders.

Quality has the highest value of eight out of 10.

The specific alternatives that correspond to the side yield a range of 11.07 miles,

which is acceptable but only a little more than half way

between the minimum 10 miles and the required of 12 miles.

Thus, the score is a little above 50 percent or 53.5 percent exactly.

The quality of the proposed design is 59.59,

which is lower than the required quality of 75.

But the reliability is above the required 65.

And therefore, the score for reliability is 100 percent.

When the value of each alternative is calculated,

it is possible to find all the attractive feasible alternatives

and to perform trade-off analysis taking into account the cost schedule,

risk and value of each alternative.

The focus is on the efficient frontier

feasible alternatives that are not dominated by any other alternative.

A dominated alternative, for example,

an alternative with a value of 80 and a cost of 100.

If this cost and alternative with higher value of 90 is available,

they alternative is a value of 80,

and a cost of 100 is dominated or inferior,

and should not be considered any further.

In many real problems,

most alternatives are dominated.

And therefore, the efficient frontier can have

the new product development team focus on few efficient alternatives.

The weights and scores model is straightforward and easy to understand.

Its major weakness is the difficulty to find the weight of each of the criteria,

a weight that truly represents the voice of

the different customers and stakeholders in different markets.

In addition, there is a need to assign a score

to each alternative with respect to each criteria.

While this may be straightforward in a criterion such as

the range of radar system that can be measured in miles,

it might be very difficult in a criterion like aesthetics or style.

Professor Thomas Saaty of

the Wharton School of Business at the University of Pennsylvania

developed a model for finding the weights and scores based on the voice of the customers.

The model is known as The Analytical Hierarchy Process or AHP,

and is explained next.

The Analytical Hierarchy Process model is based on

linear algebra and can be implemented

using general purpose software like Excel and MATLAB.

There are special software packages that implement

AHP and are user-friendly and easy to use.

One example is Expert Choice.

And the following slides are taken from Expert Choice website.

The following presentation focuses on the theory

developed by Thomas Saaty to support decision making.

The theories implemented by the web-based application of Expert Choice,

a software that supports group decision making or GDSS.

By using Expert Choice,

a group of stakeholders can share their view regarding criteria,

relative importance, and scores of possible alternatives with respect to each criteria.

The Expert Choice software integrates these different views into a model that produces

a single measure of value or benefit associated with each possible alternative.

Decisions are made in different ways.

Some decisions are based on intuition and experience of the decision maker.

Other decisions are based on magic like a crystal ball that is used for fortune-telling.

And some decisions are based on scientific analysis like

decision trees and Analytical Hierarchy Process model developed by Thomas Saaty.

In the case of new product development projects,

group decision support system,

known as GDSS, are used,

and AHP is an example of GDSS.

Although there is no guarantee the decisions based

on scientific analyses are the best decisions,

the process of using such tools is in itself

important for the creation of shared understanding among the decision makers.

A good process leads to a consensus among

team members and develops their commitment to the decisions they make as a group.

The development of a model with

The Analytical Hierarchy Process starts with a discussion on three important issues.

What exactly we want to decide on, and consequently,

who are the customers and stakeholders whose voice is an important input to the process?

What are the needs and expectations of the stakeholders,

and how these needs and expectations are

translated into a set of criteria used to judge each alternative?

What are the alternatives that should be considered?

The name analytical hierarchic process or AHP comes from

the hierarchy of a tree-like model that is used to present the data on the objective,

the criteria and the alternatives.

In this model, each entity at the higher level might have

several entities connected below it in the next level.

But each entity at the lower level can have only one entity connected above it.

This is known as a one-to-many relationship.

And the entity above is called father,

while the entities below are called sons.

So there are many sons to each father and only one father to each son.

The top level is always the objective we want to decide on,

while the lowest level is always the set of alternatives.

There might be one or more intermediate levels that represent criteria.

For example, the criteria style may be broken down into shape, dimensions and color.

The weight of relative importance of the criteria is

determined by the voice of the customers and the needs and expectations of stakeholders.

Unlike personal decisions, where a person can use

intuition to decide on the relative importance of each criteria,

in GDSS, there is a need to integrate the needs and expectations of many individuals.

And, therefore, a process supported by proper tool is needed.

Saaty proposed pairwise comparisons as a basis for assignment of weights to the criteria.

The idea is to ask the stakeholders which of the two criteria

is more important in their opinion and by how much.

The individual inputs are integrated into a single way using the AHP methodology.

In the very simple example presented in this slide,

where the objective is to select a passenger car,

one can think of a process in which stakeholders were

asked whether a liability is more important than style.

And the number of stakeholders that responded saying yes was twice the number saying no.

The input is organized in a matrix,

in which each row and each column represent one criteria.

Each cell in the matrix represent the results of a pairwise comparison

between the criteria in

the corresponding column and the criteria in the corresponding row.

On the diagonal, each criterion is compared to itself and,

therefore, the entries are equal to one,

while the value itself below the diagonal is the

reciprocal of the value of the corresponding cell above the diagonal.

Saaty shows that the eigenvalues of

the pairwise comparison matrix are

weights that represent the perfect sense of the decision makers.

By finding the eigenvalues or weights,

the value or benefit of

each alternative can be calculated and used to support the decision.

The calculated benefit of value can be used along with the cost of each alternative to

find the efficient frontier of alternatives that are not dominated by other alternatives.

The points marked by blue dots are dominated by points marked by a red cross.

As for the same or lower cost,

higher benefits can be realized,

or the same or higher benefit can be achieved at a lower cost.

Thus, in this example,

there are two alternatives on the efficient frontier marked by an

X. AHP was used successfully on numerous projects.

And the results are reported in several books and articles,

some of which are available on the internet.

There are others tools similar to AHP.

For example, 1000minds was developed in

the University of Otago in New Zealand and widely used for numerous projects.

The motives are different,

but the purpose is the same: to support group decision-making

and to select the best alternative by a well-defined process and tools.

In this model, we concentrated on the question, what to develop?

Starting with the voice of the customer and the needs and expectations of

other stakeholders and translating the information into a measure of value or benefit.

It is possible to use the efficient frontier to select the most promising ideas.

The next step is to use the most promising ideas as

a basis for a plan of a project that can deliver the product,

process or service that was selected for implementation.

The process of translating the voice of the customer into

a specific product or process design is part of scope management.

We will define two types of scope.

The product scope is defined as

the features and functions of the product selected for development.

The project scope is defined as the work that needs to be

done to deliver the product with the required features and functions.

The tools presented in this lecture can help

us in translating the voice of the customer and

the needs and expectations of stakeholders into

a specific alternative for the product scope.

The next step is to translate the alternative selected into a specific project plan.

The basic building block of such plans are the activities of the project.

We will use a modification of the model called quality function deployment, QFD,

to transform the product scope into the project scope.

QFD is a process and tool developed by

Dr Yoji Akao to transfer the voice of the customer,

or VOC, to design parameters.

QFD is a template that has several dimensions.

We will concentrate on two.

The voice of the customer or what the customer wants,

also known as the whats in QFD terminology.

The engineering aspects or how we are going to fulfill the customer requirements,

also known as the hows in QFD terminology.

In the center of the QFD model,

each row corresponds to a specific requirement,

what the customer wants,

and each column correspond to a specific engineering decision,

how we are going to do it.

The cells in the center of the matrix

represent the relationship between the corresponding row and column,

in terms of the correlation between the two.

In the modified QFD, rows represent the project activities of the project scope,

and the columns represent the voice of the customer or requirements or the product scope.

The top cell in each column represent

the equation by which the requirement is calculated.

For example, under range,

we see the radar equation that tells us that

the range of a radar system in a function of three parameters.

TP, or the transmitter power, RS,

or the receiver sensitivity,

and AG, or the antenna gain.

The values of these three parameters

are determined by the corresponding three project activities.

TP, the transmitter power,

is determined by the transmitter design.

RS, the receiver sensitivity,

is determined by the receiver design.

And AG, the antenna gain,

is determined by the antenna design.

Given this information, we know that we should focus on

these three activities in order to satisfy the requirement for range.