-
The relative importance ofbrand-packaging, price and tastein
affecting brand preferences
Jose Luis Mendez, Javier Oubina and Natalia RubioDepartment of
Finance and Marketing Research, Business Studies Faculty,
Autonoma University of Madrid, Madrid, Spain
Abstract
Purpose This paper aims to analyze the relative importance of
brand-packaging, price and taste inthe formation of brand
preference for manufacturer and store brands in food product
categories.
Design/methodology/approach The authors first perform a blind
taste test of the product usingthree brands (two manufacturer
brands and one store brand) in two categories with
differentiatedcharacteristics (cola drinks and olives stuffed with
anchovies). They then use conjoint analysis toanalyze the influence
of the intrinsic cue (taste) and the extrinsic cues (price and
brand-packaging) onconsumers preference for manufacturer and store
brands. Finally, after telling the consumers whichtaste belongs to
each brand, the authors study the influence of the extrinsic cues
on the consumersquality evaluations of the real stimuli.
Findings The results show that not knowing the brand to which
the taste tested belongs, leadsconsumers in general to order their
preferences fundamentally by taste. However, the results differ
byproduct category and consumer segment analyzed. Consumers who
evaluate the taste of store brandsas better change their
preferences more when they know which brand belongs to which taste.
Further,the change in preference when consumers know the
brand-taste correspondence is clearly greater inthe most
differentiated category.
Research limitations/implications The main limitations of this
research derive from the factorsconditioning the information. A
greater number of categories and attributes would enrich
theinformation. In addition, it would be useful to analyze more
than one store brand.
Practical implications The results obtained have interesting
implications for manufacturers andretailers concerning management
of the brands in their product portfolio and management of
theirrelationships in the distribution channel.
Originality/value The main contribution of this paper lies in
the work methodology used. Thepaper offers a comprehensive analysis
of how the relative importance of brand-packaging, price andtaste
affect brand preference for manufacturer and store brands. The
study also contributes evidenceon how the consumers knowledge of
the correspondence between brand and taste can change his orher
brand preferences, an issue of great interest for manufacturers and
distributors in managing theirproduct portfolios.
Keywords Store brand, Manufacturer brand, Preferences, Conjoint
analysis, Brands, Brand awareness,Drinks, Prices
Paper type Research paper
The current issue and full text archive of this journal is
available at
www.emeraldinsight.com/0007-070X.htm
The authors wish to acknowledge the financial support of
Ministry of Education and Science(research project ref.:
ECO2008-00488/ECON) and Madrid Regional Ministry of
Education(research project ref.: S2007/HUM-0413). The authors are
listed in alphabetical order because allcontributed equally to
developing this article.
Importance ofbrand-packaging,
price and taste
1229
Received October 2009Revised May 2010
Accepted May 2010
British Food JournalVol. 113 No. 10, 2011
pp. 1229-1251q Emerald Group Publishing Limited
0007-070XDOI 10.1108/00070701111177665
-
1. IntroductionThe main problem facing store brands is that they
are perceived by consumers asproducts of lower quality than
manufacturer brands and thus have lower purchaseloyalty (Corstjens
and Lal, 2000). The results of recent research demonstrate,
however,that there has been a progressive reduction of the quality
differential betweenmanufacturer and store brands (Apelbaum et al.,
2003; Davies and Brito, 2004 andMendez et al., 2008). As shown by
the merchandising policy applied to these brands(Fernandez and
Gomez, 1999; Hoch, 1996), retailers have been devoting
increasingeffort to marketing their own brands through the presence
of these brands in newcategories (Dunne and Narasimhan, 1999),
through active communication about thesebrands both inside the
establishment and in the communications media (Richardsonet al.,
1994), and through market segmentation of store brands (Dunne
andNarasimhan, 1999; Seok, 2001). An exemplary case is the
supermarket Carrefour inSpain, which incorporates store brands
Carrefour Seleccion, Carrefour Calidad yOrigen and Carrefour Eco
Bio as premium brands, the Carrefour brand as a valuebrand and the
brand 1 as an economy brand.
In this context, we attempt to determine:. consumers quality
evaluation of manufacturer and store brands based on the
intrinsic and extrinsic cues of both brands; and. the relative
importance of the intrinsic and extrinsic cues in consumers
preferences.
Our main contribution lies in the work methodology used: first,
we analyze theperceived quality of the brands through a blind taste
test of the product; then, we applyconjoint analysis to determine
the relative importance of different extrinsic andintrinsic cues on
the formation of the consumers brand preference; finally, we
examinethe brand quality perceived by the consumer for the real
stimuli. In this last stage, theconsumer judges the quality of each
brand by considering its true taste, price andbrand-packaging
jointly.
More specifically, the paper offers a comprehensive analysis of
the relativeimportance of how brand-packaging, price and taste
affect brand preference formanufacturer and store brands. It also
contributes evidence on how the consumersknowledge of the
correspondence between brand and taste can change his or her
brandpreferences, an issue of great interest for manufacturers and
distributors in managingtheir product portfolios.
This study presents the results obtained in an investigation
applied to two clearlydifferentiated product categories (cola
drinks and olives) for three kinds of brands(leading manufacturer
brand, second leading manufacturer brand and value storebrand). It
studies the consumers perception of one intrinsic cue (taste) and
twoextrinsic cues (brand-packaging and price). Interesting
implications for managementemerge from the study performed.
2. Literature review: quality perceived by consumers and
formation ofpreferencesConsumers use, ordering and evaluation of
product attributes provide an initialframework for investigating
which factors determine the formation of qualityperceived by
consumers in brand choice.
BFJ113,10
1230
-
The first theoretical studies proposed different classifications
of these attributes tomeasure differences in quality between
brands. Cox (1967) argues that products can beconceived as arrays
of cues and that consumers assign information values to the
cuesavailable based on their predictive value PV (the degree to
which consumersassociate this attribute with the products quality)
and confidence value CV (theconfidence that consumers have in their
own capacity to use and interpret this attributeeffectively as an
indicator of quality).
Later, Olson (1972) and Olson and Jacoby (1973) classified
product attributes asintrinsic and extrinsic to products. Intrinsic
cues represent product-related attributesthat cannot be manipulated
without altering the physical product itself. They
includeingredients, taste, freshness, texture, aroma and
nutritional value. Conversely,extrinsic cues are product-related
attributes that are not part of the physical productitself and that
may include price, brand name, advertising, labelling and
packaging.
There is evidence that extrinsic cues are more easily
recognized, integrated andinterpreted than intrinsic cues, which
are harder to process (Purwar, 1982). Thus, onecan assume that
consumers assign greater CV to extrinsic cues than to intrinsic
cues. Itis reasonable to assume, however, that in determining the
real quality of packagedgoods consumers attribute more importance
to intrinsic cues than to extrinsic ones. Itis thus realistic to
argue that the PV assigned to intrinsic cues is greater than that
forextrinsic cues (Richardson et al., 1994).
It is accepted that consumers have limited cognitive
capabilities and seek efficientmeans of processing and storing
information. Brand names have traditionally beenshown to represent
many attributes, and there is clear evidence that
consumersrecognize a brand name as an informational chunk (Jacoby
et al., 1977). Brand namesrecall store information from memory, and
this information interacts with the intrinsiccues to produce
different results (Bettman and Park, 1980).
Researchers have attempted to clarify the relative importance of
extrinsic andintrinsic cues in perceived product quality for
consumers. Olson (1972) argued thatintrinsic cues are more
important than extrinsic cues, mostly for utilitarian productssuch
as nylon hose, carpeting and envelopes. Similar analysis of food
products hasproduced contradictory results. Whereas Chung et al.
(2006) find that intrinsicattributes are more important, Richardson
et al. (1994) conclude the opposite. Holbrook(1986) and De
Chernatony and Knox (1990) observe that extrinsic attributes
mainlybrand name and packaging may be more important than intrinsic
cues for productsfor which image is important. This argument is
supported by studies ofimage-reflective products such as carbonated
bottled water, beer and colas(Christopher et al., 1987 and
Steenkamp, 1990).
Zeithaml (1988) and Steenkamp (1990) incorporate these intrinsic
and extrinsicattributes into their definitions of perceived product
quality. Zeithaml definesperceived quality as the result of an
overall process of evaluation of a product(high-level abstraction)
that integrates the information provided by a set of
objectiveattributes of the product and whose importance, as
informative inputs in the evaluationprocess, is given by a set of
factors that are situational (situation prior to purchase oract of
consumption) and personal (e.g. motivation or experience). These
attributes arenot perceived in the same way by all consumers.
Likewise, Steenkamp (1990)understands perceived product quality as
an idiosyncratic value judgment withrespect to the fitness for
consumption, which is based upon the conscious and/or
Importance ofbrand-packaging,
price and taste
1231
-
unconscious processing of quality cues in relation to relevant
quality attributes withinthe context of significant personal and
situational variables.
Quality can also be understood in terms of fitness for
consumption or satisfaction ofneeds (Steenkamp, 1990; Kotler et
al., 2006). Kotler et al. (2006) define quality as the setof
aspects and characteristics of a product and service that maintain
a relationshipwith their capacity to satisfy consumers expressed or
latent needs. The AmericanSociety for Quality Control integrates
the elements of objective and subjective qualityinto its
definition, using Kotlers concept of the term.
Several studies have focused on store brands, performing
comparative analysis oftheir quality with manufacturer brands
through the evaluation of their intrinsic andextrinsic
attributes.
On the one hand, Bellizzi et al. (1981) uses a Likert scale to
measure consumersperceptions of different intrinsic and extrinsic
cues (specifically, design, externalappearance, and overall quality
measured). They find significant differences in theevaluation of
extrinsic cues in favor of manufacturers brands. Cunningham et
al.(1982) observe that consumers evaluate both the intrinsic and
the extrinsic cues ofmanufacturer brands more positively. However,
Richardson et al. (1994) examine theeffect of intrinsic and
extrinsic cues on consumer-perceived quality for five
productcategories and find that the most valued choices have the
extrinsic cues ofmanufacturer brands, independently of their
objective quality.
On the other hand, the objective quality of manufacturer and
store brands based onevaluations by qualified technical experts has
been studied by Apelbaum et al. (2003),Davies and Brito (2004), and
Mendez et al. (2008). Davies and Brito (2004) obtain dataon
objective quality for a single product category, liquid dishwashing
detergent (basedon evaluations performed by the British Consumers
Association). They find that in thiscategory the consumer did not
pay extra for image only, as the leading brand had achemical patent
that made the product perform better. Mendez et al. (2008)
analyzeobjective quality in 90 categories of food and beverage and
28 categories of homecleaning and personal hygiene products (based
on the evaluations performed by OCUCompra Maestra). They find that
only eight (8.9 percent) categories of food andbeverage and six
(21.4 percent) categories of home cleaning and personal
hygieneproducts showed significant differences in quality between
store and manufacturerbrands, in favor of the latter.
The empirical studies developed from prior theoretical
contributions to analyze thequality of manufacturer and store
brands have primarily used variance analysis withexperimental
designs to obtain information. Nevertheless, if we consider quality
asthe set of aspects and characteristics of a product and service
that maintain arelationship with their capacity to satisfy
consumers expressed or latent needs(Kotler et al., 2006), variance
analysis does not provide information on the intensitywith which
specific attributes act, or on the role of specific levels of
attributes in theformation of consumer preferences.
In this study, we propose and apply conjoint analysis as a
technique forunderstanding:
. the importance of different attributes; and
. the preference for different levels of attributes, both of
which are key issues forimproving satisfaction of consumers real or
latent needs.
BFJ113,10
1232
-
Conjoint analysis studies the effects of the joint influence of
two or more attributes(independent variables) on consumers
preferences (dependent variable) by measuringquantitatively the
relative importance of some attributes over others. To classify a
setof alternatives in order of preference, the consumer must
perceive these alternativesindividually, not as a whole but as a
set of partial attributes or characteristics. We canthus interpret
consumers final ranking as the sum of qualities (or perceived
utilities)associated with the different properties or levels of the
attributes that characterize thebrand (Hair et al., 1998).
The application of the technique of conjoint analysis to food
products is relativelyrecent. Steenkamp (1987) was one of the first
European researchers to apply thistechnique to evaluate the quality
of ham as perceived by Dutch consumers. Steenkampused the
attributes of brand name, packaging, store and price. In the 1990s,
we findvarious studies along these lines, which examine consumers
preferences for meatproducts (Sendim, 1995), wine (Gil and Sanchez,
1997) and eggs (Ness and Gerhardy,1994), among other products.
In the current decade, interesting studies also incorporate the
technique of conjointanalysis to evaluate consumer preferences in
different food products. To examinePortuguese consumers preferences
in cheese, De Souza Monteiro and Ventura Lucas(2001) consider the
attributes of price, texture (creamy, partially aged, or fully
aged)and size of packaging. Skytte and Blunch (2001) study
consumers preferences for fishand cheese products and use the
attributes of quality, consistency, traceability, price,promotion,
and reputation, among others. For the category of yogurt, Ferjani
et al.(2009) measure brand equity using three attributes- brand
name, price and flavour. Forflavour, they do not use a taste test
but instead establish a priori flavours of vanilla,strawberry and
banana. The results of their study show the importance of brand
valuein the formation of consumers preferences for this
product.
None of the researchers mentioned incorporates a blind taste
test of the product todetermine the consumers perception of the
intrinsic attributes and to establish the effectof the experience
of consumption on consumers evaluation for different brands.
Theonly study we have found that uses conjoint analysis and blind
taste test is that ofDavies and Brito (2004). These researchers
employ conjoint analysis to examine theconsumers preferences for
manufacturer and store brands, analyzing the attributes ofbrand,
taste and price. Their study performs blind taste tests on
consumers using fourcategories of food products. Although most of
the consumers in the sample claimed thatthey usually bought the
manufacturer brand in the product categories analyzed(margarine,
breakfast cereal, crackers and mayonnaise), only a small proportion
of theseconsumers actually chose the manufacturer brand in the
blind taste test of the product.Further, very few of those who
chose the manufacturer brand continued to do so oncethey knew its
price of sale to the public. These results show that brand image is
often theonly explanation for the higher prices consumers pay to
acquire manufacturer brands.
Our research follows Davies and Brito (2004) in using conjoint
analysis and inincorporating attributes of price, taste and brand.
However, their study differs fromours, among other issues, in that
ours:
. considers brand as linked to packaging (binomial
brand-packaging); and
. uses presentation of full profile stimuli instead of trade-off
presentation.
We describe the research methodology in the following
section.
Importance ofbrand-packaging,
price and taste
1233
-
3. Methodology3.1 Techniques of analysis appliedThis study uses
the technique of conjoint analysis to examine consumers
preferencesfor two product categories (cola drinks and olives
stuffed with anchovies). Tounderstand the existence of market
segments with homogeneous preferences, weapplied a TwoStep cluster
analysis to the estimations of utility obtained in the
conjointanalysis for each individual.
3.2 Phases of analysisObtaining information to understand
consumer preferences is developed in threephases, first for the
category of cola drinks and second for the category of olives
stuffedwith anchovies.
In the first phase, the consumer is asked to evaluate the
perceived quality of thetaste of three brands through a blind taste
test. Quality was measured on a scale with arange of seven points,
where the value 1 corresponded to very poor quality and 7 tovery
good quality. The taste of each brand in the category was presented
to theconsumer as Taste 1, Taste 2 or Taste 3. The environmental
conditions of light,temperature, etc. were the same for all
consumers.
After the consumers had evaluated the taste of each brand in
both categories, weperformed the second phase, which consisted of
the application of conjoint analysis.Table I shows the attributes
and levels considered in the conjoint analysis. Theattributes were
taste as an intrinsic cue and brand-packaging and price as
extrinsiccues. The levels for the attribute taste were presented to
the consumer as Taste 1, Taste2 and Taste 3. These correspond to
the taste tested in Phase 1. For brand-packaging,we considered two
manufacturer brands the leading brand and the second leadingbrand
and one store brand (belonging to Carrefour). The consumer was
shownbrand-packaging for the product, including the name of the
brand, symbols andcolouring, but not information about the
composition of the product. Finally, the valuesof the prices were
average actual (non-promotional prices) prices for the brands
used,applied by Carrefour stores in Spains capital city,
Madrid[1].
We chose presentation of full profile stimuli[2] instead of
trade-off presentation,because consumers usually evaluate jointly
combinations of more than two attributesto form their preferences.
The full profile presentation avoids the loss of realism
Cola drinks Stuffed olives
TasteTaste 1 Gold El SerpisTaste 2 Pepsi Cola CarrefourTaste 3
Coca Cola La Espanola
PriceLeading manufacturer brand (Coca Cola/La Espanola) 0.37e
0.73eSecond Leading manufacturer brand (Pepsi Cola/El Serpis) 0.32e
0.92eStore brand (Gold/Carrefour) 0.16e 0.52e
Brand-packaging Coca Cola La EspanolaPepsi Cola El SerpisGold
(Carrefour) Carrefour
Table I.Attributes and levelsconsidered in the
conjointanalysis
BFJ113,10
1234
-
involved in showing only two attributes per card and thus gives
the results greaterpredictive validity.
We also applied an orthogonal procedure (Orthoplan available on
SPSS, version17.0) to the initial (3 3 3 27) hypothetical products
to reduce them to nineproduct concepts. We used a complementary set
of stimuli (four holdout cards) toevaluate the models
reliability[3].
The consumers were asked to order the cards according to their
preferences. Theconjoint analysis was performed using the following
models to relate the data onpreferences to the factors: for the
attributes taste and brand-packaging, we used thediscrete model
because it reflects the definition of these attributes (discrete).
For theattribute price, we used the linear model, which obtained a
better fit.
After consumers had ordered the cards in the conjoint analysis,
we began the thirdphase. In this phase, we told consumers which
taste belonged to each brand. We thenasked them to evaluate the
overall quality of the three cards that contained the
REALinformation on each brand tested, using a scale with a range of
seven points (1 verypoor quality 7 very good quality). Each card
presented one brand-packaging, itstaste and its price.
The software package used in the data treatment was SPSS
17.0.
3.3 Qualitative research: selection of categories and commercial
chainTo select the product categories and commercial chain for
analysis, we performed 20in-depth interviews of consumers who did
not subsequently participate in thequantitative research.
We chose the categories of cola drinks and anchovies stuffed
with olives because theyrepresent categories that are familiar to
consumers and they are characterized by socialconsumption. These
categories have different brand detectability in social
consumptionand different brand differentiation among the brands
that make up the category.
The category of cola drinks involves social consumption and a
manufacturer brandwith strong leadership (we confirmed that,
according to Alimarket[4], in Spain in 2008the leading manufacturer
brand, Coca Cola, had a market share of 84.6 percent; thesecond
leading manufacturer brand, Pepsi Cola, a share of 8.1 percent; and
all storebrands taken together a share of 6.3 percent). In
contrast, different manufacturerbrands of olives stuffed with
anchovies have moderate leadership (we confirmed thatin 2008 the
leading manufacturer brand, La Espanola, had a market participation
of21.9 percent, the second leading manufacturer brand, El Serpis, a
participation of 4.1percent and all store brands taken together a
participation of 67.3 percent).Consumption of olives has also
social dimension, but, unlike cola drinks, the consumerdoes not see
the brand.
We chose Carrefour and its store brand, because both the chain
and the store brandhave great brand recognition for the consumers
interviewed. This commercial chainhas a substantial presence in the
city of Madrid, with five hypermarkets, eightsupermarkets and ten
convenience stores.
3.4 Quantitative research: sample and field workThe criterion
used for the sample unit selection in the quantitative research was
thatthe person participating in the study habitually did his or her
shopping in Carrefour.Shoppers were recruited at all stores of the
chain in the city of Madrid, that is, in the
Importance ofbrand-packaging,
price and taste
1235
-
five Carrefour hypermarkets, the eight Carrefour Express
supermarkets and the tenconvenience stores, Carrefour city.
The shoppers were chosen through a stratified process according
to sex (70 percentfemales, 30 percent males) and occupation (70
percent working, 30 percent notworking). We first chose 350
shoppers from the 23 Carrefour establishments operatingin the city
of Madrid. The interviewers were graduate-level marketing students
trainedin how to approach shoppers with the characteristics of the
study. The interviewersexplained the broad outlines of the research
to the shoppers (that it consisted of tastingand evaluating brands
in some of the categories that they marked on a list). Theshoppers
chosen had to mark the categories they purchased and consumed
regularlyon a list of 15 product categories. They were asked for
their telephone number so thatwe could contact them to participate
in the study. After one week, the marketinggraduate students
contacted the consumers who had marked the categories of coladrinks
and olives stuffed with anchovies (220). Of the shoppers contacted,
130participated in the study. However, only 93 shoppers contributed
valid information forboth product categories, a slightly high
sample error for each category,^10.37 percentwith a confidence
level of 95.5 percent. For the sample size achieved, the
statisticalpower is 90 percent, higher than the minimum recommended
level of 80 percent, forsignificance levels of 0.05 and moderate
effects size (Hair et al., 1998).
To obtain the information (from the blind taste test, the
conjoint analysis and theevaluation of the stimuli with the real
information), we performed the study in anair-conditioned room in
which the environmental conditions of light, temperature, etc.were
the same for all consumers. We began by asking for information
about thecategory of cola drinks. The fieldwork for this category
was performed over 13 days,during which we obtained information
from a maximum of ten shoppers per day. Forthis category, we
obtained valid information from 100 shoppers, who weresubsequently
contacted to perform the study in the category of stuffed olives.
Thefieldwork for this category was performed in the same way as it
was for cola drinks,although we obtained valid information from
only 93 shoppers for this category. Thus,93 shoppers contributed
valid information for both product categories, cola drinks
andstuffed olives, a total of 186 preferences. 70 percent of the
respondents were female, and75 percent were employed. The average
education level was secondary school and themedian age category
35-49. The median monthly household income was between2000-3000
Euros and the average number of members of the household was
three.
4. Results4.1 Results of conjoint analysisResults for the whole
sample. Tables II and III show the results of the conjoint
analysisfor the sample of consumers in the categories of cola
drinks and stuffed olives.
The values for measures of goodness of fit for the sample
(Pearsons R and KendallTau parameters) show the high precision of
the model, not only for the original stimulibut also for the
validation stimuli. They have a value of 1 or nearly 1 and
arestatistically significant with a confidence level of 95 percent.
This result implies thatthe estimated ordering of the stimuli
presented corresponds faithfully to the realstructure of
preferences in the two product categories composing the sample
analyzed.
The stimulus most valued in both categories analyzed is taste,
before price and thebinomial brand-packaging. However, price is
more important than brand-packaging
BFJ113,10
1236
-
for stuffed olives (28.46 v. 21.39), whereas brand-packaging is
more important thanprice for cola drinks (37.33 v. 20.39).
Interestingly, in both categories only one of the brands showed
positive utilities fortaste: Coca Cola for cola drinks and El
Serpis for stuffed olives. In contrast,brand-packaging has positive
utilities in both categories of the manufacturer brands.The only
negative utilities are for store brands, in taste and
brand-packaging. Storebrands are only evaluated more favorably than
manufacturer brands in price.
With cola drinks, the preferred taste and brand-packaging are
those of the leadingbrand, Coca Cola. For stuffed olives, however,
consumers preferred the taste of thesecond leading brand, El
Serpis, while they preferred the brand-packaging of theleading
brand, La Espanola.
The difference between the manufacturer and store brands in the
range of variationof the utilities assigned to taste and brand
packaging is also interesting for each of thecategories
analyzed.
In the case of taste, the maximum range of variation (difference
between the highestand the lowest value of the brands analyzed) is
1.34 for cola drinks and 0.64 for stuffedolives; that is, based on
the attributes of their experience in consumption,
consumersevaluate the manufacturer brand better in differentiated
categories than in genericcategories.
Cola drinks (Total): 93 consumersCoca Cola Pepsi Gold
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 37.33 1.21 0.07 0.26 0.07 21.47 0.07Taste 42.28
0.79 0.07 20.24 0.07 20.55 0.07Price 20.39 0.69 0.69 0.07 1.39 0.11
2.08 0.11Constant 3.62 0.12
Notes: Pearsons R 0.99 * * *; Kendalls Tau 1.00 * * *; Kendalls
Tau 4 holdout 1.00 * *; se standarderror; ( *, * *, * * * level of
statistical significance of 0.10, 0.05 and 0.01 respectively); aIn
performing theconjoint analysis with the program SPSS 17.0, we used
the subcommand Subject. This subcommandaverages the values for
importance of the individual subjects. These averages do not
generally agreewith those calculated in the summary of utilities
obtained
Table II.Results of conjoint
analysis for the samplea
Olives stuffed with anchovies (Total): 93 consumersLa Espanola
El Serpis Carrefour
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 21.39 0.47 0.09 0.08 0.09 20.55 0.09Taste 50.15
20.13 0.09 0.39 0.09 20.25 0.09Price 28.46 1.13 2.26 0.16 1.13 0.08
3.39 0.23Constant 2.74 0.17
Notes: Pearsons R 0.99 * * *; Kendalls Tau 0.87 * * *; Kendalls
Tau 4 holdout 1.00 * *; se standarderror; *, * *, * * * level of
statistical significance of 0.10, 0.05 and 0.01 respectively; aIn
performing theconjoint analysis with the program SPSS 17.0, we used
the subcommand Subject. This subcommandaverages the values for
importance of the individual subjects. These averages do not
generally agreewith those calculated in the summary of utilities
obtained
Table III.Results of conjoint
analysis for the samplea
Importance ofbrand-packaging,
price and taste
1237
-
In the case of brand-packaging, the difference is even greater.
The maximum rangeof variation is 2.68 for cola drinks and 1.02 for
stuffed olives. As we might expect, thevariation occurs between the
leading brand and the store brand and takes a muchgreater value in
the differentiated than in the generic category.
Results for the segments. We analyzed the utilities obtained for
each individual inboth product categories to identify possible
market segments according to theirpreferences by applying TwoStep
cluster analysis. The results obtained with TwoStepcluster analysis
were validated with those obtained using Wards method.
Bothprocedures provide nearly identical results. We then applied
conjoint analysis to eachsegment. Tables IV-IX show the results of
the conjoint analysis performed on themarket segments obtained. The
values for measures of goodness of fit for the segments
Segment 1: Consumers guided by intrinsic cues who prefer the
store brand (31 consumers)Coca Cola Pepsi Gold
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 21.36 0.27 0.09 0.20 0.09 20.47 0.09Taste 54.94
0.28 0.09 21.09 0.09 0.81 0.09Price 23.70 0.88 0.88 0.08 1.76 0.154
2.65 0.23Constant 3.27 0.17
Notes: Pearsons R 0.99 * * *; Kendalls Tau 1.000 * * *; Kendalls
Tau 4 holdout 1.00 * *; se standarderror; ( *, * *, * * * level of
statistical significance 0.10, 0.05 and 0.01 respectively)
Table IV.Results of conjointanalysis for the segmentsin cola
drinks
Segment 3: Consumers guided by intrinsic cues who prefer the
leading brand (25 consumers)Coca Cola Pepsi Gold
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 14.33 0.43 0.13 0.01 0.13 20.52 0.13Taste 67.40
2.28 0.13 0.65 0.13 22.93 0.13Price 18.27 0.74 0.74 0.11 1.48 0.23
2.22 0.34Constant 3.51 0.25
Notes: Pearsons R 0.99 * * *; Kendalls Tau 0.93 * * *; Kendalls
Tau 4 holdout 0.91 * *; se standarderror; ( *, * *, * * * level of
statistical significance 0.10, 0.05 and 0.01 respectively)
Table VI.Results of conjointanalysis for the segmentsin cola
drinks
Segment 2: Brand consumers (37 consumers)Coca Cola Pepsi
Gold
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 66.25 2.52 0.08 0.42 0.08 22.95 0.08Taste 14.70
0.22 0.08 20.13 0.08 20.10 0.08Price 19.05 0.51 0.51 0.07 1.01 0.14
1.51 0.22Constant 3.99 0.16
Notes: Pearsons R 0.99 * * *; Kendalls Tau 0.94 * * *; Kendalls
Tau 4 holdout 1.00 * *; se standarderror; ( *, * *, * * * level of
statistical significance 0.10, 0.05 and 0.01 respectively)
Table V.Results of conjointanalysis for the segmentsin cola
drinks
BFJ113,10
1238
-
(Pearsons R and Kendall Tau parameters) are high and significant
for the originalstimuli and for the validation stimuli.
In the category of cola drinks, we distinguish three
segments[5]. Segment 1 is definedas consumers guided by intrinsic
cues who prefer the store brand. These consumersestablish
preferences based primarily on taste. The store brand and Coca Cola
are thetwo brands with positive utilities in this respect, although
the value for the store brand isconsiderably greater. Segment 3 is
composed of consumers guided by intrinsic cueswho prefer the
leading brand and who also determine their preferences based on
taste.However, the alternatives with positive utilities are the
manufacturer brands, and theleading brand Coca Cola has a
considerably higher value. Segment 2 is composed ofbrand consumers.
These consumers focus their preference on brand-packaging andchoose
the leading brand Coca Cola or, in its absence, the second leading
brand Pepsi.
Segment 1: Consumers guided by intrinsic cues and
brand-packaging (43 consumers)La Espanola El Serpis Carrefour
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 33.79 0.77 0.11 0.13 0.11 20.90 0.11Taste 44.94
0.33 0.11 21.19 0.11 0.86 0.11Price 21.27 0.76 1.52 0.18 0.76 0.10
2.28 0.27Constant 3.48 0.20
Notes: Pearsons R 0.99 * * *; Kendalls Tau 0.94 * * *; Kendalls
Tau 4 holdout 0.87 * *; se standarderror; ( *, * *, * * * level of
statistical significance 0.10, 0.05 and 0.01 respectively)
Table VII.Results of conjoint
analysis for the segmentsin stuffed olives
Segment 3: Price conscious consumers (15 consumers)La Espanola
El Serpis Carrefour
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 14.19 0.41 0.18 20.14 0.18 20.26 0.18Taste 19.32
20.12 0.18 0.19 0.18 20.07 0.18Price 66.49 2.82 5.64 0.31 2.82 0.16
8.46 0.47Constant 20.64 0.34
Notes: Pearsons R 0.99 * * *; Kendalls Tau 1.00 * * *; Kendalls
Tau 4 holdout 0.87 * *; se standarderror; ( *, * *, * * * level of
statistical significance 0.10, 0.05 and 0.01 respectively)
Table IX.Results of conjoint
analysis for the segmentsin stuffed olives
Segment 2: Consumers guided by intrinsic cues (35 consumers)La
Espanola El Serpis Carrefour
Attributes Importance (%) Coefficient se Utility se Utility se
Utility se
Brand-packaging 9.03 0.13 0.10 0.11 0.10 20.23 0.10Taste 68.89
20.71 0.10 2.41 0.10 21.70 0.10Price 22.08 0.91 1.82 0.18 0.91 0.09
2.73 0.27Constant 3.18 0.20
Notes: Pearsons R 0.98 * * *; Kendalls Tau 0.94 * * *; Kendalls
Tau 4 holdout 0.87 * *; se standarderror; ( *, * *, * * * level of
statistical significance 0.10, 0.05 and 0.01 respectively)
Table VIII.Results of conjoint
analysis for the segmentsin stuffed olives
Importance ofbrand-packaging,
price and taste
1239
-
For the category of stuffed olives, Segment 1 is composed of
consumers guided byintrinsic cues and brand-packaging. These
consumers grant greater importance totaste (44 percent). For them,
the store brand has the highest positive utility, followed bythe
leading brand. These consumers also grant relatively high
importance tobrand-packaging (34 percent). The leading brand is the
most valued brand, and storebrands show negative utility. Segment 2
is defined as consumers guided by intrinsiccues. These consumers
show markedly higher preference for taste and value thesecond
leading brand, El Serpis, highest. Segment 3 is composed of price
consciousconsumers. These consumers prefer store brands.
4.2 Results compared: taste test, utility of stimuli in the
conjoint analysis and evaluationof the stimuli with the real
informationTo observe how different intrinsic and extrinsic cues
affect the formation ofpreferences, we analyzed three different
aspects of consumers evaluations, accordingto the phases of the
research explained in the methodology section. The first aspect
isthe evaluation of taste quality in each of the brands of cola
drinks and stuffed olivesthrough a blind taste test of the product.
The second is the estimated utility value of thereal stimuli of
brand-packaging, price and taste, based on the results of the
conjointanalysis. The third and final aspect is the consumers
quality evaluation of the realstimuli of brand-packaging, price and
taste, once the consumer is informed of thebrand-taste
correspondence.
Results for the whole sample. Table X shows the results obtained
for each productcategory.
Cola drinks Mean SD Stuffed olives Mean SD
Blind taste test of the product (values from 1 verypoor quality
to 7 very good quality)
Blind taste test of the product (values from 1 verypoor quality
to 7 very good quality)
Taste of Coca Cola 5.01 1.62 Taste of La Espanola 4.62 1.69Taste
of Pepsi 3.86 1.50 Taste of El Serpis 4.82 1.70Taste of Gold 3.65
1.54 Taste of Carrefour 4.43 1.56
Preferences estimated for the real stimuli(utility values)
Preferences estimated for the real stimuli(utility values)
Taste, brand-packaging and pricefor Coca Cola (SIMULATION) 6.32
2.17
Taste, brand-packaging and pricefor La Espanola (SIMULATION) 5.3
2.09
Taste, brand-packaging and pricefor Pepsi (SIMULATION) 5.03
1.54
Taste, brand-packaging and pricefor El Serpis (SIMULATION) 4.34
2.58
Taste, brand-packaging and pricefor Gold (SIMULATION) 3.68
2.17
Taste, brand-packaging and pricefor Carrefour (SIMULACION) 5.33
2.32
Overall quality for the real stimuli (values from1 very poor
quality to 7 very good quality)
Overall quality for the real stimuli (values from1 very poor
quality to 7 very good quality)
Taste, brand-packaging and pricefor Coca Cola 5.43 1.48
Taste, brand-packaging and pricefor La Espanola 4.79 1.49
Taste, brand-packaging and pricefor Pepsi 4.67 1.23
Taste, brand-packaging and pricefor El Serpis 4.79 1.57
Taste, brand-packaging and pricefor Gold 3.41 1.56
Taste, brand-packaging and pricefor Carrefour 4.24 1.51
Note: SD standard deviation
Table X.Comparative analysis forsample: taste test, utilityof
the stimuli in theconjoint analysis, andevaluation of the
stimuliwith the real information
BFJ113,10
1240
-
Regarding blind taste test of the product, the store brand
receives the lowest evaluationin both categories (3.65 for cola
drinks and 4.43 for stuffed olives). Although consumersevaluate the
leading brand of cola drinks highest, they prefer the second
leading brandof stuffed olives. The range of evaluation of taste
quality is much greater for coladrinks (a difference of 1.36) than
for stuffed olives (a difference of 0.39). For cola drinks,the
taste quality evaluation for Coca Cola is clearly higher than that
of the other twochoices, which receive similar values. In contrast,
all three brands of stuffed olivesreceive similar taste quality
evaluations.
When we compare the results of the estimated utility of the real
stimuli and thequality evaluation of the real stimuli, we find a
clear influence of extrinsic cues onconsumers preferences.
For cola drinks, the second leading brand (Pepsi) and the store
brand (Gold) clearlydiffer in the estimated utility and in the
quality evaluation (even though they wereevaluated as similar in
the blind taste test of the product). We see a clear preference
forCoca Cola, followed by Pepsi and finally by Gold. Gold is the
only brand assigned avalue lower than 4 (mean value of the
scale).
For stuffed olives, the results of conjoint analysis reveal
similar utility for theleading brand and the store brand. However,
consumers relegate the store brand to lastplace once they know
which taste belongs to which brand.
Results for the segments. We performed a variance analysis to
evaluate the existenceof statistically significant differences
between the segments in the three aspectsdiscussed (blind taste
test, estimated utility of the real stimuli and quality evaluation
ofthe real stimuli). Tables XI and XII shows the results in the
segments identified.
The segments differ significantly according to the evaluations
obtained for the blindtaste test of the product, the estimated
utility of the real stimuli, and the qualityevaluation of the real
stimuli.
For cola drinks, we can see statistically significant
differences in the blind taste testof the product for the leading
brand, Coca Cola, and the store brand, Gold. Segment 1,consumers
guided by intrinsic cues who prefer store brand, assigns the
highest valuefor quality to the taste of the store brand. This is
the only segment that values the tasteof the store brand as above
average. Segment 3, consumers guided by intrinsic cueswho prefer
the leading brand, assigns a value considerably higher than that of
othersegments to the taste of Coca Cola.
When consumers order all of the stimuli presented after the
blind taste test,Segments 1 and 3 give priority to the products
taste over price and brand-packaging.Segment 1, consumers guided by
intrinsic cues who prefer the store brand, assignsthe store brand,
Gold, comparatively higher preferences. Although these
consumersprefer the taste of Gold only slightly over that of Coca
Cola in the blind taste test, therelative importance granted to
price (23.7 percent) implies that they prefer the storebrand
considerably more than Coca Cola in the conjoint analysis. Coca
Cola showscomparatively higher preferences for Segment 3, consumers
guided by intrinsic cueswho prefer the leading brand, and for
Segment 2, brand consumers.
As to the value obtained for the quality evaluation of the real
stimuli, Coca Cola isprimarily valued by Segments 2 brand consumers
and 3 consumers guided byintrinsic cues who prefer the leading
brand. For both segments, the results obtainedagree with the
preferences obtained in the conjoint analysis. In contrast, for
Segment 1,consumers guided by intrinsic cues who prefer the store
brand, the quality evaluation
Importance ofbrand-packaging,
price and taste
1241
-
Seg
men
t1
Seg
men
t2
Seg
men
t3
INT
RIN
SIC
ST
OR
EB
RA
ND
PR
EF
ER
EN
CE
BR
AN
D
INT
RIN
SIC
LE
AD
ING
BR
AN
DP
RE
FE
RE
NC
EC
ola
dri
nk
sM
ean
SD
Mea
nS
DM
ean
SD
F-S
ned
ecor
Blin
dta
ste
test
ofth
epr
oduct
(valu
esfr
om1
very
poor
qualit
yto
7ve
rygo
odqualit
yT
aste
ofC
oca
Col
a4.
451.
654.
951.
765.
800.
965.
31*
Tas
teof
Pep
si3.
711.
353.
761.
644.
201.
470.
88T
aste
ofG
old
4.45
1.31
3.76
1.54
2.48
1.05
15.0
9*
Pre
fere
nce
ses
tim
ate
dfo
rth
ere
alst
imuli
(uti
lity
valu
es)
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rC
oca
Col
a(S
IMU
LA
TIO
N)
4.70
2.70
7.23
1.05
6.96
1.50
17.5
6*
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rP
epsi
(SIM
UL
AT
ION
)4.
132.
055.
300.
715.
731.
1710
.15
*
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rG
old
(SIM
UL
AT
ION
)6.
251.
412.
471.
052.
271.
0011
2.46
*
Ove
rall
qualit
yfo
rth
ere
alst
imuli
(valu
esfr
om1
very
poor
qualit
yto
7ve
rygo
odqualit
y)T
aste
,b
ran
d-p
ack
agin
gan
dp
rice
for
Coc
aC
ola
4.65
1.47
5.68
1.49
6.04
1.02
8.10
*
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rP
epsi
4.55
1.09
4.51
1.37
5.00
1.14
1.61
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rG
old
3.84
1.49
3.70
1.65
2.44
1.08
7.59
*
Notes:
* Lev
elof
stat
isti
cal
sig
nifi
can
ceof
0.01
;S
Dst
and
ard
dev
iati
on
Table XI.Comparative analysis forsegments in cola drinks:taste
test, utility of thestimuli in the conjointanalysis, and
evaluationof the stimuli with thereal information
BFJ113,10
1242
-
Seg
men
t1
Seg
men
t2
Seg
men
t3
INT
RIN
SIC
BR
AN
DIN
TR
INS
ICP
RIC
ES
tuff
edol
ives
Mea
nS
DM
ean
SD
Mea
nS
DF
-Sn
edec
or
Blin
dta
ste
test
ofth
epr
oduct
(valu
esfr
om1
very
poor
qualit
yto
7ve
rygo
odqualit
y)T
aste
ofL
aE
span
ola
4.74
1.71
4.26
1.67
5.14
1.61
1.62
Tas
teof
El
Ser
pis
3.88
1.76
5.91
0.85
4.93
1.44
19.4
9*
**
Tas
teof
Car
refo
ur
4.60
1.71
3.89
1.41
5.29
0.91
4.86
**
*
Pre
fere
nce
ses
tim
ate
dfo
rth
ere
alst
imuli
(uti
lity
valu
es)
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rL
aE
span
ola
(SIM
UL
AT
ION
)6.
102.
184.
421.
945.
291.
057.
09*
**
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rE
lS
erp
is(S
IMU
LA
TIO
N)
3.18
2.36
6.60
1.35
2.23
0.87
43.6
2*
**
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rC
arre
fou
r(S
IMU
LA
CIO
N)
5.72
2.38
3.98
1.73
7.49
1.04
17.1
4*
**
Ove
rall
qualit
yfo
rth
ere
alst
imuli
(valu
esfr
om1
very
poor
qualit
yto
7ve
rygo
odqualit
y)T
aste
,b
ran
d-p
ack
agin
gan
dp
rice
for
La
Esp
anol
a4.
421.
555.
061.
535.
290.
912.
76*
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rE
lS
erp
is4.
401.
565.
711.
203.
711.
2013
.80
**
*
Tas
te,
bra
nd
-pac
kag
ing
and
pri
cefo
rC
arre
fou
r4.
421.
483.
771.
464.
861.
513.
29*
*
Note:
Lev
elof
stat
isti
cal
sig
nifi
can
ceof
* 0.1
0;*
* 0.0
5;*
** 0
.01;
SD
stan
dar
dd
evia
tion
Table XII.Comparative analysis for
segments in stuffedolives: taste test, utility ofthe stimuli in
the conjoint
analysis, and evaluationof the stimuli with the
real information
Importance ofbrand-packaging,
price and taste
1243
-
for the real stimuli differs from the taste evaluation and from
the preferences estimated inthe conjoint analysis. In the conjoint
analysis, this segment chooses the store brand forits quality of
taste as well as its price. Yet when consumers in this segment know
thebrands of the tastes tested, they assign the store brand the
lowest quality evaluation.Specifically, Gold obtains a value
slightly lower than 4 (mean value of the scale), whilethe
manufacturer brands Coca Cola and Pepsi obtain similar values above
the meanvalue of the scale. This result shows the importance of the
detectability of brand inquality evaluations for product categories
with highly recognized brands.
For stuffed olives, the results of the blind taste test of the
product show significantdifferences in the second leading brand, El
Serpis, and the store brand, Carrefour.Consumers in Segment 2,
consumers guided by intrinsic cues prefer the taste of ElSerpis
significantly more, finding it to be quite good. This segment
assigns aconsiderably higher value to the taste of El Serpis than
to the other two brands.Segment 3, price conscious consumers, finds
the taste of Carrefour significantlybetter, in fact quite good, but
the value that this group assigns to the taste of Carrefouris only
slightly higher than that assigned to the manufacturer brands.
In the conjoint analysis, Segment 2, consumers guided by
intrinsic cues clearlychooses El Serpis due to the importance these
consumers give to taste. Segment 3, priceconscious consumers,
chooses Carrefour for price. Segment 1, consumers guided
byintrinsic cues and brand packaging who evaluate the taste of the
leading brand, LaEspanola, and the store brand, Carrefour, as very
similar, showing similar estimatedutilities for both brands. We
must remember that for this segment the importance oftaste is 45
percent, but that of brand-packaging is also relatively high, at 34
percent.
When the consumers learn the brands of the tastes tested, the
quality evaluation ofthe real stimuli varies in some cases,
specifically in Segments 1 and 3.
In Segment 1, consumers guided by intrinsic cues and
brand-packaging,brand-packaging clearly influences quality
evaluations. The results of the conjointanalysis show a definite
preference for Carrefour and La Espanola, a result that agreeswith
the point values given to the tastes of these brands and the
importance granted to theattribute of taste. Once the brands of the
preferred tastes are known, however, consumersdo not clearly prefer
any of the three stimuli but assign all three nearly identical
values.
In Segment 3, price conscious consumers, price-brand interaction
clearlyinfluences quality evaluations. Consumers perceive a similar
taste for the threebrands and initially form their preferences
based on price. However, when they knowthat the leading brand is
not the highest priced, they give this brand the highest
qualityevaluation, followed by the store brand.
5. Conclusions and implications for marketing strategyThis paper
proposes an approach to determining consumer preference
formationaccording to consumers choice of manufacturer and store
brands. To establish whataspects of a product influence preference,
we perform a conjoint analysis with a fullprofile presentation of
stimuli.
We have seen in the categories analyzed that attributes of
experience play a majorrole in consumer preference structure after
consumption.
In general, once a product has been consumed, taste is the
strongest attributedetermining the choice of the preferred
stimulus, but price and brand-packaging alsoaffect the consumer
preference structure greatly. The results obtained in this
research
BFJ113,10
1244
-
show that the influence of price and brand-packaging are
specific to the productcategory analyzed, based on its degree of
differentiation. The influence ofbrand-packaging occurs in the most
differentiated category, whereas the effect ofprice is greater in
the generic category.
It is also important to point out that the product with the best
evaluation forbrand-packaging does not always receive the highest
taste evaluation. The agreementbetween taste and brand-packaging
occurs in the most differentiated category.Further, we see high
negative utility in taste and brand-packaging for store brands
inboth categories analyzed. Only in price do store brands achieve
their competitiveadvantage.
On the cola market, the leading brand maintains its leadership
in both taste andbrand-packaging. The consumer preferences
expressed for taste are considerablyaffected by brand-packaging.
Consumers resolve similar preference for taste in twobrands, one a
manufacturer brand and the other a store brand, in favor of
thewell-known manufacturer brand. Consumers make this choice
because they have ahigh degree of knowledge of this category due to
the social character of itsconsumption and the existence of two
very well-known manufacturer brands thatcontinually update their
packaging design.
The results obtained in this category show noticeable
differences for the consumersegments identified. The consumer
segments who prefer the leading brand, whetherfor its taste
(Segment 3) or for its brand-packaging (Segment 2), value the
overallquality of manufacturer brands consistently in their
estimated preferences. However,those who prefer the store brands
taste (Segment 1) change their preference structureentirely when
they know that the preferred taste corresponds to the store brand.
Thissegment, which represents 33.3 percent of the sample, relegates
store brands to lastplace in the overall quality evaluation.
The fact that a highly differentiated category has a consumer
segment that prefersthe store brand taste when it does not know the
extrinsic cues is interesting for storebrand management in these
categories. Store brands should make a significantmarketing effort
in highly differentiated categories to strengthen their image
andgenerate brand equity and hence consumer confidence. Only so can
they takeadvantage of their products having similar quality in its
intrinsic attributes and thuscompete with manufacturer brands in
specific segments.
Some store brands on the Spanish market, such as that of the
retailer Mercadona,generate a high degree of consumer satisfaction.
Consumers trust the quality ofMercadonas store brands and inform
other consumers of their advantages overmanufacturer brands. This
fact favors store brand loyalty, which encourages ratherthan
discourages establishment loyalty.
For stuffed olives, the preferences are less pronounced. Price
affects consumerpreference to a greater extent than
brand-packaging, and store brands are betteraccepted. This category
has a low level of differentiation and lower consumption, andone
does not usually detect the brand in social consumption.
In this category, the consumer segment (Segment 2) that shows
clear preferences forthe second manufacturer brand, El Serpis,
based on taste assigns values of overallquality and estimated
preferences in the conjoint analysis that are consistent with
thepreferences declared on the blind taste test. Brand-packaging
and price mainly affectthe consumer segments that assign higher
values to the store brand on the blind taste
Importance ofbrand-packaging,
price and taste
1245
-
test of the product (Segments 1 and 3). However, when we compare
this category tocola drinks, we find that knowing the brand-taste
correspondence does not affect thequality evaluation of the store
brand as negatively.
Consumers acceptance of store brands is usually wider in
slightly differentiatedcategories than in highly differentiated
categories. In the former, extrinsic cues alsoaffect the quality
evaluation, but consumers trust their own perceptions of
intrinsicquality more once they have tasted the product. Retail
firms should take advantage ofthis fact and encourage tasting of
the store brand in these categories. Further,promoting an image of
value could also favor the growth of store brands in
lessdifferentiated categories.
Although we have only analyzed two product categories with three
brands in each,and although we must be careful when drawing
conclusions, this study has shown:
. the clear importance of taste in the choice of mass
consumption products;
. the different influence, according to the product category, of
brand-packagingand price in establishing preferences between
manufacturer and store brandswhose taste the consumer perceives as
similar;
. the distinct importance of price and differentiation as
strategic tools for favoringstore brands according to the product
category; and
. the existence of consumer segments that justify the strategy
of segmented storebrands currently being developed by
distributors.
Among the limitations of this paper, we must consider sample
size, the number ofcategories and the attributes evaluated. The
three basic factors that made it difficult toobtain a larger sample
size were:
(1) the fact that the same consumers had to provide information
on the two productcategories examined;
(2) the time it took to obtain information due to the taste
test; and
(3) the fact that the consumers had to be regular shoppers at
Carrefour and toconsume products from both of the categories under
study regularly.
This last condition was established in order to control for
distortion in the result basedon the consumers different
experiences of the category or the store. Future researchshould
achieve a larger sample size. A higher number of categories and
attributeswould also enrich the information, although it would make
the work of those surveyedmore difficult, as they would have a
greater number of stimuli to rank.
We used only one store brand, Carrefour, a brand that has a
substantial marketpresence in Madrid. We used the information of
consumers who shop at Carrefourregularly. Further research could
include store brands belonging to different retailersor with
different market positioning, information on regular and
non-regularconsumers at the chains under analysis, and different
locations.
The final limitation of this research is the use of a
compensatory model to estimateconsumer preferences, since there may
be consumers whose levels of attributes leadthem to eliminate a
certain brand from consideration. In the future, it would be
usefulto try other, non-compensatory methods.
BFJ113,10
1246
-
Notes
1. We can observe inTable I that the price intervals between
manufacturer and store brandsvary considerably from one category to
the other. For colas, the price differences for thelowest (store
brand) and the highest-priced brands are 131.25 percent and 100
percent. Forolives, the differences are 76.92 percent and 40
percent. Several authors in the academicliterature have developed
explanations for the differences observed between manufacturerand
store brand prices by product category. Connor and Peterson (1992)
find that productdifferentiation is the main explanatory factor of
price differentials between manufacturerand store brands. The
higher the manufacturer brand differentiation, the lower the degree
ofsubstitutability between manufacturer and store brands in the
product category (Lal andNarasimhan, 1996). In categories with
highly differentiated manufacturer brands, therefore,retailers
apply higher price differentials between manufacturer and store
brands to motivatepurchase of store brands (Raju et al., 1995).
2. The Appendix (Table AI) provides an example of a card used in
the conjoint analysis.
3. Appendix (Table AI) shows the fractional factorial design and
the holdout cards used foreach of the categories analyzed.
4. Publicaciones Alimarket, S.A. is a firm that specializes in
economic and market informationon the sectors and firms in Spain.
Its journal, Alimarket Revista, publishes an annualmonograph on
commercial distribution in the consumer goods sector (food,
beverages andnon-food products). More information is available at:
www.alimarket.es
5. We first obtained two market segments: consumers guided by
intrinsic cues and brandconsumers. We considered the possibility of
applying TwoStep cluster analysis specifyingthree market segments,
and the segment of consumers guided by intrinsic cues divided
intotwo segments: consumers guided by intrinsic cues who prefer the
store brand andconsumers guided by intrinsic cues who prefer the
leading brand. We chose the solution ofthree segments because we
believe it makes more theoretical-practical sense. In the area
ofstore brands, it makes theoretical sense that there is a segment
that prefers the store brandfor its taste. There is great practical
interest in knowing what happens when this segmentrealizes that it
prefers the store brand (which is traditionally perceived as having
lowerquality and worse extrinsic cues). This is crucial in
categories such as cola drinks, in whichthere are prestigious
manufacturer brands.
References
Apelbaum, E., Gerstner, E. and Naik, P.A. (2003), The effects of
expert quality evaluationsversus brand name on price premiums,
Journal of Product and Brand Management,Vol. 12 Nos 2/3, pp.
154-65.
Bellizzi, J.A., Krueckeberg, H.F., Hamilton, J.R. and Martin,
W.S. (1981), Consumers perceptionsof national, private and generic
brands, Journal of Retailing, Vol. 57 No. 4, pp. 56-70.
Bettman, J.R. and Park, C.W. (1980), Effects of prior knowledge
and experience and phase of thechoice process on consumer decision
processes: a protocol analysis, Journal of ConsumerResearch, Vol.
7, pp. 234-48.
Christopher, M., Majaro, S. and McDonald, M. (1987), Strategy
Search, Gower, Aldershot.
Chung, J-E., Yu, J.P. and Pysarchik, D.T. (2006), Cue
utilization to assess food product quality:a comparison of
consumers and retailers in India, International Review of
Retail,Distribution and Consumer Research, Vol. 16 No. 2, pp.
199-214.
Connor, J.M. and Peterson, E.B. (1992), Market structure
determinants of national brand privatelabel price differences of
manufactured food products, The Journal of IndustrialEconomics,
Vol. 40 No. 2, pp. 157-71.
Importance ofbrand-packaging,
price and taste
1247
-
Corstjens, M. and Lal, R. (2000), Building store loyalty through
store brands, Journal ofMarketing Research, Vol. 37 No. 7, pp.
281-91.
Cox, D.F. (1967), The sorting rule model of the consumer product
evaluation process,Risk Taking and Information Handling in Consumer
Behavior, Graduate School ofBusiness Administration, Harvard
University, Boston, MA, pp. 324-69.
Cunningham, I.C., Hardy, A.P. and Imperia, G. (1982), Generic
brands versus national brandsand store brands, Journal of
Advertising Research, Vol. 22 No. 5, pp. 25-32.
Davies, G. and Brito, E. (2004), Price and quality competition
between brands and own brands:a value systems perspective, European
Journal of Marketing, Vol. 38 Nos 1/2, pp. 30-55.
De Chernatony, L. and Knox, S. (1990), How an appreciation of
consumer behavior can help inproduct testing, Journal of the Market
Research Society, Vol. 32 No. 3, pp. 329-47.
De Souza Monteiro, D.M. and Ventura Lucas, M.R. (2001), Conjoint
measurement of preferencesfor traditional cheeses in Lisbon,
British Food Journal, Vol. 103 No. 6, pp. 414-24.
Dunne, D. and Narasimhan, C. (1999), The new appeal of private
labels, Harvard BusinessReview, Vol. 77 No. 3, pp. 41-52.
Ferjani, M., Jedidi, K. and Jagpal, S. (2009), A conjoint
approach for consumer- and firm-levelbrand valuation, Journal of
Marketing Research, Vol. 46, December, pp. 846-62.
Fernandez, A. and Gomez, M. (1999), Estrategias de las marcas de
distribuidor, Distribucion yConsumo, Vol. 45, pp. 30-49.
Gil, J.M. and Sanchez, M. (1997), Consumer preferences for wine
attributes: a conjoint approach,British Food Journal, Vol. 99 No.
1, pp. 3-11.
Hair, J.F., Anderson, R.E., Tatham, R.L. and Black, W.C. (1998),
Multivariate Analysis, 5th ed.,Prentice Hall International Inc.,
Englewood Cliffs, NJ.
Hoch, S.J. (1996), How should national brands think about
private labels?, Sloan ManagementReview, Vol. 37 No. 2, pp.
89-102.
Holbrook, M.B. (1986), Aims, concepts, and methods for the
representation of individualdifferences in esthetic responses to
design features, Journal of Consumer Research, Vol. 13No. 3, pp.
337-47.
Jacoby, J., Szybillo, G. and Busato-Schach, J. (1977),
Information acquisition behavior in brandchoice situations, Journal
of Consumer Research, Vol. 3 No. 4, pp. 209-16.
Kotler, P., Lane, K., Camara, D. and Molla, A. (2006), Direccion
de Marketing, 12th ed., Pearson,Prentice Hall, Upper Saddle River,
NJ.
Lal, R. and Narasimhan, C. (1996), The inverse relationship
between manufacturer and retailermargins: a theory, Marketing
Science, Vol. 15 No. 2, pp. 132-51.
Mendez, J.L., Oubina, J. and Rubio, N. (2008), Expert quality
evaluation and price of store vsmanufacturer brands. An analysis of
the Spanish mass market, Journal of Retailing andConsumer Services,
Vol. 15 No. 3, pp. 144-55.
Ness, M. and Gerhardy, H. (1994), Consumer preferences for
quality and freshness attributes ofeggs, British Food Journal, Vol.
96 No. 6, pp. 53-278.
Olson, J.C. (1972), Cue utilization in the quality perception
process: a cognitive model and anempirical test, unpublished
doctoral dissertation, Purdue University, West Lafayette, IN.
Olson, J.C. and Jacoby, J. (1973), Cue utilization in the
quality perception process: a cognitive andempirical test, in
Venkatesan, M. (Ed.), (Ed.), Proceedings of the Third Annual
Conferenceof the Association for Consumer Research, pp. 167-79.
BFJ113,10
1248
-
Purwar, P.C. (1982), The role of price cue in product quality
perception: a comprehensive modeland an empirical investigation,
Doctoral dissertation, State University of New York,Buffalo,
NY.
Raju, J.S., Sethuraman, R. and Dhar, S.K. (1995), National
brand-store brand price differentialand store brand market share,
Pricing Strategy and Practice: An International Journal,Vol. 3 No.
2, pp. 17-24.
Richardson, P.S., Dick, A. and Jain, A.K. (1994), Extrinsic and
intrinsic cue effects on perceptionof store brand quality, Journal
of Marketing, Vol. 58 No. 4, pp. 28-36.
Sendim, A. (1995), La percepcion de calidad de la canal de
cordero ligero tipo TernascoAragones, Doctoral dissertation,
CIHEAM-IAMZ, Zaragoza.
Seok, J. (2001), Shoppers alternative choice of store and
manufacturers brand: an examinationof the attraction and compromise
effects toward retailers profit optimization, Doctoraldissertation,
University of Texas at Arlington, Arlington, TX.
Skytte, H. and Blunch, N. (2001), Food retailers buying
behaviour: an analysis in 16 Europeancountries, Journal on Chain
and Network Science, Vol. 1 No. 2, pp. 133-45.
Steenkamp, J.B.E.M. (1987), Conjoint measurement in ham quality
evaluation, Journal ofAgricultural Economics, Vol. 38 No. 3, pp.
473-80.
Steenkamp, J.B.E.M. (1990), Conceptual model of the quality
perception process, Journal ofBusiness Research, Vol. 21 No. 4, pp.
309-33.
Zeithaml, V. (1988), Consumer perceptions of price, quality and
value. A means end model andsynthesis of evidence, Journal of
Marketing, Vol. 52 No. 3, pp. 2-22.
Further reading
Anderson, N.H. (1974), Information integration theory: a brief
survey, in Krantz, D.H.,Atkinson, R.C., Luce, R.D. and Suppes, P.
(Eds), Contemporary Developments inMathematical Psychology, Vol. 2,
Freeman, San Francisco, CA.
Dick, A., Chakravarti, D. and Biehal, G. (1990), Memory based
inference during consumerchoice, Journal of Marketing Research,
Vol. 17 No. 1, pp. 82-93.
Loudon, D.L. and Della Bitta, A.J. (1995), Comportamiento del
Consumidor: Conceptos yAplicaciones, McGraw-Hill, Mexico City.
Lutz, R. (1986), Quality is as quality does: an attitudinal
perspective on consumer qualityjudgment, presentation to the
Marketing Science Institute Trustees Meeting, Cambridge,MA.
Importance ofbrand-packaging,
price and taste
1249
-
Appendix
Col
ad
rin
ks
Oli
ves
stu
ffed
wit
han
chov
ies
Bra
nd
-pac
kag
ing
Tas
teP
rice
Bra
nd
-pac
kag
ing
Tas
teP
rice
Car
d1
Gol
d1
(Gol
d)
0.16e
(Gol
d)
Car
refo
ur
1(S
erp
is)
0.52e
(Car
refo
ur)
Car
d2
Coc
aC
ola
3(C
oca
Col
a)0.
32e
(Pep
si)
La
Esp
anol
a3
(La
Esp
anol
a)0.
92e
(Ser
pis
)C
ard
3G
old
3(C
oca
Col
a)0.
37e
(Coc
aC
ola)
Car
refo
ur
3(L
aE
span
ola)
0.73e
(La
Esp
anol
a)C
ard
4C
oca
Col
a2
(Pep
si)
0.16e
(Gol
d)
La
Esp
anol
a2
(Car
refo
ur)
0.52e
(Car
refo
ur)
Car
d5
Pep
si3
(Coc
aC
ola)
0.16e
(Gol
d)
Ser
pis
3(L
aE
span
ola)
0.52e
(Car
refo
ur)
Car
d6
Gol
d2
(Pep
si)
0.32e
(Pep
si)
Car
refo
ur
2(C
arre
fou
r)0.
92e
(Ser
pis
)C
ard
7P
epsi
1(G
old
)0.
32e
(Pep
si)
Ser
pis
1(S
erp
is)
0.92e
(Ser
pis
)C
ard
8P
epsi
2(P
epsi
)0.
37e
(Coc
aC
ola)
Ser
pis
2(C
arre
fou
r)0.
73e
(La
Esp
anol
a)C
ard
9C
oca
Col
a1
(Gol
d)
0.37e
(Coc
aC
ola)
La
Esp
anol
a1
(Ser
pis
)0.
73e
(La
Esp
anol
a)C
ard
10(h
old
out)
Gol
d1
(Gol
d)
0.32e
(Pep
si)
Car
refo
ur
1(S
erp
is)
0.92e
(Ser
pis
)C
ard
11(h
old
out)
Pep
si1
(Gol
d)
0.37e
(Coc
aC
ola)
Ser
pis
1(S
erp
is)
0.73e
(La
Esp
anol
a)C
ard
12(h
old
out)
Pep
si3
(Coc
aC
ola)
0.37e
(Coc
aC
ola)
Ser
pis
3(L
aE
span
ola)
0.73e
(La
Esp
anol
a)C
ard
13(h
old
out)
Coc
aC
ola
1(G
old
)0.
16e
(Gol
d)
La
Esp
anol
a1
(Ser
pis
)0.
52e
(Car
refo
ur)
Table AI.Fractional factorialdesign used for thecategories
analyzed
BFJ113,10
1250
-
About the authorsJose Luis Mendez is Associate Professor of
Marketing Management, Autonoma University ofMadrid. His research
interests are distribution and pricing strategies. He has published
articlesin Spanish journals and in several international journals
such as the International Review ofRetail, Distribution and
Consumer Research, Journal of Product and Brand Management
andJournal of Retailing and Consumer Services, among others. He is
also author of conferenceproceedings for the EMAC, EAERCD and
EIRASS conferences, among others.
Javier Oubina is Associate Professor of Marketing, Autonoma
University of Madrid. Hisresearch interests are power relationships
in distribution channels and pricing strategies. Hispublications
include international journals such as the International Review of
Retail,Distribution and Consumer Research, Journal of Product and
Brand Management, Journal ofRetailing and Consumer Services,
International Journal of Retail and Distribution Managementand
Journal of Marketing Channels, among others; and prestigious
Spanish journals such asRevista Espanola de Investigacion de
Marketing ESIC, Informacion Comercial Espanola andEstudios Sobre
Consumo. He has presented papers at the international conferences
of EMAC,EAERCD and EIRASS.
Natalia Rubio is Associate Professor of Marketing Management,
Autonoma University ofMadrid. Her research interests are brand
management, distribution and marketing research. Shehas published
articles in the European Journal of Marketing, International
Journal of MarketResearch, International Review of Retail,
Distribution and Consumer Research, Journal ofRetailing and
Consumer Services, Journal of Marketing Management, Journal of
Product andBrand Management and International Journal of Retail and
Distribution Management. Herresearch has also appeared in
prestigious Spanish journals such as Revista Espanola
deInvestigacion de Marketing, Informacion Comercial Espanola and
Estudios Sobre Consumo. Shehas presented several papers at the
annual conferences of EMAC, EAERCD and EIRASS, amongothers. Natalia
Rubio is the corresponding author and can be contacted at:
[email protected]
Importance ofbrand-packaging,
price and taste
1251
To purchase reprints of this article please e-mail:
[email protected] visit our web site for further
details: www.emeraldinsight.com/reprints