The Effects of Culture and Socioeconomics on the Performance of Global Brand Image

Strategies
Author(s): Martin S. Roth
Source: Journal of Marketing Research, Vol. 32, No. 2 (May, 1995), pp. 163-175
Published by: American Marketing Association
Stable URL: http://www.jstor.org/stable/3152045 .
Accessed: 27/08/2013 18:02

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Marketing Association is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Marketing Research.

http://www.jstor.org

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
 I

MARTINS. ROTH*

Developing and managing brand image is an important part of a firm's

marketingprogram.However, littleresearch has been done (1) on linkingthe
use of brand image strategies to product performance or (2) on managing
brand images in global markets. The author examines the brand image-
performance linkage for consumer goods in two categories marketed
internationally.He also develops a conceptual framework that identifies
various culturaland socioeconomic environmental characteristics of foreign
markets that are hypothesized to affect brand image performance. Results
from a 10 country/60 region study indicate that cultural power distance,
culturalindividualism,and regional socioeconomics affect the performance of
functional (problem prevention and solving), social (group membership and
symbolic), and sensory (novelty, variety, and sensory gratification) brand
image strategies. The author then discusses the implications for managers
marketing brands internationallyand the directions for further research.

The Effects of Culture and Socioeconomics

on the Performance of Global Brand
Image Strategies

Brandimage managementis a criticalpartof a company's First, research should link brand image strategies and
marketingprogram.Communicatinga clearly defined brand productperformance.Althoughresearchon brandextensions
image enables consumers to identify the needs satisfied by has begun to examine the effects of using establishedversus
the brand (Park, Jaworski,and Maclnnis 1986) and differ- new brands on market share (Simon and Sullivan 1993;
entiate the brand from its competitors (DiMingo 1988; Smith and Park 1992), the relationshipbetween brandimage
Reynolds and Gutman 1984), and has been prescribedby strategiesand performancehas received little attention.Yet
both marketingpractitioners(Ogilvy 1963) and researchers brandimage strategiescan affect measuresof relativeperfor-
(Gardnerand Levy 1955) as a key to product success. In mance (e.g., marketshare)as well. Specifically,a need-based
fact, brandimage is an integralcomponentof a brand'seq- image strategy establishes a brand's position (relative to
uity, that is, the value of a brandin the minds' of consumers competition) within a product category, defining its niche
(Keller 1993). Developing a needs-based image strategy and establishing its market potential. Research has shown
providesthe foundationfor marketingprogramdevelopment that serving largerniches typically leads to betterlong-term
and enables the brandto create a clear and distinct position productperformancethan serving smallerniches within the
within its category. Although much conceptual work has same category(Lambkinand Day 1989; Montgomery1982;
been writtenon defining and managingbrandimage (Keller Romanelli 1987). Thus brand image strategy should affect
1993; Park, Jaworski, and Maclnnis 1986), very little em- product performance,because images appealing to large
pirical research exists to guide managers' image manage- niches should capturea larger share of a productcategory
ment efforts. Two important areas must be addressed by thanbrandswhose images are targetedto smallerniches.
studies examining and prescribingbrand image strategies: Second, research should address internationalas well as
(1) the effect of brand image strategies on productperfor- domestic brand image management. Given that effective
mance and (2) the management of brand images across brand image management centers on satisfying customer
global markets. needs (Park,Jaworski,and Maclnnis 1986), and that cultur-
al, social, and economic dimensionsof internationalmarkets
*MartinS. Roth is Associate Professor of Marketing,CarrollSchool of shape consumer needs (Dannhaeuser 1987; Daun 1983;
Management,Boston College. The authorthanks Russ Belk, LarryFeick, Hondrich1983; Reilly andWallendorf1987; Tanand Farley
Tim Heath, Rajiv Grover, Bernie Jaworski, Sandra Milberg, and Gerald 1987; Tse, Wong, and Tan 1988), global aspects of image
Zaltmanfor their many helpful comments. Insights providedby BartWeitz
and three anonymousJMRreviewers are also gratefullyacknowledged. managementmust be understood.Although there are strate-
gic advantagesto targetingthe same customersand satisfy-

Journal of MarketingResearch
163 Vol. XXXII(May 1995), 163-175

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
164 JOURNAL OF MARKETINGRESEARCH, MAY1995

ing the same types of needs cross-nationally(Levitt 1983), Using the normativemodel proposed by Park,Jaworski,
in many cases, the niche or positioning opportunitiesdiffer and Maclnnis (1986), I investigatethree types of brandim-
across markets(Douglas andWind 1987; Jain 1989). For ex- ages-functional, social, and sensory images. These images
ample, Levi's blue jeans creates differentbrandimages for arebased on the fulfillmentof basic consumerneeds-prob-
customers in different parts of the world. A very social, lem solving and problem prevention (functional), group
group-orientedimage is used in the United States, whereas membership and affiliation (social), and novelty, variety
a much more individualist, sexual image is maintainedin seeking, and sensory gratification(sensory)-and are con-
Europe. Similarly, Reebok varies the image of its athletic sistent with other descriptions of relationships involving
shoes on the basis of nationaland regionaldifferencesit per- persons, environment, and sociocultural systems (Mali-
ceives in consumer tastes and preferences. In the United nowski 1944), Maslow's (1954) need hierarchy,frameworks
States, Reebok tries to balanceboth lifestyle and athleticim- of motivating forces driving consumer behavior (Rossiter
ages, whereas in WesternEurope, the brandimage focus is and Percy 1987), and a typology of consumer needs devel-
more narrowlyon athletics and performance.Some brands oped from a review of behavioralscience researchon needs
have successfully targeted the same types of customers and values (Hanna 1980). It is not uncommon for brands
using the same brandimage strategiescross nationally(e.g., within a category to be uniquely positioned along these
Coca Cola, Nike), but many firms have failed international- types of needs. For example in the United States toothpaste
ly by not adaptingtheir strategiesto cultural,economic, and category,Crest's functionalimage is based on cavity, tartar,
otherdifferencesamong markets(Hill and Still 1984; Ricks and decay prevention,UltraBrite is positionedtowardsocial
1983). Therefore,it is imperativethatmanagersidentify,as- needs to provide whiter teeth and social acceptability,and
sess, and respondto global marketconditionsthatcan affect Aim's image is based on taste and sensory gratification.Al-
the size of niches for particularbrandimage strategies. though the normative model suggests firms should select
My purposeis to provide managerswith a frameworkfor only one type of need when developingbrandimages (Park,
selecting brand image strategies for internationalmarkets. Jaworski,and Maclnnis 1986), researchhas shown that (1)
Specifically, I examine the linkage among brand image firmstend not to emphasizeonly one need, and (2) strategies
strategies, cultural and socioeconomic factors, and market emphasizing one need do not necessarily lead to better fi-
share in international markets. The international brand nancial performancethan those based on multiple needs
image strategies used by consumer goods firms in two in- (Roth 1992). The findings indicate that managersfeel it is
dustries were identified, indicating the type of needs the either necessary and/or advantageous to diversify their
brandsseek to satisfy and their positioning strategyin each brandimage strategies across two or more needs. The fre-
market.Because the relative performanceof a brandposi-
quency with which firms try to be "many things to many
tion dependson the size of the niche targeted,I also investi-
people" indicates a need to better understandthe linkages
gated cultural and socioeconomic factors hypothesized to between marketenvironmentsand consumerneeds. Identi-
affect the size of the targetedniche and, therefore,the mar-
ket potentialof the image strategy.Because marketershave fying environmentalfactorsthatfavor one type of consumer
need over anotherwill help managersselect the brandimage
long noted that countries are heterogeneousand comprised
of uniqueculturaland socioeconomic regionalmarkets(e.g., strategieswith the greatestmarketpotential.
the FlandersandWalloniaregions in Belgium, and the Que- Culturaland Socioeconomic Characteristicsof
bec province in Canada;cf. Douglas and Wind 1987; Hill InternationalMarkets
and Still 1984; Jain 1989), the study includedboth a nation-
al and a more micro regional focus. Social science and managementresearch offer insights
The article is organizedas follows. The next section pre- into the effects culturaland socioeconomic characteristicsof
sents a conceptualframeworkthat links brandimage strate- foreign marketsmay have on consumer needs and the sub-
gies, culturaland socioeconomic marketcharacteristics,and sequentsuccess of brandimage strategies.Figure 1 presents
product performance.Next, I discuss the research design, a conceptual model of the potential moderatingeffects of
which involved secondarydata and surveys given to product culture and socioeconomics on the performanceof brand
managers.I then presentresults, followed by discussions of image strategies. In internationalmarkets, environmental
the managerialand theoreticalimplicationsof the findings. characteristics are likely to moderate the image-perfor-
Finally, I presentdirectionsfor furtherresearch. mance relationship.As shown, two characteristics,(1) na-
tional cultureand (2) regional socioeconomic conditions af-
CONCEPTUALFRAMEWORK fect the performanceof functional,social, and sensorybrand
Much of internationalmarketingresearchhas focused on images. In addition, because market and firm conditions
marketingmix customizationversus standardization(for a may also impact performance,the model shows likely co-
recentreview, see Aulakhand Kotabe 1993). Yet, before one variates, that is, marketexperience, extent of competition,
makes marketingmix decisions, brandimage strategiesfor and marketingmix implementationproblems,that shouldbe
international markets must be developed. Brand image managedwhen examining the effects of environmentalfac-
strategies must be designed to appropriatelyposition the tors and brandimage on performance.
product for the targeted market segments. Although man- National Culture
agers have many image and positioning alternativesfrom
which to choose, internationalmarketersmust explore how A country's culture has long been identified as an envi-
culturaland socioeconomic characteristicsaffect the perfor- ronmentalcharacteristicthat influences consumerbehavior,
mance of brandimage strategies. andthe manyaspects of a cultureaffect differentlythe needs

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
GlobalBrandImageStrategies 165

Figure1
THEEFFECTSOF CULTURAL
ANDSOCIOECONOMIC
FACTORSON THEPERFORMANCE
OF BRANDIMAGESTRATEGIES

Product
Performance
I~ *Marketshare

Market & Firm Conditions

*Marketexperience
* Competitiveproblems
* Marketingmix problems

consumers satisfy throughthe acquisitionand use of goods needs. More formally,power distanceis hypothesizedto af-
and services. Hofstede's (1984) influential work on cross- fect brandimage performanceas follows:
culturalvalue systems identifies three aspects of culturethat
can be relatedto consumerneeds and brandimages-power Hi: The effects of social brandimages on marketshare will be
greater(lower) when culturalpower distance is high (low)
distance, uncertainty avoidance, and individualism. Al- than when power distance is low (high).
though Hofstede's researchwas conductedin organizational
settings, the values he identifies have been associated with Uncertainty avoidance captures the cultural pattern of
consumer behavior as well as with work-relatedactivities seeking stability, predictability,and low stress ratherthan
(for a recent example, see Lynn,Zinkhan,and Harris1993). change and new experiences (Hofstede 1984). People in
Power distance describes the extent to which a culture high uncertaintyavoidanceculturesare risk averse,resistant
fosters social inequality. Cultures high in power distance to change and variety seeking, and have a low tolerancefor
tend to emphasize the importanceof prestige and wealth in ambiguity. For brand image management, uncertainty
shaping boundariesor vertical relationshipsbetween social avoidance can be linked to aversion to a stimulus or event.
and economic classes such as rich and poor, and superiors When consumers experience some type of aversion, they
and subordinates (Hofstede 1984; Inkeles 1960; Lenski will be motivatedto remove the aversionor at least reduceit
1966). In high power distancecultures,people seek to main- as far as possible (Rossiter and Percy 1987). For example,
tain and increasetheir power as a source of satisfaction.So- when averse to cavities or other dental hygiene problems,
cial consciousness is high, and they are motivated by the consumers will seek a toothpaste capable of reducing the
need to conform with those in their class or in classes to risk of cavities. Functionalbrand images are positioned to
which they aspire. In low power distance cultures, people preventor solve problems and, therefore,reduce perceived
are much less focused on class differences and social aspi- risk. When culturesare high in uncertaintyavoidance,con-
rations.If one relates the aspects of culturalpower distance sumers will be very focused on risk aversion and problem
to consumerneeds, then it appearsthat social brandimages solving and prevention and, thus, will be good candidates
should be the best fit for high power distance cultures be- for functional brandimages. In contrast, when uncertainty
cause people are highly motivatedby social status and affil- avoidanceis low, risk aversionis diminished,andconsumers
iation norms. On the other hand, when power distance is aremore open to varietyand novelty behaviors.Thus, in low
low, social brandimages will not match the culturalnorms uncertaintyavoidancecultures,sensory brandimages focus-
because consumers are not motivated by group-related ing on variety,novelty, and sensory gratificationare effec-

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
166 JOURNAL OF MARKETINGRESEARCH, MAY1995

tive. More formally, uncertaintyavoidance is hypothesized similar consumers within trading blocs will become more
to affect brandimage performancein the following ways: importantthan nationaldifferences.
Although culturestend to transcendentire nations, social
H2a:The effects of functionalbrandimages on marketsharewill and economic conditions often vary substantiallywithin
be greater (lower) when cultural uncertaintyavoidance is
countries. Within most countries, regions exist that differ
high (low) than when uncertaintyavoidanceis low (high).
H2b:The effects of sensory brandimages on marketshare will markedlyfrom one anotherwith regardsto income, mobility,
be greater (lower) when culturaluncertaintyavoidance is media access, employment,and othersocioeconomiccharac-
low (high) than when uncertaintyavoidanceis high (low). teristics. Sociologists often assess a region's socioeconomic
level in terms of its "modernity"-a measureof the institu-
Individualismis an aspect of culturethat pertainsto peo- tions andorganizationsthatprovidelife experiencesfor mem-
ple's tendency to value personal and individual time, free- bers of a society (Inkeles 1983; Inkelesand Smith 1974).
dom, and experiences (Hofstede 1984; Parsons and Shils One aspect of modernitythat relates to consumptionand
1951; Riesman, Glazer,and Denney 1953). In contrast,cul- the viability of brandimage strategiesis disposableincome.
tures that emphasize collectivism exhibit patternsof group Disposable income indicates the amount of resources con-
or collective thinking and acting. In other words, cultures sumers allocate to goods and services. When resources are
high in individualismtend to seek varietyand hedonisticex- limited, consumers use goods and services to satisfy their
periences, whereas collectivist culturescorrelatemore with most basic, functionalneeds. As resourcesincrease, so does
conformity and group behavior.In addition,people in high consumerwillingness to spend money on productsthat sat-
individualismculturestend not to follow social norms, but, isfy more symbolic and sensory needs. For example, the
rather,form relationships,make decisions, and initiate be- major consumption goal of Zinderois Nigerians, a culture
haviors independent of others. In terms of brand image with very limited economic resources,is nutritionalself-suf-
strategies, cultures high in individualism seem well-suited ficiency, in contrastto the more symbolic and experiential
for sensory images that emphasize variety,novelty, and in- consumption patterns of more affluent North Americans
dividual gratification.On the other hand, cultures with low (Wallendorf and Amould 1988). Hence, when socioeco-
individualism will find social brand images that reinforce nomic conditions are low, functionalbrandimage strategies
will have the most appeal.When socioeconomic conditions
group membershipand affiliation more attractive.The fol-
increase, consumptionof social and sensory goods to satis-
lowing hypotheses describe how culturalindividualismwill
affect brandimage performance: fy symbolic and hedonic needs is more likely.
Anotheraspect of modernitythatinfluencesconsumerbe-
H3a:The effects of sensorybrandimages on marketsharewill be havioris exposureto consumption.The extentto which con-
greater (lower) when cultural individualismis high (low) sumersare exposed to Western,material-orientedconsump-
than when uncertaintyavoidanceis low (high). tion cultureswill influence their attractionto specific prod-
H3b:The effects of social brandimages on marketshare will be ucts. The demonstrationeffect, a phenomenon in which
greater (lower) when cultural individualismis low (high) poorerconsumersbuy symbolic and sensory products(with
than when individualismis high (low). little functional value) to identify with consumption soci-
eties to which they have been exposed (Keyfitz 1982;
Regional Socioeconomics Nurske 1953), explains why consumersin pooreror lesser-
Although country-levelfactors, such as national culture, developed countries purchase status and hedonic goods
affecting brand image strategy are important,micro geo- when functional ones would seem more appropriate(Belk
graphicunits, that is, regions within countries,meritinvesti- 1988; Keegan, Still, and Hill 1987). In short, what con-
sumerssee, hear,and access via television, other media, au-
gation as well. Although the predominantapproachin past
internationalbrandmanagementresearchhas been to exam- tomobiles, and other forms of mobility shapes their needs
and attitudestowardthe consumptionof productsand ser-
ine country-level differences (e.g., Boote 1983; Huszagh,
vices (Belk 1988; Jain 1989; Keegan, Still, and Hill 1987;
Fox, and Day 1986; Martenson1989), national analyses of
Keyfitz 1982; O'Guinn, Lee, and Faber 1986; Weimann
marketingstrategyeffectiveness may be too broadin scope, 1984). When marketsare characterizedby low levels of re-
causing importantintra- and inter-countrysimilarities and gional modernity,consumers will not be familiar with the
differences to be overlooked. Greaterhomogeneity will be materialaspects of consumerculturenor with the ability of
found within more localized regions than in countries.(See goods and services to satisfy social and sensoryneeds. Their
anthropologicalresearchon intra-versus inter-culturalvari- lack of mobility and exposure to media will prevent them
ation in Boster 1987; Roth and Moorman 1988.) Although from learningabout the more symbolic and experientialas-
proponents of the global standardizationapproach (e.g., pects of productconsumption.Instead,productswill be val-
Levitt 1983) have espoused the growing similarity of con- ued for their functionalcapabilities,thus making functional
sumerscross-nationally,domestic managers,in fact, contin- brandimage strategiesmore suited to the environmentthan
ually look for sophisticatedways to define marketsand de- social or sensory images. As regions become more modem,
velop micromarketingstrategies, such as the integrationof however,exposure to materialcultureincreases, motivating
census and marketresearchdata bases to identify geodemo- people to both desire the goods they see being consumed in
graphic,geopsychographic,and otherarea-basedtargetmar- other culturesand to purchaseproductsthat associate them
kets within the United States (McKenna 1992; Mehrotra with other cultures and societies. Hence, social brand im-
1990). Furthermore,as tradingnationsreduce and eliminate ages that promotegroup identificationwill have greaterap-
structural,political, and economic barriers,the search for peal in high modernityregionalmarkets.However,high lev-

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
Global Brand Image Strategies 167

els of modernityalso foster more individualand self-aware- strategicperformanceacrossall of the regionsin the database.
ness values (Inglehart 1977; Yankelovich 1981), which are Step 4-Manager identification. I contacted marketing
associated with needs for self-enhancement and personal managersat each company and asked them to identify the
fulfillmentand consistent with arousaland stimulation(sen- countriesin which they marketedtheirproducts.The person
sory) need-basedbrandimages. contacted often provided the name of anothermanagerre-
In most countries, the correlationbetween per capita in- sponsible for internationalmarketingfor a particularcoun-
come and modernity is quite high (e.g., Johansson and try. I asked managersof firms marketingtheir brandin one
Moinpour 1977; Sethi 1971). Measures of mobility and or more marketswithin the 10 countriesto participatein a
Westernexposure can provide an accuratedescriptionof a marketresearchstudy examining the performanceof brand
market'sdegree of modernity,capturingboth economic and image strategiesin internationalmarkets.
demonstrationeffects on consumption.The hypothesizedre- Step 5-Survey administration. I mailed surveys cus-
lationships between regional socioeconomics (modernity) tomized for each country,a cover letter,and $1 as a token of
and brandimage performanceare summarizedas follows: gratitudeto the managers.The survey contained items for
each region within the countriesserved by the firm. If a re-
H4a:The effects of functionalbrandimages on marketsharewill
be greater (lower) when regional socioeconomics is low sponse was not received after two to three weeks, I mailed
reminderletters and questionnaires.
(high) than when regional socioeconomics is high (low).
H4b:The effects of social brandimages on marketshare will be Thirtyeight managersfrom 11 firmsreturnedusableques-
greater(lower) when regional socioeconomics is high (low) tionnaires,yielding a response rate of 33% (115 managers
than when regional socioeconomics is low (high). were mailed surveys).Many of the participantsin the study
H4c:The effects of sensorybrandimages on marketsharewill be managed a brandin multiple foreign markets.Collectively,
greater(lower) when regional socioeconomics is high (low) they reportedon 209 cases of a particularbrand'sstrategy,
than when regional socioeconomics is low (high). market,and performancein a particularregionalmarket.On
average,each managerparticipatingin the studyreportedsix
METHOD cases. To assess the degreeof non-responsebias, a sampleof
10 of the 77 managers who did not returnquestionnaires
I identified and collected environmentaldata from coun-
were contacted by phone and found not to differ from re-
tries and the regions within them that are culturallyand so-
spondentswith regardto the brandimage, covariate,or per-
cioeconomically varied. Then I administered surveys to formancesurveymeasuresdescribedsubsequently.The most
firms marketingconsumer goods in these countries and re- common reasons for not respondingwere proprietarydata
gions to determineboth the brandimages used and the re- concerns,lack of time, and lack of regional-leveldata.
sulting product performance. Collectively, these two data
sources provided the means for testing the model and hy- Measures
potheses outlined previously. Brandimages. The surveyaskedmanagersto characterize
Data Collection theirbrand'simage in each particularregionalmarketby al-
locating 100 points across three types of strategies:func-
Step 1-Socioeconomic data base. I compiled a data base tional, social, and sensory,with more points being allocated
of socioeconomic characteristicsfor 60 cities and towns to the more emphasized images (see the Appendix). Man-
within 10 countries (Argentina, Belgium, China, France,
agers could allocate 100%of the points to one brandimage
Germany,Japan, Italy, Netherlands,Peru, and Yugoslavia), (a depth strategyapproach)or allocate the points across two
which were chosen because of their social, economic, and or three of the strategies(breadthapproach).
culturaldiversity.Socioeconomic datafor the regions within Culture and socioeconomics. Cultural measures of na-
each country(the numberof regionsper countryrangedfrom tional power distance, uncertaintyavoidance, and individu-
four to nine) were drawnfrom a varietyof publicly available alism were taken from indexes developed by Hofstede
statistical data sources. Because multiple socioeconomic (1984), on the basis of his survey of over 100,000 respon-
variableswere collected, factoranalysis was used to develop dents from 66 countries.Hofstede's sample is not represen-
a scale indicatingeach regionalmarket'sdegree of socioeco- tative of the entirepopulation,but of the middle classes em-
nomics (see the Appendixfor a descriptionof the analysis). ployed in multinationalcorporations.Because we are not in-
Step 2-Survey development. My survey measured all terestedin absolutescores, but in cross-nationalculturaldif-
items at the regional level and included questions about the ferences, the indexes are valid because the sampledrespon-
brandimage strategyused, marketexperience,problemsen- dents were well-matched demographically,with the only
countered, and market share for a brand in a particularre- systematicdifferencebeing their nationality.I measuredre-
gional market. The survey was developed, pretested, and gional socioeconomics in terms of each region's modernity,
modified for administrationto marketingand productman- which was estimated as a factor score determined using
agers responsible for marketing their products in one or scoring coefficients computed from the factor loadings
more of the 10 countriesin the data base. (refer to the Appendix). The scoring coefficients were esti-
Step3-Sample development.UnitedStatesfirmsmanufac- matedusing the Andersonand Rubin (1956) method,which
turingconsumergoods in the bluejean and athleticshoe cate- producesaccurateestimatesfor use as independentvariables
gories were then identifiedand chosen because of the diverse in regressionanalyses (Lastovickaand Thamodaran1991).
brandimage strategiesused to marketthem. In addition,both Product performance. Managers indicated the brand's
of these categorieshave many United States-basedmultina- market share during the most recent annual period for the
tional competitors,thus offering an opportunityto explore brandin each regionalmarket.Marketsharewas chosen be-

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
168 JOURNAL OF MARKETINGRESEARCH, MAY1995

cause of its wide use as an indicatorof performanceamong yond three standarddeviations of the mean and, therefore,
consumerproductmarketers(e.g., Smith andPark1992), as- were not included in subsequentanalysis. The sample used
sociation with a business's relative size in its served market for hypothesis testing was thus 193 cases. Table 1 reports
(e.g., Buzzell and Gale 1987), and consistentuse in otherin- Pearson correlationsand descriptive statistics for the vari-
ternationalmarketingstudies of business performance(e.g. ables used in this study. Social and sensory images had a
Ryans 1988; Szymanski, Bharadwaj, and Varadarajan
small, positive correlation,indicatinga weak patternof joint
1993). Althoughmarketsharediffers from typical measures
of communication effectiveness (e.g., brand awareness, emphasis in the same brand image strategy. In contrast,
brand attitude), the images reported are a function of the functional image had a strong, negative correlation with
firm's entire marketingprogram-not just their advertising both social and sensory images, which indicatedthat these
(Park,Jaworski,and Maclnnis 1986) and, therefore,better image pairs tended not to be emphasizedtogether.Also, no
accountsfor the entire marketingprogram'sperformance. firms reportedusing a single-need brandimage strategy-
Covariates.The survey included three covariates:market the maximumpoints allocatedto any one strategywas 75 of
experience,extent of competition,and marketingmix prob- 100-thus, all firms incorporatedtwo or more needs into
lems. Internationalmarketingexperience is often a key de- their brandimage strategy.
terminantof internationalmarketingsuccess (Denis and De-
The level of intercorrelationsamong the independent,
pelteau 1985; Gronhaugand Graham 1987; Johanson and
Vahlne 1977). In addition,the length of time the brandhas moderator,and covariate variables was, in general, low to
been availablein the marketoften explains unstablegrowth moderate,rangingfrom .001 to .54. There were two excep-
rates and performancemeasures.In some cases, new market tions-(1) cultural individualism and regional socioeco-
entrantsdemonstraterapidperformancegrowthbefore pro- nomics (r = .714), and (2) marketingmix problemsand cul-
gressing into the growth and maturitystages of the product turalindividualism(r = -.579). Correlationsat this level sig-
life cycle. On the otherhand, some firms exportingproducts nal the possible presence of multicollinearity,which inflates
to new marketsoften encounterslow growthratesuntil they the standarderrorsof parameterestimates.
adjusttheir marketingprogramsas more marketexperience To minimize and check for the presence of multi-
is gained. Managersreportedthe numberof years the brand
collinearity,a numberof steps were taken. First, I centered
had been availablein the regional market.
all of the independent, moderatorand covariate variables
Managersalso reportedthe extent to which they encoun-
tered competition problems in each regional market.This (i.e., put them in deviationscore form so thattheirmeans are
item was measuredusing a seven-pointLikertscale. Extentof zero).1As shown by Aiken and West (1991, p. 32), center-
competitionis an importantcovariatebecausedirectcompeti- ing variablessignificantlyreduces multicollinearity,thereby
tors try to servethe same niche, andfight for marketshare.In providingunbiasedparameterestimates. Second, I calculat-
addition,competitionaffects internationalmarketingperfor- ed varianceinflation factors (VIF) to examine the extent to
mance (Denis and Depelteau 1985; Jain 1989). Hence, brand which nonorthogonality among independent, moderator,
image performancecould be negativelyaffected when com- and covariatevariablesinflates standarderrors.The average
petitiveproblemsare encounteredin a regionalmarket. VIF was 1.94, ranging from 1.32 to 2.95-well below the
Effective implementation of the marketing mix brand
cutoff of 10 recommendedby Neter, Wasserman,and Kut-
conveys to consumersthe image intendedby the firm.Firms
ner (1989, p. 409). The centering procedure and the VIF
experiencingdifficulties managingtheir marketingmix will
have problems effectively communicating the intended findings suggest that multicollinearityis not a threatto the
brandimage to consumers, and, subsequently,their perfor- substantiveconclusions to be drawnfrom the analysis dis-
mance will be adverselyaffected.The prevalenceof market- cussed subsequently.
ing implementation problems can hurt a brand's perfor- The multipleregressionmodel tested containedthe effects
mance and is, therefore, another importantcovariate. The of each independentandmoderatorvariable(i.e., threediffer-
survey included four Likert scaled questions on problems ent brandimages; and one regionaland four culturalfactors,
encounteredwith the marketingmix and enabled managers respectively),threecovariatevariables,andthe independentX
to reportthe extent of product,distribution,pricing, and ad- moderatorinteractionsfor testingeach hypothesis.The inter-
vertising and promotion problems experienced in each re- action terms indicate the moderatingeffect of the environ-
gional market.A principalcomponentsanalysis showed the
four items to load on a single factor.All items had factor mentalfactorson the performanceof the brandimage strate-
loadings and communalities greater than .4. Cronbach's gies. To control for any correlatedresidualsdue to multiple
alpha, when combining the four variables into one scale, productcategoryand within-categorybrandobservations,10
was .727. The Appendix displays the covariatemeasures. dummy variableswere created(one dummy variablefor the
Using multiple regression, I examined the hypotheses two productcategories,five dummyvariablesfor the six blue
predictingthe effects of the environmentalfactors on brand jean brands,and four dummy variablesfor the five athletic
image strategymarketshare. Specifically, I used moderated shoe brands)and includedin the regressionmodel.
regression analysis (Arnold 1982; Sharma, Durand, and
Gur-Arie1981) to examine the moderatingeffects of culture
lThis process is recommended when conducting regression analyses
and socioeconomics on the market share of brand image
(e.g., Dillon and Goldstein 1984; Marquardt1980), especially when prod-
strategies.Prior to hypothesis testing, the data were exam- uct terms such as independentX moderatorinteractions are to be used
ined for outlying cases. Sixteen cases had marketsharesbe- (Aiken and West 1991; Cronbach1987).

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
 Table 1
AND DESCRIPTIVE
CORRELATIONS STATISTICS
FOR DEPENDENT,INDEPENDENT,
MODERATOR, VA
AND COVARIATE

DEPENDENT INDEPENDENT MODERATORS

Functional Social Sensory National National
Market Brand Brand Brand Cultural Cultural National Regional Regio
Share Image Image Image Power Uncertainty Cultural Socio- Mark
(%) (%) (%) (%) Distance Avoidance Individualism economics Experi
FunctionalBrandImage .124
Social BrandImage .165 -.745
Sensory BrandImage -.318 - .829 .245
Power Distance .254 -.092 .144 .013
UncertaintyAvoidance .192 .068 .006 -.103 .502
Individualism -.153 -.538 .360 .480 -.368 .323
Regional Socioeconomics -.067 -.329 .208 .304 -.226 -.217 .714
MarketExperience .448 .134 -.001 -.194 .046 -.025 .156 .043
CompetitiveProblems -.512 .153 -.098 -.140 -.262 -.136 -.099 .030 -.50
MarketingMix Problems -.207 .361 -.167 -.391 .232 .254 -.579 -.358 -.40

Mean 11.49 48.52 18.16 33.32 54.36 80.96 58.03 0.00 14.6
S.D. 6.36 24.24 13.98 16.67 12.47 12.15 18.77 .93 5.56
Range-High 31 75 60 65 76 94 80 1.33 20
Range-Low 2 0 0 0 35 29 16 -2.53 2.2
n= 193

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
170 JOURNAL OF MARKETINGRESEARCH, MAY1995

Table 2 image strategy.Marketshare was not directly relatedto the

EFFECTSON BRANDIMAGESTRATEGY
MODERATING threenationalculturemeasures(p > .10). Marketingproducts
STANDARDIZED
PERFORMANCE: REGRESSION when regionalsocioeconomicsis high, however,does tend to
COEFFICIENTS enhancemarketshare(P = .165, p < .10).
ModeratingEffects of Cultureand Socioeconomicson
MarketShare
Brand Image Performance
Variables (n = 193)
Main effects of independentvariables The sign and significance of the independentX modera-
FunctionalBrandImage .001 tor variableinteraction, coefficients are used to accept or
Social BrandImage .117 reject the hypotheses. I accomplished furtherdiagnosis of
Sensory BrandImage -.042
statistically significant interactions via post-hoc simple
Main effects of moderators slope tests (Aiken and West 1991). Following Cohen and
National CulturalPower Distance .117 Cohen's (1983) study,I recomputedmoderatorsexhibitinga
National CulturalUncertaintyAvoidance -.074 significantinteractionwith brandimage into high, medium,
National CulturalIndividualism -.023
.165*
and low values correspondingto (1) one standarddeviation
Regional Socioeconomics
above the mean, (2) at the mean, and (3) one standarddevi-
IndependentX moderatorinteractions ation below the mean, respectively.The high, medium, or
FunctionalImage X Power Distance -.074 low value representsthe "conditionalvalue"of the modera-
Social Image X Power Distance .198** tor (Darlington 1990). I then performedsimple regressions
Sensory Image X Power Distance .175**
of marketshareon brandimage, the conditionalvalue of the
FunctionalImage X UncertaintyAvoidance -.025 moderator,and the brandimage X conditionalmoderatorin-
Social Image X UncertaintyAvoidance .094
Sensory Image X UncertaintyAvoidance .053 teractionfor the high, medium, and low moderatorvalues. I
used T-teststo assess whetherthe simple slopes differ from
FunctionalImage X Individualism .169*
Social Image X Individualism -.339*** zero, which indicated if the regression of market share on
Sensory Image X Individualism .261*** brandimage was positive or negative at differentcondition-
al values of the moderator.
FunctionalImage X Regional Socioeconomics -.203**
Social Image X Regional Socioeconomics .305*** Culturalpower distance. H1 stated that social brandim-
Sensory Image X Regional Socioeconomics .286*** ages would performwell when power distancewas high. As
exhibited in Table 2, the power distance X social brand
Main effects of covariates
MarketExperience .042
image interactionterm was positive and statisticallysignifi-
CompetitiveProblems -.174** cant (P = .198, p < .05). Although no formal hypotheses
MarketingMix Problems -.365*** were made aboutthe moderatingeffect of powerdistanceon
the performance of functional or sensory brand images,
Dummyvariables these effects were also explored and are shown in Table 2.
ProductCategory .021
Brand 1 -.339*** Power distance had a significant,positive moderatingeffect
Brand2 .001 on market share for sensory brand images (,3 = .175, p <
Brand3 .470*** .05), but did not significantlyaffect functionalimages.
Brand4 -.104 The resultsof simple slope analyses shown in Table3 pro-
Brand5 .012
Brand6 .333*** vide more insight into the moderatingeffects of power dis-
Brand7 .234** tance on brandimage. The first line in Table3 illustratesthe
Brand8 .023 brandimage beta coefficient when marketshareis regressed
Brand9 -.019 on three variables:(1) brandimage, (2) the culturalpower
.502 distancemoderatorat one standarddeviationabove its mean
AdjustedR-square
F for full model (32 d.f.) 24.86*** (high power distance), and (3) the brandimage X high cul-
turalpower distance interaction.If the regressionis signifi-
***p < .01
**p < .05 cantly different from zero, evidence exists that the brand
*p< .10 image's effect on market share is indeed moderatedat the
high conditionalvalue of culturalpower distance.For social
RESULTS brand images, when power distance is high, the effect on
market share is very positive (,3 = .501, p < .01). When
Brand Image and Moderator Main Effects power distanceis medium,the effect on marketshareis pos-
itive, but much less so than when power distance is high (P
Table 2 shows the results of the moderator regression anal- = .134, p < .05). When power distance is low, the impact on
yses. Although the hypothesis did not address main effects, marketshare is negative (,3 = -.233, p < .05). Thus, social
all independent, moderator, covariate, and dummy variable brandimages enhance performancewhen power distance is
main effects were included in the model first to more pre- high, and to a lesser extent when it is medium, whereas so-
cisely estimate the moderating (interaction) effects. The main cial images hurtsharewhen power distance is low, support-
effects of the brand images on market share were very small ing hypothesis H1.
and not statistically significant (p > .10), which indicates that The resultsin Tables2 and 3 show thatpowerdistancehas
firms do not necessarily benefit from any one type of brand a similar moderating effect on the performance of sensory

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
Global BrandImage Strategies 171

Table 3
SLOPEANALYSIS INDEPENDENT
OF SIGNIFICANT X MODERATOR
INTERACTIONS:
STANDARDIZEDREGRESSIONCOEFFICIENTS
ANDT-VALUES

Functional Social Sensory

Brand Image BrandImage BrandImage
Variables(a) a3 t-value (3 t-value 13 t-value
CulturalPower Distance X Brand Image
Image at High Power Distance .501 4.64** .452 3.47**
Image At Medium Power Distance n.s. .134 2.00* -.256 -2.92*
Image at Low Power Distance -.233 -2.18* -.259 -3.04**

CulturalIndividualismX Brand Image

Image at High Individualism .374 4.71** -.402 -2.78** .486 4.39**
Image at Medium Individualism .519 4.10** .198 2.26* -.063 -.67
Image at Low Individualism -.136 -1.51 .300 3.57** -.424 -5.40**

Regional SocioeconomicsX Brand Image

Image at High Regional Socioeconomics -.098 -1.35 .538 4.50** .468 4.27**
Image at Medium Regional Socioeconomics .046 .25 .108 1.47 -.088 -.71
Image at Low Regional Socioeconomics .215 2.43* -.282 -3.07** -.278 -4.02**
**p < .01
*p < .05

brandimages. High levels of power distance (13= .452, p < Thus, H3areceived strong support.
.01) enhance marketshare,but medium (13= -.256, p < .05) For social images, marketshare was affected negatively
and low (P = -.259, p < .01) levels reduce share.In summa- when individualismwas high (3 = -.402, p < .01), enhanced
ry, the results show that cultural power distance is a rela- when individualismwas medium (3 = .198, p < .05), and
tively strong moderatorof social and sensory brand image greatly enhancedwhen individualismwas low (1 = .300, p
performance.When a market'spower distance is high, in- < .01). Hence, H3balso received strong support.
creased use of social and sensory brandimage strategiesen- Although I made no hypothesis about the moderatingef-
hances performance,whereas limited use of social and sen- fects of individualism on functional image performance,
sory images enhancesperformancewhen culturalpowerdis- Table3 shows thatfunctionalimages are enhancedwhen in-
tance is low. dividualismis at high (3 = .374, p < .01) and medium (13=
Cultural uncertaintyavoidance. H2aand H2b stated that .519, p < .01) levels, but not affected when individualismis
functionalbrandimages would performwell in high uncer- low (13= -.136, p > .05). In summary,when culturalindi-
tainty avoidance markets and that sensory brand images vidualism is high, sensory and functionalimages should be
would perform well when uncertaintyavoidance was low. emphasized, whereas social images should not. When indi-
Althoughnot addressedby the hypotheses,the effects of un- vidualism is low, social images will have the greatestposi-
certaintyavoidanceon the performanceof social brandim- tive impact on marketshare, whereas sensory images will
ages were examined as well. The uncertaintyavoidance X hurtperformance.
brandimage interactionsfound in Table 2 were not statisti- Regional socioeconomics. H4a_c stated that in low so-
cally significant(13= .025 for functionalimage, 13= .094 for cioeconomic regions, functionalbrandimages will perform
social image, 3= .053 for sensory image; all hadp-values > well, and in high socioeconomic regions, social and sensory
.10). H2aand H2bwere not supported. brandimages will performwell. Table 2 exhibits directional
Culturalindividualism.H3astatedthat sensory brandim- and statisticalsupportfor these hypotheses. The functional
ages would performwell in high individualismcultures,and brand image X regional socioeconomics interaction was
H3bstated that social brandimage strategieswould perform negativeand statisticallysignificant(13= -.203, p < .05), and
well in cultureslow in individualism.The results in Table 2 the social and sensory brand image X regional socioeco-
show strong, positive individualismX sensory brandimage nomics interactionswere positive and statistically signifi-
interactions(13= .261, p < .01), as well as a negative indi- cant (1 = .305, p < .01 for social image; 13= .286, p < .01
vidualism X social brandimage interaction(13= -.339, p < for sensory image). The slope analyses in Table 3 provide
.01). Although not included in the hypotheses, culturalindi- furtherinsights into the moderatingeffects of regional so-
vidualism had a positive and marginallysignificant moder- cioeconomics. For functional images, low regional socioe-
ating effect on the marketshare of functionalbrandimages conomics positivly impactsmarketshare(13= .215, p < .05).
= .169, p < .10). Slope analyses illustratedin Table3 pro-
(P13 At medium and high levels of socioeconomics, functional
vide specific insights into the moderatingeffects of individ- images have little impact on share (13= .046, p > .05 at
ualism. For sensory images, high individualismhas a posi- medium socioeconomics; P = -.09, p > .05 at high socioe-
tive effect on sensory image market share (13= .486, p < conomics). Thus, H4areceives partial supportbecause low
.01). Medium individualismhas no effect on sensory image levels of regionalsocioeconomics enhancesfunctionalbrand
share (1 = -.063, p > .05), and low individualismhas a neg- image performance,but does not affect performanceat high
ative effect on sensory image share (p = -.424, p < .01). levels. For social images, marketshareis enhancedwhen re-

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
172 JOURNAL OF MARKETINGRESEARCH, MAY1995

gional socioeconomicsis high (p = .538, p < .01), not affect- Two aspects of culturehad significantimpact on the perfor-
ed when socioeconomicsis medium(P = .108, p > .05), and mance of brandimage strategies-power distance and indi-
negativelyaffectedwhen socioeconomics is low (P3= -.282, vidualism. In low power distance cultures (e.g., Germany,
p < .01), thus providingstrong supportfor H4b.The results Netherlands,Argentina)in which people are not highly fo-
also stronlysupportH4c: For sensoryimages, marketshareis cused on social roles and group affiliation,functionalbrand
enhancedwhen regionalsocioeconomicsis high (p = .468, p images that de-emphasizethe social, symbolic, sensory,and
< .01), not affected when socioeconomics is medium (P = experientialbenefits of productsaremost appropriate.When
-.088, p > .05), and adverselyaffectedwhen socioeconomics the country'sdegree of power distance is high (e.g., China,
is low (P = -.278, p < .01). In summary,low regional so- France, Belgium), social and/or sensory needs should be
cioeconomics enhancesperformancewhen functionalbrand emphasized.
images are emphasizedas opposed to social and sensory im- In countrieswith high individualismcultures(e.g., Euro-
ages. High regional socioeconomics achieves better perfor- pean countries), brand images that emphasize functional,
mance using social and sensorybrandimage strategies. variety, novelty, and experiential needs are more effective
than social image strategies. On the other hand, cultures
Main Effectsof the Covariates with low individualism (e.g., Asian countries) are more
The main effects of marketexperience,competitiveprob- amenable to social brand image strategies that emphasize
lems, and marketingmix problemsare also found in Table2. group membershipand affiliation benefits than they are to
Marketexperience, operationalizedas length of time in the sensory brandimages.
market,did not affect marketshare(p > .10), however,com- Given the strongregionaleffects on modernity,one could
petitive and marketingmix problems did. For competition expect regional culture also to have a moderatingeffect on
problems, the effect on market share was negative (P3= brand image performance.Unfortunately,cultural data for
-.174) and significantatp < .05, indicatingthatthe presence regions within countriesare not widely availableoutside of
of competitiveproblemshampersmarketshare.For market- North America (e.g., Kahle 1986). Nonetheless, furtherre-
ing mix problems,the beta coefficient was also negative (P search examining subcultureeffects may enrich our under-
= -.365) and significant at p < .01, indicating that when standingof brandimage performance.
managersexperiencedifficultiesimplementingtheirmarket-
FURTHERRESEARCH
ing programs,productperformancesuffers.In summary,the
more (fewer) competitiveand implementationproblemsex- My study provides a startingpoint for helping managers
perienced,the lower (higher)marketsharetended to be. marketproductsinternationallyso as to best align brandim-
ages with customerneeds. As with all empiricalresearch,the
DISCUSSIONAND IMPLICATIONS study and results presentedhave limitations,which provide
My study addressedimportantlimitationsof past interna- interesting avenues for further research in three specific
tional marketing management research by linking brand areas: (1) extensions of the present model to other product
image strategies,the managementof which should precede categories, (2) the impact of brandimage customizationon
marketingmix customizationand standardizationdecisions, brandequity,and (3) alternativemeasuresof brandimage.
to productperformanceacross a variety of global markets. First, the findings can, to some extent, be generalizedto
The findings suggest that environmentalcharacteristicsof consumergoods positionedon a varietyof consumerneeds.
foreign markets are (1) importantindicators of customer This study found the moderatingeffects of culture and re-
segments and marketpotentialand (2) provide insights into gional socioeconomics on marketshare for the two product
the emphasis managers should place on functional, social, categories examined after accountingfor both category and
and sensory brandimage strategies. brand effects. These product categories, however, may be
Brand image success relied on several factors. First, re- somewhatunique because they are very "American"goods
gional socioeconomics was a strong moderatorof brand that often carrystrong culturalmessages. These goods gen-
image marketshare. Most internationalmarketingresearch eratestrongcountry-of-origineffects, becauseperceptionsof
focuses on cross-national differences, but the results pre- athletic shoe and blue jean brands encompass the United
sented here imply that managers should narrowtheir geo- States' images of modemrn, Westernfashion and athleticism.
graphicalfocus to cities and towns in additionto examining Researchhas shown, in fact, that brandpurchaseintentions
entire countries. Heterogeneity within countries makes it areelevatedwhen a country'spositiveimage is closely linked
difficult for managers to develop brand image strategies with benefitsdesiredfrom a productcategory(HanandTerp-
with strong within-countryappeal, yet, the micro-market- stra1988; Roth andRomeo 1992). Here,Because most of the
ing, regional focus allows managers to develop strategies brandswere marketedabroadusing the United States brand
targetedat more homogeneous target markets.As the find- name, countryimage effects have likely occured.Yet, within
ings illustrate, emphasis on functional brand image strate- two categoriesof countryimage-sensitivebrands,managers
gies enhances performancewhen regional socioeconomics implementedunique brandimage strategies,which resulted
is low. When regional socioeconomics is high, and con- in varying market shares depending on the environmental
sumershave wide exposureand easy access to Westerncon- conditions of their foreign markets.Nonetheless, the gener-
sumerculturethroughmedia and mobility,emphasis should alizability of these findings might be limited to brandsthat
be shifted to social and sensory brandimages to maximize can be tied to a strong, positive countryimage. Furtherre-
marketshare. search on categories not so positively related to country
Second, managers can use knowledge of a market'sna- image will be useful to explore the marketingenvironment's
tional cultureto develop successful brandimage strategies. impacton brandimage strategyeffectiveness.

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
Global Brand Image Strategies 173

Second, further research can examine how changing not availablefor cities and towns except in the few countries
brandimage and marketingprogramstrategiesaffects image (such as the United States)with highly developedcensus sys-
perceptionsand brandequity.An importantfeatureof glob- tems. I compiled a data base of 17 variablesfor each region
al marketingis the mobility of consumersand the frequency to assess each of the 60 regions'level of socioeconomics.Al-
of theirexposure(via communicationvehicles) to many cul- though additionalindicatorswould be desirable,most inter-
tures-especially the Western consumption culture (Belk national data bases possess limited numbers of accurately
1988; Quelch and Hoff 1986). Managersmust remainaware measuredvariablesfor the desiredcases. The numberof and
that although a product may require brand image cus- types of variablesused here appearsto capturethe same so-
tomization(for example, shifting an existing brand'srelative cioeconomic depth and breadthas similar classification ef-
emphasis from social to more functionalneeds when enter- forts on equal or greaternumbersof cases (e.g., Jaffe 1974).
ing a new market),culturaland/orsocioeconomic shifts may I used factor analysis to identify common factors ac-
requirefurtherbrandimage adjustments.For example, as a counted for by the socioeconomic variablesand eliminated
market's level of modernity increases, issues of cross-na- variables that did not account for variance across regions.
tional brandimage consistency must be evaluatedto create Factorscores for each region were generatedfor subsequent
and maintain a clear image for consumers exposed to the analyses of the effect of environmentalfactors on brand
brandfrom various internationalsources. Where previously image performance.I identifiedfactorswith the convention-
the incidence of cross-borderexposureto a brand'smulti-re- al criteriaof eigenvalues greaterthan one and adherenceto
gional marketingprograms may have been low, increased a scree test. All variablesremainingin the final factor solu-
modernitywill make such occurrencesmore prevalent.Fail- tion met the criteriaof loading on a factorat .3 or higher,and
ure to achieve brand image consistency may adversely, in having communalitiesof .4 or higher (Acito and Anderson
the long run, affect the brand's equity and the company's 1986). I eliminated variableswith low factor loadings, low
reputationby causing too much confusion aboutthe brand's communalities,and/orhigh cross-loadingsfrom the analysis
position in the marketplace(Levitt 1983), or by negatively and found a two factor solution. The ratio of numberof re-
affecting perceptionsof productquality and value that can gions analyzed to numberof factors extractedwas greater
affect marketshare (Smith and Park 1992). Furtherstudies than 20 (60 regions + 2 factors = 30), thus indicating ade-
are needed that relate environmental changes, consumer quate sample size and a stable factor solution (Arrindelland
perceptions,and the effects of customizingimage strategies. van der Ende 1985). The first factor accounted for 42% of
A third area for furtherresearch is the measurementof the data base variance,capturingthe market'smodernity-
brand image. Here, I investigated the emphasis placed on its mobility, communicationexposure, sector employment,
three image strategies (functional, social, and sensory). growthrate, and household size. The more autos and televi-
Managershave otherpositioningalternativesat theirdispos- sions per capita, the lower the birthrate,populationgrowth,
al (see Crawford1985) that also warrantinvestigation.Fur- and percent employed in agriculture,and the fewer people
thermore, this research used survey data from managers, per dwelling, the higherthe region's modernityor socioeco-
thus, the brand image measures represents managers' as- nomic level.
sessments of the image they intendedconsumersto perceive.
The meaningconsumersattachto the brandmay in fact dif- SurveyMeasures
fer from what managementintended.To addressthis possi- Brand image. How would you characterizeyour brand's
ble validity issue, consumer perceptions of brand images image in this market?Please allocate 100 points to each of
should be measured.One procedureused in other interna- the following types of images.Allocate the most pointsto the
tional marketingresearch (e.g., Mueller 1987; Tse, Belk, most emphasized image (up to 100 points), and the fewest
and Zhou 1989) would have consumersevaluateadvertising pointsto the least emphasizedimage (as little as zero points).
contentas a means for assessing perceivedbrandimage. An-
other approachthat takes into account not just advertising Mean(S.D.)
but the entire marketingprogramuses more exploratoryin- Functional
image(problemsolving,
terviewingtechniques.For example, one-on-one interviews, problemprevention) 48.52(24.24)
in which consumers elicit image descriptions of the brand Socialimage(conveysstatus,social
and their relationshipwith it, can be used to develop brand accreditation)
approval, 18.16(13.98)
image measures (Blackston 1992; Levy 1985). These ap-
proacheswould provide richerbrandmeanings and images, Sensoryimage(providesvariety,
though data collection for subsequentquantitativeanalysis stimulation,
sensorygratification) 33.32(16.67)
would be more expensive and time consuming than survey- 100%
based methods. These techniques, when applied cross-cul-
turallyamong diverse regions, may be useful for furthering Marketexperience. For how many years has your brand
our understandingof the marketingenvironment'smoderat- been availablein this market?
ing effects on image perceptionsand brandperformance. Mean= 14.63;S.D. = 5.56
APPENDIX Extentof competition.To what extent are you experienc-
ing competitiveproblemsin this market?
Regional Socioeconomics Measure (7-point Likert scale, where 1 = no problems, 7 = many
At the regional level, socioeconomic "summary"mea- problems.)
sures, such as disposable income or GNP/capita,are simply Mean= 3.60;S.D. = .97

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
174 JOURNAL OF MARKETINGRESEARCH, MAY1995

Marketing implementation problems. To what extent are Daun,Ake (1983), "TheMaterialisticLife-Style: Some Socio-Psy-
you experiencing marketing problems in each of the follow- chological Aspects,"in ConsumerBehavior and Environmental
ing areas? Quality,L. Uusitalo, ed. New York:St. Martin'sPress, 6-16.
(7-point Likert scale, where 1 = no problems, 7 = many Denis, Jean-Emileand Daniel Depelteau (1985), "MarketKnowl-
problems.) edge, Diversificationand Export Expansion,"Journal of Inter-
national Business Studies, 16 (Fall), 77-89.
Mean (S.D.) Dillon, William R. and Matthew Goldstein (1984), Multivariate
a) Distribution 4.07 (.99) Analysis: Methods and Applications. New York:John Wiley &
3.39 (.92) Sons, Inc.
b) Pricing
c) Advertisingand promotion 3.70 (1.11) DiMingo, Edward(1988), "The Fine Art of Positioning,"Journal
d) Productcharacteristics 3.23 (.79) of Business Strategy,9 (March/April),34-38.
Douglas, Susan P. andYoramWind (1987), "The Myth of Global-
Cronbach'salpha= .727. ization,"ColumbiaJournal of WorldBusiness, 22 (4), 19-29.
Gardner,BurleighB. and Sidney J. Levy (1955), "TheProductand
Scale mean = 3.65; S.D. = .71 the Brand,"HarvardBusiness Review, 33 (March/April),33-39.
REFERENCES Gronhaug,Kjell and John L. Graham(1987), "InternationalMar-
keting Research Revisited,"in Advances in InternationalMar-
Acito, FrankandRonaldD. Anderson(1986), "ASimulationStudy keting, Vol. 2, S. Tamer Cavusgil, ed. Greenwhich, CT: JAI
of Factor Score Indeterminancy,"Journal of Marketing Re- Press, Inc., 121-37.
search, 23 (May), 111-18. Han, C. Min andVernTerpstra(1988), "Country-of-OriginEffects
Aiken, Leona S. and StephenG. West (1991), MultipleRegression: for Uni-Nationaland Bi-NationalProducts,"Journal of Interna-
Testingand InterpretingInteractions.NewburyPark,CA: Sage tional Business Studies, 19 (Summer),235-55.
Publications. Hanna,JaniceG. (1980), "ATypology of ConsumerNeeds,"in Re-
Anderson,T. W. andH. Rubin(1956), "StatisticalInferencein Fac- search in Marketing,Vol. 3, Jagdish Sheth, ed. Greenwich,CT:
torAnalysis,"in Proceedingsof the ThirdBerkleySymposiumon JAI Press, 83-104.
MathematicalStatisticsand Probability.Berkeley:Universityof Hill, John S. and RichardR. Still (1984), "Effectsof Urbanization
CaliforniaPress, 111-50. on Multinational Product Planning: Markets in LDCs,"
Arnold, Hugh J. (1982), "ModeratorVariables:A Clarificationof ColumbiaJournal of WorldBusiness, 19 (Summer),62-67.
Conceptual,Analytic, and PsychometricIssues,"Organizational Hofstede, Geert (1984), Culture's Consequences. Beverly Hills,
Behavior and HumanPerformance,29 (April), 143-74. CA: Sage Publications.
Arrindell,Willem A. and Jan van der Ende (1985), "AnEmpirical Hondrich, Karl Otto (1983), "How Do Needs Change?"in Con-
Test of the Utility of the Observation-To-Variables
Ratio in Fac- sumer Behavior and EnvironmentalQuality, L. Uusitalo, ed.
tor and ComponentsAnalysis,"AppliedPsychological Measure- New York:St. Martin'sPress, 56-73.
ment, 9 (June), 165-78. Huszagh,SandraM., RichardJ. Fox, andEllen Day (1986), "Glob-
Aulakh, Preet S. and Masaaki Kotabe (1993), "AnAssessment of al Marketing:An EmpiricalInvestigation,"ColumbiaJournal of
Theoretical and Methodological Development in International WorldBusiness, 21 (4), 31-43.
Marketing:1980-1990," Journal of InternationalMarketing, 1 Inglehart,Ronald (1977), The Silent Revolution:Changing Values
(2), 5-28. and Political Styles Among Western Publics. Princeton, NJ:
Belk, Russell W. (1988), "ThirdWorldConsumerCulture,"in Mar- PrincetonUniversityPress.
keting's Contributionto Development,Research in Marketing, Inkeles, Alex (1960), "IndustrialMan: The Relation of Status to
Suppl. 4, E. Kumcu and A. F. Firat, eds. Greenwich, CT: JAI Experience, Perception,and Value,"American Journal of Soci-
Press, 103-27. ology, 75, 208-25.
Blackston, Max (1992), "BuildingBrandEquity by Managingthe (1983), Exploring Individual Modernity. New York:
Brand's Relationships,"Journal of Advertising Research, 32 ColumbiaUniversityPress.
(May/June),79-83. and David HortonSmith (1974), Becoming Modem. Cam-
Boote, Alfred S. (1983), "PsychographicSegmentationin Europe," bridge, MA: HarvardUniversityPress.
Journal of Advertising Research, 22 (December/January), Jaffe, Eugene D. (1974), Grouping:A Strategyfor International
19-25. Marketing.New York:AmericanManagementAssociation.
Boster,JamesS., ed. (1987), IntraculturalVariation,AmericanBe- Jain, SubhashC. (1989), "Standardizationof InternationalMarket-
havioral Scientist, 31 (November/December).Newbury Park, ing Strategy:Some Research Hypotheses,"Journal of Market-
CA: Sage Publications. ing, 53 (January),70-79.
Buzzell, RobertD. and BradleyT. Gale (1987), The PIMS Princi- Johanson,Jan and Jan-ErikVahlne (1977), "The Internationaliza-
ples: Linking Strategy to Performance. New York: The Free tion Process of the Firm-A Model of KnowledgeDevelopment
Press. and IncreasingForeignMarketCommitments,"Journalof Inter-
Cohen, Jacob and PatriciaCohen (1983), AppliedMultipleRegres- national Business Studies, 8 (Spring/Summer),23-32.
sion/CorrelationAnalysisfor the Behavioral Sciences, 2nd ed. Johansson,Johny K. and Reza Moinpour (1977), "Objectiveand
Hillsdale, NJ: LawrenceErlbaum. PerceivedSimilarityof Pacific Rim Countries,"ColumbiaJour-
Crawford,Merle (1985), "A New Positioning Typology,"Journal nal of WorldBusiness, 11 (Winter),65-76.
of ProductInnovationManagement,(December),243-53. Kahle, Lynn R. (1986), "The Nine Nations of North America and
Cronbach,Lee J. (1987), "StatisticalTests for ModeratorVariables: the Value Basis of GeographicSegmentation,"Journal of Mar-
Flaws in Analysis Recently Proposed,"Psychological Bulletin, keting, 50 (April), 37-47.
102 (3), 414-17. Keegan, WarrenJ., Richard R. Still, and John S. Hill (1987),
Dannhaeuser,Norbert(1987), "MarketingSystems and RuralDe- "Transferabilityand Adaptability of Products and Promotion
velopment:A Review of ConsumerGoods Distribution,"Human Themes in MultinationalMarketing-MNCs in LDCs,"Journal
Organization,46 (2), 177-85. of Global Marketing,1 (Fall/Winter),85-103.
Darlington, R. B. (1990), Regression and Linear Models. New Keller, Kevin Lane (1993), "Conceptualizing, Measuring, and
York:McGraw-Hill. Managing Customer-BasedBrand Equity,"Journal of Market-

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions
Global Brand Image Strategies 175

ing, 57 (January),1-22. (February/March),27-38.

Keyfitz, Nathan (1982), "Development and the Elimination of Ricks, David A. (1983), Big Business Blunders. Homewood, IL:
Poverty,"Economic Developmentand CulturalChange, 30 (3), Dow Jones-Irwin.
649-70. Riesman, Daniel, N. Glazer, and R. Denney (1953), The Lonely
Lambkin, Mary and George S. Day (1989), "EvolutionaryPro- Crowd: A Study of the Changing American Character. New
cesses in CompetitiveMarkets:Beyond the ProductLife Cycle," York:Doubleday.
Journal of Marketing,53 (July), 4-20. Romanelli, Elaine (1987), "New VentureStrategies in the Micro-
Lastovicka,JohnL. and KanchanaThamodaran(1991), "Common computerIndustry,"California ManagementReview, 30 (Fall),
FactorScore Estimatesin MultipleRegressionProblems,"Jour- 160-75.
nal of MarketingResearch, 28 (February),105-12. Rossiter,John R. and LarryPercy (1987), Advertisingand Promo-
Lenski, G.E. (1966), Power and Privilege: A Theory of Social tion Management.New York:McGraw-Hill.
Stratification.New York:McGraw-Hill. Roth, MartinS. (1992), "DepthVersusBreadthStrategiesfor Glob-
Levitt,Theodore(1983), "TheGlobalizationof Markets,"Harvard al BrandImageManagement,"Journalof Advertising,21 (June),
Business Review, 61 (3), 92-102. 25-36.
Levy, Sidney (1985), "Dreams,Fairy Tales, Animals, and Cars," and ChristineMoorman(1988), "The CulturalContent of
Psychology and Marketing,2 (Summer),67-81. Cognition and the Cognitive Content of Culture:Implications
Lynn,Michael, GeorgeM. Zinkhan,andJudyHarris(1993), "Con- for Consumer Research,"in Advances in ConsumerResearch,
sumer Tipping:A Cross-CountryStudy,"Journal of Consumer Vol. 15, M. J. Houston, ed. Provo, UT: Association for Con-
Research,20 (December),478-88. sumerResearch,403-10.
Malinowski, Bronislaw (1944), A Scientific Theory of Culture. and Jean B. Romeo (1992), "MatchingProductCategory
ChapelHill: Universityof North CarolinaPress. and Country Image Perceptions:A Frameworkfor Managing
Marquardt,D. W. (1980), "YouShould StandardizePredictorVari- Country-of-OriginEffects," Journal of International Business
ables in YourRegressionModels,"Journalof the AmericanSta-
Studies, 23 (Fall), 477-97.
tistical Association, 75, 87-91.
Ryans, Adrian B. (1988), "Strategic Market Entry Factors and
Martenson,Rita (1989), "InternationalAdvertising in Cross-Cul- Market Share Achievement in Japan,"Journal of International
tural Environments,"Journal of International ConsumerMar-
Business Studies, 19 (Fall), 389-409.
keting, 2 (1), 7-18. Sethi, S. Prakash(1971), "ComparativeClusterAnalysis for World
Maslow, Abraham H. (1954), Motivation and Personality. New Markets,"Journal of MarketingResearch, 8 (August), 348-54.
York:Harper-Collins.
Sharma,Subhash,RichardM. Durand,and Oded Gur-Arie(1981),
McKenna, Shawn (1992), The Complete Guide to Regional Mar-
"IdentificationandAnalysis of ModeratorVariables,"Journal of
keting. Homewood, IL: Business One Irwin.
Mehrotra,Sunil (1990), "StrategicRegional Marketing:Two Plus MarketingResearch, 18 (August), 291-300.
Two Equals Six," Journal of AdvertisingResearch, 30 (Decem- Simon, Carol J. and Mary W. Sullivan (1993), "The Measurement
and Determinants of Brand Equity: A Financial Approach,"
ber/January),9-17.
Montgomery,CynthiaA. (1982), "The Measurementof Firm Di- MarketingScience, 12 (Winter),28-52.
versification: Some New Empirical Evidence," Academy of Smith, Daniel C. and C. Whan Park(1992), "TheEffects of Brand
Extensions on MarketShare and AdvertisingEfficiency,"Jour-
ManagementJournal, 25, 299-308.
Mueller, Barbara(1987), "Reflections of Culture:An Analysis of nal of MarketingResearch,29 (August), 296-313.
JapaneseandAmericanAdvertisingAppeals,"Journalof Adver- Szymanski, David M., Sundar G. Bharadwaj, and P. Rajan
tising Research,27 (June/July),51-59. Varadarajan(1993), "StandardizationVersus Adaptationof In-
Neter, John, William Wasserman,and Michael H. Kutner(1989), ternational Marketing Strategy: An Empirical Investigation,"
Applied Linear Regression Models, 2nd ed. Homewood, IL: Journal of Marketing,57 (October), 1-17.
RichardD. Irwin, Inc. Tan, Chin Tiong and John U. Farley (1987), "The Impact of Cul-
Nurske, Ragnar(1953), Problems of Capital Formationin Unde- turalPatternson Cognition and Intentionin Singapore,"Journal
veloped Countries.New York:Oxford UniversityPress. of ConsumerResearch, 13 (4), 540-44.
Oglivy, David (1963), Confessions of an Advertising Man. New Tse, David K., JohnK. Wong, and Chin Tiong Tan(1988), "Toward
Tork:Ballantine. Some StandardizedCross CulturalConsumptionValues,"in Ad-
O'Guinn,Thomas,Wei-Na Lee, and Ronald J. Faber(1986), "Ac- vances in ConsumerResearch,Vol.15, M. J. Houston,ed. Provo,
culturation:The Impactof DivergentPathson Buyer Behavior," UT: Association for ConsumerResearch, 387-95.
in Advances in Consumer Research, Vol. 13, R. J. Lutz., ed. , Russell W. Belk, andNan Zhou (1989), "Becominga Con-
Provo UT: Association for ConsumerResearch,579-83. sumer Society: A Longitudinal and Cross-Cultural Content
Park, C. Whan, Bernard J. Jaworski, and Deborah J. Maclnnis Analysis of PrintAds from Hong Kong, the People's Republicof
(1986), "StrategicBrand Concept- Image Management,"Jour- China, and Taiwan," Journal of Consumer Research, 15
nal of Marketing,50 (October), 135-45. (March),457-72.
Parsons,Talcottand E. A. Shils (1951), Towarda General Theory Wallendorf,Melanie and Eric J. Arnould (1988), "'My Favorite
of Action. Cambridge,MA: HarvardUniversityPress. Things': A Cross-CulturalInquiryinto ObjectAttachment,Pos-
Quelch, JohnA. and EdwardJ. Hoff (1986), "CustomizingGlobal sessiveness, and Social Linkage," Journal of Consumer Re-
Marketing,"HarvardBusiness Review, 64 (May/June),59-68. search, 14 (March),531-47.
Reilly, Michael D. andMelanieWallendorf(1987), "AComparison Weimann,Gabriel(1984), "Imagesof Life in America:The Impact
of Group Differences in Food ConsumptionUsing Household of AmericanT.V.in Israel,"InternationalJournalof Intercultur-
Refuse," Journal of Consumer Research, 14 (September), al Relations, 8, 185-97.
289-94. Yankelovich,Daniel (1981), New Rules: Searchingfor Self-Fulfill-
Reynolds,ThomasJ. and JonathanGutman(1984), "Advertisingas ment in a World Turned Upside Down. New York: Random
Image Management," Journal of Advertising Research, 24 House.

This content downloaded from 194.214.27.178 on Tue, 27 Aug 2013 18:02:31 PM

All use subject to JSTOR Terms and Conditions