SlideShare a Scribd company logo
1/7/2015
To imitate or to differentiate
as a private label brand?
The influence of different levels of visual similarity between
the package of a private label product and its national brand
alternative on grocery shoppers’ quality perceptions and
buying intentions
Rico Obdam
STUDENT NUMBER: 1813331
FINAL THESIS, MSC MARKETING, VU UNIVERSITY
SUPERVISOR: ARJEN VAN LIN
2
3
Management summary
This study addresses the following research question: what is the influence of the level and the type
of visual similarity between private label (PL) products and their national brand (NB) alternatives
on grocery shoppers’ quality perceptions of and willingness to buy PL products in a context with
an NB, a PL and a C brand (C) product choice option?
It is found that the higher the level of visual similarity between a PL product and its NB
alternative becomes, the larger grocery shoppers perceive the quality gap between these products,
in favor of the NB option. The results of the analyses suggest that the influence of the level of visual
similarity between a PL product and its NB alternative on the perceived quality gap between the PL
product and its C alternative is less straightforward. It appears that this relationship is inverted U-
shaped, instead of linear. This implies that the higher the level of visual similarity between the PL
and NB product option, the larger the perceived quality gap between the PL product and the C
alternative becomes, until a certain tipping point, after which a further increase in visual similarity
between PL and NB leads to a smaller perceived quality gap between PL and C again. While
assuming that, on average, NB, PL and C products can be roughly classified along a straight line
between the axes perceived quality and price, this finding might imply that the perceived quality
continuum between C, PL and NB is more dynamic than postulated earlier. It might be the case that
not only PL moves along the perceived quality continuum when the level of visual similarity
between PL and NB changes, but also the perceived quality of C and NB are affected.
The influence of the level of visual similarity between PL and NB products on the perceived
quality gaps between the different product options in consumers’ choice set turns out not to be
affected by the type of visual similarity, either feature similarity or theme similarity. Only the level
of visual similarity influences these quality gaps, regardless of whether a particular level of
4
similarity is achieved by imitating the shape, size or color of the package, or by imitating the inferred
attribute and underlying theme of the NB product.
The findings of this study show that the smaller the perceived quality gap between an NB
product and the PL alternative, the higher the willingness to buy the PL product becomes. Also, the
smaller the perceived quality gap between a PL product and the C alternative, the lower the
willingness to buy the PL product becomes. These results provide tentative support for the
assumption that context effects, and in particular asymmetric dominance effects, play a role in
grocery shoppers’ willingness to buy PL products. The perceived quality gaps between NB and PL
products, and PL and C products, respectively, are found to serve as mediators for the relationship
between the level of similarity between a PL product and its NB alternative and grocery shoppers’
willingness to buy the PL product.
To secure the substitution power and absolute margins of PL products, it is important that
grocery retailers position these products as decent quality alternatives for NB products, rather than
as cheap, lower quality options. The results of this study provide insight into how to achieve such a
product positioning. It turns out that marketing managers of PL brands should differentiate the
packages of their PL products from those of the NB alternatives, instead of imitating them. Imitation,
even if it is done subtly, makes grocery shoppers perceive the difference in quality between the
product options as larger, not smaller, which decreases their willingness to buy the PL product. Yet,
PL brands should still spend enough money on packaging, to create and maintain consumers’
perception that they offer high quality products, in order to prevent directly competing on price with
C alternatives. However, PLs should not invest so much in packaging that PL selling prices have to
be increased significantly to cover the increased costs. While PLs should aim to compete with NBs
directly on value for money, they should prevent their prices from becoming equal to NB prices,
thereby preventing competing with NBs solely on quality. If for any reason directly competing with
5
NB products is unfeasible, PLs are forced to compete more directly with C products. In that case,
the visuals of NB products should be subtly imitated by PLs, in order to strengthen the PL’s image
of being an (affordable) high quality brand, so that the perceived quality difference with the C
alternatives increases.
There are also implications for NBs. NB manufacturers should not fear PL brands that are
copycatting the visuals of their products. In fact, the higher the level of visual similarity between
the PL and NB product, the more beneficial it is to the perceived quality and the sales of the NB
option. Marketing managers of NB manufacturers should not differentiate the visuals of product
packages when PL brands are imitating them. If NB manufacturers encounter a PL product with
similar visuals, though it might be perceived by them as unfair, they have no reason or ground to
sue the focal PL brand.
6
Table of contents
Chapter 1 Introduction 8
Chapter 2 Conceptual framework 14
2.1 Relationship between NB manufacturers and grocery retailers 14
2.2 Product package as a quality cue 15
2.2.1 Influence of low to high visual similarity on perceived quality 15
2.2.2 Influence of blatant copycatting on perceived quality 16
2.2.3 Influence of type of similarity on perceived quality 20
2.3 Perceived quality gaps as influencers of willingness to buy PL 21
2.3.1 Influence of perceived quality gap between NB and PL on willingness to buy PL 21
2.3.2 Influence of perceived quality gap between PL and C on willingness to buy PL 25
Chapter 3 Methodology 27
3.1 Type of research 27
3.2 Experimental design 27
3.2.1 General description of experiment 27
3.2.2 Operationalization of concepts 29
3.2.3 Control variables 32
3.3 Data analysis 33
3.4 Participants 34
7
Chapter 4 Results 36
4.1 Descriptives 36
4.1.1 Willingness to buy PL product 36
4.1.2 Perceived quality gaps between product options 39
4.2 Testing conceptual model 42
4.2.1 Perceived quality gap between NB and PL product 42
4.2.2 Perceived quality gap between PL and C product 44
4.2.3 Willingness to buy PL product 46
4.2.4 Mediating role of perceived quality gaps 48
Chapter 5 Discussion 49
5.1 Discussion of results 49
5.2 Managerial implications 54
5.3 Limitations & directions for future research 56
References 59
Appendices 62
Appendix 1 Results of pre-test and resulting visual similarity stimuli
Appendix 2 Experimental survey
Appendix 3 Descriptives of participants
Appendix 4 Descriptives
Appendix 5 Output linear and logistic regression analyses
Appendix 6 Output mediation analyses
8
Chapter 1 Introduction
Grocery private label (PL) brands are brands owned, branded, and offered exclusively by grocery
retailers. In Western Europe, more than one third of supermarkets’ assortments consists of PL
products. PL brands are represented in a wide variety of grocery product categories (Planet Retail,
2007). ACNielsen (2005) found that market shares of U.S. and European PL brands are increasing
more rapidly than national brand (NB) shares in a majority of consumer package goods (CPG)
categories. Owning strong PLs poses two main advantages for grocery retailers. First, a successful
PL brand strengthens retailers’ negotiation position with NB manufacturers, because PL products
can be positioned as close substitutes to NB products, which may decrease the necessity to offer NB
alternatives (Morton & Zettelmeyer, 2004). Second, grocery retailers’ percentage profit margins are
higher for PL products than for NB products. However, selling prices for PL products are
significantly lower than for NB products, so the absolute profit margins for PL products are not
necessarily higher than for NB products. To secure the substitution power and absolute margins of
PL products, it is important that these products are positioned as decent quality alternatives, rather
than as cheap, lower quality options, so that they are indeed perceived as close substitutes and can
be sold for a higher price, yet still lower than the price of the NB alternative (Ailawadi & Harlam,
2004). Supporting this notion, several studies suggest that price and quality are both important in
determining the success of PL brands, but that quality is most important (Richardson, Dick & Jain,
1994; Hoch & Banerji, 1993).
According to Burt (2000), in the 1970s and 1980s PL brands were mainly positioned as
offering a lower quality and lower price than NB alternatives. However, over the last decades, there
has been a shift towards positioning PL brands as offering similar quality as NBs, yet for a lower
price. This shift has been accompanied by the rise of copycat PLs. Copycatting PLs aim to free ride
9
on NBs’ brand equities by imitating visual elements of NB products, such as the size, shape, color,
or texture of the packages. The phenomenon of copycatting is widespread. Sayman, Hoch & Raju
(2002) indicate that package imitation occurs in one third of all CPG categories, and is done both
blatantly and subtle. Morton & Zettelmeyer (2004) found that half the main U.S. supermarkets copy
at least one element of an NB product package for their PL brand.
Several studies have reported that blatant copycatting can be a successful strategy for
retailers, because it may lead to brand confusion. Consumers may evaluate the copycat positively
because they believe it is the original product (Howard, Kerin & Gengler, 2000; Kapferer, 1995).
Even when consumers are not confused, blatant copycats can sometimes be successful, as long as
the imitator does not have the appearance to have aspirations to become a leading brand (Warlop &
Alba, 2004). However, Van Horen & Pieters (2012a) found that when comparative evaluation
between the leading brand and the copycat is stimulated, and thus confusion is ruled out, blatant
copycats are evaluated less positive, while more subtle copycats are evaluated more positive. Results
of another study by the same authors reveal that blatant feature imitation by PLs is perceived by
consumers as unfair and unacceptable, and has a negative influence on product choice of the focal
item (Van Horen & Pieters (2012b).
Although the topic of visual brand similarity is extensively studied, the copycatting literature
has mainly addressed blatant copycatting, often ignoring more subtle forms of visual similarity
between brands. Further, there has been a strong focus on the influence of visual similarity on
consumers’ attitudes towards the copycat brand (Howard, Kerin & Gengler, 2000; Kapferer, 1995;
Warlop & Alba, 2004; Van Horen & Pieters, 2012a). The effects of copycatting on perceived quality
and product choice of the copycat product have been understudied, especially for more subtle
copycats. Van Horen & Pieters (2012b) measured the influence of visual similarity on product
choice, yet only for blatant copycatting, and without taking into account perceived quality as a
10
mediator. Thus, little is known about the influence of different levels of visual similarity on
consumers’ willingness to buy the PL copycat, and how the perceived quality gap between the PL
copycat and the NB alternative mediates this relationship. Also, none of the previous studies on
copycatting have taken into account the likely presence of low-end generic manufacturer brands, or
C-brands (Cs), in consumers’ choice set. This study addresses the influence of different levels of
visual similarity between PL and NB products on the perceived quality gap between these products,
and on grocery shoppers’ willingness to buy the PL product. Also, the moderating effect of type of
visual similarity (i.e. feature vs. theme similarity) on the visual similarity – perceived quality gap
relationship will be addressed. Further, this study takes into account possible context effects within
consumers’ choice sets as a possible explanation for product choice. It includes C products as choice
options besides NB and PL products, to design realistic choice sets of three brand types, in which
these context effects may occur.
It could be expected that the package of a grocery product serves as a quality cue, in the
sense that if visual similarity between PL products and NB products is high, the perceived quality
gap between these products becomes small. Through the eyes of the grocery shopper, this means
that the PL product dominates the NB alternative on price, while offering close to equal quality.
This situation likely leads to an asymmetric dominance effect, which should have a positive
influence on grocery shoppers’ willingness to buy the PL product (Simonson & Tversky, 1992).
However, if visual similarity between the PL product and the NB alternative becomes very high,
and the PL product becomes a blatant copy, it could be expected that the perceived quality gap
between the two products becomes larger again, and the willingness to buy the PL product decreases.
On the other hand, it could be predicted that if visual similarity between PL products and NB
products decreases, the perceived quality gap between these products becomes larger, which implies
a smaller perceived quality gap between the PL product and the C alternative, as PL shifts towards
11
C on the perceived quality continuum between C and NB. Therefore, the PL product will be more
likely to be compared to C, which dominates PL on price. This should have a negative influence on
grocery shoppers’ willingness to buy the PL product. It is expected that above predictions hold for
feature similarity. However, it could be assumed that the influence of theme similarity on the
perceived quality gap between NB and PL is positive for all levels of similarity, also for blatant
similarity, as theme similarity is less obvious and more indirect than feature similarity.
The main research question of this study is: what is the influence of the level and the type of
visual similarity between PL products and their NB alternatives on grocery shoppers’ quality
perceptions of and willingness to buy PL products in a context with an NB, a PL and a C choice
option? In order to be able to answer the main question, the following sub-questions will be
addressed:
 How does the level of visual similarity between PL products and NB products influence the
perceived quality gap between PL products and NB products and the perceived quality gap
between PL products and C products?
 What is the influence of type of visual similarity (i.e. feature similarity vs. theme similarity)
on the influence of visual similarity between PL and NB products on the perceived quality
gap between these products?
 What is the influence of the perceived quality gap between PL products and NB products on
grocery shoppers’ willingness to buy PL products, and what is the influence of the perceived
quality gap between PL products and C products on grocery shoppers’ willingness to buy
PL products?
 Do the perceived quality gaps mediate the relationship between level of visual similarity and
willingness to buy PL products?
12
The results of this study contribute to the copycat and private label versus national brand
literature by providing insight in the role of different levels of visual brand similarity, instead of
only blatant copycatting, on both quality perceptions of and willingness to buy PL products, instead
of attitude towards the PL. This is in contrast to several copycat studies, which mainly addressed
the influence of blatant copycatting on consumers’ attitudes towards the copycat product (Howard,
Kerin & Gengler, 2000; Kapferer, 1995; Warlop & Alba, 2004; Van Horen & Pieters, 2012b).
Moreover, the results may point to the existence of context effects within grocery shoppers’ choice
sets for certain levels of visual similarity between PL products and their NB alternatives. This has
not been done before in the copycatting research domain. Though this paper is somewhat related to
a study by Van Horen & Pieters (2012b), who studied the influence of blatant copycatting on product
attitude and choice, it differs in important respects. In contrast to Van Horen & Pieters’ (2012b)
paper, this study not only addresses the influence of blatant copycatting, but also the effect of more
subtle levels of visual brand similarity on willingness to buy the copycat. Further, this paper includes
perceived quality as a possible mediator between visual similarity and product choice, which Van
Horen & Pieters (2012b) did not. Also, these authors did not take into account the likely existence
of C products and context effects within consumers’ choice sets.
If visual similarity indeed serves as a quality cue, PL managers can use the results of this
study to maximize consumers’ quality perceptions of their products by visually imitating the NB
alternatives to an optimal extent. Subsequently, the selling prices, and therefore the profit margins
for PL products can be optimized accordingly. The results are also useful for PL managers by
providing insight in how to make PL products closer substitutes to NB products. The stronger the
substitution power of PL products becomes, the less dependent on NB manufacturers grocery
retailers become, and the stronger grocery retailers’ negotiation power with NB manufacturers will
be. NB manufacturers can also benefit from the study results. The outcomes shed light on which
13
levels of visual imitation by PLs might be harmful to the sales of the focal NB products, and which
similarity levels are beneficial to the sales of NB alternatives. Knowing this, NB managers will be
able to decide when they have to change the visuals of their product packages, and when they have
to maintain them. Lastly, the results may have legislative implications. For instance, if it appears
that blatant copycatting by PLs indeed increases the perceived quality gap with the NB alternative
and decreases consumers’ willingness to buy the PL product, NBs have no reason and no ground to
sue blatantly copycatting PL brands.
14
Chapter 2 Conceptual framework
2.1 Relationship between NB manufacturers and grocery retailers
NB manufacturers and grocery retailers maintain a difficult relationship. On the one hand, they are
dependent on each other, while on the other hand, they are involved in a continuous battle for
channel profit share. PL products are retailers’ main weapons. It is important for retailers to compete
on quality with NBs, in order to secure high absolute profit margins on their PL products and to
create a strong negotiation position with NB manufacturers by offering close substitutes to the NB
products (Morton & Zettelmeyer, 2004; Ailawadi & Harlam, 2004). Several studies underscore the
importance of competing on quality. Richardson, Dick & Jain (1994) found that perceived quality
is stronger related to willingness to buy PL products than perceived value for money, although the
effect of value for money is significantly positive as well. Hoch & Banerji (1993) came up with
similar results. These findings suggest that consumers often look at quality first, and subsequently
choose the cheapest option from a choice set of products with a similar perceived quality.
NBs have more resources available for product innovation, package design, and advertising
than retailers have. Hence, if retailers aim to position their PL products as high quality alternatives
to NB products, this leaves them with few other options than trying to catch up by imitating visuals
of NB products. Moreover, the more NBs keep innovating and differentiating, the harder it becomes
for PLs to compete on quality with NBs directly (Steenkamp, Van Heerde & Geyskens, 2012).
Product packages are one of the most important determinants of grocery shoppers’ product
perceptions. Therefore, an appealing product package can justify a higher selling price. However,
as soon as competing brands start copying packaging elements, the imitated leading brands often
try to differentiate again by creating a new distinct packages. This interplay between leading and
trailing brands implies an ongoing circle of package differentiation and imitation (Kapferer, 1995).
15
In some cases, retailers and NB manufacturers end up in a lawsuit, mostly when NBs sue PLs for
overly blatant copycatting. To avoid prosecution, retailers tend to avoid producing perfect one-to-
one copies. However, any less obvious forms of blatantly copycatting are still pursued by PLs
(Planet Retail, 2007).
2.2 Product package as a quality cue
2.2.1 Influence of low to high visual similarity on perceived quality
Perceived product quality can be defined as “the consumer’s judgment about a product’s overall
excellence or superiority” (Zeithaml, 1988, p. 3). It differs from objective quality in that it is the
quality of a product through the eyes of the consumer, rather than the quality according to predefined
objective criteria.
Steenkamp, Van Heerde & Geyskens (2012) found that the distinctiveness of the package of
NB products has a positive influence on the perceived quality gap between NB products and the PL
alternatives. They base their findings on perception theory. This theory implies that consumers’
product perceptions are dependent on consumers’ tendency to generalize from one product to
another, and consumers’ proficiency to discriminate between products. If a PL product is visually
similar to an NB product, product generalization is more likely to occur, whereby perceived quality
is generalized from the NB product to the PL alternative. On the other hand, if visual similarity is
low, product discrimination is more likely to occur, whereby a significant perceived quality gap
between the NB and PL products remains. This theory implies that PLs can exploit positive
associations consumers have with NB products by imitating visual elements of these products.
16
Another reason to assume that visual similarity between PL and NB product could have an
effect on the quality gap between these products is that several studies suggest that multiple factors
besides the actual physical product influences consumers’ quality perceptions. Such factors are
called extrinsic cues. According to Teas and Agarwal (2000) and Richardson, Dick & Jain (1994),
several extrinsic product cues, such as price, brand name, country name, but also packaging,
positively affect perceived product quality. Moreover, Richardson, Dick & Jain (1994) found that
extrinsic cues even have a stronger influence on perceived quality than intrinsic product
characteristics, such as ingredients. Visual brand similarity could be viewed as an extrinsic cue that
could affect quality perceptions, as it is a derivative of the extrinsic cue ‘packaging’. Visual
similarity is often achieved through packaging, by imitating elements of the packaging of leading
brand products.
On the other hand, PL products are generally lower priced than NB products, and a lower
product price is generally associated with lower quality. According to attribution theory, a relatively
low price for a PL product might point at inferior product attributes, compared to the NB option
(Sawyer & Dickson, 1984). However, if visual similarity between a PL product and the NB
alternative also serves as a quality cue, the negative influence of price on consumers’ quality
perceptions of the PL product may be offset to a large extent by a highly similar package.
2.2.2 Influence of blatant copycatting on perceived quality
Whereas previously addressed literature supports the assumption that a higher visual similarity
between PL products and their NB alternatives leads to a smaller perceived quality gap between
these products, other studies suggest that this effect reverses beyond a certain tipping point. ‘Too
high’ visual similarity, or blatant copycatting, tends to evoke negative attitudes toward the copycat
product in situations where brand confusion is ruled out, and consumers are aware that the copycat
17
is not the leading brand product. When confusion is not ruled out, blatant copycats can be successful
by triggering consumers to buy the copycat while they think it is the original (Howard, Kerin &
Gengler, 2000; Kapferer, 1995). However, brand confusion is not likely to exist in a supermarket
setting, where consumers compare NB, PL and C products simultaneously. In such a setting, a
potential PL copycat is situated next to the original NB product, and comparative evaluation by
consumers is highly likely to occur. According to Van Horen & Pieters (2012a), blatant copycats
are evaluated less positive, and more subtle copycats are evaluated more positive when comparative
evaluation is stimulated. Results from another study by the same authors show that PL products that
are a blatant copy of the NB alternative in terms of features such as color, depicted objects, package
shape, and sound of the brand name, are evaluated by grocery shoppers as unfair and unacceptable
(Van Horen & Pieters, 2012b). Further, consumers generally perceive obvious visual similarity
between a PL product and the NB original as a PL’s attempt to mislead them (Warlop & Alba,
2004).
It could be assumed that the negative attitude of consumers towards blatant copycats
transfers to their quality perceptions of blatant copycats. Consumers are likely to perceive the copy
as ‘fake’ and ‘not original’. Zaichkowsky & Simpson (1996) support this notion by suggesting that
blatant copycatting causes contrast effects, whereby consumers more critically evaluate the quality
gap between the copy and the original. This implies that ‘too high’ visual similarity between PL and
NB products leads to a larger perceived quality gap between these products. However, because
several studies suggest that moderate to high visual similarity between PL products and NB
alternatives, in contrast to blatant copycatting, leads to a smaller perceived quality gap between the
NB and PL products, the following hypothesis is derived.
18
H1: The higher the visual similarity between a PL product and the NB alternative, the smaller the
perceived quality gap between these two products becomes, until a certain tipping point, after which
a further increase in visual similarity leads to a larger perceived quality gap again. This implies a U-
shape (see figure 1)
Fig. 1: Predicted influence of different levels of visual similarity between PL and NB on the perceived quality gap
between them
Although private labels are offering more and more products that can directly compete on quality
with national brand products, NB, PL and C products can still be roughly classified along a straight
line between the axes (perceived) quality and price. NB products offer high quality for a high price,
PL products offer medium quality for a medium price, and C products offer acceptable quality for
a low price (De Wulf et al., 2005). If these relationships between the three different types of brands
still hold, it can be assumed that a larger perceived quality gap between an NB product and the PL
alternative leads to a smaller perceived quality gap between the PL product and the C alternative, as
19
PL shifts towards C on the perceived quality continuum between C and NB. Also, a smaller
perceived quality gap between NB and PL can be assumed to lead to a larger perceived quality gap
between PL and C. Because the perceived quality gap between NB and PL is assumed to be
influenced by the level of visual similarity between these products, the perceived quality gap
between PL and C should be dependent on the level of visual similarity between PL and NB.
Therefore, the predictions stated in H1 are used to come to the following hypothesis:
H2: The higher the visual similarity between a PL product and the NB alternative, the larger the
perceived quality gap between the PL product and the C alternative becomes, until a certain tipping
point, after which a further increase in visual similarity between PL and NB leads to a smaller
perceived quality gap between PL and C again. This implies an inverted U-shape (see figure 2).
Fig. 2: Predicted influence of different levels of visual similarity between PL and NB on the perceived quality gap
between PL and C
20
2.2.3 Influence of type of similarity on perceived quality
Based on the theory described above, it is predicted that the influence of the level of visual similarity
between a PL and NB product on the perceived quality gap between the PL product and its
alternatives follows an (inverted) U-shape. However, this prediction pertains to visual similarity in
general. The type of visual similarity has not been taken into account. There are two types of visual
similarity between products of different brands: feature similarity and theme similarity (Van Horen
& Pieters, 2012a). Following Van Horen & Pieters (2012a), feature similarity is defined as the
similarity between two products in terms of package shape, size, color, or texture, or in terms of
depicted objects, or the sound of the brand name. Theme similarity could be defined as the similarity
between products in terms of the semantic meaning or inferred attribute of the products. This type
of similarity involves the underlying theme or meaning of the products (e.g. the “traditional, family-
produced olive oil” theme of the Bertolli brand).
Findings by Van Horen & Pieters (2012b) show that grocery shoppers perceive blatant
feature similarity as unacceptable and unfair, while they perceive blatant theme similarity as more
acceptable and less unfair, even if they are aware of the theme copying. Thus, theme copycats are
evaluated more positively than feature copycats. According to the authors, the underlying causes of
these effects are related to the directness of consumers’ perceptual link between the copycat and the
national brand product. Both feature and theme similarity work through similarity associations with
the national brand, which causes feelings and knowledge related to the national brand to transfer to
the copying brand. Feature similarity is a form of literal similarity. Therefore, the perceptual link
between the feature copycat and the original product tends to be direct. This leads to a direct
comparison between the two products, which negatively affects the evaluation of the blatant feature
copycat. Theme similarity, however, is a more indirect type of similarity. The perceptual link
between the theme copycat and the original product is more likely to be indirect, because theme
21
similarity is more implicit, less tangible, and more focused on affective associations. Also,
underlying themes of a product are not only linked to the national brand product, but also to other
products or situations. This leads to a more indirect comparison between the theme copycat and the
imitated product, and therefore to a less negative evaluation of blatant theme copycats.
It is predicted that the negative attitude of consumers toward blatant feature copycats has a
negative influence on the quality perceptions of these products. As already mentioned, such obvious
copying is likely to cause contrast effects between the copy and the original, which makes
consumers perceive the copy as ‘fake’ and ‘not original’ (Zaichkowsky & Simpson, 1996). Such
contrast effects are less likely to occur in case of theme similarity, because such similarity is more
indirect. This leads to the hypothesis:
H3: The type of visual similarity between PL and NB products moderates the relationship between
the level of similarity between these products and the perceived quality gaps between PL and its
alternatives. Specifically, the (inverted) U-shape does exist for feature similarity, but not for theme
similarity, so that the higher the theme similarity between PL and NB, the smaller the perceived
quality gap between NB and PL, and the larger the perceived quality gap between PL and C
becomes.
2.3 Perceived quality gaps as influencers of willingness to buy PL
2.3.1 Influence of perceived quality gap between NB and PL on willingness to buy PL
Van Heerde & Geyskens (2012) studied what makes grocery shoppers willing to pay a premium
price for NB products over PL products. Their results suggest that consumers’ quality perceptions
of NB products have a positive influence on their willingness to pay a premium price for these items.
22
Correspondingly, outcomes of a study by Sethuraman & Cole (1999) indicate that the perceived
quality gap between NB products and PL alternatives is the most important driver of willingness to
pay a premium price for NB products. The reverse also holds. Lower quality perceptions for NB
products, and therefore a smaller perceived quality gap between NB and PL products lead to a lower
willingness to pay a premium price for NB products. Because willingness to pay a premium price
could be interpreted as a derivative of willingness to buy, it could be assumed that a smaller
perceived quality gap between NB and PL products leads to a lower willingness to buy NB products,
and therefore a higher willingness to buy PL products. Sinha & Batra (1999) provide additional
support for this prediction. They found that consumers are more price conscious if they perceive the
price of an NB product as unfair. This feeling of price unfairness arises when the perceived quality
gap between the NB product and the PL alternative is smaller than the size of the quality gap as
claimed by the NB. Subsequently, a higher level of price consciousness positively influences
consumers’ PL purchases.
The findings described above point to the possibility that the effect of the perceived quality
gap between an NB product and the PL alternative on grocery shoppers’ willingness to buy the PL
product is driven by context effects. A consumer’s willingness to buy a certain product is often
affected by the context in which the product evaluation takes place, for instance by the
characteristics and number of other products in a consumer’s choice set (Simonson & Tversky,
1992). As Simonson & Tversky (1992) state: “the same product may appear attractive on the
background of less attractive alternatives and unattractive on the background of more attractive
alternatives” (p. 281). This type of context effect is called tradeoff contrast. One form of tradeoff
contrast is asymmetric dominance, whereby the existence of an inferior product in the consumer’s
choice set increases the consumer’s willingness to buy the superior product. In case of a small
perceived quality gap between NB and PL, through the eyes of the consumer the quality between
23
the two products is comparable, while the price of the PL product is significantly lower. The C
product is only attractive to a minority of consumers who simply seek the lowest price and are
willing to accept a significantly lower quality. NB and PL form the preferred region in consumers’
choice sets, because these products cluster together in terms of quality and represent a two against
one ‘majority’ against the C option. As a consequence, the majority of consumers will choose one
of these two options. In case of a small perceived quality gap between NB and PL, consumers
perceive NB and PL as close to equal in quality, while PL dominates NB on price. Then, the PL
product is the superior option and the NB alternative is the inferior option. This should increase
consumers’ willingness to buy the PL product, compared to the situation where the perceived quality
gap between NB and PL is large, and NB is not an inferior option.
It should be mentioned that this is under the condition that price and quality are more or less
equally important drivers of product choice. As already indicated above, Richardson, Dick & Jain
(1994) provide evidence that this condition is fulfilled by showing that perceived quality is only a
slightly more important driver of willingness to buy PL products than value for money. Furthermore,
they found that “although PL prices are 21% lower, mean value for money perceptions are only 7%
higher than those assigned to NBs. It appears that the poor perceived quality of PLs partially offsets
the otherwise favorable reactions to their lower prices” (p. 33). This finding points to an interplay
between price and quality in determining product choice. Price appears to have a larger influence
on willingness to buy a PL product for higher levels of perceived quality of the PL product. The
results by Richardson, Dick & Jain (1994) provide additional support for the existence of
asymmetric dominance effects within consumers’ choice sets comprising an NB, PL and C product,
by suggesting that consumers first look at quality, and subsequently choose the cheapest option from
a choice set of products with a comparable level of quality. This would imply that consumers not
just seeking the lowest price (C), will compare NB and PL on quality first, and if the perceived
24
quality gap is small (e.g. due to similar packaging), will choose the PL option, because it dominates
the NB alternative on price, and price is an important driver of PL product choice when the perceived
PL quality is high.
Figure 3a illustrates the asymmetric dominance effect for a small perceived quality gap
between NB and PL (which is predicted to exist when visual similarity between PL and NB is high).
It is important to note that context effects are not directly tested in this study, but rather are used as
an argumentation for the predicted effects of certain perceived quality gaps between PL products
and their alternatives on grocery shoppers’ willingness to buy PL products.
Fig. 3: Illustration of variants of asymmetric dominance for small and large perceived quality gap between NB and PL
A review of the findings by previous studies on the role of perceived quality (gaps) on a consumer’s
willingness to buy a product, leads to the following hypothesis:
25
H4: The smaller the perceived quality gap between an NB product and its PL alternative, the higher
the willingness to buy the PL product becomes.
2.3.2 Influence of perceived quality gap between PL and C on willingness to buy PL
It can be derived from the outcomes of a study by Garretson, Fisher & Burton (2002) that grocery
shoppers with high levels of smart shopper self-perception and value consciousness are more likely
to buy PL products than non-discounted NB products. In a similar vein, it could be expected that
the same consumers are more likely to buy C products than PL products when the perceived quality
gap between PL and C is small, because C products are generally cheaper than PL alternatives (De
Wulf et al., 2005).
It could be stated that the results by De Wulf et al. (2005) actually imply the existence of
context effects. For a small perceived quality gap between PL and C, both products offer close to
equal quality, while C dominates PL on price. Thus, asymmetric dominance effects are also likely
to occur in this situation. Consumers seeking high quality regardless of the price are still likely to
buy the NB option, but an increasing number of consumers will choose option C, as it forms a
quality cluster with PL, while it has a significantly lower price. Figure 3b illustrates the assumed
asymmetric dominance effect for a large perceived quality gap between NB and PL, and thus a small
perceived quality gap between PL and C (which is predicted to exist when visual similarity between
PL and NB is non-existing or ‘too’ high). The following hypothesis can be derived:
H5: The smaller the perceived quality gap between a PL product and its C alternative, the lower the
willingness to buy the PL product becomes.
26
Because it is predicted that the level of visual similarity between PL and NB influences the perceived
quality gaps between product options, and because the perceived quality gaps influence consumers’
willingness to buy PL, it could be assumed that the perceived quality gaps are mediators of the
relationship between the level of visual similarity between PL and NB and grocery shoppers’
willingness to buy PL.
H6: The perceived quality gaps between the product options serve as mediators for the relationship
between the level of similarity between a PL product and its NB alternative and grocery shoppers’
willingness to buy the PL product.
The conceptual model illustrated in figure 4 shows the relationships that will be tested in this study.
Fig. 4: Conceptual model
27
Chapter 3 Methodology
3.1 Type of research
Primary data from an experimental survey will be used to test the hypotheses. It was necessary to
conduct an experiment, in order to be able to measure the causal influence of different levels of
visual similarity between PL and NB products on grocery shoppers’ quality perceptions of different
product options, and their willingness to buy PL products. Secondary field data from supermarkets,
for instance transaction data, could only show purchase behavior of grocery shoppers, but cannot
provide insight in the causal influence of visual brand similarity and consumers’ quality perceptions
on this behavior. This is because secondary data is collected from a non-controlled setting, without
stimuli manipulation, so that it cannot be derived that X occured before Y, and without controlling
for extraneous variables, so that no alternative explanations can be ruled out. Thus, if certain
secondary data would somehow have covered information on visual similarity and quality
perceptions, a lack of controlled conditions would make the data useless for showing cause-effect
relationships. An experimental survey, rather than a field experiment, will be conducted, because
some products will have to be manipulated in order to realistically represent PL products. This
manipulation has to be done digitally, by altering product images. Physically manipulating products
will be unfeasible.
3.2 Experimental design
3.2.1 General description of experiment
Participants of the experiment will be asked to virtually buy items from a shopping list consisting
of eight items from several product categories, as if they actually are in a physical supermarket. The
participants will be shown images of three options for each of the eight shopping list items; one NB
28
option, one PL option, and one C option. All PL options will be presented as products from the
Perfekt brand, a PL brand that is sold at several supermarket chains throughout The Netherlands.
By using only one PL brand, the experimental results cannot be influenced by participants’ potential
different preferences for different PL brands. The three product options will be shown
simultaneously, instead of one after the other, in order to simulate a setting more similar to an actual
supermarket, and to be able to account for context effects. Also, the product options will be shown
at random places relative to the other two product options (e.g. not always in a C-PL-NB order), in
order not to suggest any quality sequences. The prices of the different items will be given. Four of
the eight shopping list items will include a PL option that has a certain level of feature similarity
with the NB option. Each of the four items will represent one of the four levels of theme similarity;
no feature similarity, low feature similarity, high feature similarity, and blatant (very high) feature
similarity. The other four of the eight items from the shopping list will include a PL option that has
a certain level of theme similarity with the NB option, ranging from no theme similarity to blatant
theme similarity. The participants will have to choose (‘buy’) one of the three options (NB, PL or
C) for each item from the shopping list.
Each participant will be shown the same items and product options, so the experiment has a
2 (feature vs. theme similarity) × 4 (no vs. low vs. high vs. blatant similarity) within-subject design.
The main reason for choosing a within-subject design is that it most realistically represents a real-
life grocery shopping setting, where some products options look like each other, and others do not.
If some participants will only see highly similar NB and PL options, and some will only be shown
non-similar options, this could influence their quality perceptions of the PL products. Items from
multiple product categories were chosen as stimuli, because a shopping list design realistically
reflects real-life grocery shopping, and shopping lists normally consists of a wide variety of
products. Further, it was unfeasible to find or create PL and NB products from one single product
29
category with the required four different levels of visual similarity. The effect of the chosen PL
product categories will be controlled for by asking respondents to what extent they have a tendency
to buy PL products instead of NB or C products for each chosen product category. This control
variable will be discussed in more detail in paragraph 3.2.3 (‘Control variables’).
After the product choice tasks, the participants will be shown the same items from the
shopping list, but now will be asked to indicate how they perceive the quality of the different product
options. As with the product choice part of the experimental survey, the NB, PL and C options per
item will be shown simultaneously. The results from the perceived quality questions will be used to
calculate the perceived quality gap between PL products and their alternatives. It is important to ask
these questions after the product choice tasks, so that the product choices are not influenced by the
stimulation of participants’ consciousness of the quality of the items. The other way around,
participants might justify their previous product choices by assigning a higher quality than they
actually expect to products they have chosen during the choice tasks. However, this should not be a
major concern, because participants will choose products based on both quality perception and price,
so that the effect of product choice justification on participants’ quality perceptions might be
partially abolished by the influence of price. The experimental survey, including the product images,
can be found in Appendix 2. The example in the appendix is in English, but the actual experimental
survey was in Dutch.
3.2.2 Operationalization of concepts
Level and type of visual similarity between PL and NB product
The level of visual similarity between PL and NB products was determined by a pre-test among
seventeen respondents. One of the product sets was pre-tested among eleven respondents, after
changing the initial set after the first pre-test. The researcher collected several sets of PL and NB
30
products with seemingly different extents of visual similarity. Some products were from low-end
supermarkets Aldi and Lidl, and were manipulated by the researcher, for instance by digitally adding
a PL logo (i.e. a Perfekt logo), so that they appeared to be PL products from a high-end supermarket.
Other products were actual PL products sold by high-end supermarkets, and did not have to be
manipulated. All NB products were actual products from high-end supermarkets, and were not
manipulated. In the pre-test, for each of the collected product sets, respondents were asked to
indicate the level of visual similarity (‘look and feel’) between the PL and the NB products on a
four-item Likert scale covering not similar, a little similar, highly similar, and blatantly similar. The
four items were assigned a number, from 0 to 3, so that mean scores could be computed. The
measure scale was created by the researcher, as no appropriate visual similarity scales were found
in existing (marketing) literature.
The results from the pre-test were used to determine the actual level of visual similarity
between the collected PL and NB products, namely no visual similarity, low visual similarity, high
visual similarity, or blatant (very high) visual similarity. The level of visual similarity was regardless
of the type of similarity (feature versus theme similarity). Product sets with mean similarity scores
close to 0, 1, 2, 3, representing the four different levels of visual similarity, were suitable to be used
within the experiment.
The pre-test determined the level of visual similarity, but not the type of visual similarity
between the PL and NB products. The researcher classified the similarity between the different
products as either feature-based or theme-based, using the definition of the two types of similarity.
PL and NB products that looked mainly similar in terms of package shape, colors, size, texture, or
in terms of depicted objects, or the sound of the product name, were classified as having feature
similarity. PL and NB products that looked mainly similar in terms of the semantic meaning or
inferred attribute, or the underlying theme or meaning, were labelled as having theme similarity
31
(Van Horen & Pieters, 2012a). Each case of feature similarity will be assigned a 0, while each case
of theme similarity will be assigned a 1. Appendix 1 shows the results from the pre-test, and the
resulting visual similarity conditions, including a description per product set of what elements likely
caused the perceived level of visual similarity. The four levels of visual similarity and the two types
of visual similarity form a total of eight experimental conditions. It should be noted that the same
stimulus will be used for both the ‘no feature similarity’ and ‘no theme similarity’ conditions, so
that there are actually seven conditions.
Perceived quality gaps between NB and PL products and between PL and C products
A one-item 7-point Likert scale will be used to measure the perceived quality of each shown NB,
PL and C product. The following statement will be depicted: “All things considered I would say this
(product name) has ...” (Richardson, Dick & Jain’s, 1994, p. 31). The participants can provide an
answer between the end points ‘poor overall quality’ (coded as 0) and ‘excellent overall quality’
(coded as 6). Richardson, Dick & Jain’s (1994) used this measure scale in a grocery store setting. It
is a reflective scale, as the indicator is a manifestation of a consumer’s quality perceptions, rather
than a defining characteristic of it (MacKenzie, 2003). The perceived quality gaps between NB and
PL products and between PL and C products will be calculated by subtracting the perceived quality
scores for the PL products from those for their NB alternatives (NB – PL), and the perceived quality
scores for the C products from those for their PL alternatives (PL – C), respectively. The perceived
quality gaps are interval variables.
Willingness to buy PL product
This is the dependent variable of the conceptual model. As part of the experimental survey,
participants will have to choose one of three options for each grocery shopping list item, or,
32
condition; the NB option, the PL option, or the C option. Consumers’ willingness to buy the PL
product will be operationalized by seperating participants’ PL product option choices from NB or
C product option choices. Each case of NB or C choice will be assigned a 0, whereas each case of
PL choice will be assigned a 1. This makes ‘willingness to buy PL product’ a dichotomous variable,
where 0 represents ‘no PL choice’ and 1 represents ‘PL choice’. It should be noted that the
participants will not actually buy products, but will only indicate which products they would buy in
real-life. Thus, despite the product choice tasks in the experiment, the dependent variable is
‘willingness to buy the PL product’, and not ‘PL product purchase’. The value of the dependent
variable directly reflects consumers’ willingness to buy the PL product. It is a ratio variable, as it
has a meaningful zero (zero participants chose the PL option), and multiplication and division
between two values is meaningful.
3.2.3 Control variables
‘Relative price difference between PL and NB product’, ‘relative price difference between C and
PL product’, and ‘tendency to buy PL in product category’, are the most important variables that
will be controlled for.
Based on actual prices of collected NB, PL and C products, the average relative price
differences between PL products and their NB and C alternatives will be calculated. For the non-
manipulated products, their actual prices will be shown. For the manipulated product (products that
were collected at low-end supermarkets, and were transformed into PL stimuli), fictional prices were
defined, based on the actual prices of their C and NB alternatives, and the calculated average relative
price differences. ‘Relative price difference between PL and NB product’ will be calculated as
follows: (price PL – price NB) / price NB. The calculation of ‘relative price difference between C
and PL product’ is: (price C – price PL) / price PL. It is important to control for the influence of
33
price difference between PL and NB or C stimuli, because by controlling for this variable, it could
be determined whether participants’ product choices were actually due to a certain level of visual
similarity between PL and NB stimuli, or (partly) due to above or below average price differences
between product options.
The outcomes of the experimental survey could be influenced by participants’ grocery
shopping behavior in real-life. Participants might be used to buying PL products for certain product
categories, while having a tendency to buy mainly NB products for other product categories, for
whatever reasons. These tendencies should be controlled for, because they may influence quality
perceptions of PL stimuli products and willingness to buy PL products for certain conditions,
regardless of the level of visual similarity between PL and NB products. The stimuli products can
be classified into different product categories. Participants’ ‘tendency to buy PL products per
product category’ will be addressed by letting participants choose an answer between ‘never’ and
‘always’ (5-item Likert scale) for the following statement: “When I buy [… product category …] in
a supermarket, I choose private label brands over other brands.”
‘Income’ is another variable that will be controlled for, because it could be assumed that
consumers with high incomes are more likely to buy NB products, regardless of the prices and
visuals of PL alternatives. Further, the influence of age and gender will be taken into account.
3.3 Data analysis
Because the dependent variable (‘willingness to buy PL product’) is dichotomous –the number of
available categories is two; yes or no– logistic regression analyses will be conducted to measure the
influence of both the level of visual similarity between PL and NB products, and perceived quality
gaps between PL products and their alternatives on consumers’ willingness to buy PL products.
34
Linear regression analyses will be conducted to test the relationships between these variables.
(Inverted) U-shaped relationships between the level of visual similarity and perceived quality gaps
between products from certain brand types are predicted. Therefore, squared variables will be added
to the model to be able to test these relationships. A moderation analysis will be conducted to analyse
the influence of the type of visual similarity on the relationship between the level of visual similarity
between PL and NB and the perceived quality gap between NB and PL / PL and C.
Since standard mediation analysis cannot be conducted when the dependent variable is
dichotomous, the relevant output from the linear and logistic regression analyses will be inputted
into a ‘mediation with dichotomous outcome variables’ spreadsheet, created by Nathaniel R. Herr
(2006)1
. The relevant output from the spreadsheet calculations will be used as input for Sobel tests.2
The significance of the Z-scores derived from the Sobel tests will determine whether mediation
effects exist.
3.4 Participants
The participants of the experimental survey were Dutch grocery shoppers. Only Dutch people were
eligible to participate, because the stimuli included images of products from Dutch supermarkets,
from Dutch PL brands, and with labels with Dutch texts. Foreign consumers were likely not to know
several of the shown products or brands. To ensure the generalizability of the results, grocery
shoppers from different areas, with different incomes, and from different age groups were recruited.
In total, 178 people participated in the experimental survey. They were between 19 and 70 years
old, with an average of 33 years. 64% of the participants were female, and 36% were male. Only
around 10% percent of the participants had a household income that is far above or far below
1
http://www.nrhpsych.com/mediation/logmed.html
2
http://quantpsy.org/sobel/sobel.htm
35
average, while around 90% had non-extreme household incomes. 50% of the participants indicated
that they have an average income. A complete overview of the descriptives of the participants can
be found in Appendix 3.
36
Chapter 4 Results
4.1 Descriptives
4.1.1 Willingness to buy PL product
In total, the 178 participants of the experimental survey were submitted to seven different
conditions: one ‘no visual similarity’ condition, two ‘low similarity’ conditions (low feature
similarity, and low theme similarity), two ‘high similarity’ conditions (high feature similarity, and
high theme similarity), and two ‘blatant similarity’ conditions (blatant feature similarity, and blatant
theme similarity). When only taking into account the level of visual similarity between a PL product
and its NB alternative, regardless of the type of visual similarity and regardless of participants’
tendency to buy PL products in a specific product category, it can be derived that the relative
willingness to buy the PL option among the participants was highest when visual similarity between
the PL product and the NB alternative was low (42%), and lowest for blatant visual similarity (22%).
Also, willingness to buy the PL product is higher for the two middle levels of visual similarity (low
– 42% and high – 28%) than for the extreme levels of similarity (no – 23% and blatant – 22%).
Table 1 provides an overview of participants’ willingness to buy the PL option per level of
similarity. The complete SPSS descriptives output can be found in Appendix 4.
Table 1: Participants’ willingness to buy PL per level of visual similarity between PL and NB product
Willingness to buy PL
No Yes Total
Count % Count % N %
Level of visual similarity
between PL and NB
No visual similarity 137 77,0% 41 23,0% 178 100,0%
Low visual similarity 206 57,9% 150 42,1% 356 100,0%
High visual similarity 256 71,9% 100 28,1% 356 100,0%
Blatant visual similarity 277 77,8% 79 22,2% 356 100,0%
37
The results in table 1 are illustrated in figure 5. Although these descriptives do not prove any
relationship between level of visual similarity between PL products and their NB alternatives and
willingness to buy PL products, they suggest that an inverted U-shape relation between the two
variables exists.
Fig. 5: Participants’ willingness to buy PL product per level of visual similarity between PL and NB product
Similar results were derived when taking into account the type of visual similarity between PL and
NB products. The results in table 2, which are illustrated in figure 6, show that participants’
willingness to buy PL products across different levels of visual similarity do not deviate much
between feature and theme similarity, and that the percentage distribution of participants’
‘willingness to buy PL product’ is similar across the two types of visual similarity. For both types
of visual similarity, and similar to the general results, participants’ willingness to buy the PL product
was highest for low visual similarity and lowest for blatant visual similarity. Further, participants’
willingness to buy the PL product was higher for the two middle levels of visual similarity (low and
high) than for the extreme levels of similarity (no and blatant), regardless of type of visual similarity.
38
Table 2: Participants’ willingness to buy PL product per level and type of visual similarity between PL and NB product
Willingness to buy PL
No Yes Total
Count % Count % N %
Level of visual
similarity
between
PL and NB
No visual similarity 137 77,0% 41 23,0% 178 100,0%
Low visual similarity Type of visual
similarity
Feature 88 49,4% 90 50,6% 178 100,0%
Theme 118 66,3% 60 33,7% 178 100,0%
High visual similarity Type of visual
similarity
Feature 124 69,7% 54 30,3% 178 100,0%
Theme 132 74,2% 46 25,8% 178 100,0%
Blatant visual similarity Type of visual
similarity
Feature 137 77,0% 41 23,0% 178 100,0%
Theme 140 78,7% 38 21,3% 178 100,0%
Although relationships between variables are not properly tested yet, these descriptives provide a
first indiction that type of visual similarity does not play a role in determining the influence of level
of visual similarity between PL product and their NB alternatives on grocery shoppers’ willingness
to buy PL products.
Fig. 6: Willingness to buy PL product among participants per level of visual similarity between PL and NB product
39
4.1.2 Perceived quality gaps between product options
The results shown in table 3 and illustrated by figure 7 indicate that the mean quality gap between
NB and PL products as perceived by the participants is smallest in the ‘no visual similarity’
condition (0.21) and largest for blatant visual similarity (1.49). Further, this gap is larger for low
similarity than for high similarity (1.10 vs. 0.95). These results point at a linear, instead of a U-
shaped, relationship between level of visual similarity and the perceived quality gap between NB
and PL.
Table 3: Perceived quality gaps between product options among participants per level of visual similarity between PL
and NB product
Perceived quality gap between NB and PL
(NB – PL)
Perceived quality gap between PL and C
(PL – C)
Mean Stnd. Dev. N Mean Stnd. Dev. N
Level of visual
similarity between
PL and NB
No visual similarity ,21 1,24 178 -,17 1,24 178
Low visual similarity 1,10 1,24 356 1,12 1,20 356
High visual similarity ,95 1,19 356 ,65 1,53 356
Blatant visual similarity 1,49 1,35 356 ,64 1,54 356
The results in table 3 and the graph in figure 7 show that an inverted U-shaped relationship between
level of visual similarity and the perceived quality gap between PL and C products might exist. The
mean gap is smallest (even negative) for ‘no visual similarity’ (-0.17), highest for low visual
similarity (1.12), and reduces again for high (0.65) and blatant (0.64) visual similarity.
40
Table 4: Perceived quality gaps between product options among participants per level and type of visual similarity
between PL and NB product
Perceived quality gap between
NB and PL (NB – PL)
Perceived quality gap between
PL and C (PL – C)
Mean Stnd. Dev. N Mean Stnd. Dev. N
Level of
visual
similarity
between
PL and NB
No visual similarity ,21 1,24 178 -,17 1,24 178
Low visual similarity Type of visual
similarity
Feature 1,00 1,26 178 1,24 1,20 178
Theme 1,19 1,21 178 1,01 1,19 178
High visual similarity Type of visual
similarity
Feature ,96 1,17 178 ,55 1,60 178
Theme ,94 1,22 178 ,76 1,45 178
Blatant visual
similarity
Type of visual
similarity
Feature 1,32 1,38 178 ,36 1,55 178
Theme 1,65 1,30 178 ,94 1,47 178
The quality gaps between the products in the different experimental conditions as perceived by the
participants are similar across the two types of visual similarity, as shown in figures 8 and 9. Again,
the results for each type of visual similarity are comparable to the general results. The only exception
is the perceived quality gap between PL and C products for blatant visual similarity, which the
participants perceived on average as much larger for theme similarity (0.94) than for feature
similarity (0.36).
41
Fig. 7: Perceived quality gaps between product options among participants per level of visual similarity between PL
and NB product
Fig. 8: Perceived quality gap between NB and PL product among participants per level and type of visual similarity
between PL and NB product
42
Fig. 9: Perceived quality gap between PL and C product among participants per level and type of visual similarity
between PL and NB product
4.2 Testing conceptual model
4.2.1 Perceived quality gap between NB and PL product
Linear regression analyses were conducted to test the relationships between the independent
variables ‘level of visual similarity between PL and NB product’ and ‘type of visual similarity
between PL and NB product’, and the conceptual mediating variables, ‘perceived quality gap
between NB and PL product’ and ‘perceived quality gap between PL and C product’, controlling
for ‘tendency to buy PL in product category’. The outcomes of the linear regression analysis point
at a positive relationship between ‘level of visual similarity between PL and NB product’ and
‘perceived quality gap between NB and PL product’ (NB – PL) (β = .660, p < .01). Further, the
influence of the squared variable ‘level of visual similarity between PL and NB product ²’ on this
43
perceived quality gap turns out to be non-significant (β = -.069, p = 0.105). These results do not
support hypothesis 1. It was expected that the higher the visual similarity between a PL product and
the NB alternative becomes, the smaller the perceived quality gap between these two products
becomes, until a certain tipping point, after which a further increase in visual similarity leads to a
larger perceived quality gap again. However, the results do not support a U-shaped relationship
between the level of visual similarity and the perceived quality gap between NB and PL. Figure 10
shows the linearity of the relationship. A U-shaped relationship neither was found for specific types
of visual similarity (i.e. feature or theme similarity) (β = -.019, p = .595), in contrast to hypothesis
3, which predicted that a U-shape would exist in case of feature similarity, but not in case of theme
similarity. Table 5 provides an overview of the linear regression analysis outcomes. The complete
output of all analyses can be found in Appendix 5.
Table 5: Results linear regression analysis, DV = perceived quality gap between NB and PL product
Variables β Std. Error Sig.
(Constant) .562 .294 .056
Level of visual similarity .660 .156 .000**
Type of visual similarity .350 .242 .148
Level of visual similarity ² -.069 .042 .105
Level of visual similarity ² × Type of visual similarity
Tendency to buy PL in product category
-.019
-.198
.036
.036
.595
.000**
N=178; *
p < .05; **
p < .01
44
Due to multicollinearity issues, SPSS excluded the variable ‘level of visual similarity × type of
visual similarity’ from the analysis. When conducting the analysis without the squared variables
included, ‘level of visual similarity × type of visual similarity’ was not excluded, and its influence
on the dependent variable could be addressed. ‘Level of visual similarity × type of visual similarity’
turns out not to significantly influence the perceived quality gap between NB and PL products
(p = .593) (see Appendix 5). This outcome implies that the type of visual similarity between PL and
NB products does not moderate the relationship between the level of visual similarity and the
perceived quality gap between NB and PL products.
Fig. 10: Relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap
between NB and PL product’
4.2.2 Perceived quality gap between PL and C product
The results of another linear regression analysis, shown in table 6, indicate that the relationship
between the level of visual similarity between PL and NB product options and the perceived quality
gap between the PL and C options (PL – C) is positive overall (β = 1.077, p < .01). However, the
influence of the squared variable ‘level of visual similarity ²’ on this quality gap appears to be
significant as well (β = -.256, p < .01), pointing at an inverted U-shaped relationship between the
45
level of visual similarity between a PL product and its NB alternative and the perceived quality gap
between the PL products and its C alternative. These results support hypothesis 2. A plot of this
relationship is illustrated in figure 11. However, this inverted U-shaped relationship appears to exist
irrespective of the type of visual similarity (p = .970), so no support for H3 was found.
Table 6: Results linear regression analysis, DV = perceived quality gap between PL and C
Variables β Std. Error Sig.
(Constant) -.297 .337 .379
Level of visual similarity 1.077 .179 .000**
Type of visual similarity .475 .277 .087
Level of visual similarity ² -.256 .048 .000**
Level of visual similarity ² × Type of visual similarity
Tendency to buy PL in product category
-.002
-.086
.041
.041
.970
.035*
N=178; *
p < .05; **
p < .01
Fig. 11: Relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap
between PL and C product’
46
Again, due to issues with multicollinearity, the variable ‘level of visual similarity × type of visual
similarity’ was excluded from the initial analysis. Excluding the squared variables from the model,
and including ‘level of visual similarity × type of visual similarity’, its influence on the DV could
be addressed in an additional linear regression analysis. ‘Level of visual similarity × type of visual
similarity’ appears to significantly influence the perceived quality gap between PL and C products
(β = .506, p < .01) (see Appendix 5). This outcome would imply that the type of visual similarity
between PL and NB products moderates the relationship between the level of visual similarity
between PL and NB products and the perceived quality gap between PL and C products. However,
this effect is overruled by the non-significance of the influence of ‘level of visual similarity ² × type
of visual similarity’ (p = .970).
4.2.3 Willingness to buy PL product
Logistic regression analyses were conducted to determine the influence of the independent and
conceptual mediating variables on the dependent variable, ‘willingness to buy PL product’. Results
of the logistic regression analysis (see table 7) show that the perceived quality gap between NB and
PL product options has a negative effect on consumers’ willingness to buy the PL product (β = .653,
p < .01). This outcome supports hypothesis 4, which predicted that the smaller the perceived quality
gap between an NB product and the PL alternative, the higher the willingness to buy the PL product
becomes. Further, a positive relationship was found between ‘perceived quality gap between PL and
C product’ and ‘willingness to buy PL product’, in support of hypothesis 5. This hypothesis
predicted that a smaller perceived quality gap between a PL product and the C alternative leads to a
lower willingness to buy the PL product.
47
Table 7: Results logistic regression analysis, DV = willingness to buy PL product Reported as ‘β (Std. Error)’
Variables Block 1 Block 2
(Constant) -1.642 (1.001) -1.984 (1.116)
Level of visual similarity 1.533 (1.117) .815 (1.252)
Type of visual similarity 1.245 (1.393) .737 (1.552)
Level of visual similarity × Type of visual similarity
Level of visual similarity ²
Tendency to buy PL in product category
Relative price difference between PL and NB
Relative price difference between C and PL
Perceived quality gap between NB and PL
Perceived quality gap between PL and C
-.505 (.528)
-.269 (.123)*
.115 (.068)
.058 (.067)
-.701 (.803)
-.406 (.587)
-.143 (.137)
.159 (.078)*
.032 (.075)
-.391 (.896)
-.325 (.081)**
.653 (.070)**
N=178; *
p < .05; **
p < .01
The control variables ‘age’ (p = .127), ‘gender’ (p = .084), ‘household income’ (p = .960), and
‘relative price difference between PL and NB’ (p = .663) turned out not to have a significant
influence on grocery shoppers’ willingness to buy PL products, and were therefore excluded.
‘Tendency to buy PL in product category’ only has a significant effect when the perceived quality
gap variables are added to the model (β = .159, p < .05) ‘Level of visual similarity ² × type of visual
similarity’ and ‘relative price difference between C and PL’ were excluded from the logistic analysis
by SPSS, but separate analyses including these variables show that both do not have a significant
influence on ‘willingness to buy PL’ (p = .328 and p = .663, respectively) (see Appendix 5).
48
4.2.4 Mediating role of perceived quality gaps
Because the dependent variable ‘willingness to buy PL product’ is dichotomous, standard mediation
analysis could not be conducted. Therefore, an alternative mediation analysis was conducted. The
relevant output from the linear and logistic regression analyses was inputted into a ‘mediation with
dichotomous outcome variables’ spreadsheet, created by Nathaniel R. Herr (2006). The relevant
output from the spreadsheet calculations was used as input for two Sobel tests. The results from the
Sobel tests indicate that both ‘perceived quality gap between NB and PL product’ (Z = -2.911,
p < .01) and ‘perceived quality gap between PL and C product’ (Z = 5.056, p < .01) mediate the
relationship between the level of visual similarity between a PL product and its NB alternative and
grocery shoppers’ willingness to buy the PL option. These outcomes support hypothesis 6. The
results from the logistic regression analyses, shown in table 7, show that ‘level of visual similarity
between PL and NB product’ has a significant direct effect on ‘willingness to buy PL product’, but
that this relationship becomes non-significant when the perceived quality gap variables are added
to the model. Combining these results with the outcomes of the Sobel tests, it can be concluded that
the effect of the level of visual similarity between a PL product and its NB alternative on consumers’
willingness to buy the PL product is indirect, as it works through the perceived quality gaps between
the different product options, and full mediation exists. The output from both the spreadsheet and
the Sobel tests can be found in Appendix 6.
49
Chapter 5 Discussion
5.1 Discussion of results
This study addresses the influence of the level and the type of visual similarity between PL products
and their NB alternatives on grocery shoppers’ quality perceptions of and willingness to buy PL
products in a context with an NB, a PL and a C choice option.
Controlling for consumers’ tendency to buy PL products in specific product categories, it is
found that the higher the level of visual similarity between a PL product and its NB alternative
becomes, the larger grocery shoppers perceive the quality gap between these products, in favor of
the NB option. The linearity of this relationship implies that even subtle forms of copycatting by
PLs can have a negative effect on how consumers perceive the quality of the imitating PL product
compared to the imitated NB product. Latter outcome is the opposite to findings by Steenkamp, Van
Heerde & Geyskens (2012), who found that the distinctiveness of an NB product package has a
positive influence on the perceived quality gap between the NB product and its PL alternative, and
therefore a higher level of visual similarity between PL and NB leads to a smaller quality gap. They
use perception theory as an explanation for their findings by stating that consumers’ product
perceptions are dependent on their tendency to generalize from one product to another, and their
ability to discriminate between products. They suggest that consumers’ tendency to generalize from
one product to another is stronger for higher levels of visual similarity between the products, and
that their ability to discriminate between products is lower for higher levels of visual similarity.
However, the same theory might explain why the opposite of Van Heerde & Geyskens’ (2012)
findings appears to be true. It could be the case that the higher the distinctiveness of a PL product
package, or the lower the level of visual similarity between the PL product and its NB alternative,
the more difficult it becomes for grocery shoppers to compare the two product options on quality,
50
the smaller the perceived quality gap between NB and PL becomes. Visual similarity between PL
and NB, even when subtle, makes each of the two products less differentiated, which makes
comparison between the products easier. This might stimulate consumers to compare the quality of
the two products more closely, which could lead to the perception that the NB option is ‘real’ and
the PL option is ‘fake’, increasing the perceived quality gap between the products.
The finding that blatant visual similarity between a PL product and its NB alternative leads
to a larger perceived quality gap between the products than moderate to high visual similarity is in
line with this newly proposed idea, and also complies to findings by Van Horen & Pieters (2012a),
who indicate that subtle copycats are evaluated more positive than blatant copycats, as long as the
copycat and the original product are evaluated simultaneously. The outcome also reinforces results
from other studies, which show that consumers possess a negative attitude towards blatant copycats
(Van Horen & Pieters, 2012b), and that this negative attitude transfers to their quality perceptions
of the product compared to the original (Zaichkowsky & Simpson, 1996).
The results of the analyses suggest that the influence of the level of visual similarity between
a PL product and its NB alternative on the perceived quality gap between the PL product and its C
alternative is less straightforward. It appears that this relationship is inverted U-shaped, instead of
linear. This implies that the higher the level of visual similarity between the PL and NB product
option, the larger the perceived quality gap between the PL product and the C alternative becomes,
until a certain tipping point, after which a further increase in visual similarity between PL and NB
leads to a smaller perceived quality gap between PL and C again. These findings imply that PLs can
improve the perceived quality of their products compared to their C alternatives by visually imitating
NB products, as long as the similarity is more or less subtle, and not too obvious.
51
It is remarkable that the relationship between ‘level of visual similarity between PL and NB
product’ and ‘perceived quality gap between NB and PL product’ turns out to be linear, while the
relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality
gap between PL and C product’ turns out to be non-linear. It is assumed that, on average, NB, PL
and C products can be roughly classified along a straight line between the axes perceived quality
and price, with NB products offering high quality for a high price, PL products offering medium
quality for a medium price, and C products offering acceptable quality for a low price (De Wulf et
al., 2005). As a result, while drafting the hypotheses of this study, it was assumed that a larger
perceived quality gap between an NB product and its PL alternative automatically leads to a smaller
perceived quality gap between the PL product and the C alternative, as PL shifts towards C on the
perceived quality continuum between C and NB. Also, a smaller perceived quality gap between NB
and PL was assumed to lead to a larger perceived quality gap between PL and C. Because the
perceived quality gap between NB and PL was correctly assumed to be influenced by the level of
visual similarity between these products, the perceived quality gap between PL and C should be
dependent on the level of visual similarity between PL and NB. Following this reasoning, the shapes
of the relationships between ‘level of visual similarity between PL and NB product’ and ‘perceived
quality gap between NB and PL product’, and between ‘level of visual similarity between PL and
NB product’ and ‘perceived quality gap between PL and C product’ should be symmetrical to each
other (i.e. being positive vs. negative, or being U-shaped vs. inverted U-shaped). However, they are
not, as they are positive versus inverted U-shaped. While still assuming that, on average, NB, PL
and C products can be roughly classified along a straight line between the axes perceived quality
and price, this finding might imply that the perceived quality continuum between C, PL and NB is
more dynamic than postulated earlier. It might be the case that not only PL moves along the
perceived quality continuum when the level of visual similarity between PL and NB changes, but
52
also the perceived quality of C and NB are affected. This might explain why, in contrast to the
predictions, the perceived quality gap between NB and PL, and the pereived quality gap between
PL and C both become larger when the visual similarity between PL and NB changes from low to
moderately high. For instance, it could be that in that particular situation NB moves to the right
(higher perceived quality), and C moves to the left (lower perceived quality), while PL moves only
slightly to the left. In that case, the reason why C shifts to the left might be that PL gains the
appearance of an A product when advancing its packaging while subtly imitating an NB brand, so
that the lower quality of C, through the eyes of consumers, becomes more obvious.
The influence of the level of visual similarity between PL and NB products on the perceived
quality gaps between the different product options in the choice set are not affected by the type of
visual similarity, either feature similarity or theme similarity. Thus, only the level of visual similarity
influences these quality gaps, regardless of whether a particular level of similarity is achieved by
imitating the shape, size or color of the package, or by imitating the inferred attribute and underlying
theme of the NB product. This outcome is not consistent with the results of a study by Van Horen
& Pieters (2012b). These authors state that consumers evaluate theme copycats more positively than
feature copycats, because theme similarity is a more indirect form of similarity compared to feature
similarity, and therefore leads to a more indirect perceptual link between the copycat and the NB
product, so that theme similarity is perceived as less unfair. Although this study uses different
outcome variables than Van Horen & Pieters (2012b) did (perceived quality gaps vs. general
evaluation) to test the influence of the type of visual similarity between a copycat and its NB
alternative, the contradicting findings of the two studies are unexpected. Namely, in this study the
same ‘theme similarity’ stimuli as in Van Horen & Pieter’s (2012b) study were used for two of the
four ‘theme similarity’ experimental conditions: ‘high theme similarity’ and ‘blatant theme
similarity’. It might be that the influence of the type of visual similarity between two products is
53
highly dependent on the number or characteristics of alternative product options in the consumer’s
choice set. In this study, a choice set of three product options is used, including an NB, a PL, and a
C product. In Van Horen & Pieter’s (2012b) experiment, the manipulated choice set consists of two
product options: an NB product and a non-PL copycat product.
The findings of this study show that the smaller the perceived quality gap between an NB
product and the PL alternative, the higher the willingness to buy the PL product becomes. Also, the
smaller the perceived quality gap between a PL product and the C alternative, the lower the
willingness to buy the PL product becomes. These results are in line with the hypotheses, and
provide tentative support for the assumption that context effects, and in particular asymmetric
dominance effects (Simonson & Tversky, 1992), play a role in grocery shoppers’ willingness to buy
PL products. When the perceived quality gap between NB and PL is small, the consumer believes
that the quality between the two product options is comparable, while PL is significantly cheaper.
Thus, consumers perceive NB and PL as close to equal in quality, while PL dominates NB on price.
This means that the PL product is the superior option and the NB alternative is the inferior option
within the preferred region in the consumer’s mind consisting of PL and NB. It is presumed that this
context effect drives consumers’ willingness to buy the PL product. In a similar vein, when the
perceived quality gap between PL and C is small, the consumer believes that the quality between
the two product options is comparable, while C is significantly cheaper. Thus, consumers perceive
PL and C as close to equal in quality, while C dominates PL on price. This means that the C product
is the superior option and the PL alternative is the inferior option within the preferred region in the
consumer’s mind consisting of C and PL, which assumedly discourages grocery shoppers to choose
the PL product.
The perceived quality gaps between NB and PL products, and PL and C products,
respectively, serve as mediators for the relationship between the level of similarity between a PL
54
product and its NB alternative and grocery shoppers’ willingness to buy the PL product, as expected.
Thus, the level of similarity influences the quality gaps, which at their turn influence consumers’
willingness to buy the PL option.
The relative price differences between PL and NB products, and C and PL products,
respectively, appear not to influence grocery shoppers’ willingness to buy PL items. However, no
extreme prices were used in the experimental stimuli. Undoubtedly, as relative price differences
between product options become larger than usual, this will have an effect on consumers’ buying
intentions.
5.2 Managerial implications
The results of this study have several implications for both grocery retailers and NB manufacturers,
as well as legislative implications. As mentioned, PLs are important to grocery retailers in two
different ways. First, a strong PL brand strengthens retailers’ negotiation position with NB
manufacturers, because these manufacturers may realize that retailers can replace some of their NB
products by PL products that are positioned as close substitutes. Second, the percentage profit
margins are higher for PL products than for NB products, mainly because of the lower costs
associated to packaging and advertising. Yet, selling prices for PL products are lower than for NB
products, so the absolute profit margins for PL products are not necessarily higher than for NB
products. To secure the substitution power and absolute margins of PL products, it is important that
grocery retailers position these products as decent quality alternatives for NB products, rather than
as cheap, lower quality options. The findings of this study provide insight into how PL products can
be positioned as decent quality alternatives for more expensive NB products.
55
It turns out that marketing managers of PL brands should differentiate the packages of their
PL products from those of the NB alternatives, instead of imitating them. Imitation, even if it is done
subtly, makes grocery shoppers perceive the difference in quality between the product options as
larger, not smaller, which decreases their willingness to buy the PL product. At the same time, PL
brands should still spend enough money on packaging, to create and maintain consumers’ perception
that they offer high quality products, in order to prevent directly competing on price with C
alternatives. However, PLs should not invest so much in packaging that PL selling prices have to be
increased significantly to cover the increased costs. While PLs should aim to compete with NBs
directly on value for money, they should prevent their prices from becoming equal to NB prices,
thereby preventing competing with NBs solely on (perceived) quality.
If for any reason directly competing with NB products is unfeasible, PLs are forced to
compete more directly with C products. In that case, the visuals of NB products should be subtly
imitated by PLs, in order to strengthen the PL’s image of being an (affordable) high quality brand,
so that the perceived quality difference with the C alternatives increases. This will make consumers
perceive the PL product and C product as being close to equal in price, while they will perceive the
PL product as superior in quality compared to the C option, which decreases the likelihood that they
will choose the C alternative.
The outcomes of this study imply that NB manufacturers should not fear PL brands that are
copycatting the visuals of their products, because copycat PLs appear not to harm NB sales. In fact,
the higher the level of visual similarity between the PL and NB product, the more beneficial it is to
the perceived quality and the sales of the NB option. Contrary to intuition, marketing managers of
NB manufacturers should not differentiate the visuals of product packages when PL brands are
imitating them. The costs savings derived from a lower frequency of package differentiation will
56
lead to higher profits by NB brands. The saved money can also be used to lower NB selling prices,
which makes the PL alternatives less attractive options in terms of value for money.
There are also legislative implications. If NB manufacturers encounter PL products with
similar visuals, they have no reason and no ground to sue the focal PL brand, as the copycatting in
fact benefits them. Although NB manufacturers will likely perceive copycatting by PLs as unfair,
they should ignore them, rather than fight them in court.
5.3 Limitations & directions for future research
It is important to pay some attention to the validity and reliability of this study, especially since it
yields some unexpected findings. First, the pre-test that was conducted to test the suitability of the
experimental stimuli had a sample size of seventeen. The rather small sample size may have led to
a relatively low reliability of the pre-test, and therefore to a relatively low validity of the
experimental conditions. This may have affected the results from the experiment. Second, only
(manipulated) PL products from the ‘Perfekt’ brand were used as stimuli, while participants’ attitude
towards the Perfekt brand compared to other PL brands was not controlled for. Participants’ attitude
towards this specific PL brand may have affected the validity of the results. Third, all PL product
stimuli were digitally manipulated by adding a Perfekt logo. Some participants may have noted that
some product stimuli were ‘fake’, which might have influenced their responses. Fourth, regarding
the reliability of the study results, all participants were Dutch, so generalization of the results might
be problematic. Fifth, the sample size of 178 participants is rather small, and the results could have
been more reliable if more people participated in the experiment. The limiations described above
mainly represent generalizability issues. Future research should determine the robustness of the
results of this study. Sixth, asymmetric dominance effects are postulated to drive the effect of the
57
perceived quality gaps between NB and PL, and PL and C, respectively, on consumers’ willingness
to buy PL products. However, such context effects were not directly tested. It is only assumed that
these effects play a role in the relationship between the mentioned variables. Future studies should
directly test the influence of such context effects. Finally, although this study controls for the
influence of relative price differences between product options, this study does not explicitly take
into account different types of PL products (i.e. economy PL products, standard PL products, and
premium PL products). Results of a study by Geyskens, Gielens & Gijsbrechts (2010) show that the
introduction of economy PLs cannibalizes standard PLs, and the introduction of premium PLs
cannibalizes both standard and premium PLs, driven by price-quality context effects. Also, economy
PL introductions benefit mainstream-quality NBs because these NBs become a middle or
compromise option in terms of quality in the consumers’ choice set. These findings confirm that
consumers assign different quality levels to different PL product types. Future research should
explore whether the influence of the level of visual similarity between a PL product and its NB
alternative on grocery shoppers’ willingness to pay the PL option is dependent on the type of PL
product. For instance, it could be expected that a copycat economy PL product is perceived as more
‘fake’ than a copycat premium PL product, as an economy PL product is perceived as a low quality
option anyway. A copycat premium PL product might benefit from its appearance of being able to
compete on quality with the NB alternative.
There are also several avenues for future research that are not directly derived from the
limitations. Some possible explanations for unexpected results were given in paragraph 5.1,
‘Discussion of results’. Future studies should explore whether these explanations are correct. First,
if the relationship between the level of visual similarity between a PL and NB product and the
perceived quality gap between these products indeed appears to be linear and positive, the perception
theory explanation provided in paragraph 5.1 should be tested. Second, futures studies should
58
address the (non-)symmetry of the relationships between ‘level of visual similarity between PL and
NB product’ and ‘perceived quality gap between NB and PL product’, and between ‘level of visual
similarity between PL and NB product’ and ‘perceived quality gap between PL and C product’. This
study found that the shapes of these relationships are asymmetrical, in contrast to what was expected.
Future studies should test whether they are indeed asymmetrical, and if so, test whether this
asymmetry is due to the dynamics within the perceived quality continuum between C, PL and NB,
as postulated in paragraph 5.1. Third, an avenue for future study is the moderating role of the type
of visual similarity on the effect of the level of visual similarity between PL and NB on the perceived
quality gaps between different product options and on consumers’ the willingness to buy PL
products. This study found that the type of visual similarity does not play a significant role, in
contrast to a previous findings. Future research should provide decisive insights into this issue.
Finally, a direction for study could be the influence of different types of PL products
59
References
ACNielsen (2005), The Power of Private Label 2005: A Review of Growth Trends Around the World.
New York: ACNielsen.
Ailawadi, K. L., & Harlam, B. (2004). An empirical analysis of the determinants of retail margins:
the role of store-brand share. Journal of Marketing, 68(1), 147-165.
Garretson, J. A., Fisher, D., & Burton, S. (2002). Antecedents of private label attitude and national
brand promotion attitude: similarities and differences. Journal of Retailing, 78(2), 91-99.
Hoch, S. J., & Banerji, S. (1993). When do private labels succeed. Sloan Management Review, 34.
Horen, F. V., & Pieters, R. (2012a). When high-similarity copycats lose and moderate-similarity
copycats gain: the impact of comparative evaluation. Journal of Marketing Research, 49(1), 83-91.
Horen, F. V., & Pieters, R. (2012b). Consumer evaluation of copycat brands: The effect of imitation
type. International Journal of Research in Marketing, 29(3), 246-255.
Howard, D. J., Kerin, R. A., & Gengler, C. (2000). The effects of brand name similarity on brand
source confusion: Implications for trademark infringement. Journal of Public Policy &
Marketing, 19(2), 250-264.
60
Kapferer, J. N. (1995). Brand confusion: empirical study of a legal concept. Psychology &
Marketing, 12(6), 551-568.
MacKenzie, S. B. (2003). The dangers of poor construct conceptualization. Journal of the Academy
of Marketing Science, 31(3), 323-326.
Morton, F. S., & Zettelmeyer, F. (2004). The strategic positioning of store brands in retailer–
manufacturer negotiations. Review of Industrial Organization, 24(2), 161-194.
Planet Retail (2007), Private Label Trends Worldwide. London: Planet Retail Ltd.
Richardson, P. S., Dick, A. S., & Jain, A. K. (1994). Extrinsic and intrinsic cue effects on perceptions
of store brand quality. The Journal of Marketing, 28-36.
Sawyer, A. G., & Dickson, P. R. (1984). Psychological perspectives on consumer response to sales
promotion. Research on sales promotion: Collected papers, 1-21.
Sayman, S., Hoch, S. J., & Raju, J. S. (2002). Positioning of store brands. Marketing Science, 21(4),
378-397.
Simonson, I., & Tversky, A. (1992). Choice in context: Tradeoff contrast and extremeness
aversion. Journal of marketing research, 29(3), 281-295.
61
Steenkamp, J. B. E., Van Heerde, H. J., & Geyskens, I. (2010). What makes consumers willing to
pay a price premium for national brands over private labels?. Journal of Marketing Research, 47(6),
1011-1024.
Teas, R. K., & Agarwal, S. (2000). The effects of extrinsic product cues on consumers’ perceptions
of quality, sacrifice, and value. Journal of the Academy of marketing Science, 28(2), 278-290.
Warlop, L., & Alba, J. W. (2004). Sincere flattery: Trade-dress imitation and consumer
choice. Journal of Consumer Psychology, 14(1), 21-27.
Wulf, K. D., Odekerken-Schröder, G., Goedertier, F., & Van Ossel, G. (2005). Consumer
perceptions of store brands versus national brands. Journal of Consumer Marketing, 22(4), 223-232.
Zaichkowsky, J. L., & Simpson, R. N. (1996). The effect of experience with a brand imitator on the
original brand. Marketing Letters, 7(1), 31-39.
Zeithaml, V. A. (1988). Consumer perceptions of price, quality, and value: a means-end model and
synthesis of evidence. The Journal of marketing, 52, 2-22.
62
Appendices
Appendix 1 Results of pre-test and resulting visual similarity stimuli
Blatant feature similarity (mean pre-test score = 2.8, N=17)
Elements likely causing perceived level of visual similarity:
- Package shape (round)
- Depicted items (cow, countryside, blue sky)
- Image type (cartoon)
- Colours (brown, green, blue, white)
High feature similarity (mean pre-test score = 2.2, N=17)
Elements likely causing perceived level of visual similarity:
- Package shape (small, long)
- Package material (carton)
- Depicted items (cows, grass, countryside, Dutch flag, dish with yogurt)
- Image type (cartoon)
- Colours (dark green)
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis
Obdam_Rico_1813331_final_thesis

More Related Content

What's hot

Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...
Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...
Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...
SKIM
 
2012 showrooming article
2012 showrooming article2012 showrooming article
2012 showrooming article
ndiamond2425
 
Sample Market Research & Analysis
Sample Market Research & AnalysisSample Market Research & Analysis
Sample Market Research & Analysis
Rubayet Hassan
 
Umberla
UmberlaUmberla
Colgate palmolive company
Colgate palmolive companyColgate palmolive company
Colgate palmolive company
Prateek Jain
 
New Strategies in Building Profitable Brand Extensions
New Strategies in Building Profitable Brand ExtensionsNew Strategies in Building Profitable Brand Extensions
New Strategies in Building Profitable Brand Extensions
Datamonitor Consumer
 
I038079084
I038079084I038079084
I038079084
inventionjournals
 
A study on brand personality orientation for lee jeans
A study on brand personality orientation  for lee jeansA study on brand personality orientation  for lee jeans
A study on brand personality orientation for lee jeans
Arijit Basu
 
Review of literature spl
Review of literature splReview of literature spl
Review of literature spl
sreedharm
 
The Effects of the Determinants of Customer Satisfaction on Brand Loyalty
The Effects of the Determinants of Customer Satisfaction on Brand LoyaltyThe Effects of the Determinants of Customer Satisfaction on Brand Loyalty
The Effects of the Determinants of Customer Satisfaction on Brand Loyalty
Samaan Al-Msallam
 
Hingorani
HingoraniHingorani
Market Research Anti Fog Mirrors Business Proposal
Market Research Anti Fog Mirrors Business ProposalMarket Research Anti Fog Mirrors Business Proposal
Market Research Anti Fog Mirrors Business Proposal
Huzefa Chini
 
msl_2014_296
msl_2014_296msl_2014_296
msl_2014_296
Shabnam Kh
 
Chapter 10 crafting the brand positioning
Chapter 10 crafting the brand positioningChapter 10 crafting the brand positioning
Chapter 10 crafting the brand positioning
Dondon Corpuz
 

What's hot (14)

Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...
Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...
Win-Win – Give telecoms customers the freedom they want and they’ll reward yo...
 
2012 showrooming article
2012 showrooming article2012 showrooming article
2012 showrooming article
 
Sample Market Research & Analysis
Sample Market Research & AnalysisSample Market Research & Analysis
Sample Market Research & Analysis
 
Umberla
UmberlaUmberla
Umberla
 
Colgate palmolive company
Colgate palmolive companyColgate palmolive company
Colgate palmolive company
 
New Strategies in Building Profitable Brand Extensions
New Strategies in Building Profitable Brand ExtensionsNew Strategies in Building Profitable Brand Extensions
New Strategies in Building Profitable Brand Extensions
 
I038079084
I038079084I038079084
I038079084
 
A study on brand personality orientation for lee jeans
A study on brand personality orientation  for lee jeansA study on brand personality orientation  for lee jeans
A study on brand personality orientation for lee jeans
 
Review of literature spl
Review of literature splReview of literature spl
Review of literature spl
 
The Effects of the Determinants of Customer Satisfaction on Brand Loyalty
The Effects of the Determinants of Customer Satisfaction on Brand LoyaltyThe Effects of the Determinants of Customer Satisfaction on Brand Loyalty
The Effects of the Determinants of Customer Satisfaction on Brand Loyalty
 
Hingorani
HingoraniHingorani
Hingorani
 
Market Research Anti Fog Mirrors Business Proposal
Market Research Anti Fog Mirrors Business ProposalMarket Research Anti Fog Mirrors Business Proposal
Market Research Anti Fog Mirrors Business Proposal
 
msl_2014_296
msl_2014_296msl_2014_296
msl_2014_296
 
Chapter 10 crafting the brand positioning
Chapter 10 crafting the brand positioningChapter 10 crafting the brand positioning
Chapter 10 crafting the brand positioning
 

Similar to Obdam_Rico_1813331_final_thesis

private label
private labelprivate label
private label
poojanagrut
 
The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...
The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...
The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...
AJHSSR Journal
 
Paper 1 (2).pdf
Paper 1 (2).pdfPaper 1 (2).pdf
Paper 1 (2).pdf
RupamLonsane1
 
Comparison on ads or not? Influence of Referent on Advertising Effectiveness
Comparison on ads or not? Influence of Referent on Advertising EffectivenessComparison on ads or not? Influence of Referent on Advertising Effectiveness
Comparison on ads or not? Influence of Referent on Advertising Effectiveness
Waqas Tariq
 
The State of Native Advertising Disclosure: The Consumer View
The State of Native Advertising Disclosure: The Consumer ViewThe State of Native Advertising Disclosure: The Consumer View
The State of Native Advertising Disclosure: The Consumer View
TripleLift
 
The role of content - nielsen
The role of content - nielsenThe role of content - nielsen
The role of content - nielsen
Genaro Bardy
 
Albert
AlbertAlbert
Albert
Md Alam
 
Marketing Chapter 20
Marketing Chapter 20Marketing Chapter 20
Marketing Chapter 20
WanBK Leo
 
Consumer behavior for unexpected prices
Consumer behavior for unexpected pricesConsumer behavior for unexpected prices
Consumer behavior for unexpected prices
Krishna Bhawsar
 
Final eco. assignment (1)
Final eco. assignment (1)Final eco. assignment (1)
Final eco. assignment (1)
SHIKHAR MAURYA
 
The Influence of Product Quality, Price and Brand Image On Customer Satisfact...
The Influence of Product Quality, Price and Brand Image On Customer Satisfact...The Influence of Product Quality, Price and Brand Image On Customer Satisfact...
The Influence of Product Quality, Price and Brand Image On Customer Satisfact...
International Journal of Business Marketing and Management (IJBMM)
 
Strategic Brand Management 1
Strategic Brand Management 1Strategic Brand Management 1
Strategic Brand Management 1
rishistd
 
Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...
Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...
Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...
IJMCERJournal
 
Brand loyalty, Brand Relationship and Brand Equity - Introduction to Branding
Brand loyalty, Brand Relationship and Brand Equity - Introduction to BrandingBrand loyalty, Brand Relationship and Brand Equity - Introduction to Branding
Brand loyalty, Brand Relationship and Brand Equity - Introduction to Branding
Rai University Ahmedabad
 
Mediation of brand equity on word of mouth a study with special reference to ...
Mediation of brand equity on word of mouth a study with special reference to ...Mediation of brand equity on word of mouth a study with special reference to ...
Mediation of brand equity on word of mouth a study with special reference to ...
Dr. V.K. Hamza
 
Brand Relationship Quality as a Formative Third-order Construct – Findings of...
Brand Relationship Quality as a Formative Third-order Construct – Findings of...Brand Relationship Quality as a Formative Third-order Construct – Findings of...
Brand Relationship Quality as a Formative Third-order Construct – Findings of...
CBR Conference
 
MILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTA
MILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTAMILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTA
MILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTA
IAEME Publication
 
project
projectproject
Pricing
PricingPricing
Pricing
Nevo Hadas
 
New microsoft office word document
New microsoft office word documentNew microsoft office word document
New microsoft office word document
gurubhai2012
 

Similar to Obdam_Rico_1813331_final_thesis (20)

private label
private labelprivate label
private label
 
The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...
The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...
The Effect of Brand Equity on Consumer Buyer Decisions : A Case Study in Groc...
 
Paper 1 (2).pdf
Paper 1 (2).pdfPaper 1 (2).pdf
Paper 1 (2).pdf
 
Comparison on ads or not? Influence of Referent on Advertising Effectiveness
Comparison on ads or not? Influence of Referent on Advertising EffectivenessComparison on ads or not? Influence of Referent on Advertising Effectiveness
Comparison on ads or not? Influence of Referent on Advertising Effectiveness
 
The State of Native Advertising Disclosure: The Consumer View
The State of Native Advertising Disclosure: The Consumer ViewThe State of Native Advertising Disclosure: The Consumer View
The State of Native Advertising Disclosure: The Consumer View
 
The role of content - nielsen
The role of content - nielsenThe role of content - nielsen
The role of content - nielsen
 
Albert
AlbertAlbert
Albert
 
Marketing Chapter 20
Marketing Chapter 20Marketing Chapter 20
Marketing Chapter 20
 
Consumer behavior for unexpected prices
Consumer behavior for unexpected pricesConsumer behavior for unexpected prices
Consumer behavior for unexpected prices
 
Final eco. assignment (1)
Final eco. assignment (1)Final eco. assignment (1)
Final eco. assignment (1)
 
The Influence of Product Quality, Price and Brand Image On Customer Satisfact...
The Influence of Product Quality, Price and Brand Image On Customer Satisfact...The Influence of Product Quality, Price and Brand Image On Customer Satisfact...
The Influence of Product Quality, Price and Brand Image On Customer Satisfact...
 
Strategic Brand Management 1
Strategic Brand Management 1Strategic Brand Management 1
Strategic Brand Management 1
 
Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...
Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...
Perceived Quality and Competitive Advantage in Beer Products In Kabale Distri...
 
Brand loyalty, Brand Relationship and Brand Equity - Introduction to Branding
Brand loyalty, Brand Relationship and Brand Equity - Introduction to BrandingBrand loyalty, Brand Relationship and Brand Equity - Introduction to Branding
Brand loyalty, Brand Relationship and Brand Equity - Introduction to Branding
 
Mediation of brand equity on word of mouth a study with special reference to ...
Mediation of brand equity on word of mouth a study with special reference to ...Mediation of brand equity on word of mouth a study with special reference to ...
Mediation of brand equity on word of mouth a study with special reference to ...
 
Brand Relationship Quality as a Formative Third-order Construct – Findings of...
Brand Relationship Quality as a Formative Third-order Construct – Findings of...Brand Relationship Quality as a Formative Third-order Construct – Findings of...
Brand Relationship Quality as a Formative Third-order Construct – Findings of...
 
MILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTA
MILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTAMILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTA
MILLENNIALS PERCEPTION TO HIGH CLASS CINEMA CGV JAKARTA
 
project
projectproject
project
 
Pricing
PricingPricing
Pricing
 
New microsoft office word document
New microsoft office word documentNew microsoft office word document
New microsoft office word document
 

Obdam_Rico_1813331_final_thesis

  • 1. 1/7/2015 To imitate or to differentiate as a private label brand? The influence of different levels of visual similarity between the package of a private label product and its national brand alternative on grocery shoppers’ quality perceptions and buying intentions Rico Obdam STUDENT NUMBER: 1813331 FINAL THESIS, MSC MARKETING, VU UNIVERSITY SUPERVISOR: ARJEN VAN LIN
  • 2. 2
  • 3. 3 Management summary This study addresses the following research question: what is the influence of the level and the type of visual similarity between private label (PL) products and their national brand (NB) alternatives on grocery shoppers’ quality perceptions of and willingness to buy PL products in a context with an NB, a PL and a C brand (C) product choice option? It is found that the higher the level of visual similarity between a PL product and its NB alternative becomes, the larger grocery shoppers perceive the quality gap between these products, in favor of the NB option. The results of the analyses suggest that the influence of the level of visual similarity between a PL product and its NB alternative on the perceived quality gap between the PL product and its C alternative is less straightforward. It appears that this relationship is inverted U- shaped, instead of linear. This implies that the higher the level of visual similarity between the PL and NB product option, the larger the perceived quality gap between the PL product and the C alternative becomes, until a certain tipping point, after which a further increase in visual similarity between PL and NB leads to a smaller perceived quality gap between PL and C again. While assuming that, on average, NB, PL and C products can be roughly classified along a straight line between the axes perceived quality and price, this finding might imply that the perceived quality continuum between C, PL and NB is more dynamic than postulated earlier. It might be the case that not only PL moves along the perceived quality continuum when the level of visual similarity between PL and NB changes, but also the perceived quality of C and NB are affected. The influence of the level of visual similarity between PL and NB products on the perceived quality gaps between the different product options in consumers’ choice set turns out not to be affected by the type of visual similarity, either feature similarity or theme similarity. Only the level of visual similarity influences these quality gaps, regardless of whether a particular level of
  • 4. 4 similarity is achieved by imitating the shape, size or color of the package, or by imitating the inferred attribute and underlying theme of the NB product. The findings of this study show that the smaller the perceived quality gap between an NB product and the PL alternative, the higher the willingness to buy the PL product becomes. Also, the smaller the perceived quality gap between a PL product and the C alternative, the lower the willingness to buy the PL product becomes. These results provide tentative support for the assumption that context effects, and in particular asymmetric dominance effects, play a role in grocery shoppers’ willingness to buy PL products. The perceived quality gaps between NB and PL products, and PL and C products, respectively, are found to serve as mediators for the relationship between the level of similarity between a PL product and its NB alternative and grocery shoppers’ willingness to buy the PL product. To secure the substitution power and absolute margins of PL products, it is important that grocery retailers position these products as decent quality alternatives for NB products, rather than as cheap, lower quality options. The results of this study provide insight into how to achieve such a product positioning. It turns out that marketing managers of PL brands should differentiate the packages of their PL products from those of the NB alternatives, instead of imitating them. Imitation, even if it is done subtly, makes grocery shoppers perceive the difference in quality between the product options as larger, not smaller, which decreases their willingness to buy the PL product. Yet, PL brands should still spend enough money on packaging, to create and maintain consumers’ perception that they offer high quality products, in order to prevent directly competing on price with C alternatives. However, PLs should not invest so much in packaging that PL selling prices have to be increased significantly to cover the increased costs. While PLs should aim to compete with NBs directly on value for money, they should prevent their prices from becoming equal to NB prices, thereby preventing competing with NBs solely on quality. If for any reason directly competing with
  • 5. 5 NB products is unfeasible, PLs are forced to compete more directly with C products. In that case, the visuals of NB products should be subtly imitated by PLs, in order to strengthen the PL’s image of being an (affordable) high quality brand, so that the perceived quality difference with the C alternatives increases. There are also implications for NBs. NB manufacturers should not fear PL brands that are copycatting the visuals of their products. In fact, the higher the level of visual similarity between the PL and NB product, the more beneficial it is to the perceived quality and the sales of the NB option. Marketing managers of NB manufacturers should not differentiate the visuals of product packages when PL brands are imitating them. If NB manufacturers encounter a PL product with similar visuals, though it might be perceived by them as unfair, they have no reason or ground to sue the focal PL brand.
  • 6. 6 Table of contents Chapter 1 Introduction 8 Chapter 2 Conceptual framework 14 2.1 Relationship between NB manufacturers and grocery retailers 14 2.2 Product package as a quality cue 15 2.2.1 Influence of low to high visual similarity on perceived quality 15 2.2.2 Influence of blatant copycatting on perceived quality 16 2.2.3 Influence of type of similarity on perceived quality 20 2.3 Perceived quality gaps as influencers of willingness to buy PL 21 2.3.1 Influence of perceived quality gap between NB and PL on willingness to buy PL 21 2.3.2 Influence of perceived quality gap between PL and C on willingness to buy PL 25 Chapter 3 Methodology 27 3.1 Type of research 27 3.2 Experimental design 27 3.2.1 General description of experiment 27 3.2.2 Operationalization of concepts 29 3.2.3 Control variables 32 3.3 Data analysis 33 3.4 Participants 34
  • 7. 7 Chapter 4 Results 36 4.1 Descriptives 36 4.1.1 Willingness to buy PL product 36 4.1.2 Perceived quality gaps between product options 39 4.2 Testing conceptual model 42 4.2.1 Perceived quality gap between NB and PL product 42 4.2.2 Perceived quality gap between PL and C product 44 4.2.3 Willingness to buy PL product 46 4.2.4 Mediating role of perceived quality gaps 48 Chapter 5 Discussion 49 5.1 Discussion of results 49 5.2 Managerial implications 54 5.3 Limitations & directions for future research 56 References 59 Appendices 62 Appendix 1 Results of pre-test and resulting visual similarity stimuli Appendix 2 Experimental survey Appendix 3 Descriptives of participants Appendix 4 Descriptives Appendix 5 Output linear and logistic regression analyses Appendix 6 Output mediation analyses
  • 8. 8 Chapter 1 Introduction Grocery private label (PL) brands are brands owned, branded, and offered exclusively by grocery retailers. In Western Europe, more than one third of supermarkets’ assortments consists of PL products. PL brands are represented in a wide variety of grocery product categories (Planet Retail, 2007). ACNielsen (2005) found that market shares of U.S. and European PL brands are increasing more rapidly than national brand (NB) shares in a majority of consumer package goods (CPG) categories. Owning strong PLs poses two main advantages for grocery retailers. First, a successful PL brand strengthens retailers’ negotiation position with NB manufacturers, because PL products can be positioned as close substitutes to NB products, which may decrease the necessity to offer NB alternatives (Morton & Zettelmeyer, 2004). Second, grocery retailers’ percentage profit margins are higher for PL products than for NB products. However, selling prices for PL products are significantly lower than for NB products, so the absolute profit margins for PL products are not necessarily higher than for NB products. To secure the substitution power and absolute margins of PL products, it is important that these products are positioned as decent quality alternatives, rather than as cheap, lower quality options, so that they are indeed perceived as close substitutes and can be sold for a higher price, yet still lower than the price of the NB alternative (Ailawadi & Harlam, 2004). Supporting this notion, several studies suggest that price and quality are both important in determining the success of PL brands, but that quality is most important (Richardson, Dick & Jain, 1994; Hoch & Banerji, 1993). According to Burt (2000), in the 1970s and 1980s PL brands were mainly positioned as offering a lower quality and lower price than NB alternatives. However, over the last decades, there has been a shift towards positioning PL brands as offering similar quality as NBs, yet for a lower price. This shift has been accompanied by the rise of copycat PLs. Copycatting PLs aim to free ride
  • 9. 9 on NBs’ brand equities by imitating visual elements of NB products, such as the size, shape, color, or texture of the packages. The phenomenon of copycatting is widespread. Sayman, Hoch & Raju (2002) indicate that package imitation occurs in one third of all CPG categories, and is done both blatantly and subtle. Morton & Zettelmeyer (2004) found that half the main U.S. supermarkets copy at least one element of an NB product package for their PL brand. Several studies have reported that blatant copycatting can be a successful strategy for retailers, because it may lead to brand confusion. Consumers may evaluate the copycat positively because they believe it is the original product (Howard, Kerin & Gengler, 2000; Kapferer, 1995). Even when consumers are not confused, blatant copycats can sometimes be successful, as long as the imitator does not have the appearance to have aspirations to become a leading brand (Warlop & Alba, 2004). However, Van Horen & Pieters (2012a) found that when comparative evaluation between the leading brand and the copycat is stimulated, and thus confusion is ruled out, blatant copycats are evaluated less positive, while more subtle copycats are evaluated more positive. Results of another study by the same authors reveal that blatant feature imitation by PLs is perceived by consumers as unfair and unacceptable, and has a negative influence on product choice of the focal item (Van Horen & Pieters (2012b). Although the topic of visual brand similarity is extensively studied, the copycatting literature has mainly addressed blatant copycatting, often ignoring more subtle forms of visual similarity between brands. Further, there has been a strong focus on the influence of visual similarity on consumers’ attitudes towards the copycat brand (Howard, Kerin & Gengler, 2000; Kapferer, 1995; Warlop & Alba, 2004; Van Horen & Pieters, 2012a). The effects of copycatting on perceived quality and product choice of the copycat product have been understudied, especially for more subtle copycats. Van Horen & Pieters (2012b) measured the influence of visual similarity on product choice, yet only for blatant copycatting, and without taking into account perceived quality as a
  • 10. 10 mediator. Thus, little is known about the influence of different levels of visual similarity on consumers’ willingness to buy the PL copycat, and how the perceived quality gap between the PL copycat and the NB alternative mediates this relationship. Also, none of the previous studies on copycatting have taken into account the likely presence of low-end generic manufacturer brands, or C-brands (Cs), in consumers’ choice set. This study addresses the influence of different levels of visual similarity between PL and NB products on the perceived quality gap between these products, and on grocery shoppers’ willingness to buy the PL product. Also, the moderating effect of type of visual similarity (i.e. feature vs. theme similarity) on the visual similarity – perceived quality gap relationship will be addressed. Further, this study takes into account possible context effects within consumers’ choice sets as a possible explanation for product choice. It includes C products as choice options besides NB and PL products, to design realistic choice sets of three brand types, in which these context effects may occur. It could be expected that the package of a grocery product serves as a quality cue, in the sense that if visual similarity between PL products and NB products is high, the perceived quality gap between these products becomes small. Through the eyes of the grocery shopper, this means that the PL product dominates the NB alternative on price, while offering close to equal quality. This situation likely leads to an asymmetric dominance effect, which should have a positive influence on grocery shoppers’ willingness to buy the PL product (Simonson & Tversky, 1992). However, if visual similarity between the PL product and the NB alternative becomes very high, and the PL product becomes a blatant copy, it could be expected that the perceived quality gap between the two products becomes larger again, and the willingness to buy the PL product decreases. On the other hand, it could be predicted that if visual similarity between PL products and NB products decreases, the perceived quality gap between these products becomes larger, which implies a smaller perceived quality gap between the PL product and the C alternative, as PL shifts towards
  • 11. 11 C on the perceived quality continuum between C and NB. Therefore, the PL product will be more likely to be compared to C, which dominates PL on price. This should have a negative influence on grocery shoppers’ willingness to buy the PL product. It is expected that above predictions hold for feature similarity. However, it could be assumed that the influence of theme similarity on the perceived quality gap between NB and PL is positive for all levels of similarity, also for blatant similarity, as theme similarity is less obvious and more indirect than feature similarity. The main research question of this study is: what is the influence of the level and the type of visual similarity between PL products and their NB alternatives on grocery shoppers’ quality perceptions of and willingness to buy PL products in a context with an NB, a PL and a C choice option? In order to be able to answer the main question, the following sub-questions will be addressed:  How does the level of visual similarity between PL products and NB products influence the perceived quality gap between PL products and NB products and the perceived quality gap between PL products and C products?  What is the influence of type of visual similarity (i.e. feature similarity vs. theme similarity) on the influence of visual similarity between PL and NB products on the perceived quality gap between these products?  What is the influence of the perceived quality gap between PL products and NB products on grocery shoppers’ willingness to buy PL products, and what is the influence of the perceived quality gap between PL products and C products on grocery shoppers’ willingness to buy PL products?  Do the perceived quality gaps mediate the relationship between level of visual similarity and willingness to buy PL products?
  • 12. 12 The results of this study contribute to the copycat and private label versus national brand literature by providing insight in the role of different levels of visual brand similarity, instead of only blatant copycatting, on both quality perceptions of and willingness to buy PL products, instead of attitude towards the PL. This is in contrast to several copycat studies, which mainly addressed the influence of blatant copycatting on consumers’ attitudes towards the copycat product (Howard, Kerin & Gengler, 2000; Kapferer, 1995; Warlop & Alba, 2004; Van Horen & Pieters, 2012b). Moreover, the results may point to the existence of context effects within grocery shoppers’ choice sets for certain levels of visual similarity between PL products and their NB alternatives. This has not been done before in the copycatting research domain. Though this paper is somewhat related to a study by Van Horen & Pieters (2012b), who studied the influence of blatant copycatting on product attitude and choice, it differs in important respects. In contrast to Van Horen & Pieters’ (2012b) paper, this study not only addresses the influence of blatant copycatting, but also the effect of more subtle levels of visual brand similarity on willingness to buy the copycat. Further, this paper includes perceived quality as a possible mediator between visual similarity and product choice, which Van Horen & Pieters (2012b) did not. Also, these authors did not take into account the likely existence of C products and context effects within consumers’ choice sets. If visual similarity indeed serves as a quality cue, PL managers can use the results of this study to maximize consumers’ quality perceptions of their products by visually imitating the NB alternatives to an optimal extent. Subsequently, the selling prices, and therefore the profit margins for PL products can be optimized accordingly. The results are also useful for PL managers by providing insight in how to make PL products closer substitutes to NB products. The stronger the substitution power of PL products becomes, the less dependent on NB manufacturers grocery retailers become, and the stronger grocery retailers’ negotiation power with NB manufacturers will be. NB manufacturers can also benefit from the study results. The outcomes shed light on which
  • 13. 13 levels of visual imitation by PLs might be harmful to the sales of the focal NB products, and which similarity levels are beneficial to the sales of NB alternatives. Knowing this, NB managers will be able to decide when they have to change the visuals of their product packages, and when they have to maintain them. Lastly, the results may have legislative implications. For instance, if it appears that blatant copycatting by PLs indeed increases the perceived quality gap with the NB alternative and decreases consumers’ willingness to buy the PL product, NBs have no reason and no ground to sue blatantly copycatting PL brands.
  • 14. 14 Chapter 2 Conceptual framework 2.1 Relationship between NB manufacturers and grocery retailers NB manufacturers and grocery retailers maintain a difficult relationship. On the one hand, they are dependent on each other, while on the other hand, they are involved in a continuous battle for channel profit share. PL products are retailers’ main weapons. It is important for retailers to compete on quality with NBs, in order to secure high absolute profit margins on their PL products and to create a strong negotiation position with NB manufacturers by offering close substitutes to the NB products (Morton & Zettelmeyer, 2004; Ailawadi & Harlam, 2004). Several studies underscore the importance of competing on quality. Richardson, Dick & Jain (1994) found that perceived quality is stronger related to willingness to buy PL products than perceived value for money, although the effect of value for money is significantly positive as well. Hoch & Banerji (1993) came up with similar results. These findings suggest that consumers often look at quality first, and subsequently choose the cheapest option from a choice set of products with a similar perceived quality. NBs have more resources available for product innovation, package design, and advertising than retailers have. Hence, if retailers aim to position their PL products as high quality alternatives to NB products, this leaves them with few other options than trying to catch up by imitating visuals of NB products. Moreover, the more NBs keep innovating and differentiating, the harder it becomes for PLs to compete on quality with NBs directly (Steenkamp, Van Heerde & Geyskens, 2012). Product packages are one of the most important determinants of grocery shoppers’ product perceptions. Therefore, an appealing product package can justify a higher selling price. However, as soon as competing brands start copying packaging elements, the imitated leading brands often try to differentiate again by creating a new distinct packages. This interplay between leading and trailing brands implies an ongoing circle of package differentiation and imitation (Kapferer, 1995).
  • 15. 15 In some cases, retailers and NB manufacturers end up in a lawsuit, mostly when NBs sue PLs for overly blatant copycatting. To avoid prosecution, retailers tend to avoid producing perfect one-to- one copies. However, any less obvious forms of blatantly copycatting are still pursued by PLs (Planet Retail, 2007). 2.2 Product package as a quality cue 2.2.1 Influence of low to high visual similarity on perceived quality Perceived product quality can be defined as “the consumer’s judgment about a product’s overall excellence or superiority” (Zeithaml, 1988, p. 3). It differs from objective quality in that it is the quality of a product through the eyes of the consumer, rather than the quality according to predefined objective criteria. Steenkamp, Van Heerde & Geyskens (2012) found that the distinctiveness of the package of NB products has a positive influence on the perceived quality gap between NB products and the PL alternatives. They base their findings on perception theory. This theory implies that consumers’ product perceptions are dependent on consumers’ tendency to generalize from one product to another, and consumers’ proficiency to discriminate between products. If a PL product is visually similar to an NB product, product generalization is more likely to occur, whereby perceived quality is generalized from the NB product to the PL alternative. On the other hand, if visual similarity is low, product discrimination is more likely to occur, whereby a significant perceived quality gap between the NB and PL products remains. This theory implies that PLs can exploit positive associations consumers have with NB products by imitating visual elements of these products.
  • 16. 16 Another reason to assume that visual similarity between PL and NB product could have an effect on the quality gap between these products is that several studies suggest that multiple factors besides the actual physical product influences consumers’ quality perceptions. Such factors are called extrinsic cues. According to Teas and Agarwal (2000) and Richardson, Dick & Jain (1994), several extrinsic product cues, such as price, brand name, country name, but also packaging, positively affect perceived product quality. Moreover, Richardson, Dick & Jain (1994) found that extrinsic cues even have a stronger influence on perceived quality than intrinsic product characteristics, such as ingredients. Visual brand similarity could be viewed as an extrinsic cue that could affect quality perceptions, as it is a derivative of the extrinsic cue ‘packaging’. Visual similarity is often achieved through packaging, by imitating elements of the packaging of leading brand products. On the other hand, PL products are generally lower priced than NB products, and a lower product price is generally associated with lower quality. According to attribution theory, a relatively low price for a PL product might point at inferior product attributes, compared to the NB option (Sawyer & Dickson, 1984). However, if visual similarity between a PL product and the NB alternative also serves as a quality cue, the negative influence of price on consumers’ quality perceptions of the PL product may be offset to a large extent by a highly similar package. 2.2.2 Influence of blatant copycatting on perceived quality Whereas previously addressed literature supports the assumption that a higher visual similarity between PL products and their NB alternatives leads to a smaller perceived quality gap between these products, other studies suggest that this effect reverses beyond a certain tipping point. ‘Too high’ visual similarity, or blatant copycatting, tends to evoke negative attitudes toward the copycat product in situations where brand confusion is ruled out, and consumers are aware that the copycat
  • 17. 17 is not the leading brand product. When confusion is not ruled out, blatant copycats can be successful by triggering consumers to buy the copycat while they think it is the original (Howard, Kerin & Gengler, 2000; Kapferer, 1995). However, brand confusion is not likely to exist in a supermarket setting, where consumers compare NB, PL and C products simultaneously. In such a setting, a potential PL copycat is situated next to the original NB product, and comparative evaluation by consumers is highly likely to occur. According to Van Horen & Pieters (2012a), blatant copycats are evaluated less positive, and more subtle copycats are evaluated more positive when comparative evaluation is stimulated. Results from another study by the same authors show that PL products that are a blatant copy of the NB alternative in terms of features such as color, depicted objects, package shape, and sound of the brand name, are evaluated by grocery shoppers as unfair and unacceptable (Van Horen & Pieters, 2012b). Further, consumers generally perceive obvious visual similarity between a PL product and the NB original as a PL’s attempt to mislead them (Warlop & Alba, 2004). It could be assumed that the negative attitude of consumers towards blatant copycats transfers to their quality perceptions of blatant copycats. Consumers are likely to perceive the copy as ‘fake’ and ‘not original’. Zaichkowsky & Simpson (1996) support this notion by suggesting that blatant copycatting causes contrast effects, whereby consumers more critically evaluate the quality gap between the copy and the original. This implies that ‘too high’ visual similarity between PL and NB products leads to a larger perceived quality gap between these products. However, because several studies suggest that moderate to high visual similarity between PL products and NB alternatives, in contrast to blatant copycatting, leads to a smaller perceived quality gap between the NB and PL products, the following hypothesis is derived.
  • 18. 18 H1: The higher the visual similarity between a PL product and the NB alternative, the smaller the perceived quality gap between these two products becomes, until a certain tipping point, after which a further increase in visual similarity leads to a larger perceived quality gap again. This implies a U- shape (see figure 1) Fig. 1: Predicted influence of different levels of visual similarity between PL and NB on the perceived quality gap between them Although private labels are offering more and more products that can directly compete on quality with national brand products, NB, PL and C products can still be roughly classified along a straight line between the axes (perceived) quality and price. NB products offer high quality for a high price, PL products offer medium quality for a medium price, and C products offer acceptable quality for a low price (De Wulf et al., 2005). If these relationships between the three different types of brands still hold, it can be assumed that a larger perceived quality gap between an NB product and the PL alternative leads to a smaller perceived quality gap between the PL product and the C alternative, as
  • 19. 19 PL shifts towards C on the perceived quality continuum between C and NB. Also, a smaller perceived quality gap between NB and PL can be assumed to lead to a larger perceived quality gap between PL and C. Because the perceived quality gap between NB and PL is assumed to be influenced by the level of visual similarity between these products, the perceived quality gap between PL and C should be dependent on the level of visual similarity between PL and NB. Therefore, the predictions stated in H1 are used to come to the following hypothesis: H2: The higher the visual similarity between a PL product and the NB alternative, the larger the perceived quality gap between the PL product and the C alternative becomes, until a certain tipping point, after which a further increase in visual similarity between PL and NB leads to a smaller perceived quality gap between PL and C again. This implies an inverted U-shape (see figure 2). Fig. 2: Predicted influence of different levels of visual similarity between PL and NB on the perceived quality gap between PL and C
  • 20. 20 2.2.3 Influence of type of similarity on perceived quality Based on the theory described above, it is predicted that the influence of the level of visual similarity between a PL and NB product on the perceived quality gap between the PL product and its alternatives follows an (inverted) U-shape. However, this prediction pertains to visual similarity in general. The type of visual similarity has not been taken into account. There are two types of visual similarity between products of different brands: feature similarity and theme similarity (Van Horen & Pieters, 2012a). Following Van Horen & Pieters (2012a), feature similarity is defined as the similarity between two products in terms of package shape, size, color, or texture, or in terms of depicted objects, or the sound of the brand name. Theme similarity could be defined as the similarity between products in terms of the semantic meaning or inferred attribute of the products. This type of similarity involves the underlying theme or meaning of the products (e.g. the “traditional, family- produced olive oil” theme of the Bertolli brand). Findings by Van Horen & Pieters (2012b) show that grocery shoppers perceive blatant feature similarity as unacceptable and unfair, while they perceive blatant theme similarity as more acceptable and less unfair, even if they are aware of the theme copying. Thus, theme copycats are evaluated more positively than feature copycats. According to the authors, the underlying causes of these effects are related to the directness of consumers’ perceptual link between the copycat and the national brand product. Both feature and theme similarity work through similarity associations with the national brand, which causes feelings and knowledge related to the national brand to transfer to the copying brand. Feature similarity is a form of literal similarity. Therefore, the perceptual link between the feature copycat and the original product tends to be direct. This leads to a direct comparison between the two products, which negatively affects the evaluation of the blatant feature copycat. Theme similarity, however, is a more indirect type of similarity. The perceptual link between the theme copycat and the original product is more likely to be indirect, because theme
  • 21. 21 similarity is more implicit, less tangible, and more focused on affective associations. Also, underlying themes of a product are not only linked to the national brand product, but also to other products or situations. This leads to a more indirect comparison between the theme copycat and the imitated product, and therefore to a less negative evaluation of blatant theme copycats. It is predicted that the negative attitude of consumers toward blatant feature copycats has a negative influence on the quality perceptions of these products. As already mentioned, such obvious copying is likely to cause contrast effects between the copy and the original, which makes consumers perceive the copy as ‘fake’ and ‘not original’ (Zaichkowsky & Simpson, 1996). Such contrast effects are less likely to occur in case of theme similarity, because such similarity is more indirect. This leads to the hypothesis: H3: The type of visual similarity between PL and NB products moderates the relationship between the level of similarity between these products and the perceived quality gaps between PL and its alternatives. Specifically, the (inverted) U-shape does exist for feature similarity, but not for theme similarity, so that the higher the theme similarity between PL and NB, the smaller the perceived quality gap between NB and PL, and the larger the perceived quality gap between PL and C becomes. 2.3 Perceived quality gaps as influencers of willingness to buy PL 2.3.1 Influence of perceived quality gap between NB and PL on willingness to buy PL Van Heerde & Geyskens (2012) studied what makes grocery shoppers willing to pay a premium price for NB products over PL products. Their results suggest that consumers’ quality perceptions of NB products have a positive influence on their willingness to pay a premium price for these items.
  • 22. 22 Correspondingly, outcomes of a study by Sethuraman & Cole (1999) indicate that the perceived quality gap between NB products and PL alternatives is the most important driver of willingness to pay a premium price for NB products. The reverse also holds. Lower quality perceptions for NB products, and therefore a smaller perceived quality gap between NB and PL products lead to a lower willingness to pay a premium price for NB products. Because willingness to pay a premium price could be interpreted as a derivative of willingness to buy, it could be assumed that a smaller perceived quality gap between NB and PL products leads to a lower willingness to buy NB products, and therefore a higher willingness to buy PL products. Sinha & Batra (1999) provide additional support for this prediction. They found that consumers are more price conscious if they perceive the price of an NB product as unfair. This feeling of price unfairness arises when the perceived quality gap between the NB product and the PL alternative is smaller than the size of the quality gap as claimed by the NB. Subsequently, a higher level of price consciousness positively influences consumers’ PL purchases. The findings described above point to the possibility that the effect of the perceived quality gap between an NB product and the PL alternative on grocery shoppers’ willingness to buy the PL product is driven by context effects. A consumer’s willingness to buy a certain product is often affected by the context in which the product evaluation takes place, for instance by the characteristics and number of other products in a consumer’s choice set (Simonson & Tversky, 1992). As Simonson & Tversky (1992) state: “the same product may appear attractive on the background of less attractive alternatives and unattractive on the background of more attractive alternatives” (p. 281). This type of context effect is called tradeoff contrast. One form of tradeoff contrast is asymmetric dominance, whereby the existence of an inferior product in the consumer’s choice set increases the consumer’s willingness to buy the superior product. In case of a small perceived quality gap between NB and PL, through the eyes of the consumer the quality between
  • 23. 23 the two products is comparable, while the price of the PL product is significantly lower. The C product is only attractive to a minority of consumers who simply seek the lowest price and are willing to accept a significantly lower quality. NB and PL form the preferred region in consumers’ choice sets, because these products cluster together in terms of quality and represent a two against one ‘majority’ against the C option. As a consequence, the majority of consumers will choose one of these two options. In case of a small perceived quality gap between NB and PL, consumers perceive NB and PL as close to equal in quality, while PL dominates NB on price. Then, the PL product is the superior option and the NB alternative is the inferior option. This should increase consumers’ willingness to buy the PL product, compared to the situation where the perceived quality gap between NB and PL is large, and NB is not an inferior option. It should be mentioned that this is under the condition that price and quality are more or less equally important drivers of product choice. As already indicated above, Richardson, Dick & Jain (1994) provide evidence that this condition is fulfilled by showing that perceived quality is only a slightly more important driver of willingness to buy PL products than value for money. Furthermore, they found that “although PL prices are 21% lower, mean value for money perceptions are only 7% higher than those assigned to NBs. It appears that the poor perceived quality of PLs partially offsets the otherwise favorable reactions to their lower prices” (p. 33). This finding points to an interplay between price and quality in determining product choice. Price appears to have a larger influence on willingness to buy a PL product for higher levels of perceived quality of the PL product. The results by Richardson, Dick & Jain (1994) provide additional support for the existence of asymmetric dominance effects within consumers’ choice sets comprising an NB, PL and C product, by suggesting that consumers first look at quality, and subsequently choose the cheapest option from a choice set of products with a comparable level of quality. This would imply that consumers not just seeking the lowest price (C), will compare NB and PL on quality first, and if the perceived
  • 24. 24 quality gap is small (e.g. due to similar packaging), will choose the PL option, because it dominates the NB alternative on price, and price is an important driver of PL product choice when the perceived PL quality is high. Figure 3a illustrates the asymmetric dominance effect for a small perceived quality gap between NB and PL (which is predicted to exist when visual similarity between PL and NB is high). It is important to note that context effects are not directly tested in this study, but rather are used as an argumentation for the predicted effects of certain perceived quality gaps between PL products and their alternatives on grocery shoppers’ willingness to buy PL products. Fig. 3: Illustration of variants of asymmetric dominance for small and large perceived quality gap between NB and PL A review of the findings by previous studies on the role of perceived quality (gaps) on a consumer’s willingness to buy a product, leads to the following hypothesis:
  • 25. 25 H4: The smaller the perceived quality gap between an NB product and its PL alternative, the higher the willingness to buy the PL product becomes. 2.3.2 Influence of perceived quality gap between PL and C on willingness to buy PL It can be derived from the outcomes of a study by Garretson, Fisher & Burton (2002) that grocery shoppers with high levels of smart shopper self-perception and value consciousness are more likely to buy PL products than non-discounted NB products. In a similar vein, it could be expected that the same consumers are more likely to buy C products than PL products when the perceived quality gap between PL and C is small, because C products are generally cheaper than PL alternatives (De Wulf et al., 2005). It could be stated that the results by De Wulf et al. (2005) actually imply the existence of context effects. For a small perceived quality gap between PL and C, both products offer close to equal quality, while C dominates PL on price. Thus, asymmetric dominance effects are also likely to occur in this situation. Consumers seeking high quality regardless of the price are still likely to buy the NB option, but an increasing number of consumers will choose option C, as it forms a quality cluster with PL, while it has a significantly lower price. Figure 3b illustrates the assumed asymmetric dominance effect for a large perceived quality gap between NB and PL, and thus a small perceived quality gap between PL and C (which is predicted to exist when visual similarity between PL and NB is non-existing or ‘too’ high). The following hypothesis can be derived: H5: The smaller the perceived quality gap between a PL product and its C alternative, the lower the willingness to buy the PL product becomes.
  • 26. 26 Because it is predicted that the level of visual similarity between PL and NB influences the perceived quality gaps between product options, and because the perceived quality gaps influence consumers’ willingness to buy PL, it could be assumed that the perceived quality gaps are mediators of the relationship between the level of visual similarity between PL and NB and grocery shoppers’ willingness to buy PL. H6: The perceived quality gaps between the product options serve as mediators for the relationship between the level of similarity between a PL product and its NB alternative and grocery shoppers’ willingness to buy the PL product. The conceptual model illustrated in figure 4 shows the relationships that will be tested in this study. Fig. 4: Conceptual model
  • 27. 27 Chapter 3 Methodology 3.1 Type of research Primary data from an experimental survey will be used to test the hypotheses. It was necessary to conduct an experiment, in order to be able to measure the causal influence of different levels of visual similarity between PL and NB products on grocery shoppers’ quality perceptions of different product options, and their willingness to buy PL products. Secondary field data from supermarkets, for instance transaction data, could only show purchase behavior of grocery shoppers, but cannot provide insight in the causal influence of visual brand similarity and consumers’ quality perceptions on this behavior. This is because secondary data is collected from a non-controlled setting, without stimuli manipulation, so that it cannot be derived that X occured before Y, and without controlling for extraneous variables, so that no alternative explanations can be ruled out. Thus, if certain secondary data would somehow have covered information on visual similarity and quality perceptions, a lack of controlled conditions would make the data useless for showing cause-effect relationships. An experimental survey, rather than a field experiment, will be conducted, because some products will have to be manipulated in order to realistically represent PL products. This manipulation has to be done digitally, by altering product images. Physically manipulating products will be unfeasible. 3.2 Experimental design 3.2.1 General description of experiment Participants of the experiment will be asked to virtually buy items from a shopping list consisting of eight items from several product categories, as if they actually are in a physical supermarket. The participants will be shown images of three options for each of the eight shopping list items; one NB
  • 28. 28 option, one PL option, and one C option. All PL options will be presented as products from the Perfekt brand, a PL brand that is sold at several supermarket chains throughout The Netherlands. By using only one PL brand, the experimental results cannot be influenced by participants’ potential different preferences for different PL brands. The three product options will be shown simultaneously, instead of one after the other, in order to simulate a setting more similar to an actual supermarket, and to be able to account for context effects. Also, the product options will be shown at random places relative to the other two product options (e.g. not always in a C-PL-NB order), in order not to suggest any quality sequences. The prices of the different items will be given. Four of the eight shopping list items will include a PL option that has a certain level of feature similarity with the NB option. Each of the four items will represent one of the four levels of theme similarity; no feature similarity, low feature similarity, high feature similarity, and blatant (very high) feature similarity. The other four of the eight items from the shopping list will include a PL option that has a certain level of theme similarity with the NB option, ranging from no theme similarity to blatant theme similarity. The participants will have to choose (‘buy’) one of the three options (NB, PL or C) for each item from the shopping list. Each participant will be shown the same items and product options, so the experiment has a 2 (feature vs. theme similarity) × 4 (no vs. low vs. high vs. blatant similarity) within-subject design. The main reason for choosing a within-subject design is that it most realistically represents a real- life grocery shopping setting, where some products options look like each other, and others do not. If some participants will only see highly similar NB and PL options, and some will only be shown non-similar options, this could influence their quality perceptions of the PL products. Items from multiple product categories were chosen as stimuli, because a shopping list design realistically reflects real-life grocery shopping, and shopping lists normally consists of a wide variety of products. Further, it was unfeasible to find or create PL and NB products from one single product
  • 29. 29 category with the required four different levels of visual similarity. The effect of the chosen PL product categories will be controlled for by asking respondents to what extent they have a tendency to buy PL products instead of NB or C products for each chosen product category. This control variable will be discussed in more detail in paragraph 3.2.3 (‘Control variables’). After the product choice tasks, the participants will be shown the same items from the shopping list, but now will be asked to indicate how they perceive the quality of the different product options. As with the product choice part of the experimental survey, the NB, PL and C options per item will be shown simultaneously. The results from the perceived quality questions will be used to calculate the perceived quality gap between PL products and their alternatives. It is important to ask these questions after the product choice tasks, so that the product choices are not influenced by the stimulation of participants’ consciousness of the quality of the items. The other way around, participants might justify their previous product choices by assigning a higher quality than they actually expect to products they have chosen during the choice tasks. However, this should not be a major concern, because participants will choose products based on both quality perception and price, so that the effect of product choice justification on participants’ quality perceptions might be partially abolished by the influence of price. The experimental survey, including the product images, can be found in Appendix 2. The example in the appendix is in English, but the actual experimental survey was in Dutch. 3.2.2 Operationalization of concepts Level and type of visual similarity between PL and NB product The level of visual similarity between PL and NB products was determined by a pre-test among seventeen respondents. One of the product sets was pre-tested among eleven respondents, after changing the initial set after the first pre-test. The researcher collected several sets of PL and NB
  • 30. 30 products with seemingly different extents of visual similarity. Some products were from low-end supermarkets Aldi and Lidl, and were manipulated by the researcher, for instance by digitally adding a PL logo (i.e. a Perfekt logo), so that they appeared to be PL products from a high-end supermarket. Other products were actual PL products sold by high-end supermarkets, and did not have to be manipulated. All NB products were actual products from high-end supermarkets, and were not manipulated. In the pre-test, for each of the collected product sets, respondents were asked to indicate the level of visual similarity (‘look and feel’) between the PL and the NB products on a four-item Likert scale covering not similar, a little similar, highly similar, and blatantly similar. The four items were assigned a number, from 0 to 3, so that mean scores could be computed. The measure scale was created by the researcher, as no appropriate visual similarity scales were found in existing (marketing) literature. The results from the pre-test were used to determine the actual level of visual similarity between the collected PL and NB products, namely no visual similarity, low visual similarity, high visual similarity, or blatant (very high) visual similarity. The level of visual similarity was regardless of the type of similarity (feature versus theme similarity). Product sets with mean similarity scores close to 0, 1, 2, 3, representing the four different levels of visual similarity, were suitable to be used within the experiment. The pre-test determined the level of visual similarity, but not the type of visual similarity between the PL and NB products. The researcher classified the similarity between the different products as either feature-based or theme-based, using the definition of the two types of similarity. PL and NB products that looked mainly similar in terms of package shape, colors, size, texture, or in terms of depicted objects, or the sound of the product name, were classified as having feature similarity. PL and NB products that looked mainly similar in terms of the semantic meaning or inferred attribute, or the underlying theme or meaning, were labelled as having theme similarity
  • 31. 31 (Van Horen & Pieters, 2012a). Each case of feature similarity will be assigned a 0, while each case of theme similarity will be assigned a 1. Appendix 1 shows the results from the pre-test, and the resulting visual similarity conditions, including a description per product set of what elements likely caused the perceived level of visual similarity. The four levels of visual similarity and the two types of visual similarity form a total of eight experimental conditions. It should be noted that the same stimulus will be used for both the ‘no feature similarity’ and ‘no theme similarity’ conditions, so that there are actually seven conditions. Perceived quality gaps between NB and PL products and between PL and C products A one-item 7-point Likert scale will be used to measure the perceived quality of each shown NB, PL and C product. The following statement will be depicted: “All things considered I would say this (product name) has ...” (Richardson, Dick & Jain’s, 1994, p. 31). The participants can provide an answer between the end points ‘poor overall quality’ (coded as 0) and ‘excellent overall quality’ (coded as 6). Richardson, Dick & Jain’s (1994) used this measure scale in a grocery store setting. It is a reflective scale, as the indicator is a manifestation of a consumer’s quality perceptions, rather than a defining characteristic of it (MacKenzie, 2003). The perceived quality gaps between NB and PL products and between PL and C products will be calculated by subtracting the perceived quality scores for the PL products from those for their NB alternatives (NB – PL), and the perceived quality scores for the C products from those for their PL alternatives (PL – C), respectively. The perceived quality gaps are interval variables. Willingness to buy PL product This is the dependent variable of the conceptual model. As part of the experimental survey, participants will have to choose one of three options for each grocery shopping list item, or,
  • 32. 32 condition; the NB option, the PL option, or the C option. Consumers’ willingness to buy the PL product will be operationalized by seperating participants’ PL product option choices from NB or C product option choices. Each case of NB or C choice will be assigned a 0, whereas each case of PL choice will be assigned a 1. This makes ‘willingness to buy PL product’ a dichotomous variable, where 0 represents ‘no PL choice’ and 1 represents ‘PL choice’. It should be noted that the participants will not actually buy products, but will only indicate which products they would buy in real-life. Thus, despite the product choice tasks in the experiment, the dependent variable is ‘willingness to buy the PL product’, and not ‘PL product purchase’. The value of the dependent variable directly reflects consumers’ willingness to buy the PL product. It is a ratio variable, as it has a meaningful zero (zero participants chose the PL option), and multiplication and division between two values is meaningful. 3.2.3 Control variables ‘Relative price difference between PL and NB product’, ‘relative price difference between C and PL product’, and ‘tendency to buy PL in product category’, are the most important variables that will be controlled for. Based on actual prices of collected NB, PL and C products, the average relative price differences between PL products and their NB and C alternatives will be calculated. For the non- manipulated products, their actual prices will be shown. For the manipulated product (products that were collected at low-end supermarkets, and were transformed into PL stimuli), fictional prices were defined, based on the actual prices of their C and NB alternatives, and the calculated average relative price differences. ‘Relative price difference between PL and NB product’ will be calculated as follows: (price PL – price NB) / price NB. The calculation of ‘relative price difference between C and PL product’ is: (price C – price PL) / price PL. It is important to control for the influence of
  • 33. 33 price difference between PL and NB or C stimuli, because by controlling for this variable, it could be determined whether participants’ product choices were actually due to a certain level of visual similarity between PL and NB stimuli, or (partly) due to above or below average price differences between product options. The outcomes of the experimental survey could be influenced by participants’ grocery shopping behavior in real-life. Participants might be used to buying PL products for certain product categories, while having a tendency to buy mainly NB products for other product categories, for whatever reasons. These tendencies should be controlled for, because they may influence quality perceptions of PL stimuli products and willingness to buy PL products for certain conditions, regardless of the level of visual similarity between PL and NB products. The stimuli products can be classified into different product categories. Participants’ ‘tendency to buy PL products per product category’ will be addressed by letting participants choose an answer between ‘never’ and ‘always’ (5-item Likert scale) for the following statement: “When I buy [… product category …] in a supermarket, I choose private label brands over other brands.” ‘Income’ is another variable that will be controlled for, because it could be assumed that consumers with high incomes are more likely to buy NB products, regardless of the prices and visuals of PL alternatives. Further, the influence of age and gender will be taken into account. 3.3 Data analysis Because the dependent variable (‘willingness to buy PL product’) is dichotomous –the number of available categories is two; yes or no– logistic regression analyses will be conducted to measure the influence of both the level of visual similarity between PL and NB products, and perceived quality gaps between PL products and their alternatives on consumers’ willingness to buy PL products.
  • 34. 34 Linear regression analyses will be conducted to test the relationships between these variables. (Inverted) U-shaped relationships between the level of visual similarity and perceived quality gaps between products from certain brand types are predicted. Therefore, squared variables will be added to the model to be able to test these relationships. A moderation analysis will be conducted to analyse the influence of the type of visual similarity on the relationship between the level of visual similarity between PL and NB and the perceived quality gap between NB and PL / PL and C. Since standard mediation analysis cannot be conducted when the dependent variable is dichotomous, the relevant output from the linear and logistic regression analyses will be inputted into a ‘mediation with dichotomous outcome variables’ spreadsheet, created by Nathaniel R. Herr (2006)1 . The relevant output from the spreadsheet calculations will be used as input for Sobel tests.2 The significance of the Z-scores derived from the Sobel tests will determine whether mediation effects exist. 3.4 Participants The participants of the experimental survey were Dutch grocery shoppers. Only Dutch people were eligible to participate, because the stimuli included images of products from Dutch supermarkets, from Dutch PL brands, and with labels with Dutch texts. Foreign consumers were likely not to know several of the shown products or brands. To ensure the generalizability of the results, grocery shoppers from different areas, with different incomes, and from different age groups were recruited. In total, 178 people participated in the experimental survey. They were between 19 and 70 years old, with an average of 33 years. 64% of the participants were female, and 36% were male. Only around 10% percent of the participants had a household income that is far above or far below 1 http://www.nrhpsych.com/mediation/logmed.html 2 http://quantpsy.org/sobel/sobel.htm
  • 35. 35 average, while around 90% had non-extreme household incomes. 50% of the participants indicated that they have an average income. A complete overview of the descriptives of the participants can be found in Appendix 3.
  • 36. 36 Chapter 4 Results 4.1 Descriptives 4.1.1 Willingness to buy PL product In total, the 178 participants of the experimental survey were submitted to seven different conditions: one ‘no visual similarity’ condition, two ‘low similarity’ conditions (low feature similarity, and low theme similarity), two ‘high similarity’ conditions (high feature similarity, and high theme similarity), and two ‘blatant similarity’ conditions (blatant feature similarity, and blatant theme similarity). When only taking into account the level of visual similarity between a PL product and its NB alternative, regardless of the type of visual similarity and regardless of participants’ tendency to buy PL products in a specific product category, it can be derived that the relative willingness to buy the PL option among the participants was highest when visual similarity between the PL product and the NB alternative was low (42%), and lowest for blatant visual similarity (22%). Also, willingness to buy the PL product is higher for the two middle levels of visual similarity (low – 42% and high – 28%) than for the extreme levels of similarity (no – 23% and blatant – 22%). Table 1 provides an overview of participants’ willingness to buy the PL option per level of similarity. The complete SPSS descriptives output can be found in Appendix 4. Table 1: Participants’ willingness to buy PL per level of visual similarity between PL and NB product Willingness to buy PL No Yes Total Count % Count % N % Level of visual similarity between PL and NB No visual similarity 137 77,0% 41 23,0% 178 100,0% Low visual similarity 206 57,9% 150 42,1% 356 100,0% High visual similarity 256 71,9% 100 28,1% 356 100,0% Blatant visual similarity 277 77,8% 79 22,2% 356 100,0%
  • 37. 37 The results in table 1 are illustrated in figure 5. Although these descriptives do not prove any relationship between level of visual similarity between PL products and their NB alternatives and willingness to buy PL products, they suggest that an inverted U-shape relation between the two variables exists. Fig. 5: Participants’ willingness to buy PL product per level of visual similarity between PL and NB product Similar results were derived when taking into account the type of visual similarity between PL and NB products. The results in table 2, which are illustrated in figure 6, show that participants’ willingness to buy PL products across different levels of visual similarity do not deviate much between feature and theme similarity, and that the percentage distribution of participants’ ‘willingness to buy PL product’ is similar across the two types of visual similarity. For both types of visual similarity, and similar to the general results, participants’ willingness to buy the PL product was highest for low visual similarity and lowest for blatant visual similarity. Further, participants’ willingness to buy the PL product was higher for the two middle levels of visual similarity (low and high) than for the extreme levels of similarity (no and blatant), regardless of type of visual similarity.
  • 38. 38 Table 2: Participants’ willingness to buy PL product per level and type of visual similarity between PL and NB product Willingness to buy PL No Yes Total Count % Count % N % Level of visual similarity between PL and NB No visual similarity 137 77,0% 41 23,0% 178 100,0% Low visual similarity Type of visual similarity Feature 88 49,4% 90 50,6% 178 100,0% Theme 118 66,3% 60 33,7% 178 100,0% High visual similarity Type of visual similarity Feature 124 69,7% 54 30,3% 178 100,0% Theme 132 74,2% 46 25,8% 178 100,0% Blatant visual similarity Type of visual similarity Feature 137 77,0% 41 23,0% 178 100,0% Theme 140 78,7% 38 21,3% 178 100,0% Although relationships between variables are not properly tested yet, these descriptives provide a first indiction that type of visual similarity does not play a role in determining the influence of level of visual similarity between PL product and their NB alternatives on grocery shoppers’ willingness to buy PL products. Fig. 6: Willingness to buy PL product among participants per level of visual similarity between PL and NB product
  • 39. 39 4.1.2 Perceived quality gaps between product options The results shown in table 3 and illustrated by figure 7 indicate that the mean quality gap between NB and PL products as perceived by the participants is smallest in the ‘no visual similarity’ condition (0.21) and largest for blatant visual similarity (1.49). Further, this gap is larger for low similarity than for high similarity (1.10 vs. 0.95). These results point at a linear, instead of a U- shaped, relationship between level of visual similarity and the perceived quality gap between NB and PL. Table 3: Perceived quality gaps between product options among participants per level of visual similarity between PL and NB product Perceived quality gap between NB and PL (NB – PL) Perceived quality gap between PL and C (PL – C) Mean Stnd. Dev. N Mean Stnd. Dev. N Level of visual similarity between PL and NB No visual similarity ,21 1,24 178 -,17 1,24 178 Low visual similarity 1,10 1,24 356 1,12 1,20 356 High visual similarity ,95 1,19 356 ,65 1,53 356 Blatant visual similarity 1,49 1,35 356 ,64 1,54 356 The results in table 3 and the graph in figure 7 show that an inverted U-shaped relationship between level of visual similarity and the perceived quality gap between PL and C products might exist. The mean gap is smallest (even negative) for ‘no visual similarity’ (-0.17), highest for low visual similarity (1.12), and reduces again for high (0.65) and blatant (0.64) visual similarity.
  • 40. 40 Table 4: Perceived quality gaps between product options among participants per level and type of visual similarity between PL and NB product Perceived quality gap between NB and PL (NB – PL) Perceived quality gap between PL and C (PL – C) Mean Stnd. Dev. N Mean Stnd. Dev. N Level of visual similarity between PL and NB No visual similarity ,21 1,24 178 -,17 1,24 178 Low visual similarity Type of visual similarity Feature 1,00 1,26 178 1,24 1,20 178 Theme 1,19 1,21 178 1,01 1,19 178 High visual similarity Type of visual similarity Feature ,96 1,17 178 ,55 1,60 178 Theme ,94 1,22 178 ,76 1,45 178 Blatant visual similarity Type of visual similarity Feature 1,32 1,38 178 ,36 1,55 178 Theme 1,65 1,30 178 ,94 1,47 178 The quality gaps between the products in the different experimental conditions as perceived by the participants are similar across the two types of visual similarity, as shown in figures 8 and 9. Again, the results for each type of visual similarity are comparable to the general results. The only exception is the perceived quality gap between PL and C products for blatant visual similarity, which the participants perceived on average as much larger for theme similarity (0.94) than for feature similarity (0.36).
  • 41. 41 Fig. 7: Perceived quality gaps between product options among participants per level of visual similarity between PL and NB product Fig. 8: Perceived quality gap between NB and PL product among participants per level and type of visual similarity between PL and NB product
  • 42. 42 Fig. 9: Perceived quality gap between PL and C product among participants per level and type of visual similarity between PL and NB product 4.2 Testing conceptual model 4.2.1 Perceived quality gap between NB and PL product Linear regression analyses were conducted to test the relationships between the independent variables ‘level of visual similarity between PL and NB product’ and ‘type of visual similarity between PL and NB product’, and the conceptual mediating variables, ‘perceived quality gap between NB and PL product’ and ‘perceived quality gap between PL and C product’, controlling for ‘tendency to buy PL in product category’. The outcomes of the linear regression analysis point at a positive relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between NB and PL product’ (NB – PL) (β = .660, p < .01). Further, the influence of the squared variable ‘level of visual similarity between PL and NB product ²’ on this
  • 43. 43 perceived quality gap turns out to be non-significant (β = -.069, p = 0.105). These results do not support hypothesis 1. It was expected that the higher the visual similarity between a PL product and the NB alternative becomes, the smaller the perceived quality gap between these two products becomes, until a certain tipping point, after which a further increase in visual similarity leads to a larger perceived quality gap again. However, the results do not support a U-shaped relationship between the level of visual similarity and the perceived quality gap between NB and PL. Figure 10 shows the linearity of the relationship. A U-shaped relationship neither was found for specific types of visual similarity (i.e. feature or theme similarity) (β = -.019, p = .595), in contrast to hypothesis 3, which predicted that a U-shape would exist in case of feature similarity, but not in case of theme similarity. Table 5 provides an overview of the linear regression analysis outcomes. The complete output of all analyses can be found in Appendix 5. Table 5: Results linear regression analysis, DV = perceived quality gap between NB and PL product Variables β Std. Error Sig. (Constant) .562 .294 .056 Level of visual similarity .660 .156 .000** Type of visual similarity .350 .242 .148 Level of visual similarity ² -.069 .042 .105 Level of visual similarity ² × Type of visual similarity Tendency to buy PL in product category -.019 -.198 .036 .036 .595 .000** N=178; * p < .05; ** p < .01
  • 44. 44 Due to multicollinearity issues, SPSS excluded the variable ‘level of visual similarity × type of visual similarity’ from the analysis. When conducting the analysis without the squared variables included, ‘level of visual similarity × type of visual similarity’ was not excluded, and its influence on the dependent variable could be addressed. ‘Level of visual similarity × type of visual similarity’ turns out not to significantly influence the perceived quality gap between NB and PL products (p = .593) (see Appendix 5). This outcome implies that the type of visual similarity between PL and NB products does not moderate the relationship between the level of visual similarity and the perceived quality gap between NB and PL products. Fig. 10: Relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between NB and PL product’ 4.2.2 Perceived quality gap between PL and C product The results of another linear regression analysis, shown in table 6, indicate that the relationship between the level of visual similarity between PL and NB product options and the perceived quality gap between the PL and C options (PL – C) is positive overall (β = 1.077, p < .01). However, the influence of the squared variable ‘level of visual similarity ²’ on this quality gap appears to be significant as well (β = -.256, p < .01), pointing at an inverted U-shaped relationship between the
  • 45. 45 level of visual similarity between a PL product and its NB alternative and the perceived quality gap between the PL products and its C alternative. These results support hypothesis 2. A plot of this relationship is illustrated in figure 11. However, this inverted U-shaped relationship appears to exist irrespective of the type of visual similarity (p = .970), so no support for H3 was found. Table 6: Results linear regression analysis, DV = perceived quality gap between PL and C Variables β Std. Error Sig. (Constant) -.297 .337 .379 Level of visual similarity 1.077 .179 .000** Type of visual similarity .475 .277 .087 Level of visual similarity ² -.256 .048 .000** Level of visual similarity ² × Type of visual similarity Tendency to buy PL in product category -.002 -.086 .041 .041 .970 .035* N=178; * p < .05; ** p < .01 Fig. 11: Relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between PL and C product’
  • 46. 46 Again, due to issues with multicollinearity, the variable ‘level of visual similarity × type of visual similarity’ was excluded from the initial analysis. Excluding the squared variables from the model, and including ‘level of visual similarity × type of visual similarity’, its influence on the DV could be addressed in an additional linear regression analysis. ‘Level of visual similarity × type of visual similarity’ appears to significantly influence the perceived quality gap between PL and C products (β = .506, p < .01) (see Appendix 5). This outcome would imply that the type of visual similarity between PL and NB products moderates the relationship between the level of visual similarity between PL and NB products and the perceived quality gap between PL and C products. However, this effect is overruled by the non-significance of the influence of ‘level of visual similarity ² × type of visual similarity’ (p = .970). 4.2.3 Willingness to buy PL product Logistic regression analyses were conducted to determine the influence of the independent and conceptual mediating variables on the dependent variable, ‘willingness to buy PL product’. Results of the logistic regression analysis (see table 7) show that the perceived quality gap between NB and PL product options has a negative effect on consumers’ willingness to buy the PL product (β = .653, p < .01). This outcome supports hypothesis 4, which predicted that the smaller the perceived quality gap between an NB product and the PL alternative, the higher the willingness to buy the PL product becomes. Further, a positive relationship was found between ‘perceived quality gap between PL and C product’ and ‘willingness to buy PL product’, in support of hypothesis 5. This hypothesis predicted that a smaller perceived quality gap between a PL product and the C alternative leads to a lower willingness to buy the PL product.
  • 47. 47 Table 7: Results logistic regression analysis, DV = willingness to buy PL product Reported as ‘β (Std. Error)’ Variables Block 1 Block 2 (Constant) -1.642 (1.001) -1.984 (1.116) Level of visual similarity 1.533 (1.117) .815 (1.252) Type of visual similarity 1.245 (1.393) .737 (1.552) Level of visual similarity × Type of visual similarity Level of visual similarity ² Tendency to buy PL in product category Relative price difference between PL and NB Relative price difference between C and PL Perceived quality gap between NB and PL Perceived quality gap between PL and C -.505 (.528) -.269 (.123)* .115 (.068) .058 (.067) -.701 (.803) -.406 (.587) -.143 (.137) .159 (.078)* .032 (.075) -.391 (.896) -.325 (.081)** .653 (.070)** N=178; * p < .05; ** p < .01 The control variables ‘age’ (p = .127), ‘gender’ (p = .084), ‘household income’ (p = .960), and ‘relative price difference between PL and NB’ (p = .663) turned out not to have a significant influence on grocery shoppers’ willingness to buy PL products, and were therefore excluded. ‘Tendency to buy PL in product category’ only has a significant effect when the perceived quality gap variables are added to the model (β = .159, p < .05) ‘Level of visual similarity ² × type of visual similarity’ and ‘relative price difference between C and PL’ were excluded from the logistic analysis by SPSS, but separate analyses including these variables show that both do not have a significant influence on ‘willingness to buy PL’ (p = .328 and p = .663, respectively) (see Appendix 5).
  • 48. 48 4.2.4 Mediating role of perceived quality gaps Because the dependent variable ‘willingness to buy PL product’ is dichotomous, standard mediation analysis could not be conducted. Therefore, an alternative mediation analysis was conducted. The relevant output from the linear and logistic regression analyses was inputted into a ‘mediation with dichotomous outcome variables’ spreadsheet, created by Nathaniel R. Herr (2006). The relevant output from the spreadsheet calculations was used as input for two Sobel tests. The results from the Sobel tests indicate that both ‘perceived quality gap between NB and PL product’ (Z = -2.911, p < .01) and ‘perceived quality gap between PL and C product’ (Z = 5.056, p < .01) mediate the relationship between the level of visual similarity between a PL product and its NB alternative and grocery shoppers’ willingness to buy the PL option. These outcomes support hypothesis 6. The results from the logistic regression analyses, shown in table 7, show that ‘level of visual similarity between PL and NB product’ has a significant direct effect on ‘willingness to buy PL product’, but that this relationship becomes non-significant when the perceived quality gap variables are added to the model. Combining these results with the outcomes of the Sobel tests, it can be concluded that the effect of the level of visual similarity between a PL product and its NB alternative on consumers’ willingness to buy the PL product is indirect, as it works through the perceived quality gaps between the different product options, and full mediation exists. The output from both the spreadsheet and the Sobel tests can be found in Appendix 6.
  • 49. 49 Chapter 5 Discussion 5.1 Discussion of results This study addresses the influence of the level and the type of visual similarity between PL products and their NB alternatives on grocery shoppers’ quality perceptions of and willingness to buy PL products in a context with an NB, a PL and a C choice option. Controlling for consumers’ tendency to buy PL products in specific product categories, it is found that the higher the level of visual similarity between a PL product and its NB alternative becomes, the larger grocery shoppers perceive the quality gap between these products, in favor of the NB option. The linearity of this relationship implies that even subtle forms of copycatting by PLs can have a negative effect on how consumers perceive the quality of the imitating PL product compared to the imitated NB product. Latter outcome is the opposite to findings by Steenkamp, Van Heerde & Geyskens (2012), who found that the distinctiveness of an NB product package has a positive influence on the perceived quality gap between the NB product and its PL alternative, and therefore a higher level of visual similarity between PL and NB leads to a smaller quality gap. They use perception theory as an explanation for their findings by stating that consumers’ product perceptions are dependent on their tendency to generalize from one product to another, and their ability to discriminate between products. They suggest that consumers’ tendency to generalize from one product to another is stronger for higher levels of visual similarity between the products, and that their ability to discriminate between products is lower for higher levels of visual similarity. However, the same theory might explain why the opposite of Van Heerde & Geyskens’ (2012) findings appears to be true. It could be the case that the higher the distinctiveness of a PL product package, or the lower the level of visual similarity between the PL product and its NB alternative, the more difficult it becomes for grocery shoppers to compare the two product options on quality,
  • 50. 50 the smaller the perceived quality gap between NB and PL becomes. Visual similarity between PL and NB, even when subtle, makes each of the two products less differentiated, which makes comparison between the products easier. This might stimulate consumers to compare the quality of the two products more closely, which could lead to the perception that the NB option is ‘real’ and the PL option is ‘fake’, increasing the perceived quality gap between the products. The finding that blatant visual similarity between a PL product and its NB alternative leads to a larger perceived quality gap between the products than moderate to high visual similarity is in line with this newly proposed idea, and also complies to findings by Van Horen & Pieters (2012a), who indicate that subtle copycats are evaluated more positive than blatant copycats, as long as the copycat and the original product are evaluated simultaneously. The outcome also reinforces results from other studies, which show that consumers possess a negative attitude towards blatant copycats (Van Horen & Pieters, 2012b), and that this negative attitude transfers to their quality perceptions of the product compared to the original (Zaichkowsky & Simpson, 1996). The results of the analyses suggest that the influence of the level of visual similarity between a PL product and its NB alternative on the perceived quality gap between the PL product and its C alternative is less straightforward. It appears that this relationship is inverted U-shaped, instead of linear. This implies that the higher the level of visual similarity between the PL and NB product option, the larger the perceived quality gap between the PL product and the C alternative becomes, until a certain tipping point, after which a further increase in visual similarity between PL and NB leads to a smaller perceived quality gap between PL and C again. These findings imply that PLs can improve the perceived quality of their products compared to their C alternatives by visually imitating NB products, as long as the similarity is more or less subtle, and not too obvious.
  • 51. 51 It is remarkable that the relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between NB and PL product’ turns out to be linear, while the relationship between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between PL and C product’ turns out to be non-linear. It is assumed that, on average, NB, PL and C products can be roughly classified along a straight line between the axes perceived quality and price, with NB products offering high quality for a high price, PL products offering medium quality for a medium price, and C products offering acceptable quality for a low price (De Wulf et al., 2005). As a result, while drafting the hypotheses of this study, it was assumed that a larger perceived quality gap between an NB product and its PL alternative automatically leads to a smaller perceived quality gap between the PL product and the C alternative, as PL shifts towards C on the perceived quality continuum between C and NB. Also, a smaller perceived quality gap between NB and PL was assumed to lead to a larger perceived quality gap between PL and C. Because the perceived quality gap between NB and PL was correctly assumed to be influenced by the level of visual similarity between these products, the perceived quality gap between PL and C should be dependent on the level of visual similarity between PL and NB. Following this reasoning, the shapes of the relationships between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between NB and PL product’, and between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between PL and C product’ should be symmetrical to each other (i.e. being positive vs. negative, or being U-shaped vs. inverted U-shaped). However, they are not, as they are positive versus inverted U-shaped. While still assuming that, on average, NB, PL and C products can be roughly classified along a straight line between the axes perceived quality and price, this finding might imply that the perceived quality continuum between C, PL and NB is more dynamic than postulated earlier. It might be the case that not only PL moves along the perceived quality continuum when the level of visual similarity between PL and NB changes, but
  • 52. 52 also the perceived quality of C and NB are affected. This might explain why, in contrast to the predictions, the perceived quality gap between NB and PL, and the pereived quality gap between PL and C both become larger when the visual similarity between PL and NB changes from low to moderately high. For instance, it could be that in that particular situation NB moves to the right (higher perceived quality), and C moves to the left (lower perceived quality), while PL moves only slightly to the left. In that case, the reason why C shifts to the left might be that PL gains the appearance of an A product when advancing its packaging while subtly imitating an NB brand, so that the lower quality of C, through the eyes of consumers, becomes more obvious. The influence of the level of visual similarity between PL and NB products on the perceived quality gaps between the different product options in the choice set are not affected by the type of visual similarity, either feature similarity or theme similarity. Thus, only the level of visual similarity influences these quality gaps, regardless of whether a particular level of similarity is achieved by imitating the shape, size or color of the package, or by imitating the inferred attribute and underlying theme of the NB product. This outcome is not consistent with the results of a study by Van Horen & Pieters (2012b). These authors state that consumers evaluate theme copycats more positively than feature copycats, because theme similarity is a more indirect form of similarity compared to feature similarity, and therefore leads to a more indirect perceptual link between the copycat and the NB product, so that theme similarity is perceived as less unfair. Although this study uses different outcome variables than Van Horen & Pieters (2012b) did (perceived quality gaps vs. general evaluation) to test the influence of the type of visual similarity between a copycat and its NB alternative, the contradicting findings of the two studies are unexpected. Namely, in this study the same ‘theme similarity’ stimuli as in Van Horen & Pieter’s (2012b) study were used for two of the four ‘theme similarity’ experimental conditions: ‘high theme similarity’ and ‘blatant theme similarity’. It might be that the influence of the type of visual similarity between two products is
  • 53. 53 highly dependent on the number or characteristics of alternative product options in the consumer’s choice set. In this study, a choice set of three product options is used, including an NB, a PL, and a C product. In Van Horen & Pieter’s (2012b) experiment, the manipulated choice set consists of two product options: an NB product and a non-PL copycat product. The findings of this study show that the smaller the perceived quality gap between an NB product and the PL alternative, the higher the willingness to buy the PL product becomes. Also, the smaller the perceived quality gap between a PL product and the C alternative, the lower the willingness to buy the PL product becomes. These results are in line with the hypotheses, and provide tentative support for the assumption that context effects, and in particular asymmetric dominance effects (Simonson & Tversky, 1992), play a role in grocery shoppers’ willingness to buy PL products. When the perceived quality gap between NB and PL is small, the consumer believes that the quality between the two product options is comparable, while PL is significantly cheaper. Thus, consumers perceive NB and PL as close to equal in quality, while PL dominates NB on price. This means that the PL product is the superior option and the NB alternative is the inferior option within the preferred region in the consumer’s mind consisting of PL and NB. It is presumed that this context effect drives consumers’ willingness to buy the PL product. In a similar vein, when the perceived quality gap between PL and C is small, the consumer believes that the quality between the two product options is comparable, while C is significantly cheaper. Thus, consumers perceive PL and C as close to equal in quality, while C dominates PL on price. This means that the C product is the superior option and the PL alternative is the inferior option within the preferred region in the consumer’s mind consisting of C and PL, which assumedly discourages grocery shoppers to choose the PL product. The perceived quality gaps between NB and PL products, and PL and C products, respectively, serve as mediators for the relationship between the level of similarity between a PL
  • 54. 54 product and its NB alternative and grocery shoppers’ willingness to buy the PL product, as expected. Thus, the level of similarity influences the quality gaps, which at their turn influence consumers’ willingness to buy the PL option. The relative price differences between PL and NB products, and C and PL products, respectively, appear not to influence grocery shoppers’ willingness to buy PL items. However, no extreme prices were used in the experimental stimuli. Undoubtedly, as relative price differences between product options become larger than usual, this will have an effect on consumers’ buying intentions. 5.2 Managerial implications The results of this study have several implications for both grocery retailers and NB manufacturers, as well as legislative implications. As mentioned, PLs are important to grocery retailers in two different ways. First, a strong PL brand strengthens retailers’ negotiation position with NB manufacturers, because these manufacturers may realize that retailers can replace some of their NB products by PL products that are positioned as close substitutes. Second, the percentage profit margins are higher for PL products than for NB products, mainly because of the lower costs associated to packaging and advertising. Yet, selling prices for PL products are lower than for NB products, so the absolute profit margins for PL products are not necessarily higher than for NB products. To secure the substitution power and absolute margins of PL products, it is important that grocery retailers position these products as decent quality alternatives for NB products, rather than as cheap, lower quality options. The findings of this study provide insight into how PL products can be positioned as decent quality alternatives for more expensive NB products.
  • 55. 55 It turns out that marketing managers of PL brands should differentiate the packages of their PL products from those of the NB alternatives, instead of imitating them. Imitation, even if it is done subtly, makes grocery shoppers perceive the difference in quality between the product options as larger, not smaller, which decreases their willingness to buy the PL product. At the same time, PL brands should still spend enough money on packaging, to create and maintain consumers’ perception that they offer high quality products, in order to prevent directly competing on price with C alternatives. However, PLs should not invest so much in packaging that PL selling prices have to be increased significantly to cover the increased costs. While PLs should aim to compete with NBs directly on value for money, they should prevent their prices from becoming equal to NB prices, thereby preventing competing with NBs solely on (perceived) quality. If for any reason directly competing with NB products is unfeasible, PLs are forced to compete more directly with C products. In that case, the visuals of NB products should be subtly imitated by PLs, in order to strengthen the PL’s image of being an (affordable) high quality brand, so that the perceived quality difference with the C alternatives increases. This will make consumers perceive the PL product and C product as being close to equal in price, while they will perceive the PL product as superior in quality compared to the C option, which decreases the likelihood that they will choose the C alternative. The outcomes of this study imply that NB manufacturers should not fear PL brands that are copycatting the visuals of their products, because copycat PLs appear not to harm NB sales. In fact, the higher the level of visual similarity between the PL and NB product, the more beneficial it is to the perceived quality and the sales of the NB option. Contrary to intuition, marketing managers of NB manufacturers should not differentiate the visuals of product packages when PL brands are imitating them. The costs savings derived from a lower frequency of package differentiation will
  • 56. 56 lead to higher profits by NB brands. The saved money can also be used to lower NB selling prices, which makes the PL alternatives less attractive options in terms of value for money. There are also legislative implications. If NB manufacturers encounter PL products with similar visuals, they have no reason and no ground to sue the focal PL brand, as the copycatting in fact benefits them. Although NB manufacturers will likely perceive copycatting by PLs as unfair, they should ignore them, rather than fight them in court. 5.3 Limitations & directions for future research It is important to pay some attention to the validity and reliability of this study, especially since it yields some unexpected findings. First, the pre-test that was conducted to test the suitability of the experimental stimuli had a sample size of seventeen. The rather small sample size may have led to a relatively low reliability of the pre-test, and therefore to a relatively low validity of the experimental conditions. This may have affected the results from the experiment. Second, only (manipulated) PL products from the ‘Perfekt’ brand were used as stimuli, while participants’ attitude towards the Perfekt brand compared to other PL brands was not controlled for. Participants’ attitude towards this specific PL brand may have affected the validity of the results. Third, all PL product stimuli were digitally manipulated by adding a Perfekt logo. Some participants may have noted that some product stimuli were ‘fake’, which might have influenced their responses. Fourth, regarding the reliability of the study results, all participants were Dutch, so generalization of the results might be problematic. Fifth, the sample size of 178 participants is rather small, and the results could have been more reliable if more people participated in the experiment. The limiations described above mainly represent generalizability issues. Future research should determine the robustness of the results of this study. Sixth, asymmetric dominance effects are postulated to drive the effect of the
  • 57. 57 perceived quality gaps between NB and PL, and PL and C, respectively, on consumers’ willingness to buy PL products. However, such context effects were not directly tested. It is only assumed that these effects play a role in the relationship between the mentioned variables. Future studies should directly test the influence of such context effects. Finally, although this study controls for the influence of relative price differences between product options, this study does not explicitly take into account different types of PL products (i.e. economy PL products, standard PL products, and premium PL products). Results of a study by Geyskens, Gielens & Gijsbrechts (2010) show that the introduction of economy PLs cannibalizes standard PLs, and the introduction of premium PLs cannibalizes both standard and premium PLs, driven by price-quality context effects. Also, economy PL introductions benefit mainstream-quality NBs because these NBs become a middle or compromise option in terms of quality in the consumers’ choice set. These findings confirm that consumers assign different quality levels to different PL product types. Future research should explore whether the influence of the level of visual similarity between a PL product and its NB alternative on grocery shoppers’ willingness to pay the PL option is dependent on the type of PL product. For instance, it could be expected that a copycat economy PL product is perceived as more ‘fake’ than a copycat premium PL product, as an economy PL product is perceived as a low quality option anyway. A copycat premium PL product might benefit from its appearance of being able to compete on quality with the NB alternative. There are also several avenues for future research that are not directly derived from the limitations. Some possible explanations for unexpected results were given in paragraph 5.1, ‘Discussion of results’. Future studies should explore whether these explanations are correct. First, if the relationship between the level of visual similarity between a PL and NB product and the perceived quality gap between these products indeed appears to be linear and positive, the perception theory explanation provided in paragraph 5.1 should be tested. Second, futures studies should
  • 58. 58 address the (non-)symmetry of the relationships between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between NB and PL product’, and between ‘level of visual similarity between PL and NB product’ and ‘perceived quality gap between PL and C product’. This study found that the shapes of these relationships are asymmetrical, in contrast to what was expected. Future studies should test whether they are indeed asymmetrical, and if so, test whether this asymmetry is due to the dynamics within the perceived quality continuum between C, PL and NB, as postulated in paragraph 5.1. Third, an avenue for future study is the moderating role of the type of visual similarity on the effect of the level of visual similarity between PL and NB on the perceived quality gaps between different product options and on consumers’ the willingness to buy PL products. This study found that the type of visual similarity does not play a significant role, in contrast to a previous findings. Future research should provide decisive insights into this issue. Finally, a direction for study could be the influence of different types of PL products
  • 59. 59 References ACNielsen (2005), The Power of Private Label 2005: A Review of Growth Trends Around the World. New York: ACNielsen. Ailawadi, K. L., & Harlam, B. (2004). An empirical analysis of the determinants of retail margins: the role of store-brand share. Journal of Marketing, 68(1), 147-165. Garretson, J. A., Fisher, D., & Burton, S. (2002). Antecedents of private label attitude and national brand promotion attitude: similarities and differences. Journal of Retailing, 78(2), 91-99. Hoch, S. J., & Banerji, S. (1993). When do private labels succeed. Sloan Management Review, 34. Horen, F. V., & Pieters, R. (2012a). When high-similarity copycats lose and moderate-similarity copycats gain: the impact of comparative evaluation. Journal of Marketing Research, 49(1), 83-91. Horen, F. V., & Pieters, R. (2012b). Consumer evaluation of copycat brands: The effect of imitation type. International Journal of Research in Marketing, 29(3), 246-255. Howard, D. J., Kerin, R. A., & Gengler, C. (2000). The effects of brand name similarity on brand source confusion: Implications for trademark infringement. Journal of Public Policy & Marketing, 19(2), 250-264.
  • 60. 60 Kapferer, J. N. (1995). Brand confusion: empirical study of a legal concept. Psychology & Marketing, 12(6), 551-568. MacKenzie, S. B. (2003). The dangers of poor construct conceptualization. Journal of the Academy of Marketing Science, 31(3), 323-326. Morton, F. S., & Zettelmeyer, F. (2004). The strategic positioning of store brands in retailer– manufacturer negotiations. Review of Industrial Organization, 24(2), 161-194. Planet Retail (2007), Private Label Trends Worldwide. London: Planet Retail Ltd. Richardson, P. S., Dick, A. S., & Jain, A. K. (1994). Extrinsic and intrinsic cue effects on perceptions of store brand quality. The Journal of Marketing, 28-36. Sawyer, A. G., & Dickson, P. R. (1984). Psychological perspectives on consumer response to sales promotion. Research on sales promotion: Collected papers, 1-21. Sayman, S., Hoch, S. J., & Raju, J. S. (2002). Positioning of store brands. Marketing Science, 21(4), 378-397. Simonson, I., & Tversky, A. (1992). Choice in context: Tradeoff contrast and extremeness aversion. Journal of marketing research, 29(3), 281-295.
  • 61. 61 Steenkamp, J. B. E., Van Heerde, H. J., & Geyskens, I. (2010). What makes consumers willing to pay a price premium for national brands over private labels?. Journal of Marketing Research, 47(6), 1011-1024. Teas, R. K., & Agarwal, S. (2000). The effects of extrinsic product cues on consumers’ perceptions of quality, sacrifice, and value. Journal of the Academy of marketing Science, 28(2), 278-290. Warlop, L., & Alba, J. W. (2004). Sincere flattery: Trade-dress imitation and consumer choice. Journal of Consumer Psychology, 14(1), 21-27. Wulf, K. D., Odekerken-Schröder, G., Goedertier, F., & Van Ossel, G. (2005). Consumer perceptions of store brands versus national brands. Journal of Consumer Marketing, 22(4), 223-232. Zaichkowsky, J. L., & Simpson, R. N. (1996). The effect of experience with a brand imitator on the original brand. Marketing Letters, 7(1), 31-39. Zeithaml, V. A. (1988). Consumer perceptions of price, quality, and value: a means-end model and synthesis of evidence. The Journal of marketing, 52, 2-22.
  • 62. 62 Appendices Appendix 1 Results of pre-test and resulting visual similarity stimuli Blatant feature similarity (mean pre-test score = 2.8, N=17) Elements likely causing perceived level of visual similarity: - Package shape (round) - Depicted items (cow, countryside, blue sky) - Image type (cartoon) - Colours (brown, green, blue, white) High feature similarity (mean pre-test score = 2.2, N=17) Elements likely causing perceived level of visual similarity: - Package shape (small, long) - Package material (carton) - Depicted items (cows, grass, countryside, Dutch flag, dish with yogurt) - Image type (cartoon) - Colours (dark green)