Using Semiotics To Better Understand the Nature and Extent of Copycatting By Everyday Supermarket Brands
Blythe Forde
Master’s Thesis, Graduate School of Communication Master’s Program, Communication Science
University of Amsterdam
Date Submitted: 31.01.2018 Supervisor: Suzanne de Bakker
Abstract
A copycat brand is one who attempts to imitate the visual appearance of a leading brand with the aim of exploiting positive associations to become more easily acceptable as a feasible alternative. By focusing the attention on everyday consumer based goods, the overarching aim of this thesis was to investigate the nature and extent to which brands sold in
supermarkets imitate leading brands on a semiotic level. With much of the focus previously dedicated to how consumers might perceive copycats, through a range of fictitiously
generated scenarios, how they truly present themselves in the real world is a first of its kind. By comparing two of the largest supermarket retail chains in The Netherlands, Jumbo and Albert Heijn, the product packaging of forty subordinate brands in ten different product categories were analysed and scored based on their relative strength of similarity to a leading brand. Overall, it was uncovered that in the presence of a leading brand, subordinate brands apply copycatting techniques of a literal nature to a greater extent than semantic techniques. However, significant differences were not identified between the two main non-leading brand types, generics and normal competitors. Despite the highly subjective nature of the analysis, this was an insightful first step in identifying the true to life imitation tactics by brands in the real world and is hoped to set the foundation for further rationalising the mindset of the copycat brand.
Using Semiotics To Better Understand the Nature and Extent Of Copycatting By Everyday Supermarket Brands
Have you ever been to the supermarket with the urgency to buy shampoo, without a specific brand in mind? Then finding yourself not knowing where to start as you scan over the various rows of shelves? You only need one bottle of shampoo. And you have better things to do with your time. Yet, in deciding on ‘the one’, you must eliminate an array of competitors, each with their own visuals and texts, imagery and promises, selling points and special offers, similarly demanding that their shampoo is the one you must try. So which shampoo do you buy? A brand you have used in the past? Or one you recognised in a recent campaign? Maybe it has an established reputation. Perhaps in your mind, shampoo is shampoo and if you can find a cheaper alternative to a well-known brand, then why not pocket the change. So instead, do you try a brand you have not tried? If so, what precisely wins you over? Could it be that its distinct look captures your attention? Or maybe its no frills approach will suffice. Alternatively, do all brands look the same? By copying each other in blatant ways. Or are some imitations more subtle in a devious attempt to trick you? A trick that might only become apparent upon realising your very own hairdo disaster!
The Psychology Of Copycats
With such vulnerable and competitive markets, it is not unusual for competitor brands to leverage designs from existing leader brands in an attempt to shortcut their way to
grabbing the consumer’s attention and enticing them in. A copycat brand can be any brand who attempts to imitate the visual appearance (including name, logo, and/or package design) of a leading brand with the aim of exploiting the latter’s positive associations and marketing efforts, that are likely to be transferred and infused into the representation of the copycat brand and thus be more easily acceptable as a feasible alternative (Warlop & Alba, 2004; van Horen & Pieters, 2012; Miceli & Pieters, 2009). Ultimately, imitation strategies can be
effective as they activate associations that remind consumers of something they know (i.e., the leading brand) (van Horen & Pieters, 2013).
Brands who copy are essentially “...conning the consumer into, at best, feeling some halo effect from the real thing, or, at worst, buying it by mistake.” (Marketing Society Forum, 2010, p.22). A product that resembles the overall appearance of a branded product, by
combining some or all of its visual features, may effectively mislead the consumer into believing that the lookalike has the same quality, the same source of innovation and the same set of values as the original brand (Burt & Davis, 1999). The basis of imitation is that
consumers generalise the similarity between exterior physical features to infer similarity of product quality or functionality (Collins-Dodd & Zaichkowsky, 1999). But consumers who consciously purchase a lower priced imitation may experience greater value if the product turns out to be as good as the leading brand it imitates. As a consequence, this might not only impact leading brands through a loss in sales, but may irreparably damage their reputation by changing consumers' attitudes about its uniqueness (Collins-Dodd & Zaichkowsky, 1999). Copycatting Works Through Semiotics At Two Distinct Levels
With the psychology and underlying motives of the copycat in mind, the way in which brands imitate might also be of relevant importance. According to the European Brands Association (AIM), a copycat brand either looks just like the real thing, or possesses certain features of the real thing (Burt & Davis, 1999). Using the communicative concept of semiotics, which defines how signs and symbols are used to create and convey meaning (McQuail, 2010), there are two theoretical levels by which imitation can occur. Literal copycats involve brands who imitate low-level, descriptive and concrete visual or perceptual attributes, such as colours, shape, sizes, lettering, names, and graphical elements, or other distinctive marks (Miceli & Pieters, 2009; Foxman, Muehling & Berger, 1990) (see also Figures 1 and 2). Alternatively, semantic or theme based copycats involve brands imitating
higher-order, more abstract themes communicated by a leading brand's visual appearance (such as benefits, goals, or usage context) and is often represented by pictorials and communication of brand positioning (Miceli & Pieters, 2009) (see also Figures 3 and 4).
Figure 1. Boots’ Fruit Essences applies a literal Figure 2. Aldi’s Sainte Etienne beer applies a literal adaptation of Meijer’s Herbal Essences bottle adaptation of Stella Artois’ packaging shape, design design by emulating the curved/slim shape and and colour scheme (i.e. white with red/silver versus pink/green colour scheme, and by using a white with red/gold) and highly similar logo design highly similar product name (i.e. ‘Herbal (i.e. emblem shape, placement and application of Essences’ versus ‘Fruit Essence’). colour scheme).
Figure 3. Dollarama’s product name, Meteor, Figure 4. Aldi’s product name, Choco Break demonstrates a semantic adaptation of Mars’ demonstrates a semantic adaptation of Kit Kat’s galaxy themed name (i.e. the planet Mars). marketing slogan ‘Have a break, have a Kit Kat’, not evident on Kit Kat’s packaging design.
Research Gap and Relevance
Previous research undertaken in the field of brand copycatting (van Horen & Pieters, 2012; van Horen & Pieters, 2013; Miceli & Pieters, 2009; Le Roux, Bobrie & Thébault, 2014; Huddleston, 2016) has so far only focused on comparing actual brand leaders to a
range of fictitious brand examples, that don’t actually exist, but were specifically created for that category to generate a desired effect and/or response (see Figure 5). These fictitious brands, in comparison to a leading brand, usually varied by the type of semiotics applied (i.e. literal and/or semantic) and by established degrees of similarity or difference (i.e. low, medium or high), with a common purpose to understand how consumers might perceive, interact and/or relate to these imitation types in an experimental setting.
Figure 5. Van Horen and Pieters’ (2012) brand stimuli used in their study on the effect of imitation type on consumer perceptions. The comparison quadrangle on the left shows fictitious variations of Bertolli’s spreadable butter. While the comparison quadrangle on the right is made of up fictitious variations of Milka chocolate. A ‘feature copycat’ applies literal semiotics, whereas a ‘theme copycat’ applies semantic based semiotics.
This thus leaves a critical oversight into the actual efforts, lengths and means by which brands go to, to portray themselves in the real world. In reflecting on these
hypothetical imitation scenarios, the goal of this thesis was to recognise that the nature and extent of copycatting really occurring is a fundamental, yet previously overlooked first step in understanding the mindset and motivations of the brand. Such that, while it is good to gain insight into how consumers might perceive a copycat brand, through generating fictitiously designed imitations, what value is there to gain if in fact copycatting does not occur in the way that we envisaged? Are brands imitating brands in rather lazy and unimaginative ways, through the more literal and concrete perceptual attributes like colours, shapes and words? Or are they a bit more devious and sneaky, imitating the higher order, semantic based themes
and concepts, such as benefits, goals, or context? More specifically, how does copycatting, through the imitation of distinct brand packaging design, truly play out? Do brands really copy each other more literally or semantically, or a combination of the two? Mapping this out hopes to provide clear evidence that this practice is prevalent and persists, and to then
determine if and what might be the implications of this in the future of branding consumer based goods.
As fast moving consumer goods (FMCG) or consumer packaged goods (CPG) make up the majority of common everyday brands, the supermarket therefore serves as the ideal setting for this type of copycatting to occur. Here, you have an interesting competing mix: leading brands (brands who take up the highest market share and are often well-known to the everyday consumer), normal competitor brands (brands competing directly with a leading brand) and supermarket generics (brands owned and registered by the supermarket retailer by which they are stocked). By focusing the attention on everyday consumer based goods, the overarching aim of this thesis was to better understand the extent to which brands sold in supermarkets might imitate leading brands at different semiotic levels. This thesis therefore aims to answer the overarching research question...
To what extent do brands sold in supermarkets imitate leading brands on a semiotic level (literal versus semantic)?
Theoretical Framework
Establishing an Identifiable Leading Brand Through Distinct Trade Dress
In 1960, the American Marketing Association defined a brand as, “...a name, term, sign, symbol, or design, or a combination of them which is intended to identify the goods or services of one seller or a group of sellers and to differentiate them from those of
competitors.” (AMA, 1960, cited in Heding, Knudtzen & Bjerre, 2009, p. 9), whereas brand identity can be considered as the visual picture and meaning derived from the visual impact
of the brand (Zaichkowsky, 2010). For a brand to be effectively identified, it must be able to “...express the particular vision and uniqueness of the brand.” (Heding, et. al., 2009, p. 13).
The identity elements are the concrete parts that tell the consumer what the brand means and contribute to distinct consumer perceptions of various brands in the marketplace, helping to differentiate brands from competitors (Zaichkowsky, 2010). Brand differentiation, or identity, may be achieved through a well-known brand name, logo, typeface, symbol, colour, shape, distinctive design of the package, as well as unique product and benefit descriptions. Brand names in a product class should be unique in terms of sound,
pronunciation, spelling and meaning. Unique shapes of containers, packages or products also help create a distinctive identity. Finally, colours are remembered far more than shapes and are the first point of identification of the brand (Zaichkowsky, 2010). “Strong and unique links to one source is what complements and protects brands from being interchangeable in the marketplace.” (Zaichkowsky, 2010, p. 548).
Trade dress, in the form of product packaging, "embodies that arrangement of
identifying characteristics or decorations connected with a product...[that] make[s] the source of the product distinguishable from another and...promote[s] its sale." (Sandberg, 2009, p. 10). The traditional role of trade dress has been to protect, contain and deliver the product to the retail shelf, gain the attention of the consumer at point-of-purchase, convey a strong distinctive brand identity, and swiftly communicate the brand's features, quality, and value (Underwood, 2003). Acting as a silent salesman (Huddleston, 2016), differentiation by package design (or novel packages) is considered to significantly impact reward anticipation and choice (Zaichkowsky, 2010). “Products with unique and innovative packaging are more likely to pop off the shelf and attract shoppers’ attention.” (Huddleston, 2016, p. 2). By attracting and persuading consumers through an array of cues, product packaging has strong communicative power in creating brand identity, becoming a vehicle for meaning creation
and delivery (Huddleston, 2016), an expression of brand personality and individuality
(Underwood, 2003), as well as signifying some standard of quality, performance and function (Loken, Ross & Hinkle, 1986).
Understanding Copycatting by Subordinate Brands Through Semiotics So while being distinctive should not be difficult, it may be expensive to
communicate (Zaichkowsky, 2010). Copycatting is thus a pervasive (van Horen & Pieters, 2012) and deliberate (van Horen & Pieters, 2013) strategy in shortcutting to perceived product quality and benefit (Huddleston, 2016). According to Le Roux, et.al. (2014),
copycatting techniques can occur at two semiotic levels. Literal similarity of brands is based on the consumer's evaluation of common characteristics and corresponds to imitation
strategies that may be based on the number of common letters, or similar sequences of letters, for example. On the other hand, semantic similarity is based on meaning and aims at
activating higher-level semantic signification in order to create inferences about the imitator, based on copied attributes. This strategy corresponds to an imitation of the concept rather than an exact copy of the product (Le Roux, et.al., 2014).
Literal versus semantic. Van Horen and Pieters (2012) argue that copycats most often imitate the distinctive perceptual features of leader brands (i.e. visual characteristics, text and sounds), since literal imitations exclusively associated with the leader brand are directly linked to that leading brand, and will thus immediately activate a clear representation of that brand. Literal copycats are “directly linked to the leader brand and almost immediately activate a (positive) image, whereas theme copycats are only indirectly linked to the leader brand.” (van Horen & Pieters, 2012, p. 247).
This premise has even greater strength in a context of uncertainty (van Horen & Pieters, 2013). Here, it is proposed that the evaluation and choice of blatant copycats is critically dependent on the uncertainty of quality elicited by the decision context. More
specifically, it is predicted that in an uncertain context, people in fact like, instead of dislike, blatant (literal) copycats. It is thought that situations of uncertainty, where consumers have no knowledge of the brands that are on the market, how they perform and how they compare to other brands, induce unpleasant feelings and these prompt coping responses to reduce them. When the situational context induces uncertainty about product quality, a copycat is likely to serve as an effective uncertainty-reducing device, signalling safety and security. When time is limited and there is little opportunity to extensively search for information and compare alternatives, consumers are likely to search for simple, low-effort cues to assess quality in order to reduce uncertainty. Such quality cues reside, for instance, in the price, packaging, and brand name of a product (see Figure 6). “Its similarity in package design will remind consumers of something known (i.e., the leader brand) and the associated warm glow of familiarity feels good to consumers who are uncertain (van Horen & Pieters, 2013, p. 56).
Figure 6. Van Horen and Pieters’ (2013) brand stimuli used in their study on the effects of similar-dissimilar imitations on consumer perceptions in situations of uncertainty. Using Red Bull as a leading brand from the energy drinks category, direct copycat brand ‘Bull Fighter’ and a differentiated brand were used to base comparisons.
Furthermore, when consumers are in a situation in which they have no knowledge of the brands that are available and how these perform, they are likely to make inferences about missing quality information using similarity to known brands for this. Familiarity will serve
as a cue for product quality in such a way that similarity in package design will be used to infer similarity of quality, performance, and reliability. In summary, when copycats are blatant and consumers are aware that a copycat strategy is being used, feelings of familiarity will infuse and dominate decision-making. “When feeling uncertain, the warm glow of familiarity seems to compensate for ‘morality’. [And thus], this comfort of familiarity
overrides the readiness to penalise the copycat for the blatant imitation strategy being used...” (van Horen & Pieters, 2013, p. 62).
Therefore, one might expect that literal copycats are better equipped than semantic copycats to free-ride effectively on that of a leading brand; literal, blatant copying would seem a much simpler and instantly effective copying technique. Thus, in the fast moving consumer world of supermarket based product and selection, it is likely that this line of reasoning makes sense for literal imitation to reign as a more popular copycat strategy. Therefore, in basing the first half of the first hypothesis, it was considered that:
H1a: In the presence of a leading brand, subordinate brands apply copycatting techniques of a literal nature to a greater extent than semantic techniques.
Semantic versus literal. In direct contrast, it has been alternatively considered that any positive association from a leader brand to its copycat can be more effective when the copycatting is done subtly, and the imitation is not as obvious and blatant (van Horen & Pieters, 2013). When distinguishing between consumer perception of literal (i.e. attribute based) similarity and semantic (i.e. theme based) similarity in meaning that is associated with a brand’s look and feel, it has been suggested that consumers are not so easily tricked by explicit hijacking of distinctive perceptual features. But when such tactics are not as explicitly evident, consumers appear to more easily fall prey to the persuasive appeal of copycats when these piggyback on the success of leader brands in a more implicit and less distinctive manner by imitating their themes (van Horen & Pieters, 2012). Ultimately, subtler
copycats pose an interesting threat to leading brands because such tactics are more often undetected by consumers and subsequently more difficult to identify.
This approach enables the more semantic based copycatters a free ride to success, allowing subtle copycats to “coattail” on the leading brand and similarly fly under the radar of trademark legislation (van Horen & Pieters, 2012). When imitation is more literal (i.e. more obvious and blatant), copycat evaluation is negative, because such obvious imitation attempts are perceived as unacceptable and inappropriate, and consequently disliked (van Horen & Pieters, 2013). Yet when imitation is more semantic and less obvious, consumers consider imitation of meaning as more acceptable and less unfair (van Horen & Pieters 2012).
In van Horen and Pieters’ (2012) study of this concept, three variations of well-known Swiss brand, Milka Chocolate, were created containing traits that were either an imitation of a literal feature (e.g. by switching out brand name ‘Milka’ for ‘Lecha’), an imitation of the theme (e.g. by replicating the iconic Swiss mountain graphic), or a completely differentiated brand entirely (i.e. no literal or theme imitation included) (see Figure 7). In asking consumers of their willingness to buy each respective imitation variant, they uncovered that imitation of an underlying meaning or theme (semantic) was evaluated more positively than an imitation of a distinctive (literal) feature, or no feature at all. It was argued that while semantic based imitation is harder to detect, when it becomes known, it is regarded more favourable in comparison to literal based imitation tactics (van Horen & Pieters, 2012).
Figure 7. Copycat re-creations of Milka Chocolate used in an experiment by van Horen and Pieters (2012) on the effects of different imitation types (literal versus semantic versus entirely differentiated) on consumer perceptions. A ‘feature copycat’ applies literal semiotics, whereas a ‘theme copycat’ applies semantic based semiotics.
In further understanding the effects of literal versus semantic semiotic approaches to perceptions of brand similarity, Le Roux, et.al. (2016) similarly generated simulated brands that were either identical, similar or dissimilar by varying degrees and combinations to a well-known leading brand. For example, a counterfeit brand stimulus containing minor changes in the expression and intensity of well-known perfume brand, ‘CK One’ cologne, depicted both highly literal similarities (e.g. changing the brand name to ‘OK One’) and semantic variants (i.e. similar bottle shape and colour). While a counterfeit brand stimulus of greater distinction was more literally dissimilar (e.g. brand name changed to ‘KC2’) and held more marked semantic differences (i.e. greater variations in bottle shape and colour) (see Figure 8). Given this, it was revealed that stimuli closer in terms of name and appearance were considered suspicious and subsequently categorised as imitations/counterfeits. Whereas consumers who detected imitation attempts that were farther in terms of name and
Figure 8. Stimuli used to test perceptions of copycat variants to CK One cologne in an experiment by Le Roux, et.al. (2016), focusing on the effects of different imitation types (literal versus semantic) at varying levels of similarity (identical, similar, different).
Therefore, theme-based copycats work when they share few directly comparable visual attributes but still express the same abstract message as the leading brand, thus constituting more subtle copycatting strategies (Miceli & Pieters, 2009). Thus, in basing the second half of the first hypothesis, it was predicted that brands would prefer to adopt a more integral and less obvious approach. Therefore:
H1b:In the presence of a leading brand, subordinate brands apply copycatting techniques of a semantic nature to a greater extent than literal techniques.
Copycatting by Generics Versus Normal Competitors
Generic supermarket brands, defined for the purpose of this thesis, as a consumer product produced by or on behalf of a supermarket distributor and sold under the distributor's own name or trademark through the distributor's own outlet (Tran Tram, Wood & Orozco Gomez, 2010), is typified by the fact that it bears the brand of the retailer which sells it. Yet in testing the first hypothesis, the term subordinate brand, determined for the purposes of this thesis simply as any brand other than a leading brand in a given product category, is rather
broad and suggests little discrimination between the various classes or sub-types of brands that exist in a supermarket context. Therefore, it might also be important to distinguish the copycatting practices between generics and other subordinate brands (i.e. any brand not affiliated with a supermarket chain) because a different set of motivations and/or incentives may apply, thus influencing the nature and extent to which copycatting practices take place by different subordinate brand types.
Generic brands have been part of society for more than thirty years and were traditionally known for being both cheaper and lower quality options (Tran Tram, et. al., 2010). By using simple packaging and design, by giving up part of the quality and by doing without advertising, they could offer the same product at reduced prices and place their own products in the most visible areas in order to promote price comparison and sales (Castañer Camps, Coll i Barneda, Gould Gavidia & Hernández Uptegrove, 2012). Generic supermarket brands, as a component of a supermarket retailer’s brand, are recognised as a dimension of retailer image. Thus, successful generic branding strategies can establish a competitive advantage and may be effective in changing customers’ awareness, association, and perceived quality about a retailer. Providing consumers with reasonable value by offering generics may be an effective way to draw loyal consumers (Choi & Huddleston, 2014). Additionally, retailer resources such as buying power and portion of shelf space contribute to their ability to offer their generics at a lower price than normal manufacturer brands and allows supermarket retailers to offer variety in a wide range of retail categories, increasing profit margin and market share. Finally, it is thought that by reinforcing the same generic brand name over a wide array of product categories across the store strengthens awareness and recall of the supermarket retailer’s brand and can facilitate the consumer’s decision-making (Choi & Huddleston, 2014).
have boosted their comparability to many leading brands in their field. Manufacturers of generic brands have since invested heavily in their design and marketing in order to be perceived as having a quality product comparable to that of its competitor (Castañer Camps, et.al., 2012). But it seems that generic supermarket brands might be better off investing less in distinctive design and more into looking the same; while the assumption that distinctive packaging as a significant driver of perceived quality and consumer motivation to pay a premium price persists, it may not always be applicable.
To demonstrate, Huddleston’s (2016) study showed that for generic brands, distinctive packaging was not in fact effective in building a higher quality perception or increased price expectation. In her study, generic brand packaging attributes were manipulated by levels of similarity versus distinctiveness to certain packaging design features, as well as levels of similarity versus distinctiveness to certain product description features (e.g. through product claims and benefits mentioned in the text) (see Figure 9). While distinctive generic brand packaging design and content did positively influence consumer perception towards the retailer, distinctive packaging design and content did not positively influence purchase intentions. Such that, distinct packaging design and content was not a sufficient extrinsic cue to persuade consumers that quality and price of a generic label was equal to that of a national brand leader (Huddleston, 2016).
Figure 9. Stimuli created based on Proctor & Gamble’s Head & Shoulders brand in an experiment by Huddleston (2016) in studying the effects of different imitation types (similar/dissimilar packaging design versus
similar/dissimilar product descriptions) on consumer perceptions.
While the previously mentioned studies by van Horen and Pieters (2012) and Le Roux, et.al. (2016) did not focus specifically on generics, Huddleston’s (2016) findings using generics was similarly consistent in its theme; brands who go to greater distinctive efforts earn more consumer respect, but overall, has little negative impact on actual brand purchase intentions. This in practice puts distinct leading brands at risk; since the efforts of
supermarket retailers in creating distinctly designed generics is of little relevance to
consumers, in terms of actual purchase intentions, not only will generics save on innovation costs (normally needed to create a distinctly differentiated brand), they may more easily take up market share at the expense of their competitors, and ultimately, the leading brand.
Therefore, in substantiating the second hypothesis, it was considered that generic supermarket brands have much to gain through copying leading brands, as they are situated in a unique vantage point in their ability to drive, determine and exclude the way in which their brands, as compared to non-affiliated competitors (termed ‘normal competitors’ for the purpose of this thesis) are presented and marketed in their stores. Such that:
H2:In the presence of a leading brand, generic supermarket brands apply copycatting techniques (literal and semantic combined) to a greater extent than normal competitor brands.
Literal and Semantic Copycatting by Generics Versus Normal Competitors
Just as it was considered a restriction for subordinate brands in hypothesis 1b, the extra thought, time and resources normally needed to extract strategically based subtle
elements of conceptual themes from a leading brand (as opposed to literal techniques) may be even more pronounced for generic brands (compared to normal competitors) given their lower price point generated through reduced production costs. And just as in hypothesis 2, it was the similarities and not differences that were considered to work better to a generic brand’s advantage, in terms of building greater brand presence and market share. In
combining both lines of thinking, it may be plausible to suggest that the more blatant literal imitation techniques will be a more appealing, cost effective and thus prevalent copycat strategy by generic brands compared to normal competitors. Conversely, since semantic imitation takes a lot more thought and investment in an effort to replicate in less obvious ways, this technique would be less prevalent by generics compared to normal competitors. Therefore:
H3a: In the presence of a leading brand, generic supermarket brands apply
copycatting techniques of a literal nature to a greater extent than normal competitor brands; H3b: Conversely, in the presence of a leading brand, generic supermarket brands apply copycatting techniques of a semantic nature to a lesser extent than normal competitor brands.
Theoretical Model and Tests
To summarise the theoretical framework discussed and hypotheses drawn, the following model (Figure10) has been devised:
Figure 10. Theoretical model – the nature and extent of copycatting by subordinate brands
In addition to the three main hypotheses identified in this theoretical model, a series of sub-questions have been generated for subsequent testing (see Tables 1, 2 and 3). These sub-questions focus on identifying similar or contrasting patterns that might exist between that of generics and normal competitor brands, as well when distinguishing between specific types of semiotic based imitation, such as those that might focus solely on text based
techniques (e.g. imitations of words and letters and/or key messages) or techniques dedicated to graphic representation (e.g. imitations of colours, font styles and/or visual interpretation).
Table 1
Hypotheses 1 group of tests
Test # In the presence of a leading brand,
?[subordinate / generic supermarket / normal competitor] brands...
...apply
[copycatting (cc) techniques]
of a ?[literal / semantic] nature...
...to a ?[greater / lesser] extent than... ...?[literal / semantic] cc techniques. Main Hypotheses
H1a All subordinate brands cc techniques - literal greater semantic
H1b All subordinate brands cc techniques - semantic greater literal
Research Sub-Questions (SQ)
SQ1a Generic Supermarket brands cc techniques - literal greater semantic
SQ1b Normal Competitor brands cc techniques - literal greater semantic
SQ1d Generic Supermarket brands text based cc techniques - literal greater semantic – text based
SQ1e Normal Competitor brands text based cc techniques - literal greater semantic – text based
SQ1f All subordinate brands graphicsa based cc techniques - literal greater semantic – graphics based
SQ1g All subordinate brands graphicsb based cc techniques - literal greater semantic – graphics based
SQ1h Generic Supermarket brands graphicsb based cc techniques - literal greater semantic – graphics based
SQ1i Normal Competitor brands graphicsb based cc techniques - literal greater semantic – graphics based
SQ1j All subordinate brands graphicsc based cc techniques - literal greater semantic – graphics based
a. Copycatting graphic features via the literal imitation of illustration and imagery
b. Copycatting graphic features via the literal imitation of illustration and imagery, colours and/or font style
Table 2
Hypothesis 2 group of tests
Test # In the presence of a leading brand,
?[generic supermarket / normal competitor] brands...
...apply
[copycatting (cc) techniques], literal & semantic combined...
...to a ?[greater / lesser] extent than... ...?[generic supermarket / normal competitor] brands. Main Hypothesis
H2 Generic Supermarket brands cc techniques – literal & semantic greater Normal Competitor brands
Research Sub-Questions (SQ)
SQ2a Generic Supermarket brands text based cc techniques – literal & semantic greater Normal Competitor brands
Table 3
Hypotheses 3 group of tests
Test # In the presence of a leading brand,
?[generic supermarket / normal competitor] brands...
...apply
[copycatting (cc) techniques]
of a ?[literal / semantic] nature...
...to a
?[greater / lesser]
extent than...
...?[generic supermarket / normal competitor]
brands.
Main Hypotheses
H3a Generic Supermarket brands cc techniques – literal greater Normal Competitor brands
H3b Generic Supermarket brands cc techniques –semantic lesser Normal Competitor brands
Research Sub-Questions (SQ)
SQ3a Generic Supermarket brands text based cc techniques – literal greater Normal Competitor brands
SQ3b Generic Supermarket brands graphics based cc techniques – literal greater Normal Competitor brands
Methodology
This thesis will aim to develop a tool that can evaluate and measure the degrees of similarity of literal and semantic characteristics that are depicted from a range of actual brands that exist within product categories. It will include a combination of leading and subordinate brands (made up of both generics and normal competitors) and will focus on a range of characteristic types which represent each semiotic level. For example, literal attributes will be measured by the obvious characteristics such as logo wording, positioning, font style, tagline and colours. While semantic attributes will be measured by their overall look and feel and the kinds of themes, concepts and messages they represent. Brands will be coded and scored according to these characteristics, while a leading brand’s attributes will be used as the benchmark for comparison. All similarities observed will be assessed on the basis of the strength of the comparisons made. This approach will now be discussed in more detail in the following sections.
Data Sample
Selection and screening. The focus of this thesis is to draw comparisons between current leading supermarket brands and their subordinates in a specific product category and to determine the extent and nature of how semiotics (literal versus semantic) are applied to imitate these leading brands. Given that the key differentiating feature of this thesis is to determine how this plays out in the real world, the two biggest leading supermarket chains in The Netherlands were selected for data collection: Albert Heijn and Jumbo. These specific supermarket chains are the first and second largest chains in the Dutch market respectively (Angloinfo, 2018) and were within the top one hundred of the world’s largest retail
supermarket and grocery store chains, ranking 6th and 40th, respectively, in terms of revenue (The Balance, 2017). These chains were chosen based on their high similarity of product selection and range, including their mix of leading, generic and competitor brands, their
broad and prevalent geographical reach and comparable price points. The store site locations, Albert Heijn in Amsterdam-Zuid and Jumbo in Amsterdam-Noord, were chosen based on their larger size, scale and produce offering compared to stores located in the more central, built up neighbourhoods where space is often a limitation. Other chains currently part of the Dutch market, including Dirk van den Broek, Aldi and Lidl, were initially considered for selection, but they were each found to offer a limited range of brands (i.e. offering more generic brands over leading or competitor brands), had less geographical reach and offered price points at lower ends of the scale, and were thus eliminated.
In order to obtain a good cross-section of brands across a range of product categories for comparison, all product categories within each supermarket chain were systematically screened and determined for inclusion in the analysis, based on a selection of criteria created specifically for the purpose of this thesis. This entailed walking down each supermarket aisle and observing each category in its entirety and then checking the range of products within these against the screening criteria for potential inclusion. Each product category had to meet all screening criteria (see Table 4).
Table 4
Supermarket product category selection - Screening Criteria (SC)
SC# Screening Criteria per Product Category Qualified ()
1 Product must have complete or partial packaging that is branded (i.e. cannot be vegetable, fruit, bread or meat, either wrapped or unwrapped)
2 Must be able to identify a clear leading brand (base this on an initial observational assumption and subsequently confirm via market research)
3 Must identify that the assumed leading brand is not a generic brand
4 Must identify at least one generica and/or normal competitorb brand in addition to the leading brand
5 Using an observed judgement, identify at least one reasonable similarity in appearance between the leading brand and one subordinate brand (generic and/or normal competitor)
6 Leading brand is present in both supermarket chains
7 Leading brand has at least two subordinate brands which can be either generic, a normal competitor, or a combination of both (i.e. a minimum of three brands per product category)
8 Product category has sub-categories excluded (e.g. cannot include both classic cola and diet cola varieties in the one cola category)
9 Product category must offer the same type of products in terms of substance and form (e.g. clothing detergent provided in liquid as opposed to tablet form)
10 Product category must serve the same purpose (e.g. disinfectant that is used on all surfaces opposed to specific surfaces only)
a. Generic – A generic brand is a product that is owned and manufactured by the supermarket chain and has this stated on the packaging in the form of the supermarket logo and/or contact information on the back of the label.
b. Normal competitor – A normal competitor brand is a product that has no association to the supermarket chain in which it is being stocked, or in terms of ownership. This can be confirmed by checking the manufacturer information on the back of the label.
Product categories which were initially anticipated to meet each of the screening criteria were photographed in their entirety, totalling twenty-two different product categories. Once all photos were taken, comparisons within product categories across both supermarket chains were further checked against the screening criteria to ensure their inclusion. This additional check lead to a total of twelve product categories subsequently not meeting at least one of the screening criteria and were further eliminated from the final sample. Examples of this included photographing the same product category in a competitor chain to discover the leading brand was not present (i.e. not stocked). Or, later realising it was too difficult to match sub-categories precisely, such as in tea flavour varieties (e.g. pairing ‘chai vanilla’ flavour with ‘smooth chai’ did not qualify as a direct product comparison). Refer to Appendix A for the full list of eliminated product categories and explanations.
Establishing leading brands. To confirm the leading brands for each product category, various online searches were conducted to seek out top brand listings, such as Brand Finance’s Global Top 500 2017 (Brand Finance, 2017). These lists were used to verify the highest ranked brand in terms of either revenue, recognition, pioneering status and/or longevity in the market. For example, in the body wash category, only Nivea and Dove ranked within Brand Finance’s Global Top 500 2017 list, with Nivea (#228) ranking higher than Dove (#347). Therefore, Nivea was positioned as the leading brand for this category. Similarly, in the cola category, Coca-Cola (#27) ranked higher than Pepsi (#67) (Brand Finance, 2017).
Establishing subordinate brands. A subordinate branded variable was defined as any brand that was not a leading brand in that particular product category (i.e. all comparison brands). Of all subordinate brands, a generic supermarket brand was defined as any
subordinate brand that was owned and manufactured by the supermarket chain, as stated on the packaging, in the form of a supermarket logo and/or contact information (e.g. either
Jumbo or Albert Heijn). Whereas a normal competitor brand was defined as any remaining subordinate brand that had no association to the supermarket chain in which it was stocked and in terms of ownership.
Final product categories. Upon completion of the screening process and establishment of brand types within each category, a remaining total of ten product categories, each with a varying number of subordinate brands (ranging from two to eight subordinate brands per category), were identified as sufficiently meeting the screening criteria and were subsequently included for data coding and analysis. Product categories included cola original (excluding diet and flavoured forms), olive oil (specifically extra virgin), body wash (specifically with coconut), chocolate spread (e.g. Nutella) and corn flakes. The final list of product categories, including leading and subordinate brands per category, can be seen in Table 5. A final reference list is also provided in Appendix B.
Final Sample - Leading and subordinate brands (inclusive of generics and normal competitors) per product category Cola
Brand: Coca-Cola Jumbo ah ah Basic Highway Pepsi
Product: Coca-Cola - Original Taste Cola Authentic Taste Classic Cola ah Basic Cola Cola Pepsi
Company: Coke Jumbo Supermarkten B.V. Albert Heijn B.V. Albert Heijn B.V. Refresco Benelux Pepsi-Cola
Chocolate Milk
Brand: Chocomel ah Jumbo Bonomel Tony’s Chocolonely
Product: Chocomel Choco Melk Chocolade melk Chocolade Melk chocolademelk
Company: Friesland Campina Albert Heijn B.V. Jumbo Supermarkten B.V. Consalac Tony’s Chocolonely
Pasta Sauce - Basil
Brand: Bertolli Jumbo ah Barilla Grand’Italia Jamie Oliver
Product: Basilico Basilicum Pasta Saus Basilico Basilico Tomato & Basil Pasta Sauce
Olive Oil - Extra Virgin
Brand: Bertolli ah ah Basic Jumbo Carbonell
Product: Olio Extra Vergine Di Oliva Olijf Olie Extra Vierge Extra Virgin Olive Oil Extra Vierge Olijf Olie Extra Virgen Company: Bertolli Albert Heijn B.V. Albert Heijn B.V. Jumbo Supermarkten B.V. Carbonell
Brand: Filippo Berio Iliada Monini Pietro Coricelli
Product: Extra Virgin Olive Oil Extra Vierge Olijfolie Olio Extra Vergine Di Oliva Olio di Oliva Company: Salov North America Ambrosia Monini S.P.A. Pietro Coricelli S.P.A.
Chocolate Spread
Brand: Ferrero ah Basic Jumbo Penotti
Product: Nutella Chocospread Hazelnoot choco pasta Original Penotti
Company: Ferrero B.V. Albert Heijn B.V. Jumbo Supermarkten B.V. Penotti
Corn Flake Cereal
Brand: Kellogg’s ah Albert Heijn ah Basic Jumbo
Product: Corn Flakes Special Flakes Corn Flakes Corn Flakes
Margarine - Light
Brand: Becel ah Jumbo Brio Zeeuws Meisje
Product: Becel Light Bewust Light Bewust Light Brio Light Zeeuws Meisje Light
Company: Unilever Albert Heijn B.V. Jumbo Supermarkten B.V. Unilever Unilever
Body Wash - Coconut
Brand: Nivea Jumbo Dove Fa
Product: care & coconut Coconut Douchecreme Nourishing Secrets cream & oil
Company: Beiersdorf Jumbo Supermarkten B.V. Unilever Schwarzkopf & Henkel
Ground Coffee
Brand: Douwe Egberts Jumbo Perla Rood
Product: Aroma Rood aroma traditioneel Huisblends Aroma Rood
Company: Jacobs Douwe Egberts Jumbo Supermarkten B.V. Albert Heijn B.V. UCC Coffee Benelux B.V.
Fabric Softener - Pink
Brand: Robijn Jumbo ah
Product: Robijn Wasverzachter Zachte Was Was zacht Company: Unilever Jumbo Supermarkten B.V. Albert Heijn B.V.
Brand Characteristics. Based on the final given sample, a total of N = 50 brands were to be coded for analysis, inclusive of N = 10 leading brands (20%) and N = 40
subordinate brands (80%). Subordinate brands were evenly sourced from Jumbo (N = 20) and Albert Heijn (N = 20). Of all subordinate brands, 46% (N = 23) was generic supermarket owned and 34% (N = 17) was a normal competitor. The majority of all brands were owned and manufactured in The Netherlands (68%), while 26% was sourced from other parts of Europe (including 10% from Italy) and 6% derived from the United States (e.g. Coca-Cola). In comparison, only 30% of leading brands was owned and manufactured in The Netherlands (N = 10), whereas this applied to 77.5% of subordinate brands (N = 40), inclusive of all (N = 23) generic brands. In terms of shelf positioning, 40% (N = 4) of leading brands were
allocated prime positions, whereas this applied to only 22% (N = 9) of subordinate brands. But of the subordinate brands that were generic (i.e. supermarket owned), a somewhat greater proportion (30.4%; N = 7) were allocated prime positions.
The average price point for all brands in total was M = €2.61 (SD = €1.98), with leading brands located above this average at M = €2.93 (SD = €1.56) and subordinate brands lower at M = €2.53 (SD = €2.08). Of the subordinate brands, the average price point for generic brands sat lowest at M = €2.11 (SD = €1.65), while the average price point for normal competitor brands sat even higher than that of leading brands at M = €3.08 (SD = €2.49) (see Figure 11).
Figure 11. Average price point by brand type grouping
Data Coding
The objective of the data sample was to determine how similar subordinate brands (either generic or normal competitor) were in comparison to the leading brand in their category, on both a semantic and literal level. Therefore, semantic and literal characteristics will only be coded for subordinate brands, based on their level of similarity to the leading brand. Given this, out of a total of fifty (N = 50) brands across ten product categories (thus, by eliminating the ten leading brands), a total of forty (N = 40) subordinate brands were to be coded on both the semantic and literal aspects to their leading brand counterpart. The
literature on this topic was used in determining the variables that aim to encapsulate each semiotic attribute, based on how literal and semantic techniques were defined and which set of attributes they overarched.
Defining variables: Literal. Literal copycatting techniques were defined as the descriptive visual elements of the low-level, more literal, common characteristics and concrete perceptual attributes which may include package shape, size and colour, label print style and layout, lettering, such as the number of common letters or similar sequences of letters, names and other graphical elements (Miceli & Pieters, 2009; Foxman, et.al., 1990; Le
Roux, et.al., 2014). Therefore, literal coding variables involved making comparisons from a leading to a subordinate brand based on precise descriptions of visual and textual attributes.
Based on this, the final variables that were determined covered comparing logo names (in terms of sound and lettering), brand font style, colour, size, composition and positioning on the label, as well as that of the product name, product description, taglines and claims. Additionally, the main packaging colour and colour combinations that were used, package shape and size, graphic or imagery representation (in terms of describing precisely the image you see) and other descriptive attributes including ingredients, safety and environmental recommendations, historical references (e.g. established 1865) and manufacturer information were compared. A working example of some of the literal characteristics that were
established can be seen in Figure 12, while the complete list of literal attributes for analysis can be found in Appendix C, accounting for thirty-three items in total.
Figure 12. Identifying the literal characteristics from a product visual
Defining variables: Semantic. Semantic copycatting techniques was determined to work via particular themes represented by pictorial and textual communication of brand positioning. In this way, it contained much higher order, more abstract themes that can be
communicated via a brand’s visual appearance, such as benefits, goals or usage context. Ultimately, semantic similarity is based on meaning and is rather an imitation of the concept than an exact copy of the product (Miceli & Pieters, 2009; Le Roux, et.al., 2014). As a result, a range of conceptual based variables were created to measure semantic techniques requiring more subjective thought on behalf of the coder because comparisons were made based on their perception of the message rather than noting down precisely what is seen. For example, a cartoon illustration of a slice of bread with Nutella on it would literally be described as ‘a slice of bread with Nutella spread on top’. But a semantic interpretation based on the visual depiction of the concept might be determined as ‘a delicious, fun snack for children’.
Based on the packaging visual as a whole, semantic variables, therefore, involved determining and comparing the target market and brand positioning of the product (e.g. is it aimed at children or adults?, males/females?, low/middle/high social classes?), the visual depiction (as described in the previous Nutella example) and key message (e.g. ‘The original Taste Coca-Cola’ may suggest that Coke’s key message is to emphasise its historically original tried and true recipe). In addition, a total of five variables were adapted from Aaker’s (1997) and Thomas and Sekar’s (2008) brand personality traits, including sincerity,
excitement, competence, sophistication and ruggedness (Aaker, 1997; Thomas & Sekar, 2008). The objective here was to compare the extent to which a subordinate brand, in the presence of a leading brand, matched in terms of portraying specific personality dimensions. For example, ruggedness can be described as appearing ‘outdoorsy’, ‘masculine’ and ‘tough’. Therefore, a corn flake cereal which graphically depicts a bowl of corn flakes, a sunrise and a rooster (both outdoor elements) may portray more ruggedness than a bowl of corn flakes in a sterile setting. A working example of some of the semantic characteristics that were
established can be seen in Figure 13, while the complete list of semantic attributes for analysis can be found in Appendix D, accounting for eight items in total.
Figure 13. Identifying the semantic characteristics from a product visual. Note: the arrows each extend from a non-descript central area, since semantic characteristics are based on the visual in its entirety and not one specific characteristic.
Pre-coding data extraction. For each of the literal and semantic variables, all relevant textual detail was extracted from the product images, entered into a chart and translated into English (e.g. Dutch brand Albert Heijn’s light margarine ‘Bewust Light’ translates to ‘Consciously Light’ in English). The purpose of this was to provide the coder with an easy cross reference of information to make like for like comparisons. An extract of the chart can be seen in Table 7.
Table 7
Coding Reference Spreadsheet – Cross comparing data extracted from the product
packaging of chocolate milk brands (leading brand versus subordinate)
Characteristic Leading Brand - Data Subordinate Brand - Data
General
Brand Name: Chocomel Jumbo
Product Name: Chocomel - halfvol Chocolade melk - halfvol Brand Owner: Friesland Campina Jumbo Supermarkten B.V.
Retail Chain: Jumbo Jumbo
Product Category: Chocolate milk Chocolate milk
Brand Type: Leading Subordinate - generic
Shelf Positioning: Prime Average
Country of Origin: Belgium The Netherlands
Literal
Logo Name: Chocomel Chocolade melk
Logo Font Style: Italics, sentence case, medium Plain, sentence case, medium
thickness thickness
Logo Font Colour: Navy blue White
Logo Composition: Sits against yellow background, Sits against a brown graphic,
over one line over two lines
Logo Positioning: Horizontal, lower top third Horizontal, lower top third Logo Size: One-fifth of package size One-fifth of package size Semantic
Sincerity: Friendly, sentimental, original, Friendly, honest, down-to-earth,
honest family-oriented
Target Market: All ages, gender neutral, middle Children, gender neutral, middle
class class
Key Message: The original chocolate milk Can be served in a variety of ways, has nutritional value Rating scale. To score and account for each variable defined, comparisons between leading and subordinate brands were to be rated on a scale from 1 (identical) to 5 (different). This degree of similarity scale was adapted from the ‘Typology of counterfeiting and
imitation forms and modalities’ tool (Le Roux, et. al., 2016) by which characteristics of product appearance, in terms of their design or packaging, were rated as either ‘Identical’, ‘Similar’ or ‘Different’ (see Appendix E). In wanting to expand the levels of similarity somewhat, van Horen and Pieters’ (2012) brand stimuli grading levels was also incorporated, in which they defined brands based on varying levels of similarity (low, moderate or high) (see Appendix F). The rating scale can be seen in Table 6.
Table 6
Degrees of similarity (DS) between a subordinate brand and its leading brand
DS # DS Rating Description 1 Identical 2 High Similarity 3 Moderate Similarity 4 Low Similarity 5 Different
Rating allocation: all semiotics. While general coding variables (e.g. brand type, country of original, shelf positioning, etc.) were coded for all (N = 50) brands (i.e. both leading and subordinate), semiotics (i.e. both literal and semantic) were coded for subordinate brands only (N = 40), based on their comparison to the leading brand in their respective product category. Subsequently, each literal and semantic variable was rated according to its variable definition using the rating scale. For example, if the packaging shape and size of a Jumbo branded cola can (i.e. the subordinate brand) is precisely the same as the packaging shape and size of a Coca-Cola cola can (i.e. the leading brand), then the packaging shape and size variable for Jumbo Cola will be rated as a 1 (i.e. identical) (see Figure 14).
Figure 14. The literal elements here in both the Coca-Cola and Jumbo Cola brands respectively, such as packaging shape, size and even material composition are precisely the same and would subsequently rate as identical features. Additionally, the main packaging colour (red) and logo colour (white) would also rate as identical features. Therefore, rating scores of 1 will be allocated to Jumbo Cola for such variables.
In another example, if the main imagery of a Bertolli olive oil bottle (i.e. the leading brand) depicts a man out in the vineyard picking olives, while competitor brand, Carbonell (i.e. the subordinate brand), depicts a woman out in the vineyard picking olives, then the semantic visual depiction for Carbonell might rate as a 2 (i.e. high similarity), as with subtle differences apparent, it very closely represented the concept of olives being hand-picked (see Figure 15 below). Yet when comparing personality traits such as ruggedness (a
predominantly male personality trait) and sophistication (a predominantly female personality trait) between these olive oil brands, a rating of a 3 or a 4 may apply here (i.e. having
moderate to low similarity), since they each respectively depict a male worker in the field in contrast to a female (see Figure 15 below).
Figure 15. The visual depiction here in both the Bertolli and Carbonell brands, respectively, are highly similar: with their arms raised up to the olive trees, each depict that the olives are hand-picked and naturally grown. There are subtle differences, including the gender of the central characters, the precise actions taking place and the landscapes being portrayed, which may represent Italy and Spain respectively.
Alternatively, if the logo colour of leading body wash brand, Nivea, is white on a blue background, and the logo of Jumbo’s generic body wash brand is red on a brown background, then this would be rated as a 5 (i.e. different). Finally, if one product (either the leading or subordinate brand, or both) does not display a presence of a particular element (i.e. not all brands include a tagline or make an historical reference, for example), then that particular variable is not coded, since no comparison can be made.
Codebook. Having defined all semiotic variables (both literal and semantic) and the rating scale to be applied, a codebook was developed detailing all relevant rules, processes and definitions with the overall mindset that any person not related to the study could easily comprehend the information provided and subsequently proceed in coding all required variables to a satisfactory level. The codebook details the units of analysis, general coding rules, variable definitions, including relevant instructions and working examples on how to
code each semiotic item for analysis, as well as an interpretation of the rating scale.
Coder training and inter-coder reliability. With the codebook developed, official coder training was undertaken to ensure all coders had understood and were in alignment with all coding requirements and variable definitions. An inter-coder reliability (ICR) test was performed using the online software dfreelon.org in order to calculate Krippendorff’s Alpha (Kalpha) by two independent coders via a convenient sample of the first subordinate brand that appeared in the first five product categories listed in sequence. This equated to 12.5% (N = 5) of the total sample size of all subordinate brands (N = 40). Three indices were used to calculate reliability scores: percent agreement, Kappa (κ) (used in all nominal level measures, applying to two or more coders working independently and accounting for
agreement by chance) and Krippendorff’s Alpha (α) (which adjusts for small sample biases) (Riffe, Lacy & Fico, 2014).
The complete list of individual intercoder reliability scores are reported in Appendix G, while Table 8 summarises the average ICR scores per grouping category. To obtain this, grouping categories were created so that variables representing similar aspects could be combined into a single score. For example, variables focusing solely on the imitation of literal based colours (e.g. logo font colour, main packaging colour, colour combinations, etc.) or variables focusing solely on the imitation of semantic based personality traits (e.g.
Table 8
Intercoder Reliability – Average Scores per Categorya
Variable Category Agreement (%) Kappa (κ) Alpha (α)
General Variables
General 97.14% .96 .96
Literal Variables
Words & Lettering 85% .78 .78
Logo Graphic Style 65% .49 .46
Text Position / Composition / Size 51% .29 .33
Colours 60% .50 .47
Graphic Illustration / Imagery 27% .05 .10
Packaging Elements 77% .63 .65
Package Attributes 87% .76 .79
Semantic Variables
Personality 16% .03 -.19
Conceptual Interpretation 40% .18 .08
a. Grey shaded variables indicate unacceptable intercoder reliability scores
While an alpha score of .8 indicates adequate reliability, variables with alphas as low as .67 were considered as acceptable for drawing tentative conclusions for this thesis (Riffe, et. al. 2014). For all ‘general’ based variables (e.g. country of origin), the average level of agreement stood at 97.14% (α = .96), indicating adequate reliability. Of the literal based grouping categories, variables relating to words and lettering (e.g. logo name) and packaging attributes (e.g. package shape) achieved acceptable reliability scores (α = .78 and α = .79 respectively), while all remaining literal based grouping categories sat below the acceptable threshold. Additionally, both semantic based grouping categories (personality traits and
conceptual interpretation) fell below the acceptable threshold (α = -.19 and α = .08
respectively). Further leniency of reliability scores might be achieved by reducing down the variation within the scale, but a level of strictness in reporting the results was preferred. The scores though did raise the highly interpretive nature of deriving meaning from a product visual, and as a consequence, further refinements to the codebook were undertaken, including more detailed definitions and simplified explanations. The finalised codebook is located at the end in Appendix M.
Data Preparation
Independent variables. The independent variables to be used for analysis will be inclusive of all subordinate brands (i.e. all comparison brands). In addition, comparisons between the two types of subordinates that were identified; generic supermarket brands and normal competitor brands will be made.
Dependent variables. In total, thirty-three literal variables and eight semantic variables were coded for, totalling forty-one different semiotic measures. All thirty-three literal variables were computed into one new variable: ‘LiterallAll’ by obtaining the average of all literal scores present. Similarly, all eight semantic variables were computed into one new variable: ‘SemanticAll’ by obtaining the average of all semantic scores present. Finally, to obtain a complete semiotic score, all thirty-three literal variables and eight semantic variables were computed into one new variable: ‘LiteralSemanticAll’ by obtaining the average of all semiotic based scores present.
Using the grouping categories previously created to generate average intercoder reliability (Appendix H), all thirty-three individual literal variables and eight individual semantic variables were condensed further into a range of sub-groups by computing the average scores present for the variables making up that particular sub-group. All literal variables were condensed down to seven sub-group variables, while two sub-group variables
were created for semantic attributes (see Appendix I).
Finally, an additional four new variables were computed involving further
combinations of similar attributes both within and between semiotic types. This included two variations of graphic based attributes of a literal nature; ‘LiteralGraphicBased’ was computed by obtaining the average of all scores present for ‘FontStyle’, ‘Colours’ and ‘Graphics’. While ‘LiteralGraphicsAll’ additionally included ‘TextPosiCompSize’. ‘TextAll’ was a variable computed from all sub-groups across both literal and semantic attributes that related to the imitation of text (i.e. ‘WordsLetters’ from literal based techniques was combined with ‘KeyMessage’ from semantic based techniques). Similarly, ‘GraphicsBasedAll’ was
computed from a combination of graphic based imitation techniques of a literal nature (e.g. colours, graphics, font style) with graphic based imitation of a semantic nature (e.g. visual depiction). The complete breakdown of the new text and graphics based variables can be viewed in Appendix J.
Tests. For each of the three main hypotheses, a series of sub-questions were
generated, tested and reported. Paired t-tests analyses were run for the hypothesis 1 group of tests. With the main hypothesis testing the comparison between literal and semantic
techniques for all subordinate brands, additional tests were run to determine how this differed for generics and normal competitors. Additionally, similar analyses were conducted focusing solely on text and graphic based techniques (i.e. literal versus semantic). For the hypothesis 2 group of tests, independent t-test analyses were conducted to compare the extent to which semiotics (both literal and semantic techniques combined) differed between generics and normal competitor brands. Similarly, analyses were conducted that focused solely on text and graphic based techniques (i.e. literal and semantic combined). Finally, for the hypothesis 3 group of tests, again, independent t-test analyses were conducted to compare the extent to which literal techniques differed between generics and normal competitor brands, and
separately, the extent to which semantic techniques differed between generics and normal competitor brands. Analyses which focused solely on text and graphic based techniques (literal or semantic) were also conducted. See Appendix K for the complete reference list of tests conducted.
Results
Hypothesis 1: Extent of Literal Versus Semantic Copycatting Techniques by Subordinate Brands
This section focuses on the comparison between the extent of literal versus semantic copycatting techniques utilised by subordinate brands in the presence of a leading brand. In addition, analyses were conducted to see if similar patterns could be observed solely by generic supermarket brands, as well as normal competitor brands. Finally, specific types of literal versus semantic techniques were also compared. A results matrix for this section can be found in Table 9.
Using a paired t-test analysis, there was shown to be a significant difference in the mean scores for copycatting techniques of a literal nature (M = 2.89, SD = .40) and copycatting techniques of a semantic nature (M = 2.68, SD = .59); t(39) = 2.84, p<.01 by subordinate brands. These differences in means indicate that in the presence of a leading brand, subordinate brands apply copycatting techniques of a literal nature to a greater extent than semantic techniques, thus confirming hypothesis 1a. Therefore, the opposite assumption, that in the presence of a leading brand, subordinate brands apply copycatting techniques of a semantic nature to a greater extent than literal techniques was disproved, and thus hypothesis 1b was rejected.
To determine if a similar result occurred solely for generic brands, a paired t-test analysis showed no significant difference in the mean scores for copycatting techniques of a literal nature (M = 2.87, SD = .40) versus copycatting techniques of a semantic nature (M = 2.71, SD = .63); t(22) = 1.41, p>.05. Thus, the assumption that generic brands who copycat leading brands apply techniques of a literal nature to a greater extent than semantic
techniques was disproved. But when conducting the same test for normal competitor brands, the mean scores for copycatting techniques of a literal nature (M = 2.91, SD = .40) and
copycatting techniques of a semantic nature (M = 2.62, SD = .55) was significant, t(16) = 3.04, p<.01. Thus confirming the assumption that normal competitor brands who copycat leading brands apply copycatting techniques of a literal nature to a greater extent than semantic techniques.
To narrow down further on the types of semiotics used, it was decided to focus on specific copycatting techniques used at both literal and semantic levels. By focusing on text based attributes, it was similarly considered that brands who copycat leading brands would apply copycatting techniques of a literal nature (i.e. via the imitation of words and letters) to a greater extent than semantic techniques (i.e. via the imitation of key messages). Using a paired t-test analysis, there was shown to be a significant difference in the mean scores for text based copycatting techniques of a literal nature (M = 3.22, SD = .85) and text based copycatting techniques of a semantic nature (M = 2.75, SD = .78); t(39) = 3.00, p<.01 by subordinate brands, thus confirming this assumption.
To determine if a similar result occurred solely for generic brands, a paired t-test analysis similarly showed a significant difference in the mean scores for text based
copycatting techniques of a literal nature (M = 3.31, SD = .78) and copycatting techniques of a semantic nature (M = 2.70, SD = .82); t(22) = 3.03, p<.01 by generic brands, thus
confirming this assumption. Yet in contrast, a paired t-test analysis comparing mean scores for text based copycatting techniques of a literal nature (M = 3.11, SD = .94) and text based copycatting techniques of a semantic nature (M = 2.82, SD = .73) by normal competitor brands was not significant, t(16) = 1.13, p>.05.
Graphic based attributes were also a specific focus and it was similarly considered that brands who copycat leading brands using graphic based attributes, apply copycatting techniques of a literal nature (e.g. via the imitation of illustration and imagery) to a greater extent than semantic techniques (i.e. via the imitation of visual depiction). Tests were
