Onkvisit S., Shaw J. International Marketing: Analysis and Strategy

Подождите немного. Документ загружается.

Demographics, in spite of being apparent and

easily understood, should not be readily embraced

in terms of conceptual equivalence without an

examination of the varying frames of reference. A

demographic variable such as sex is universal, and

a question of this nature can be used in a cross-

cultural study.Age, on the other hand, is not always

considered in the same way – the Chinese include

the time during pregnancy in their age. Educational

level, likewise, does not have the same meaning

everywhere.The meaning of primary school, grade

school, secondary school, high school, college, and

university varies greatly.A primary school can range

from four years to eight years, depending on

the country. In some countries, a college may be

nothing more than a vocational school. A college

education in one country may not at all be equiva-

lent to that in another. It is thus wise in many

instances to ask about the number of years of

schooling attended by the respondent.

Let us consider the measurement of poverty.The

measurement process usually comprises three steps:

(1) choosing an indicator of individual economic

welfare (which tends to be household expenditure

or income over a certain period, adjusted for dif-

ferences in household size and cost of living), (2)

setting a poverty line (which deﬁnes the level of

welfare deemed necessary for an individual to

escape poverty), and (3) identifying an aggregate

poverty measure (one which summarizes the infor-

mation from the ﬁrst two steps).

The three steps mentioned are anything but easy

because the process has a number of contentious

issues.There are several conceptual difﬁculties asso-

ciated with each step. In the case of the household

welfare indicator, the disputes include such issues

as: should the indicator be consumption or income,

what should each one include, and whether and

how should it consider cost-of-living differences?

Regarding the poverty line, at what overall level

should it be set, and should the level vary in real

terms by subgroup or date? Finally, with regard to

the aggregate poverty measure, how should it treat

inequality among the poor? One has to wonder

whether economists are well equipped to answer

these questions or whether behavioral scientists

(e.g., psychologists) can do a better job by focusing

on individuals’ perceptions of their well-being.

The measurement issues and debates mentioned

above play an important role in the debate over

globalization’s impact on poverty and inequality.

While some point out that the proportion of people

living in extreme poverty in the developing world

fell sharply in 1990s, others argue that globalization

has created greater poverty. Some contend that

income equality has been rising in the world,

whereas others argue the opposite. Such conﬂicting

claims stem not only from differences in data and

methods used but also from the important concep-

tual distinctions discussed above.The opposing posi-

tions of proglobalizers and antiglobalizers represent

a question of interpretation.

One kind of conceptual equivalence a researcher

should pay close attention to is functional equiv-

alence. A particular object may perform varying

functions or may satisfy different needs in different

countries. A bicycle is a recreation device in some

countries and a basic transportation device in oth-

ers.Antifreeze is used to prevent freezing of engine

coolant in cold countries but to prevent overheating

in countries with warm climates. A hot milk-based

drink is perceived as having restful, relaxing, and

sleep-inducing properties in the United Kingdom;

Thai consumers, in contrast, view the same drink as

stimulating, energy giving, and invigorating. Thais

consume hot milk as either a substitute for or a sup-

plement to breakfast.Therefore, a valid comparison

requires the use behavior to have been developed in

response to similar problems in different cultures.

Use differences must be built into the measuring

instrument if meaningless results are to be avoided.

Deﬁnitional or classiﬁcation equivalence

is another type of conceptual equivalence that

requires careful treatment.This factor involves the

way in which an object is deﬁned or categorized –

either perceptually by consumers or ofﬁcially by law

or government agencies. In addition, even demo-

graphic characteristics are subject to the problem of

deﬁnitional or classiﬁcation equivalence.Age is one

such example. Persons in the same age group in

226

MARKET RESEARCH AND INFORMATION SYSTEM

different countries are not necessarily at the same

stage of life or family life. When a boy becomes a

man or a legal adult depends on the deﬁnition used

by a particular country. The age of legal adulthood

may vary anywhere from the very early teens to 21

years. In India, a “boy” of only 13 or 14 years of age

can legally marry a girl who is a few years younger.

As a result, chronologically identical age groups

within two or more countries do not necessarily

lead to comparable equivalent groups. It is thus not

possible to standardize age groupings and have the

same deﬁnitional or classiﬁcation equivalence on an

international basis.

Instrument equivalence

In devising and using a measure, a researcher should

distinguish between two kinds of measuring instru-

ments: emic and etic. Emic instruments are tests

constructed to study a phenomenon within one

culture only. Etic instruments, on the other hand,

are those that are “culture universal” or “culture

independent.” As such, when properly translated,

etic measuring instruments may be used in other

cultures.

International marketing research often necessi-

tates some measurement adaptation to overcome

the problems of reliability and validity. The rest of

this section on measurement is devoted to the prob-

lems and adjustments of measurement in terms of

conceptual equivalence, instrument equivalence,

linguistic equivalence (translation), response style,

measurement timing, and external validity.

A measuring instrument should be evaluated

in terms of its international suitability. Let us con-

sider the CETSCALE instrument which is used to

measure consumer ethnocentrism. Consumers who

are ethnocentric do not like to purchase foreign

products because such purchases are unpatriotic and

harm the domestic economy and employment.The

instrument was originally validated with samples of

American consumers only. Subsequently, the instru-

ment was applied to samples from France, Japan,

Germany, and the USA.The CETSCALE was found

to be a reliable measure across the four countries.

It also has some degree of validity as evidenced by

the consistent pattern of correlations across each

country’s sample.

An investigation of product-country images,

lifestyles, and ethnocentric behaviors of Turkish

consumers detected the robustness of the

CETSCALE, thus lending further support to the

earlier studies which were conducted in Western

countries.

Another study investigated whether the

CETSCALE could be used as a measure of con-

sumer ethnocentricism on an international basis. In

order to assess the instrument’s reliability and valid-

ity in as many countries and different conditions as

possible, the study focused on Spanish consumers’

ethnocentric tendencies.The ﬁndings indicated that

the scale measured a unidimensional construct and

that the degree of measurement error was quite

acceptable.

In consumer research, the most popular rating

scales are the Likert scale and the semantic differ-

ential scale. When a Likert scale is employed, a

respondent is asked to respond by indicating agree-

ment (or disagreement) and the relative intensity of

such an agreement to each item or question. The

degree of agreement (or disagreement) usually

ranges from “strongly agree” to “strongly disagree.”

The semantic differential scale measures the

meaning of an object to a respondent who is asked

to rate that concept on a series of bipolar rating

scales (i.e., the extremes of each scale employ

adjectives of opposite meanings).

A researcher must keep in mind that the Likert

and semantic differential scales,though proven to be

satisfactory in measuring behavior and opinion in

the USA, may not be understood or may not elicit

the same manner of response in other markets. One

study found signiﬁcant main effects due to country

(China, South Korea, Japan, and the USA) and type

of scale used (Likert and semantic differential) as

well as the interaction effects. The Likert and

semantic differential scales appear to have emic

(culture-bound) properties and should be treated as

culture-speciﬁc instruments.

The nature of the measurement scale can affect

responses. Asking whether a particular object is a

1111

10111

20111

45111

227

MARKET RESEARCH AND INFORMATION SYSTEM

10 on a scale of 1 to 10 may be meaningful to

American respondents but may not be understood

by respondents in most other countries where this

rating scheme is uncommon. It is imperative that

scales be tailor-made and tested carefully in each

culture in terms of relevance and appropriateness.

One issue that may pose a problem is whether a

scale should be in balance. A balanced scale is one

that has the same number of choices for the positive

side as well as for the negative side of the scale.

Should scales have a positive or negative skew? It is

possible that people in certain cultures may tend to

rate positively. As such, should a researcher offer a

larger number of alternatives among the positive

categories to achieve discrimination?

Another issue related to rating scales is the

number of choices (possible answers) provided with

each question.A 7-point scale, for example,may not

yield more information than a 5-point scale in the

USA, but the former may prove useful elsewhere.

In general, a higher-point scale is needed when a

group is relatively homogeneous, since more infor-

mation is required to distinguish among members

of the group.

According to one study, the 3-category scales

were more equivalent than the 5-category scales.

This ﬁnding raises several interesting questions.“Do

scales with fewer categories pose fewer equivalence

problems? Is it more difﬁcult to systematically

bias scales with fewer categories because choices

are constrained? If so, it would create a tradeoff

between designing items to capture variance more

accurately and designing items less susceptible to

response bias.”

Content validity should be routinely exam-

ined. Content validity is critical when a measuring

instrument developed for one population is going

to be used with another population. A test has

content validity when it consists of items or ques-

tions judged to be representative of the speciﬁed

universe of content. A test to measure the spelling

ability of ﬁfth-grade students should comprise

words which students at that level of education are

supposed to have learned. If a test were to contain

a relatively large number of sports terms, for

example, it would not have content validity because

it would be unfair to those not interested in sports.

One researcher has constructed a scale to measure

Guanxi, a behavior that is pervasive in China.

Naturally, the items used must reﬂect the typical

activities in Chinese society.

Standardized tests, such as IQ, SAT, and GMAT

tests, are somewhat controversial due to content

validity or the lack of it. In the USA, such tests are

criticized by the poor and various ethnic minorities

for being biased because the questions purportedly

reﬂect the experience of white, middle-class

Americans. As such, the questions are not repre-

sentative of the universe of content of interest.The

problem is greatly ampliﬁed when US colleges

require foreign students to take these standardized

tests to determine admission eligibility.

Uniformity is usually desirable, but that should

be achieved at the expense of content validity. A

researcher must keep in mind that some questions

are culturally speciﬁc, making them difﬁcult for for-

eigners, no matter how acculturated or smart they

may be, to understand. Some of the questions in

US are meaningful to Americans but incomprehen-

sible to Britons. A foreign language can complicate

matters even more. It would be unreasonable to say

that non-English-speaking children are unintelligent

just because they cannot answer questions written

in English.

A researcher must remember that the content

validity of an instrument depends on the purpose of

the study as well as on the groups of people under

investigation.The universe of content can vary from

group to group and from culture to culture.

Standardized tests, therefore, may not necessarily

work. Identical questions do not guarantee compa-

rable data from different countries, and some vari-

ations in questions may be necessary.

Linguistic equivalence

Linguistic differences can easily invalidate the results

of a study. There is no question that poor transla-

tion works against sound research methodology.

Particularly disadvantageous are imprecise literal

228

MARKET RESEARCH AND INFORMATION SYSTEM

translations made without regard to the intended

purpose or meaning of the study. Linguistic

equivalence must be ensured when cross-cultural

studies are conducted in different languages.

The goal of linguistic equivalence requires the

researcher to pay close attention to potential trans-

lation problems. According to Sekaran, translators

should pay attention to idiomatic vocabulary, gram-

matical and syntactical differences in languages, and

the experiential differences between cultures as

expressed in language.

Vocabulary equivalence

requires a translation that is equivalent to the orig-

inal language in which the instrument was devel-

oped. Idiomatic equivalence becomes a problem

when idioms or colloquialisms unique to one lan-

guage cannot be translated properly into another.

Grammatical and syntactical equivalence poses a

problem when long passages must be translated.

With regard to experiential equivalence, care

should be taken that accurate inferences in a given

statement are drawn by respondents from various

cultures. As in the case of one study exploring

strategic alliances in the emerging Latin American

countries, the survey instrument was developed in

English and was translated into Portuguese and

Spanish by native speakers who were business pro-

fessors. Back translation was later performed on

each local language instrument until conceptual and

functional equivalence was achieved.

There are several translation techniques that

can be used: back translation, parallel-blind trans-

lation, committee approach, random probe, and

decentering.

With back translation, the research

question is translated by one translator and then

translated back into the source language by

another translator. Any discrepancy between the

ﬁrst and last research questions indicates translation

problems.

In a parallel-blind translation, the question is

translated by several individuals independently, and

their translated statements are then compared.

The committee approach differs from the parallel-blind

technique in the sense that the former permits com-

mittee members to discuss the research questions

with one another during the translation.

A random probe involves placing probes at random

locations in both the source and translated question

during pretesting in order to ensure that the respon-

dents understand questions in the same way. In

decentering, both the source version and target

version are viewed as open to modiﬁcation. If trans-

lation problems are recognized for the source

document, then it should be modiﬁed to be more

easily translatable. Consequently, the source ques-

tion becomes more lucid and precise. For example,

the statement “I am an aerobic instructor” may be

changed to “I am a dance-exercise teacher.”

Regardless of the method of translation, there

will always be concepts that cannot be translated

into certain languages or that cannot be asked in a

meaningful way in certain cultures. 7Up’s “uncola”

slogan is an example of such a concept. Some lan-

guage purists may even contend that linguistic com-

parability is an unattainable goal. In any case, none

of the translation methods can guarantee linguistic

equivalence, but the recognition of potential trans-

lation difﬁculties should at least minimize problems.

When Gillette globally launched Venus, a triple-

blade razor for women, the product hit twenty-nine

countries simultaneously.

A global campaign must

pay attention to local languages and requirements;

so a website was designed to serve as a global tem-

plate for ﬁfteen regional sites.Since a package might

contain as many as three languages, the company set

up a translation war room in London ﬁlled with

designers, translators, and lawyers. Because the

promotional phrases of “soft, protective cushions”

and “reveal the goddess in you” carried signiﬁ-

cance nuances, the translators were required to

express back the core idea in their native language.

In the case of the word “goddess,” Gillette chose

the lower case so as to avoid a cultural or religious

controversy.

Response style

Response styles are “tendencies to respond system-

atically to questionnaire items on some basis other

than what items were speciﬁcally designed to

measure.”

Some common response styles are:

1111

10111

20111

45111

229

MARKET RESEARCH AND INFORMATION SYSTEM

acquiescence, extreme responding, use of the

middle response category on the rating scales, and

socially desirable responding (see Exhibit 8.2). Such

response styles, by contaminating respondents’

answers to substantive questions, threaten the

validity of empirical ﬁndings of both domestic

and international marketing research. One large,

representative sample of consumers from eleven

European Union countries showed systematic

effects of response styles on the scale scores as a

function of two scale characteristics (proportion of

reverse-scored items and extent of deviation of scale

mean from the midpoint of the response scale).The

correlations between the scales could be biased

upward or downward depending on correlation

between response style components. Furthermore,

a secondary analysis of a comprehensive collection

of measurement scales found many scales to fail to

control adequately for response styles.

People from different cultures have different

response styles when dealing with the same ques-

tions.Asian people, for example, are generally very

polite and may avoid making negative statements

about a product. In such a case, a researcher may

have to employ an even-number scale (e.g., yes/no)

in order to receive an objective response.The even-

number scale,however, may create another problem

in the sense that it may elicit an opinion that does

not truly exist.A researcher may then want to con-

sider whether to use an odd-number scale so that

respondents have an option of making a neutral

response.

Another issue concerning response style varia-

tion is the willingness – or lack of it – on the part

of a respondent to take an extreme position, espe-

cially when such a position is not a popular one. It

may thus be appropriate to adjust for respondents’

extreme positions and/or response style by using

“normalized” scores rather than raw scores.

More often than not, one must be sensitive to the

logic of a language. English-speaking people, for

example, use the words yes and no in a manner that

differs from the way non-English-speaking people

do. Thus, “yes” is not always “yes,” and sometimes

“no” may mean “yes” instead. The English and

Western languages require the answer to be “yes”

when the answer is afﬁrmative. It is “no” when the

answer is negative. For other languages, however,

“yes” or “no” refers to whether a person’s answer

afﬁrms or negates the question asked. Consider the

question,“Don’t you like it?”An American who does

not like the item in question would say “no,” but a

non-Westerner would probably say “yes,” indicating

agreement with the question. For the non-

Westerner, the statement “Yes, I don’t like it” is per-

fectly logical.Therefore, it is a good idea to phrase

questions in simple, positive terms. Questions with

negatives should be avoided.

It should be noted that an interviewer is part

of the measuring instrument. As such, the charac-

230

MARKET RESEARCH AND INFORMATION SYSTEM

Acquiescence: the tendency to agree with items

regardless of content. Also called agreement ten-

dency, yea-saying, or positivity.

Disacquiescence: the tendency to disagree with

items regardless of content. Also called disagreement

tendency, nay-saying, or negativity.

Extreme response: the tendency to endorse the

most extreme response categories regardless of

content.

Response range: the tendency to use a narrow or

wide range of response categories around the mean

response.

Midpoint: the tendency to use the middle scale

category regardless of content.

Noncontingent: the tendency to respond to items

carelessly, randomly, or nonpurposefully.

Source:

Adapted from Hans Baumgartner and Jan-Benedict

E.M. Steenkamp, “Response Styles in Marketing Research:

A Cross-National Investigation,”

Journal of Marketing

Research

38 (May 2001): 145.

EXHIBIT 8.2 RESPONSE STYLES

teristics of the interviewer may interact with a

respondent’s response style.As shown in one study,

response quality was affected by interaction effects

of respondent and interviewer gender and ethnic-

ity.

Overall,Anglos and female interviewers were

able to generate more responses.Also in support of

the general deference theory, both Hispanic and

Anglo respondents deferred to an interviewer

from a different background when asked about the

interviewer’s culture but not when queried about

noncultural questions.

Measurement timing

A cross-national study can be done simultaneously,

sequentially,or independently.A researcher may ini-

tially think it advantageous for a study to be con-

ducted simultaneously in the countries of interest.

But simultaneous studies in different markets may

present problems of data comparability, especially

due to seasonal factors. Consider a study of soft

drink consumption, which is generally highest in

the summer. Winter in Canada occurs when it is

summer in Argentina. Simultaneous studies in such

a case would yield invalid results.

History is a speciﬁc external event that may

affect the outcome or respondent behavior within

a study. Acting as an extraneous variable, history

complicates cross-national studies. For example, an

important election (i.e., history) held in one

country and not in another may considerably affect

the results of a study. Likewise, sales in the USA

during the Christmas season are much higher than

sales during the rest of the year, whereas sales in

China and Hong Kong tend to be higher than

normal during the period of the Chinese New Year.

Another complicating factor can be a product’s

life-cycle stage.The same product may be in differ-

ent stages of its life cycle in different countries.

Because of the phenomenon of an international

product’s life cycle, less developed countries may

lag behind more advanced nations in adopting the

product. The sales of the product in two or more

countries for the same period may not be compara-

ble, casting doubt on the wisdom of conducting

simultaneous studies. Because of this problem, it

may be better in some cases to conduct “simultane-

ous” studies when the countries are in the same

product-life-cycle stage. In addition, it may be

prudent to conduct sequential studies for a partic-

ular stage in the product’s life cycle when the

product is not introduced at the same time in the

countries of interest.

External validity

There are two major types of validity: internal and

external. So far this chapter has only discussed

internal validity. A study is said to have internal

validity when it accurately measures the charac-

teristic or behavior of interest.

However, a researcher is prudent to ask whether

the ﬁndings concerning a particular sample in a

particular study will hold true for subjects who did

not take part in the study. External validity is

concerned with the generalization of research

results to other populations. Ordinarily, there is a

limit on how far research ﬁndings can be general-

ized. Consequently, the ﬁndings may not be applic-

able to other groups or populations, other products,

other cities, other countries, or other cultures.

Consider the health concerns about the use of

saccharin. Many studies have provided evidence for

a link between saccharin and cancer. The scientiﬁc

evidence has shown that saccharin is a carcinogen

(cancer-causing substance) and that it may act as

a cancer catalyst in the sense that it assists the

carcinogen to work. Internal validity has never

been much of an issue because studies have repeat-

edly proven that saccharin indeed causes cancer, at

least in laboratory rats. Critics of this evidence have

concentrated on the issue of external validity.

According to the critics, producing a disease in

rats is not a valid basis for predicting the effect of

saccharin on humans. The US Food and Drug

Administration (FDA), however, has dismissed this

criticism by pointing out that the analysis is not

accompanied by any proof that the mechanism

causing cancer in rats would be different in humans.

As far as the FDA is concerned, the studies in

1111

10111

20111

45111

231

MARKET RESEARCH AND INFORMATION SYSTEM

question are both internally and externally valid.

Ironically, after several decades, new evidence

seems to indicate that saccharin is safe for con-

sumption after all.

The FDA, surprisingly, has reacted conserva-

tively when test results submitted have used foreign

subjects. Although the FDA has accepted results

using rats (whether American or not) as surrogates

for human beings, it had long shown great reluc-

tance in accepting foreigners as being similar to

Americans in terms of the impact of drug treat-

ment. New FDA rules now permit a drug company

to submit the results of clinical studies done abroad

to support claims of safety and efﬁcacy. Such studies

are permitted as long as they have been conducted

by competent foreign researchers, can be validated,

and are applicable to the US population.

External validity is generally not a problem when

the matter of concern is physiological in nature.The

same cannot be said about psychological matters.

Thus, people may have similar demographics but

diverse attitudes and behavior. The behavior of one

consumer group or the general behavior of people

in one country should not simply be extended to

another group or people. It may not be reasonable

to assume that a study of a marketing problem in

one country or the results obtained within one

culture will also be applicable to a marketing

problem or people in other cultures. Because of

differences between cultures, the problem of exter-

nal validity is ampliﬁed when marketers engage in

international marketing research.

Since research in the area of satisfaction has

focused primarily on American subjects, the degree

of generalization of the results must be evaluated.

Toward this end, one study tests the basic disconﬁr-

mation of expectations model in a very different

culture (Taiwan).

Based on the matched samples

of American and Taiwanese respondents, the results

conﬁrm the generalizability of the model in this par-

ticular Asian culture. Along the same lines, 100

American and 100 Chinese Singaporean customers

were used to study whether the Fishbein behavioral

intention model could be used to explain their retail

bargaining behavior.While the Chinese have a higher

level of bargaining intention as well as a more com-

petitive bargaining style, the two groups’ bargaining

attitudes and subjective norms do not differ.

Another study evaluates the applicability of the

Narver and Slater market orientation scale in the

context of service ﬁrms in Central Europe’s transi-

tion economies.

The survey measures the levels of

market orientation in 205 business-to-business ser-

vices companies and 141 consumer services com-

panies in Hungary, Poland, and Slovenia.The results

are consistent with the prediction found in pre-

dominantly Western marketing literature, thus

establishing the scale’s reliability and validity.

Service ﬁrms with the higher levels of market ori-

entation are: (1) more often found in turbulent,

rapidly changing markets, (2) more likely to pursue

longer term market-building goals (vs. short-term

efﬁciency objectives), (3) more likely to pursue dif-

ferentiated positioning through superior levels of

service, and (4) better performers on both ﬁnancial

and market-based criteria.

MARKETING INFORMATION SYSTEM

A marketing information system (MIS) is an inte-

grated network of information designed to provide

marketing managers with relevant and useful infor-

mation at the right time and place for planning,

decision making,and control.As such, the MIS helps

management identify opportunities, become aware

of potential problems, and develop marketing plans.

The MIS is an integral part of the broader manage-

ment information system. For example, Benetton’s

stores around the world are linked by computer.

When an item is sold, its color is noted. The data

collected make it possible for Benetton to deter-

mine the shade and amount of fabric to be dyed each

day, enabling the ﬁrm to respond to color trends

very quickly.

In spite of computer and other advanced tech-

nologies, “dark age” methods of data collection

and maintenance are still prevalent. In many parts

of the world, a knowledge and application of

modern management systems is nonexistent. In

many ofﬁces, scores of desks are crammed together

232

MARKET RESEARCH AND INFORMATION SYSTEM

in the same room. Each employee may have his

or her own unique and disorganized system for

ﬁling documents and information. New employees

inherit these ﬁling and accounting systems and

modify them to ﬁt their needs. The unindexed

ﬁling system, a long honored custom, makes each

employee practically indispensable since no one

knows how to ﬁnd a document that has been ﬁled

by someone else. Such problems are not conﬁned

only to less developed countries. Advanced nations

such as some European countries and Japan are still

struggling with the automation of their information

systems.

There is often a misconception that an MIS must

be automated or computerized. Although many

ﬁrms’ systems are computerized, it is possible for

a company to set up and use a manual system that

can later be computerized if desired. With

modern technology and the availability of affordable

computers, it seems quite worthwhile for an inter-

national ﬁrm to install a computer-based informa-

tion system. Yet no one should assume that the

computer is a panacea for all system problems,espe-

cially if ﬂaws are designed into the MIS. A poorly

designed system, whether computerized or not,

will never perform satisfactorily.

System development

For the MIS to achieve its desired purpose, the

system must be carefully designed and developed.

Development involves the three steps of system

analysis, design, and implementation. System

analysis involves the investigation of all users’

information needs. The relevant parties must be

contacted to determine the kinds of information

they need, when it is needed, and the suitable

format through which the information is made

available. Because information is not cost free, it

may not be feasible to satisfy all kinds of informa-

tion needs.The beneﬁt of the information provided

must be compared with the cost of obtaining and

maintaining it. Only when the beneﬁt is greater

than the cost can a particular information need be

accommodated.

System design should be the next major con-

sideration. System design transforms the various

information requirements into one or more plans

that clearly specify the procedures and programs in

obtaining, recording, and analyzing marketing data.

Alternative or competing plans are developed and

compared, and the most suitable one is ultimately

selected.

The ﬁnal step comprises system implementa-

tion.The chosen system is installed and checked to

make certain that it functions as planned. Both those

who operate the system and those who use it must

be trained, and their comments should be evaluated

to ensure the smooth operation of the system. Even

after implementation, the system should continue

to be monitored and audited. In this way, manage-

ment can make certain that the system serves

the needs of all users properly while preventing

unqualiﬁed persons from gaining access to the

system. Security Paciﬁc Bank, for example, lost

$10.3 million when a consultant was able to obtain

an electronics fund transfer code and use it to

deposit money into his Swiss bank account.

Desirable characteristics

For the MIS to be effective and efﬁcient, it should

possess certain characteristics. In general, the sys-

tem should be user-oriented, systematic, expand-

able, comprehensive, ﬂexible, integrated, reliable,

timely, and controllable. Marketing and environ-

mental information should be routinely received,

evaluated, and continuously updated.

As implied above, certain characteristics are

universal in the sense that all information systems,

whether large or small and whether domestic or

global, should possess them. However, unlike a

domestic MIS, an international marketing informa-

tion system (IMIS) needs to satisfy additional crite-

ria in order to ensure that the system can effectively

serve a company’s international marketing strategy.

Some of these criteria are time independence,

location independence, cultural and linguistic com-

patibility, legal compatibility, standardization/uni-

formity, ﬂexibility, and integration.An IMIS should

1111

10111

20111

45111

233

MARKET RESEARCH AND INFORMATION SYSTEM

be time independent by providing around-the-clock

services. Being location independent means that the

system must be capable of allowing submission and

use of data at the various strategic points globally.

Cultural knowledge and linguistic capability, when-

ever possible, should be built into the software and

system. Naturally, the implementation of an IMIS

must conform to local laws. To assure uniformity,

certain kinds of information need to be standard-

ized. Yet some degree of ﬂexibility is required as

well since a good information system should be

user oriented. Finally, given the number of users,

countries, and locations involved, an IMIS must be

designed to allow data integration.

According to the CEO of Peoplesoft, there are

twelve imperatives for a real-time enterprise: (1)

standardization of business processes, (2) pure

Internet architecture, (3) minimizing customiza-

tions, (4) holding software vendors accountable, (5)

accommodating multiple databases, (6) highly scal-

able applications, (7) multilingual, multicurrency,

(8) interoperability between vendors, (9) embed-

ded business analytics, (10) fewer vendors, broader

product lines, (11) change management, and (12)

CIO.

Regardless of size, each company needs to

have the capability to operate on a global basis.

Therefore it is important that, whenever possible,

business processes should be standardized across

subsidiaries, geographies, and divisions.

When operating in multiple countries, a

company may have a number of databases (or sub-

databases). Such databases should be deﬁned by

international standards so as to allow efﬁcient access

and data comparison. While such factors as lan-

guage, consumption habits, and retail trade struc-

ture may continue to maintain a distinct local

character, the uniform methods allow for a quick

examination of market situations across the world

or a particular region.

A company that has only one website for the

world needs to design the site that takes national or

cultural differences into account.The needs of local

consumers must be accommodated. To improve

service, the company must design the registration

and order forms to accept foreign addresses and

currencies.A US website thus must add a “country”

ﬁeld and set the postal code to accept more than ﬁve

alphanumeric digits, while not making “state” a

mandatory ﬁeld.

Subsystems

The MIS consists of several systems: internal report-

ing,marketing research, and marketing intelligence.

The internal reporting subsystem is vital to the

system because a company handles a great deal

of information on a daily basis. The marketing

department has sales reports.The consumer service

department receives consumers’ praises and com-

plaints. The accounting department routinely gen-

erates and collects such information as sales orders,

shipments, inventory levels, promotional costs, and

so on. All of these types of internally generated

information should be kept and made available to all

concerned and affected parties.

For externally generated information, the MIS

should consist of two subsystems. One of these is

the marketing research subsystem.The activities of

this subsystem have already been discussed exten-

sively.The other subsystem is the marketing intelli-

gence or environmental scanning subsystem. The

responsibility of this subsystem is to track environ-

mental changes or trends. This subsystem collects

data from salespersons, distributors, syndicated

research services, government agencies, and from

publications about technology, social and cultural

norms, the legal and political climate, economic

conditions, and competitors’ activities.

The implementation of the MIS must conform to

local laws (see It’s the Law 8.1). Many countries,

concerned with citizens’ privacy, have laws that

restrict free ﬂow of information. In England, data

users are required by the Data Protection Act to reg-

ister with the Ofﬁce of Data Protection Registrar;

otherwise, heavy ﬁnes are levied. Computer users

must state how personal information was obtained

and how it will be used. Furthermore, British resi-

dents have the right to see personal data about

themselves. Similarly, Quebec’s privacy legislation

restricts the activities of direct marketers who

234

MARKET RESEARCH AND INFORMATION SYSTEM

target the French-speaking province. The law

requires marketers to notify customers that a ﬁle is

being created on them and to explain why a data-

base record has been set up. Consumers must be

given a “reasonable opportunity” to remove their

names from mailing lists.

The MIS should be designed to do more than data

collection and maintenance. It should go beyond

data collection by adding value to the data so that

the information will be of most use to users. The

MIS thus requires an analytical component that is

responsible for conceptually and statistically analyz-

ing the data. This component may even go a step

further by offering conclusions and recommenda-

tions based on the analysis of data. In the case of

Shiseido Co., a new computer network ﬁnally links

factories, sales staff, and 16,000 stores, and enables

managers to more accurately forecast demand for

its beauty products. As a result, the company has

reduced inventory by one-third.

1111

10111

20111

45111

235

MARKET RESEARCH AND INFORMATION SYSTEM

There are two underlying problems with the quality of

international names and addresses.One problem has to

do with storing and manipulating such data in tradi-

tional domestic customer databases. The other prob-

lem focuses on ﬁnding the address standardization

tools that are capable of handling international data.

Fortunately, there are several products and services

that are designed to clean up international addresses.

It is important to deal with data at the reformat-

ting stage. In Singapore, customer names are from

three different cultures: Chinese,Malay (Islamic), and

Western. Chinese names are presented in last–middle–

ﬁrst order, while Islamic names (e.g., Mohammed bin

Abdullah) are approximately (but not exactly) last–

ﬁrst order. For Western-style names, they are ﬁrst–

middle–last order. Caribbean addresses need to be

understood and parsed as either American style or

Spanish style.

To accommodate international data, a typical US

record layout will have to be modiﬁed. For example,

the length of ﬁelds must be extended to accept longer

names and addresses.The phone and fax numbers ﬁeld

should include a country code and handle more than

ten digits. In the case of the address ﬁelds, they may

have to be lengthened beyond Address1 and Address2

to include Address3 and perhaps even Address4. In

addition, the layout needs to support the expanded

character and/or Unicode character sets. The rules

behind the postal code ﬁeld must be evaluated to allow

for alphabetical characters and spaces. Additional

ﬁelds are required to accommodate country name,

region name, and geographic code (e.g., Asia, Sales

Territory 1, etc.). It is desirable to consider parsing

ﬁrst name, last name, and multiple last names (e.g.,

Hispanic names). Finally, be sure to accommodate

multiple international currencies.

A ﬁrm’s database should be ﬂexible enough to

accommodate each country’s preferred style in main-

taining names and addresses. Concerning addresses, in

France, Germany, and Italy, the postcode usually pre-

cedes the town on a single line. In Britain, the post-

code should follow the country name. If the country

name is omitted, the postcode follows the town or city.

In France and the United Kingdom, the house number

precedes the street name. In Germany, Italy, and other

European countries, it is the opposite.

Regarding address windows, the window should

go on the left-hand side in the case of Britain, the

Netherlands, and Ireland. But in France, Sweden,

Belgium,and Germany, the right-hand side is the norm.

Regarding ﬁeld sizes, ﬂexibility is a necessity. The

German language may require up to one-third more

characters than the English-language equivalent.

Sources:

“Cher or Chere? (And a Few Other Format

Issues),”

DM News

, April 11, 1994, 29; “Ode to

International Merge/Purge,”

Anchorline

,Anchor Computer’s

Newsletter, Summer 1998; “Long-Range Missives,”

Direct

March 1, 2002, 33–5.

MARKETING STRATEGY 8.1 HOW TO ADDRESS THE ADDRESS PROBLEMS