1

THE IMPACT OF MARKETING MIX ON CUSTOMER SATISFACTION:

A CASE STUDY DERIVING CONSENSUS RANKINGS
FROM BENCHMARKING

MOHAMAD NASIR SALUDIN, AMY POH AI LING, AHMAD MAHIR RAZALI

ABSTRACT

This paper takes a cautionary stance to the impact of marketing mix on customer satisfaction, via a case study deriving consensus
rankings from benchmarking on retail stores in Malaysia. Field research was conducted in Tesco Stores (Malaysia) Sdn Bhd,
Carrefour of Magnificent Diagraph Sdn. Bhd., Giant of Dairy Farm International, and the homegrown retail store, Mydin Mohamed
Holdings Berhad. With increasing globalization, local retailers find themselves having to compete with large foreign players by
targeting niche markets. We build a model in deriving consensus rankings from benchmarking base on the marketing mix model, the
traditional marketing paradigm, embodied in the well-known Marketing Mix frame work proposed by Borden and popularized as the
4Ps (Product, Price, Place, Promotion) by McCarthy. The marketing mix is the lens through which the contemporary customer
perceives value in retail stores on 4Ps is examined. From the model, we analyze what is the best practice among the four elements
derived from a consensus ranking, a ranking method to identify the best in class. The analysis will mainly depend on the outcome of
what customer perceive towards the four marketing tactics. This paper discusses the introduction and use of a methodology for project
ranking in Retail store and, in particular, illustrates the use of a particular solution method called ELECTRE. A goal of this research
was to introduce a more objective methodology for the multicriteria outranking methodology as an alternative and more sustainable
approach for benchmarking analysis in marketing sector.

Keywords: Marketing mix, Customer satisfaction, Retailing, Benchmarking, Multi-criteria decision-making, ELECTRE methods

ABSTRAK

Kertas ini yang merupakan satu kajian kes pemerolehan darjat konsensus daripada penandarasan telah berjaya membukitkan bahawa
terdapatnya kesan campuran pemasaran terhadap kepuasan pelanggan. Kajian penyelidikan telah dijalankan di Tesco Stores
(Malaysia) Sdn Bhd, Carrefour of Magnificent Diagraph Sdn. Bhd., Giant of Dairy Farm International dan emporium tempatan,
Mydin Mohamed Holdings Berhad. Dengan pembangunan globalisasi yang pesat, syarikat peruncitan tempatan terpaksa bersaing
dengan pelabur asing. Sehubungan itu, satu model telah dibangunkan untuk memperoleh darjat konsensus daripada penandarasan
berdasarkan model campuran pemasaran, yang merupakan paradigma pemasaran traditional dengan mempraktiskan 4P iaitu produk,
harga, promosi dan tempat/pengedaran. Daripada model campuran pemasaran, kami telah membina hubungan analitik antara kepuasan
pelanggan dengan model campuran pasaran iaitu 4Ps dan seterusnya penandarasan daripadanya serta menjana respon yang positif
daripada penilaian yang diperoleh. Kertas ini memperbincangkan penggunaan kaedah pemerolehan darjat di syarikat peruncitan serta
mempaparkan kebagusan kaedah penyelesaian Electre.

Katakunci: Campuran pemasaran, kepuasan pelanggan, peruncitan, menandaras, pembuatan keputusan multi-criteria, kaedah
ELECTRE
 2

1. Introduction

To excel and flaunt as a market leader in an ultramodern era and a globalize world where we barely can
catch up with the changes, the organizations must strive not only to improve but also to commit into a
continuous improvement climate, to harvest from its marketing strategies especially marketing mix model,
benchmarking and company quality policy. Malaysia retail industry has been showing upward trends for
quite some time. Growth in this sector is particularly spurring by the changing buying patterns of
consumers and rising per capita income in the country.

Ranking and selecting projects is a relatively common, yet often difficult task. It is complicated because
there is usually more than one dimension for measuring the impact of each project and more than one
decision maker. This paper considers a real application of project selection for the marketing mix element,
using an approach called ELECTRE.

The ELECTRE method has several unique features not found in other solution methods; these are the
concepts of outranking and indifference and preference thresholds. The ELECTRE method is explained and
applied to the project selection problem using SPSS (Statistical Package for the Social Sciences)
application. Results show that ELECTRE was well received by the decision makers and, importantly,
provided sensible and straightforward rankings.

Our contribution is to show the potential in Marketing mix model in deriving a consensus ranking in
benchmarking. According to the feedback from the respondents, we dynamically rank out the best element
to be benchmark.

The decision problem faced by management has been translated into our market research problem in the
form of questions that define the information that is required to make the decision and how this information
obtained. Thus, in this paper, the decision problem regarding the marketing mix four Ps is translated into a
research problem. The corresponding research problem is to assess whether the market would accept the
consensus rankings derive from benchmarking result from the impact of marketing mix on customer
satisfaction using a multi-criteria decision making outranking methodology.

2. Literature Review

The decision problem faced by management has been translated into our market research problem in the
form of questions that define the information that is required to make the decision and how this information
obtained. Thus, in this paper, the decision problem regarding the marketing mix four Ps is translated into a
research problem. The corresponding research problem is to assess whether the market would accept the
consensus rankings derive from benchmarking result from the impact of marketing mix on customer
satisfaction using a multi-criteria decision making outranking methodology.

The project ranking problem is, like many decision problems, challenging for at least two reasons. First,
there is no single criterion in marketing mix model which adequately captures the effect or impact of each
element; in other words, it is a multiple criteria problem. Second, there is no single decision maker; instead
the project ranking requires a consensus from a group of decision makers. (Henig and Buchanan and
Buchanan et al.)

Henig and Buchanan and Buchanan et al. have argued that good decisions come from good decision
process and suggest that where possible the subjective and objective parts of the decision process should be
separated. This separation enables the decision making process to move away from being unnecessarily
subjective and toward a more objective orientation. A decision problem can be conceived as comprising
two components; a set of objectively defined alternatives and a set of subjectively defined criteria. The
relationship between the alternatives and the criteria is described using attributes, which are the objective
and measurable features of alternatives, attributes form the bridge between the alternatives and the criteria.
In Illustration 3.1 the alternative-attribute-criteria mappings are illustrated.
 3

Outranking relations, in most methods, are built using a concordance-discordance principle. More
complexity and flexibility are required in the processing of efficient alternatives. And it is the solutions, not
the criteria, which the marketing management is interested in.

Although it is not clearly stated in Simon (1977), we think that one of the main functions of review is
learning and we believe that the best support that could be provided to organizations would be for learning.
In many cases, we have observed that decision is treated as a one shot game whereas most decisions are
more or less repetitive. Human memory has some known biases and, for that reason, cannot accurately
analysis decisions ex post.

However, very little seems to have been done in this domain up to now. There are many possibilities related
to learning, review and ex post analysis. First, in some sense, a decision maker can learn the effect of the
assignment he has given to the weights. Similarly, in outranking methods, the decision maker can learn to
modify concordance and discordance factors (Roy and Skalka, 1985; Vetschera, 1986).

Most of the failures arise because one does not take into account that a decision maker makes a decision
according to a set of items (e.g., his preferences) that does not intervene explicitly in the decision making
process itself but constrains it. This is what we call contextual knowledge.

Let us also remind that, in the framework of decision making, due to the prominent look-ahead component
(Pomerol, 1995), the subjective and contextual data play an important role. Moreover, due to the
incompleteness of the model, especially during the evaluation phases (Lévine and Pomerol, 1995), among
the elements facilitating the cooperation are explanations and contextual knowledge, and the need to make
them explicit and shared both by the system and the user (Brezillon and Abu-Hakima, 1995) and Brézillon
(1996).

3. Research Methodology

3.1 Recognizance Survey

A recognisance survey was carried out in order to locate the most suitable site for the research. The section
take into consideration sites in Selangor area. Selangor is Malaysia's most populous state, with the nation's
biggest conurbation, the Klang Valley. Selangor's geographical position in the center of Peninsular
Malaysia contributed to the state's rapid development as Malaysia's transportation and industrial hub.
Selangor has a population of 4,736,100 (2005 estimate); the state's ethnic composition consisted of Malays
41%, Chinese 37%, Indians 19% and other ethnic groups 3%. The selected data collection sites are Tesco
Saujana Impian Kajang, Carrefour Alamanda Putrajaya, Giant Bukit Tinggi and Mydin Kajang.

3.2 Research Instrument

The research objectives and frame of reference was defined beforehand, including the questionnaire's
context of time, budget, manpower, intrusion and privacy. A non-comparative Likert scaling techniques was
used. The level of measurement of a variable in mathematics and statistics is a classification that was
proposed in order to describe the nature of information contained within numbers assigned to objects and,
therefore, within the variable.

The questionnaire is divided into 4 sections: customer information, marketing mix model, customer
perception and motivating factor. Variables that are measured only nominally are also called categorical
variables. The demography variables measured at a nominal level in Section 1 include gender, ethnic,
marital status, age and how often do the respondents shop at the specific retail store.

A typical test item in a Likert scale is a statement. The respondent is asked to indicate his or her degree of
agreement with the statement or any kind of subjective or objective evaluation of the statement. In Section
2, a six-point scale is used in a forced choice method where the middle option of "Neither agree nor
 4

disagree" is not available. The questions comprise four elements such as product, price, promotions,
place/distribution; six questions are allocated for each of the 4Ps.

Section 3 evaluates customer’s perception using the same scale as practice in Section 2 where Section 4, the
last part of the questionnaire measure the factor that motivates respondents the most to patronize the
specific retail store using the nominal measurement. We choose simple random sampling in the research
for conceptually; simple random sampling is the simplest of the probability sampling techniques. It requires
a complete sampling frame, which may not be available or feasible to construct for large populations. Even
if a complete frame is available, more efficient approaches may be possible if other useful information is
available about the units in the population.

3.3 Illustration of Research Framework

Illustration 1 Attribute – 4P’s – Retail Stores Mapping

The illustration of Attribute - 4P’s - Retail Stores Mapping was built to sprout a better understanding on our
study framework. It elucidates the main idea of how we determine the targeted attribute of the 4Ps and
generate it in the questionnaire to meet out objectives. The relationship between the marketing mix, 4ps
with the attributes lies in each P element were elucidate clearly linking to the four selected retail stores,
namely Carrefour, Giant, Tesco and Mydin.
 5

Once everybody agrees about the family of criteria, assuming that the alternatives are known, it remains to
complete the decision matrix, i.e., to evaluate each alternative according to the criteria. This evaluation
theoretically depends on the posterior aggregation procedure, but this fact is generally ignored by the
designers so that the assessment is generally independent of the aggregation procedure.

The system can support a direct assessment method, showing graphically to the decision making, the
position of the various alternatives or transforming a pair wise comparison into a numerical (normalized)
scale as, for example, in the so-called "Analytical Hierarchical Process"(AHP) (Saaty, 1980).

In the framework of multi-attribute utility, the utilities of a given alternative, regarding each attribute, are
jointly cardinal. They have consequently to be jointly evaluated (Pomerol & Barba-Romero, 1993). In this
case, due to the difficulty either to verify the probabilistic independence or to help the decision maker to
jointly evaluate the alternatives by solvability or by the mid-preference point method, the support of a
Multicriteria Decision Making methodology should be very useful.

3.4 Data Collection

The study was conducted in a Selangor area, the most populous state in Malaysia with approximately 4.19
million residents. At the time of the study, four retail stores were chosen as the research sites. The data were
collected by means of questionnaire. Households were the target of the research during the surveyed period.
First appointment was conducted with the personal in-charge in each retail store to request cooperation and
approval for data collection and survey respond via formal letters from the Department of Mathematical
Sciences, Faculty of Science and technology, National University of Malaysia.

Field research was conducted in Tesco Saujana Impian Kajang, Carrefour Alamanda Putrajaya, Giant Bukit
Tinggi and Mydin Mart Kajang. A simple random sample of 214 household’s respondents was obtained
from each of the four retail stores; sum up a total of 856 respondents data.

In our framework, we can think about objectives as aspiration levels defined for each criterion or
alternately as very general goals. We manage to expose the relationship between the marketing mix, 4ps
with the attributes lies in each P, it was elucidate clearly link to the four selected retail stores.

The retail stores management uses the information so obtained to understand the needs of individuals in the
marketplace, and to create strategies and marketing plans. The process of measurement is central to
quantitative research because it provides the fundamental connection between empirical observation and
mathematical expression of quantitative relationships.

3.5 Data Analysis and Interpretation

The retail market place promotes continuous improvement to survive in a turbulent environment. It does so
by creating, acquiring and transferring knowledge and modifying its behavior to reflect new knowledge.
For that, benchmarking is the search for industry best practices that leads to superior performance (Camp,
1989).

The benchmarking measurement of the retail stores considers a set of indicators and for this reason assumes
the configuration of a multi-criteria analysis. The literature on retail stores and marketing mix model has
identified four major underlying criteria essential to take place in the market place. They are as follows:

ATT1 : Product Attribute

ATT2 : Price Attribute
ATT3 : Promotions Attribute
ATT4 : Place/Distribution Attribute
 6

Multi-criteria benchmarking analysis of comparing the four retail stores (Tesco, Carrefour, Giant and
Mydin) poses many problems. Since the “dominance” relation is usually not verified, there is not a “best in
class organization”. Generally, an organization will show better performance on the basis of some
indicators and worse performance on the basis of some others: “there is no single performance management
enterprise system which is best in class across all areas” (Sharif, 2002, p. 76). However, in the absence of a
superior “best in class” dominating organization, one cannot “search for industry best practice that leads to
superior performance”, and thus cannot apply benchmarking analysis as advocated.

The “best in class” is the organization with the maximum averaged value, computed by averaging the
scores assigned to all the organizations on the basis of all the criteria. Moreover, this paper illustrates the
advantages, in terms of flexibility and realism, connected to the application of the multi-criteria outranking
methodology as an alternative and more suitable approach for benchmarking analysis of retail stores. That
is, the aim of this paper is to show the contribution of the multi-criteria outranking methodology to the
valuation of the retail stores in the market place in terms of benchmarking analysis. It enables the
benchmarking of organizational learning capability without the necessity of an aggregate indicator obtained
by averaging all scores assigned to the organizations on the basis of the different criteria.

Consider four retail stores:

R1 : Tesco
R2 : Mydin
R3 : Carrefour
R4 : Giant

This averaging methodology is the peculiarity and the main disadvantage of the traditional approach, that
is, the aim of this chapter is to show the contribution of the multi-criteria outranking methodology to the
valuation of the impact of marketing mix on customer satisfaction of the four retail stores (Tesco,
Carrefour, Giant, Mydin) in terms of benchmarking analysis. The application of outranking approach
enables, unlike the traditional analysis, the benchmarking of the impact of marketing mix without the
necessity of an aggregate indicator obtained by averaging all scores assigned to the organizations on the
basis of the different criteria. Finally, the following section discusses the contribution of the outranking
multi-criteria methodology to the benchmarking analysis of the impact of marketing mix on customer
satisfaction.

3.6 Benchmarking and Outranking-Satisfying Methodology

The outranking methodology is a family of algorithms developed by Operational Research (Roy, 1985;
Vincke, 1992; Roy and Bouyssou, 1993; Pomerol and Barba-Romero, 2000). Of these, Electre I method
will be introduced here. The input of the Electre I method is represented by a multi-criteria matrix as in
Table 1, surrounded by a line containing the weights that the decision making assigns to each criterion.

Table 1 Multicriteria matrix (Electre I)

ATT1 ATT2 (Price) ATT3 ATT4

(Product) (Promotion) (Place/Distribution)
R1 (Tesco) 4.42 3.94 3.97 3.90
R2 (Mydin) 3.91 3.73 3.42 2.95
R3 (Carrefour) 4.10 3.60 3.71 3.70
R4 (Giant) 3.90 4.02 3.76 3.92
Weight 1/4 1/4 1/4 1/4
 7

From Table 1, the retail stores’ positioning is generated and shown in the table below:

Table 2 Retail stores Positioning Table

ATT1 ATT2 ATT3 ATT4

(Product) (Price) (Promotion) (Place/Distribution)

1st Tesco Giant Tesco Giant

2nd Carrefour Tesco Giant Tesco
3rd Mydin Mydin Carrefour Carrefour
4th Giant Carrefour Mydin Mydin

Table 3 Retail Stores’ Ranking Table

Attributes Retail Stores’ Ranking

1st 2nd 3rd 4th

ATT1 (Product) Tesco Carrefour Mydin Giant

ATT2 (Price) Giant Tesco Mydin Carrefour
ATT3 (Promotion) Tesco Giant Carrefour Mydin
ATT4 (Place/Distribute) Giant Tesco Carrefour Mydin

Average ( RN ) = [ ATT1 ( RN ) + ATT2 ( RN ) + ATT3 ( RN ) + ATT4 ( RN )]/4

Now, let us consider R2 and R3 . Taking into account the values in Table 4 it is evident that R3 is better
than R2 for three criteria out of four (Marketing Model 4Ps). That is:

ATT1 ( R3 ) = 4.10 > ATT1 ( R2 ) = 3.91

ATT3 ( R3 ) = 3.71 > ATT3 ( R2 ) = 3.42
ATT4 ( R3 ) = 3.70 > ATT4 ( R2 ) = 2.95

Three criteria {1, 3, and 4} agree in considering R3 better than R2 . Only one criterion {2} considers R2
better than R3 . That is:

ATT2 ( R2 ) = 3.73 > ATT2 ( R3 ) = 3.60

 8

Interpreting the same procedure for all the other pairs of retail companies will obtain the Table 4.

Table 4 Multicriteria Matrix

ATT1 ATT2 ATT3 ATT4

(Product) (Price) (Promotion) (Place/Distribution)

R1 (Tesco) 4.42 3.94 3.97 3.90

R2 (Mydin) 3.91 3.73 3.42 2.95
R3 (Carrefour) 4.10 3.60 3.71 3.70
R4 (Giant) 3.90 4.02 3.76 3.92
Weight 1/4 1/4 1/4 1/4

Table 5 Matrix of Concordance Subsystems ( J c )

R1 R2 R3 R4
R1 {1,2,3,4} {1,2,3,4} {1,3}
R2 Ø {2} {1}
R3 Ø {1,3,4} {1}
R4 {2,4} {2,3,4} {2,3,4}

The generic element J c ( Ri , R j ) of the matrix of Table5 is given by:

J c ( Ri , R j ) = {j �J = ATTi ( Ri ) � ATT j ( R j )}; where: J = {1, 2, 3, 4}

Taking into account the weights assigned to the various criteria, a concordance index can be calculated for
each pair of company ( Ri , R j ):

C ( Ri ; R j ) = � j�J c
Kj ;

Where: K j is the weight assigned to the jth criterion.

For example, for the pair ( R3 , R2 ) we have:

C ( R3 , R2 ) = K1 + K 3 + K 4 = 1/4 + 1/4 + 1/4 = 0.75 (75 percent)

We therefore have a majority of criteria of 75 percent in favor of R3 with respect to R2 . Iterating the same
procedure for other pairs or organizations, we obtain the concordance matrix of Table 6.
 9

Table 6 Concordance Matrix

R1 R2 R3 R4
R1 1 1 0.50
R2 0 0.25 0.25
R3 0 0.75 0.25
R4 0.50 0.75 0.75

The concordance indicator in Table 6 varies between 0 and 1. It is equal to 1 only if there is unanimity or a
majority of criteria that are 100 percent in favor of Ri with respect to R j . In order to decide on the
superiority of one retail company with respect to another, the decision maker should set a concordance
threshold C*. Generally, it is chosen to be a majority greater than or equal to 75 percent (simple majority
tightened), that is: C* �0.75 (75 percent).

Taking into account the database of Table 6 and the concordance threshold C* we have the following
concordance test:

1 if C ( Ri ; R j ) �C*
Tc ( Ri , R j ) =

0 if otherwise

The results of concordance test are shown in Table 7.

Table 7 Outcomes of Concordance Test

R1 R2 R3 R4
R1 1 1 0
R2 0 0 0
R3 0 1 0
R4 0 1 1

The Electre I methodology considers another step: the construction of discordance test in order to take into
account of an excessive “distance” (dissimilarity) between the scores ATT j ( R j ) and ATTi ( Ri ). The
discordance test Td is fulfilled if the distance:
D ( R j , Ri ) = max [ ATT j ( R j ) - ATTi ( Ri )];

does not exceed discordance threshold D*. In order to simplify the analysis we suppose that the test of
discordance is fulfilled by all pairs ( Ri , R j ).

The ideas behind the test of discordance may be summarized as follows. The outranking methods consists
in examining the validity of the proposition “a outranks b”. The concordance test “measures” the arguments
in favor of saying so, but there may be arguments strongly against that assertion (discordant criteria). The
“discordant voices” can be viewed as vetoes.
 10

There is a veto against declaring that a outranks b if b is so much better than a on same criterion that it
becomes disputable or even meaningless to pretend that a might be better overall than b. The logic of the
test of discordance is quite similar to that on which statistical tests are based. Here as well, conventional
levels of significance (like the famous 5 percent rejection intervals) are widely used. The decision maker
decides the discordance threshold, that is he decides whether a hypothesis must be rejected or not.

If the discordance test is not passed alternatives a and b are said incomparable. They are too different to be
compared. For instance, the comparison of a Rolls-Royce with a small cheap car is meaningless because
the Rolls-Royce is incomparably better on many criteria but is also incomparably more expensive. Another
example, concerns the comparison of projects that involve the risk of loss of human life. Should one prefer
a more expensive project with a lower risk or a less expensive one with higher risk? One may advance that
the projects are too different to be compared.

Taking into account both the concordance and the discordance test we construct a binary outranking
relation S. Given two generic retail companies ( Ri , R j ) we say that Ri outranks R j if and only if the
concordance test and the discordance test are fulfilled, that is:

Ri S R j if and only if Tc and Td fulfilled.

Because we suppose that the discordance test is passed by all pairs ( Ri , R j ) the outranking relation S
coincides with the outcomes of concordance test of Table 4.27. That is:

Ri S R j if and only if Tc fulfilled.

The relation S may be represented by the graph of Figure 1.

Figure 1 Graph of S from Table 4.27 (C* �75 percent)

R1
Tesco
R3 R2
Giant Mydin
R4
Giant

Now, R3 is the “ 2nd worst in class” and R2 is the “worse in class”. But R1 and R4 are not comparable
structures: neither R3 outranks R4 nor the opposite. This is another important difference arising from the
refusal of the ordering based on the average benchmarking.
 11

3.7 Benchmarking On Customer Satisfaction

Benchmarking has consequences which are beyond the process itself: it reforms all the levels of the
company; modifies the process of manufacture of the product leads(drives); also reforms the hierarchical
organization of the company, the product itself, and the state of mind of the employees.

Through benchmarking, we get better understanding of the customer because it is based on the reality of
the market estimated in an objectivist way and a better economic planning of the purposes and the
objectives to achieve in the company for they are centered on what takes place outside controlled and
mastered. The management will get a better increase of the productivity, resolution of the real problems and
understanding of the processes and what they produce.

Table 8 Product Benchmarking towards customer satisfaction

We conclude that Tesco rank the highest on customer satisfaction towards product and it shall be the
benchmark or guiding star for other retail stores. Mydin, Carrefour and Giant need to benchmark Tesco’s
product strategy and improve to compete in the market.

Table 9 Price Benchmarking towards customer satisfaction

We conclude that Giant rank the highest on customer satisfaction towards price and it shall be the
benchmark or guiding star for other retail stores. It proofs that Giant’s “Everyday low price strategy” is a
success. Tesco rank the second with mean value of 3.94, still in the competition mood with Giant. Mydin
and Carrefour need to benchmark Giant’s pricing strategy and improve to compete in the market.
 12

Table 10 Promotion Benchmarking towards customer satisfaction

We conclude that Tesco rank the highest again on customer satisfaction towards promotion and it shall be
the benchmark or guiding star for other retail stores. Tesco promotion strategy is well organized and
effective; customers are aware of the latest promotion from the newspaper, flyers and promotion booklet.
Giant and Carrefour are a little bit behind with the mean value of 3.76 and 3.71. Mydin rank the last, it need
to benchmark Tesco’s promotion strategy, revise on its promotion strategy and improve to compete in the
competitive market.

Table 11 Place/ Distribution Benchmarking towards customer satisfaction

We conclude that Tesco rank the highest on customer satisfaction towards place and distribution and it shall
be the benchmark or guiding star for other retail stores. The other three retail stores having very close mean
value. Meaning the customer satisfaction towards place and distribution in four retail stores are well
perceived. From table 11, we conclude that Giant rank the second, followed by Carrefour and Mydin. They
are to benchmark Mydin’s, and improve on its place and distribution strategy.

4. Discussion on SWOT analysis

4.1 Strength

Multi-criteria decision-making (MCDM) is one of the most widely used decision methodologies in the
sciences, business, government and engineering worlds. MCDM methods can help to improve the quality
of decisions by making the decision-making process more explicit, rational, and efficient. It is not a
coincidence that a simple search (for instance, by using google.com) on the web under the key words
“multi criteria decision making” returns more than one million hits. In a decisional process the making of
choices derives from complex hierarchical comparisons among alternative options, which are often based
on conflictual criteria, a large number of external variables plays a relevant role in orienting decision-
 13

making. The strength of multi-criteria decision-making methods (MDMM) are to aid decision-makers to be
consistent with fixed ‘general’ objectives; to use representative data and transparent assessment procedures
and to help the accomplishment of decisional processes, focusing on increasing its efficiency. The Electre I
method, in which the criteria of the set of decisional alternatives are compared by means of a binary
relationship, defined as ‘outranking relationship’, are more ‘flexible’ than the ones based on a multi-
objective approach.

4.2 Weakness

An intriguing problem with decision-making methods which rank a set of alternatives in terms of a number
of competing criteria is that oftentimes different methods may yield different answers (rankings) when they
are fed with exactly the same numerical data. Thus, the issue of evaluating the relative performance of such
methods is naturally raised. This, in turn, raises the question how can one evaluate the performance of such
methods? Since it is practically impossible to know which one is the best alternative for a given decision
problem, some kind of testing procedures need to be determined.

4.3 Opportunity

In this paper, a new approach has been carried out for the use of the ELECTRE: ELimination Et Choix
Traduisant la REalité (ELimination and Choice Expressing the REality) model in marketing mix selection.
This work shows that ELECTRE can be used successfully in deriving a consensus ranking in benchmarking
to select the best in class.

4.4 Threat

In outranking approaches, the inaccuracy of the data can be modeled through the indifference and
preference threshold (so-called pseudocriteria). Of course, threshold must be assessed for each criterion and
for each problem separately.

5. Conclusion

It is not simply enough to identify the strengths, weaknesses, opportunities, and threats of the Electre
outranking method. In applying the SWOT analysis it is necessary to minimize or avoid both weaknesses
and threats. Weaknesses should be looked at in order to convert them into strengths. Likewise, threats
should be converted into opportunities. Lastly, strengths and opportunities should be matched to optimize
the potential of a firm. Applying SWOT in this fashion can obtain leverage for a company (Marketing
Strategy, 1998).

Sensitivity analysis showed that, in general, the project rankings were considerably more sensitive to
changes in the performances than they were to changes in the thresholds or weights. This is helpful and
means that within a relatively wide band of preference, the same projects are considered important. Further,
it requires the individual project sponsors to make the effort and ensure that the performance data is both
accurate and defensible. As can be seen, the marketing manager should have rough outline of potential
marketing activities that can be used to take advantage of capabilities and convert weaknesses and threats.
However, at this stage, there will likely be many potential directions for the managers to pursue. Due to the
limited resources that most firms have, it is difficult to accomplish everything at once. The manager must
prioritize all marketing activities and develop specific goals and objectives for the marketing plan
(Contemporary Marketing, 1992).

Outranking methods make it possible to deal with multicriteria benchmarking and avoid the shortcomings
of the traditional methods based on the average aggregate monocriterion. If applied to the measurement of
learning capability, they are a complete alternative to the traditional approach. They can support the
behavioral theory of organizational analysis initiated by H. Simon (Biggiero and Laise, 2003a, b). In fact,
even though H. Simon does not explicitly discuss the problem of criteria multiplicity nor does he apply
outranking methods, the behavioral theory is nonetheless perfectly comparable with them. The “levels of
 14

aspirations” hypothesized by Simon can be associated with the threshold of concordance and discordance
test.

The lower the threshold assigned to the concordance test the lower the aspiration levels will be and hence
the more the satisfying solutions will be. Outranking methods thus constitute a new and robust base on
which to found the entire edifice of the behavioral theory of benchmarking applied to measurement of
learning capability. They are a valid alternative to traditional methods, since they are equally rational and
rigorous without suffering from its shortcomings and application limitations.

6. Directions for Further Research

The relationships between customer satisfaction and behavioral outcomes are probably much more
complex than initially assumed. This study has looked only at a limited part of the puzzle of how customer
satisfaction translates into behavioral outcomes. In what way consumer characteristics moderate the
relationship between satisfactions and repurchase behavior is likely to be contingent on the product or
service category and the buying and usage process for that category. Other consumer characteristics not
included in this study, such as a propensity for variety seeking behavior or a recreational shopping
orientation, could potentially be important in many retail industries. Further research on how the effects of
satisfaction on behavior is moderated by different consumer characteristics would advance customer
satisfaction research as well as be of great managerial significance.

7. Scope and Limitation of the Study

The setting selected for conducting this marketing research was focus only on three multinational retail
stores and a homegrown retail store due to time constrain. Field researches were conduct in Tesco Stores
(Malaysia) Sdn Bhd, Carrefour of Magnificent Diagraph Sdn. Bhd., Giant of Dairy Farm International, and
the homegrown retail store, Mydin Mohamed Holdings Berhad.

The collection of primary data was based on a survey of 856 respondents who visit each respective retail
outlets, the number in the sample limited due to the restrictions of time to complete the project and
resources to support it.
 15

REFERENCES

Barlon, K. (2006) "The concept of the marketing mix" Presentation on marketing management, vol 1,
September, 2006, pp 2-7-Oulu university –Finland.
Belton, V., and Stewart, T. J., (2001), .Chapter 8: Outranking Methods,. Multiple Criteria Decision
Analysis: An Integrated Approach, Kluwer Academic Publishers, Boston, MA, USA.
Berry, Leonard L. and Ian H. Wilson (1977), Retailing: the next ten years. Journal of Retailing, 53 (Fall),
5–28.
Borden, N. (1964) "The concept of the marketing mix" Journal of Advertising Research, vol 4, June 1964,
pp 2-7.
Bose, U., Davey, A.M. and Olson, D.L. (1997) .Multi-attribute utility methods in group decision making:
Past applications and potential for inclusion in GDSS., Omega, 25, 691-706.
Brans, J.P. and Vincke, Ph. (1985) "A preference ranking organization method", Management Science, 31,
647-656.
Brownlie, D. and Saren, M. (1992), “The four Ps of the marketing concept: prescriptive, polemical,
permanent, and problematical”, European Journal of Marketing, Vol. 26 No. 4, pp. 34-47.
Domenico Laise. Benchmarking and learning organizations: ranking methods to identify “best in class”.
Benchmarking: An International Journal; Volume: 11 Issue: 6; 2004 Research Paper.
Doole, I. and Lowe, R. (1999), International Marketing Strategy, International Thompson Business Press.
London.
Douglas, S. P., & Craig, S. C. (1983), International marketing research, Englewood Cliffs, NJ: Prentice-
Hall, Inc.
Fernandez, I.P. McCarthy, T. Rakotobe-Joel. 2001. Benchmarking: An International Journal. Volume 8.
Issue 4
Figueira, J., Greco, S. and Ehrgott, M. (Eds.) (2004) Multiple Criteria Decision Analysis: State of the Art
Surveys, Springer, and New York.
Frazier, Gary L. and Tasadduq A. Shervani (1992), Multiple channels of distribution and their impact on
retailing. In R. A. Peterson (Ed.), The future of U.S. retailing: an agenda for the 21st century (pp.
217–238). New York: Quorum Books.
Glynn, Carroll J., Susan Herbst, Garrett J. O'Keefe, and Robert Y. Shapiro, Public Opinion (1999)
Government of Malaysia (2001), Eighth Malaysia Plan 2001-2005. Kuala Lumpur: Percetakan Nasional
Malaysia Berhad
Groonroos, C. (1997), “Keynote paper: from marketing mix to relationship marketing – towards a paradigm
shift in marketing”, Management Decision, Vol. 35 No. 4, pp. 322-39.
Hobbs, B.F., P. Meier, (2000), Energy Decisions and the Environment: A Guide to the Use of Multicriteria
Methods, Kluwer Academic Publishers, Boston, MA, USA.
Leyva-López, J-C. and Fernández-González, E. (2003) .A new method for group decision support based on
ELECTRE III methodology., European Journal of Operational Research, 148, 14-27.
Linkov, I., Varghese, A., Jamil, S., Seager, T.P., Kiker, G. and Bridges, T. (2004) .Multi-criteria decision
analysis: A framework for structuring remedial decisions at the contaminated sites., In: Linkov, I.
and Ramadan, A.B. (Eds.) Comparative Risk Assessment and Environmental Decision Making,
Springer, New York, pp. 15-54.
Louise Boulter. 2003. Benchmarking: An International Journal. Volume 10. Issue 6
Naresh K. Malhotra. Marketing Research: An Applied Orientation. International Edition. Fourth Edition.
2004
Rogers, M.G., M. Bruen and L.-Y. Maystre, (1999), .Chapter 3: The Electre Methodology,. Electre and
Decision Support, Kluwer Academic Publishers, Boston, MA, USA.
Triantaphyllou, E. (2000) Multi-Criteria Decision Making Methods: A Comparative Study, Kluwer
Academic Publishers, Dordrecht.
Triantaphyllou, E. and Sanchez, A. (1997) "A sensitivity analysis approach for some deterministic multi-
criteria decision making methods", Decision Sciences, 28, 151-194.