Green is the new color of
the season
An experimental study investigating the effects of
sustainability branding for established brands.
Master Thesis in Persuasive Communication 1/30/2015
Name: Sarah Semiramis Saffnauer
Student Number: 10602453
Email: s.saffnauer@gmx.de
Abstract
Drawing from associative network theory, this experimental study investigated the
effects of a sustainability branding strategy versus a regular branding strategy on consumer
responses (brand attitude, brand equity, green brand equity and willingness to pay a price
premium) towards an established fashion brand. It was furthermore investigated how
sustainability self-identity and prior brand attitude moderate this relationship. Contrary to
expectations the sustainability strategy had a negative effect on brand attitude, but a positive
on brand equity and willingness to pay a price premium. Green brand equity was the only
response measure that was moderated by sustainability self-identity. Furthermore,
sustainability self-identity negatively influenced the relationship between prior brand attitude
and brand attitude. However, the proposed three-way interaction between branding strategy,
prior brand attitude and sustainability self-identity was not found. The present data suggests
that established companies should only cautiously use sustainability branding. Limitations
Introduction
Sustainability sells: Consumers are willing to pay a premium price for ethically
produced products (Pelsmacker, Driesen and Rayp, 2005), and a marketing strategy focusing
on the “green” aspects of the product increases consumers’ favorable attitude (Hartmann,
Ibanez, & Sainz, 2005). Also the fashion industry, the focus industry in this study, tries to
fight its problems by focusing on sustainability. Sustainable fashion becomes increasingly
important in juxtaposition with fast fashion, that is, cheaply produced apparel with a short life
cycle. How frequent fast fashion scandals lead to serious image damage is depicted by the
media and consumer outcry after a factory in Bangladesh collapsed in which Western fast
fashion companies produced their clothes (The Guardian, February 24, 2014) or after
messages reporting inhumane working conditions were found stitched into clothes of a fast
fashion brand (The Guardian, June 25, 2014). These (PR) catastrophes require fashion brands
to distance themselves from the bad reputation that comes with fast fashion. One possible
solution is a sustainability branding strategy.
But sustainability is a broad concept. There are various definitions of sustainability;
one most frequently cited is from the WCED, which described sustainability as “development
that meets the needs of the present without compromising the ability of future generations to meet their own needs” (WCED 1987, p. 43). Vermeir and Verbeke (2008) explain
sustainability as a construct consisting of economic (fair price for producer and consumer),
ecological (care for natural environment), and social (socially acceptable production
processes) aspects. Thus, sustainability does not only incorporate “green”, that is
environmentally friendly produced products, but also related concepts like fair trade.
According to this definition sustainably produced fashion is characterized, for
circumstances. But a sustainability branding strategy also has to take the target, the
consumer, into account. After analyzing behavioral data Gilg, Barr, and Ford (2005) found
that there is no average environmental person but that the range is from committed
environmentalists, over mainstream and occasional environmentalist, to
non-environmentalists. Also, Van der Werff, Steg and Keizer (2013a) found that it depends on the
level of environmental self-identity for someone to be intrinsically motivated to act
pro-environmentally. It seems therefore counter-intuitive that everyone would react in the same
way to a sustainability branding strategy. That is, whereas an environmentalist may react
positively towards sustainably branded fashion a non-environmentalist in contrast might not
care at all because he is not intrinsically motivated to do so.
Besides the consumer’s level of environmentalism, another consumer characteristic
that is important for the effect of sustainability branding strategy on consumer response is the
attitude the consumer holds towards the brand prior to exposure to the sustainability strategy.
A sustainability branding strategy might change a negative attitude in a positive direction if
the consumer is concerned with sustainability issues. Hodges and Wilson (1993) explain that
an attitude change can occur when presented information is incongruent to the held attitude;
to resolve this incongruence the attitude is changed. This would mean that, for instance, if a
consumer holds a negative attitude towards a fashion brand and is generally concerned about
sustainability issues and is then confronted with a sustainability branding strategy he might
change his attitude in a positive direction to resolve the incongruence between his negative
attitude towards the brand and his positive attitude towards the sustainability branding.
There has been abundant research into green advertising and branding but this often
comprised only fictional or new brands or just an industry without specific focus on any
brand (e.g., Lee, Hsu, Han, & Kim, 2010; Kalafatis, Pollard, East, & Tsogas, 1999; Lee,
address this issue the present study uses an established brand, namely the Swedish fashion
house H&M, as an example. The global retailer comes as a natural choice since it is one of
the biggest fashion producers, which indicates that the brand is known and thus, that people
have an attitude towards it. Additionally, it also uses a sustainability branding strategy for
one of its smaller collections, the Conscious collection, which increases the ecological
validity of the study (H&M, 2013).
The current study aims to link the effects of sustainability branding to the personality
trait of sustainability self-identity and to prior brand attitude. More precisely, it is
investigated how sustainability branding shapes the perception of an established brand and if
this effect is moderated by the sustainability self-identity of the consumer and the attitude
held prior to exposure to the strategy.
To the author’s knowledge there is still no clear answer on how a sustainability
branding strategy works for an established company, whether it works differently for people
interested in the issue and whether their prior brand attitude plays a role in the relationship.
First, this paper can offer valuable insights for PR departments of fashion companies, as it
investigates if sustainability strategies contribute to the company’s reputation. Second, and
more importantly, scientifically supported success of sustainability strategies could help to
increase the production of sustainable products, which is very desirable from a societal point
of view since they come with sustainable production, fair trade and improved working
conditions.
Theoretical Framework
Sustainability branding and associative network theory
Companies are facing increasing legislative and societal pressure to produce in a way
pressure is to put the focus on the sustainable actions of the company and to brand the
company as sustainable. Derived from the sustainability definition of Vermeir and Verbeke
(2008) mentioned above, sustainability branding would encompass brand communication that
focuses, for instance, on the environmental performance or on the fair trade activities of a
brand and its products. Brands, more generally, are defined as associative network in the
consumer’s long-term memory (Franzen, 2009). This indicates that if the brand is perceived as sustainable it should share the same favorable associations that are linked to its
sub-concepts like environmentally friendly or fair trade behavior in the memory network.
One way to explain how a brand can in the end profit from sustainability branding is
by taking a closer look at how brands are stored in the memory. The associative network
theory describes how the memory works (Plaut, 1995; Anderson, 1983). Information is stored
in nodes and different nodes are interlinked. Nodes which are frequently activated together
share more links than nodes that represent information that is not commonly activated
together. If a cue activates one of the nodes the activation can spread from the activated node
to the other nodes that are closely linked to it and make the stored information more
accessible. This indicates that through the node network brands have certain meanings
attached to them. The brand Chanel might, for instance, be closely linked to luxury, and a
brand that is branded as sustainable should belong to a network that has the same sustainable
focus.
Studies have shown that sustainable behavior is linked in memory with the feeling of
pride (Onwezen, Bartels, & Antonides, 2014) or the belief of doing the right thing (Arvola,
Vassallo, Dean, Lampila, Saba, Lähteenmäki, & Shepherd, 2008). So if people behave in a
sustainable manner, the node of sustainable behavior is activated and the activation can
spread to the nodes of pride and doing the right thing, as a result, these feelings are elicited in
node network. Thus, when the person, for instance, sees the brand or thinks of it not only is
the node of the brand elicited but also its network, i.e., a feeling of pride or the belief of doing
the right thing.
The positive feelings that are elicited when confronted with a sustainably branded
product are a prerequisite for a positive brand attitude. Hartmann et al. (2005) found that
green branding leads to more favorable attitudes towards the brand. This supports once again
the findings of Onwezen et al. (2014) and Arvola et al. (2008) that sustainable behavior is
linked to pride and the feeling of doing the right thing. Thus acting sustainable, i.e., buying a
sustainable brand elicits these positive feelings and in turn leads to a more favorable attitude
towards the brand. Arguably, green branding is not the same as sustainability branding but
since Vermeir and Verbeke (2008) describe “green” or ecological awareness in their
sustainability definition as sub-concept of sustainability similar effects are expected –
especially as most consumers do not know the difference.
These positive feelings that can be linked to the brand through sustainability branding
are also desirable from a brand equity point of view. Brand equity, the value of a brand, can
be defined in various ways. From a financial perspective brand equity describes the cash
flows that a company receives additionally compared to when the products are sold
unbranded (Simon & Sullivan, 1993). Keller (1993) describes consumer based brand equity
with the difference in response towards the marketing of a brand when the consumer knows
the brand versus when he does not know it. This means that consumers react differently to,
for instance, a T-shirt with a well-known logo than to a shirt without this logo. This concept
is based on brand awareness and brand image, meaning that for high brand equity the brand
does not only need to be known, but consumers also need to have a positive attitude towards
sustainability branding. Thus, sustainability branding influences brand attitude positively
which in turn influences brand equity.
Brand equity describes the general value of a brand. To put sustainability into focus
Chen (2010) incorporates the idea of environmentally friendly branding into the concept of
brand equity. He introduced the new concept of green (i.e., environmentally friendly) brand
equity and describes it as “a set of brand assets and liabilities about green commitments and
environmental concerns linked to a brand, its name and symbol that add to or subtract from the value provided by a product or service.” (p. 310). This means that the green value of a
brand is constructed in the consumers mind through knowledge about, for instance,
environmentally friendly production processes, or the usage of organic materials. But at the
same time the green value of the brand can be perceived as very low when nothing or only
negative things are known about its environmental performance. Thus, the greener a brand is
perceived by the consumer the more valuable it is in green brand equity terms. Green brand
equity is closely related to brand equity, however unlike brand equity it only captures a small
aspect of the whole brand, therefore it cannot be assumed that the same relationship between
sustainability branding, brand attitude and brand equity applies. It is therefore expected that
sustainability branding directly influences green brand equity and does not take the detour
over brand attitude.
Sustainable claims do not only influence the value of and attitude towards a brand by
activating positive feelings but they also influence the consumer’s willingness to pay a price
premium. As de Pelsmacker et al. (2005) found in their study on willingness to pay for
fair-trade coffee, consumers are willing to pay on average 10 percent more for a coffee with a
fair-trade label on it than for a conventional coffee. A small number of respondents, the so
called fair-trade lovers, were ready to pay a price premium of 36 percent. Of course the brand
ready to pay more for a branded product than for an unbranded one (Lassar, Mittal, &
Sharma, 1995). However, this relationship goes over the scope of the study and is not further
taken into account.
These different research examples show that sustainability or its sub-concepts
influence the consumer in his perception of the brand and his actual behavior. In the present
study about the Conscious collection of H&M and their sustainable actions, it is therefore
expected that a sustainability branding strategy leads to a more favorable consumer response
than a regular branding strategy that does not involve sustainability.
Hypothesis 1: The sustainability branding strategy leads to a higher willingness to
pay a price premium, to higher green brand equity, and to a more favorable brand
attitude which in turn leads to higher brand equity than the regular branding strategy.
Sustainability self-identity
The sustainability self-identity of the consumer influences, how sustainable a
consumer behaves. A related concept that has been discussed in connection with
sustainability branding and marketing is environmental self-identity. It describes according to
Van der Werff, Steg and Keizer (2013b) the extent to which a person sees himself as
environmentally friendly. However, environmental self-identity does not fully grasp the
meaning of sustainability: the concepts of social and economic sustainability are disregarded.
Thus, a person with a strong sustainability self-identity is expected to put emphasis on for
instance organic food, use of green energy, and fair trade. But since environmental
self-identity is a sub-concept of sustainability self-self-identity it is expected that the effects which are
related to environmental self-identity also hold true for sustainability self-identity.
A number of studies tested environmental self-identity for instance as a mediator
consideration, and environmental preferences, intentions, and behavior (Van der Werff et al.,
2013b), or as an obligation-based intrinsic motivation that influences environmentally
friendly behavior (Van der Werff et al., 2013a). Whitmarsh and O’Neill (2010) found
environmental self-identity as a significant predictor for a number of environmentally
friendly behaviors ranging from waste reduction, regular water and domestic energy
conservation, to eco-shopping and eating. It can be derived from these studies that
environmental self-identity is a personality trait and hence influences inwardly how
consumers make decisions.
How this personality trait influences behavior should also be discussed in connection
to the associative network theory. The theory suggests that people act environmentally
friendly to elicit positive feelings that are connected in the memory to this kind of behavior
(Plaut, 1995; Anderson, 1983). Additionally, as derived from Van der Werff et al.’s (2013a;
2013b) and Withmarsh’s (2010) research people can have the personality trait of
environmental self-identity. This trait shows itself through an intrinsic motivation to act
pro-environmentally. So people with strong environmental self-identity do not only act
environmentally friendly to elicit feelings of pride and doing the right thing (Onwezen et al.,
2014); Arvola et al., 2008) but also because they are intrinsically motivated to do so, because
sustainable behavior is part of their personality. Consequently, a brand which people with
strong environmental self-identity perceive as sustainable can trigger these positive feelings
and should be additionally preferred due to their personality.
Why people follow their personality with actions can be explained in a number of
ways. People can act according to their self-identity to prevent an anticipated cognitive
dissonance (Akerlof, & Dickens, 1982) or to avoid anticipated guilt (Van der Werff et al.,
2013a). People want to be consistent, thus not acting sustainable while having a strong
inconsistency, or to guilt. Thus, people with strong sustainability self-identity prefer to act
accordingly. However, people can also carry out sustainable behavior because they anticipate
the pride that they will feel when acting according to their environmental self-identity
(Onwezen et al., 2014). In sum, people with a strong sustainability self-identity are more
likely than people with a weak sustainability self-identity to have a positive attitude towards a
sustainable brand, to value it higher and also to be willing to pay a higher price for it because
of reasons like consistency with their personality, anticipated guilt when not acting according
to the personality or anticipated pride when acting accordingly.
De Pelsmacker et al.’s (2005) study took a closer look at the relationship between
sustainability self-identity and willingness to pay a price premium for a sustainable product.
People with a strong sustainability self-identity are closely linked to the fair-trade lovers de
Pelsmacker et al. (2005) identified. The researchers clustered the respondents in their study in
fair-trade lovers who show a clear preference for the fair-trade label, fair-trade likers, whose
preferences are balanced but still with the fair-trade label as the most important one, then
flavor lovers for whom flavor is the most important choice criteria, and brand lovers who
have a strong focus on the brand they are purchasing. The two latter ones are rather negligent
towards fair-trade. Fair-trade lovers were ready to pay up to a 36 percent price premium for
fair-trade coffee and fair-trade likers were still ready to pay 17 percent more. The flavor
lovers and brand lovers were ready to pay only four and three percent more, respectively.
Thus it can be derived that a sustainability self-identity positively influences consumers’
willingness to pay a price premium for the sustainable product, because they not only elicit
positive feelings that are linked in the memory to sustainable behavior but they also avoid
guilt and a cognitive dissonance in doing so.
As mentioned above (Withmarsh & O’Neill, 2010; Van der Werff et al., 2013a; Van
behavior and thus it is expected that sustainability self-identity has the same effects. Based on
the previous research reviewed, this study expects to find that the positive influence of a
sustainability branding strategy on brand attitude, brand equity, green brand equity, and
willingness to pay a price premium is boosted by a high sustainability self-identity. A low
sustainability self-identity is not expected to lead to unfavorable evaluations of sustainably
branded products, because consumers have learnt through the environmental trend in society
to associate green with good. But it is expected that the positive effect of a green branding
strategy is magnified for people with a high sustainability self-identity. Therefore it is
hypothesized:
Hypothesis 2: The effect of the sustainability branding strategy is moderated by
sustainability self-identity. People with high sustainability self-identity will report a
higher brand attitude, brand equity, green brand equity, and willingness to pay a price
premium after exposure to the sustainability branding strategy than people with low
sustainability self-identity. No moderation effect by sustainability self-identity is
expected in the control group.
Interaction between condition, prior brand attitude, and sustainability self-identity
As described in the introduction, companies like H&M suffer from scandals linked to
fast fashion. Not every consumer might therefore have a positive attitude towards the
Swedish fashion house. Hartmann et al. (2005) found in their study that green branding leads
to a more favorable brand attitude, however, they did not investigate if people were positive
or negative towards the brand before being exposed to the green strategy. This gives rise to
questions like whether sustainability branding only works when people are not negatively
predisposed towards the brand, or whether sustainability self-identity plays a role when a
Already in 1955 Osgood and Tannenbaum explained how attitude change may arise.
They argue that attitude change depends on whether congruent or incongruent information
qualifies the prior attitude. The attitude changes to resolve incongruity, and moves towards
congruity. A person has, for instance, a negative attitude towards H&M. He is then
confronted with information about the sustainable behavior of the brand, which is important
to the person because he has a strong sustainability self-identity. Thus the person experiences
an incongruity between his negative attitude towards the fashion retailer and his positive
attitude towards the information he received. Consequently, the attitude changes in the
positive direction to resolve this incongruity. Hodges and Wilson (1993) also summarize
reasons for attitude change. They report that attitude change occurs when reasons come to
mind that are incongruent with the hold attitude. To resolve this conflict the attitude is
changed. Both studies report that an attitude change does not occur when the new information
is not incongruent. Congruent information supports the already hold attitude but does not
change it.
The rationale of the two studies can be applied to the present study to qualify the
proposed three-way interaction of prior brand attitude, sustainability self-identity, and
condition. A positive attitude towards H&M is not challenged by a sustainability branding
strategy, because the valence of the information is also positive. Sustainability self-identity
should not qualify this relationship because even people who are not interested in sustainable
actions are not against them. Even if they are sceptic due to their low sustainability
self-identity they should not be motivated enough to challenge the information because no
incongruity forces them to do so. The only change expected in this model is that a negative
prior brand attitude gets challenged by the sustainability strategy which is qualified by a high
sustainability self-identity. If sustainability concerns are not important to the person there is
between his negative attitude and the presented sustainability branding strategy. However, if
sustainability concerns are important to the person there is reason for attitude change because
the incongruity between the sustainability branding strategy and the hold negative attitude
must be resolved. As a result the person’s attitude shifts from a negative to a more positive
direction. Thus the following hypotheses are derived:
Hypothesis 3.1: If a person who holds a negative prior brand attitude and who has a
high sustainability self-identity is presented with the sustainability branding strategy
then his attitude changes in a positive direction.
Hypothesis 3.2: No attitude change is expected when the person has either (1) a
positive prior brand attitude, high sustainability self-identity and is presented with the
sustainability branding strategy; (2) a positive prior brand attitude, high sustainability
self-identity and is presented with the regular branding strategy; (3) a negative prior
brand attitude, high sustainability self-identity and is presented with the regular
branding strategy; (4) a positive prior brand attitude, low sustainability self-identity
and is presented with the sustainability branding strategy; (5) a positive prior brand
attitude, low sustainability self-identity and is presented with the regular branding
strategy (6) a negative prior brand attitude, low sustainability self-identity and is
presented with the sustainability branding strategy; or (7) a negative prior brand
attitude, low sustainability self-identity and is presented with the regular branding
strategy.
Table 1 offers an overview over the different hypotheses.
Hypotheses Prior brand attitude Sustainability self-identity Branding Strategy Attitude change
H3.1 - + Sustainable Attitude change
H3.2.2 + + Regular No attitude change
H3.2.3 - + Regular No attitude change
H3.2.4 + - Sustainable No attitude change
H3.2.5 + - Regular No attitude change
H3.2.6 - - Sustainable No attitude change
H3.2.7 - - Regular No attitude change
Table 1: Hypotheses 3 regarding attitude change
The conceptual model offers an overview of all Hypotheses made. Hypothesis 1
describes the main effect of condition on consumer response (for brand equity mediated via
brand attitude). Hypothesis 2 shows how sustainability self-identity moderates this main
effect and Hypothesis 3 depicts the proposed three-way interaction between condition, prior
brand attitude and sustainability self-identity.
Fig. 1: Conceptual model
Method
Participants
187 people participated in the study, 58.8 percent of the participants were female and
the age ranged between 20 and 60 years. The mean age was M = 25.28, SD = 4.94. The
participants were from 27 different countries with Germans being the biggest group (60.4
Design and procedure
The study used a 2 (stimulus: sustainable vs. regular) by 2 (sustainability self-identity:
low vs. high) by 2 (prior brand attitude: low vs. high) between subjects factorial design. After
the participants gave informed consent they were asked demographical questions (gender,
age, and their country of origin). This was followed by the measurement of prior brand
attitude. In order to prevent hinting participants that the study is about the fashion brand
H&M they also had to state their attitude towards the fashion brands Zara, Mango and
Abercrombie & Fitch. H&M was placed second between Zara and Mango in the list.
Participants were then asked to indicate their sustainability self-identity. This was followed
by the presentation of the stimulus material.
Participants were randomly assigned to either the sustainability branding strategy or
the regular branding strategy. After reading the information they were asked to indicate their
brand equity, their green brand equity and their willingness to pay a price premium.
Depending on which stimulus they were presented with they either had to indicate their
willingness to pay a price premium for a pullover from the sustainability collection or from
the regular collection. The last question was the second brand attitude measure. Here
participants only had to indicate their attitude towards H&M. Participants were thanked for
their participation and were given the opportunity to write a comment about the study.
Stimulus material
The stimulus material for the sustainability branding strategy consisted of information
about the H&M Conscious collection, their commitments and one example action that H&M
does to follow the commitments. Also two pictures with male and female models wearing
clothes from the Conscious collection were incorporated. The material for the regular
branding strategy consisted of information about H&M’s business concept, the collections
both phrased as commitments and were kept as similar as possible. Also, the same two
pictures as in the sustainability stimulus were incorporated in the regular stimulus. To
enhance ecological validity the presented information about the sustainability collection was
from H&M’s official website; all used pictures were from official H&M campaigns. To keep
the stimuli as similar as possible the material for the regular branding strategy was phrased by
the researcher but only with information found on the H&M website. Also the amount of text
was similar with 192 words in the stimulus for the sustainability branding strategy and 198
words for the regular one.
Fig. 2: Stimulus material for sustainability branding strategy (left) and for regular branding
strategy (right)
To ensure that the sustainability strategy is perceived as more sustainable than the
regular strategy the stimuli were pretested. One group evaluated the sustainability stimulus on
a seven-point Likert scale (totally disagree – totally agree) regarding the three concepts hip
(economic sustainability), environmentally friendly, more beneficial to the environment than
other brands (environmental sustainability), socially acceptable production processes (social
sustainability)), the other group did the same for the regular branding strategy. The concepts
hip and durable were retrieved from Madrigal and Boush (2008), the items measuring
sustainability were created by the researcher with respect to the sustainability definition by
Vermeir and Verbeke (2008). Only differences for the concept of sustainability are important
for this research, however to avoid producing desirability effects when only asking about this
concept the other two concepts were incorporated as well. The respondents also had to
indicate their gender, age, and nationality.
Twenty students participated in the pretest. Eighty percent were female, the mean age
was M = 21.5 years, SD = 2.37, and 60 percent of the participants were Dutch. The principal
component analysis revealed one factor with an eigenvalue of 2.53 and an explained variance
of 63.23 percent. The item “more beneficial to the environment than other brands” was
removed to increase the scale reliability from Cronbach’s alpha = .78 to Cronbach’s alpha =
.80. Equal variances could be assumed for the independent sample T-test with F = 1.88, p =
.186 for Levene’s test for equality of variances. There was no significant difference in the
scores for the sustainable stimulus (M = 3.90, SD = 0.83) and the regular stimulus (M = 3.06,
SD = 1.20); t(18)= -1.80, p = .089. However the means go in the proposed direction and the
insignificant results could be due to the small sample size.
However, as a consequence of the pretest the sustainable stimulus was made stronger
through reporting higher numbers for H&M’s sustainable actions (e.g. 55.8 percent of
sustainable produced cotton was increased to 65.8 percent); also the instruction in the
questionnaire was made clearer that people should use the presented information to guide
Measures
(Prior) Brand attitude was measured on a seven-point semantic differential scale. The
items used to describe the brand were good – bad, like very much – dislike very much, and
pleasant – unpleasant (Gardner, 1985). The brand attitude measured prior to exposure to the
stimulus revealed one factor (eigenvalue = 2.58) in the principal component analysis with an
explained variance of 86.12 percent. The scale’s reliability was excellent with a Cronbach’s
alpha of α = .92.
The brand attitude construct that was measured after the presentation of the stimulus
also had one factor (eigenvalue = 2.73) in the principal component analysis and explained
91.01 percent of variance. The scale was very reliable with a Cronbach’s alpha of α = .95.
Brand equity was measured with the overall brand equity scale introduced by Yoo and
Donthu (2001). The construct was measured with four items on a seven-point scale (totally
disagree – totally agree): “It makes sense to buy H&M instead of any other brand, even if
they are the same”, “Even if another brand has the same features as H&M, I would prefer to buy H&M”, “If there is another brand as good as H&M, I prefer to buy H&M”, and “If another brand is not different from H&M in any way, it seems smarter to purchase H&M”.
This specific brand equity scale was chosen for this research because the measurement of the
construct “Green brand equity” by Chen was based on this scale. The principal component
analysis revealed one factor with an eigenvalue of 3.26 and an explained variance of 81.42
percent. Reliability was excellent with a Cronbach’s alpha of α = .92.
Green brand equity was measured on a seven-point scale (totally disagree – totally
agree) with four items: “It makes sense to buy H&M instead of another brand because of its
environmental commitments, even if they are the same”, “Even if another brand has the same environmental features as H&M, I would prefer to buy H&M”, “If there is another brand’s environmental performance as good as H&M’s, I prefer to buy H&M”, and “If the environmental concern of another brand is not different from that of H&M in any way, it
seems smarter to purchase H&M”. The items were retrieved from Chen (2010). The
principal component analysis revealed one factor with an eigenvalue of 3.03 and an explained
variance of 75.62 percent. The reliability was good with Cronbach’s alpha of α = .89.
Like in the article of Strahilevitz (1999) willingness to pay a price premium was not
measured in absolute numbers but in ratios. Respondents were either presented with a piece
of apparel either framed as from the regular or from the sustainable collection depending on
in which condition they were in. The price for the apparel was 30 Euros. The respondents
than read that H&M plans to raise its prices due to the risen costs of their cotton supply. They
were asked to indicate on a scale from 0 – 100 percent how much more they were willing to
spend for this piece of apparel.
Sustainability self-identity was measured on a seven-point scale (totally disagree –
totally agree) with the items of the construct environmental self-identity. The items were
“Acting environmentally friendly is an important part of who I am”, “I am the type of person
who acts environmentally friendly”, and “I see myself as an environmentally friendly person”. The scale was retrieved from Van der Werff et al. (2013b). Additionally, the scale
was complemented with four items from the fair trade scale from Tanner and Wölfing Kast
(2003). “Solidarity with third-world countries is important to me”, “I would refrain from
buying products if I were not sure whether the workers were fairly paid”, “When buying a product, I pay attention to fair trade labels”, and “I would be willing to pay a higher price to support small growers from third-world countries”. Overall, the chosen items should capture
the concepts of environmental, social, and economic sustainability. The principal component
factor analysis with Varimax rotation revealed two components with an eigenvalue above 1.
The first component (eigenvalue = 4.36) and the second component (eigenvalue = 1.13)
explained a total of 87.45 percent of variance. The scale was very reliable with a Cronbach’s
Results
Treatment check
94 participants were assigned to the regular condition, 93 to the sustainability
branding condition. Participants in the regular condition took on average M = 44.10 seconds,
SD = 36.12, to look at the stimulus and participants in the sustainability condition took
around M = 81.03 seconds, SD = 231.15, to look at the sustainability stimulus. The
differences were not significant, t(185) = -1.53, p = .128. When the outlier of 2251.18
seconds in the sustainability condition was excluded from the analysis the mean was reduced
to M = 57.44, SD = 41.25. These differences were now significant, t(184) = -2.35, p = .020.
Even though the word count was almost equal the stimulus in the regular condition seems to
have been a faster read.
Main effect of condition on consumer response
Hypothesis 1 proposed a main effect of condition on brand attitude, green brand
equity, willingness to pay a price premium, and an indirect effect on brand equity mediated
via brand attitude. However, the conducted MANOVA revealed no significant differences for
condition on brand attitude and green brand equity. The evaluation of brand attitude in the
sustainability condition (M = 3.42, SD = 1.42) and the regular condition (M = 3.64, SD =
1.63), F(1, 185) = 0.99, p = .321 did not differ significantly. As did not green brand equity
(Sustainability: M = 3.23, SD = 1.50; Regular: M = 3.07, SD = 1.35), F(1, 185) = 0.58, p =
.447. But the MANOVA did reveal that the willingness to pay a price premium differed
significantly between the sustainability condition and the regular condition. Respondents who
were exposed to the sustainability stimulus were willing to pay a significantly higher price
premium for a pullover from the sustainable collection (M = 1.63, SD = 1.14) than
the regular collection (M = 1.16, SD = 1.10), F(1, 185) = 8.13, p = .005. Respondents were
willing to pay 16 percent more for a pullover from the sustainable collection; this means they
would pay 34.80 Euros for the presented pullover (price set at 30 Euros). In contrast,
respondents in the regular condition would only pay 12 percent more, which are 33.60 Euros
for a pullover from the regular collection (price also set at 30 Euros).
The proposed indirect effect of condition on brand equity via brand attitude was
investigated with Process Model 4 by Hayes (2012). A bootstrapping analyses with 1000
samples revealed no mediation through brand attitude (indirect effect = 0.10, SE = 0.11, 95%
CI [-0.09, 0.32]). In a next step, prior brand attitude was controlled. The bootstrapped
mediation analysis with 1000 samples (Process Model 4) now revealed a significant indirect
effect = .09, SE = .06, 95% CI [.01; .26]. Thus, when prior brand attitude is held constant,
condition has a positive indirect influence on brand equity via brand attitude.
Prior brand attitude was controlled for the relationship between condition and brand
attitude, green brand equity, and willingness to pay a price premium. A MANCOVA with
prior brand attitude as covariate revealed a significant effect of condition on brand attitude
measured after exposure to the stimulus, F(1, 184) = 6.54, p = .011. Unexpectedly, people in
the regular condition reported a significantly higher brand attitude towards H&M (M = 3.64,
SD = 1.63) than people in the sustainability condition (M = 3.42, SD = 1.42). No significant
differences were found for condition on green brand equity when controlling for prior brand
attitude, F(1, 184) = 0.89, p = .347. Controlling for prior brand attitude had no influence on
the willingness to pay a price premium.
Despite the significant effects for willingness to pay a price premium and brand
equity, Hypothesis 1 can only be partially accepted due to the insignificant effect of condition
Moderation effect of sustainability self-identity
Hypothesis 2 proposed a moderation effect of sustainability self-identity on the
relationship between condition and consumer response. This was investigated for brand
attitude, green brand equity, and willingness to pay a price premium with Process Model 1 by
Hayes (2012). There was no significant interaction for condition and sustainability
self-identity on brand attitude (interaction condition x sustainability self-self-identity, p = .452; Model
summary: ΔR2 = .02, F(3, 183) = 0.52, p = .668), and also not for willingness to pay a price premium (interaction condition x sustainability self-identity, p = .169; Model summary:
ΔR2
= .07, F(3, 183) = 5.95, p < .001, significant positive influence of condition, p = .009).
The proposed moderation of sustainability self-identity on the indirect effect of
condition on brand equity was investigated with Process Model 7 (Hayes, 2012). A
bootstrapping analyses with 1000 samples revealed no moderated mediation (index = 0.04,
SE = 0.04, 95% CI [-0.01, 0.16]).
There was, however, a significant moderation of the relationship between condition
and green brand equity by sustainability self-identity when prior brand attitude was
controlled for, ΔR2 = .24, F(4, 182) = 16.61, p < .001. When sustainability self-identity is low, i.e., one standard deviation below the mean, there is a non-significant negative
relationship between condition and green brand equity, b = 0.41, 95% CI [1.03, 0.21], t =
-1.29, p = .198. Also at the mean value of sustainability self-identity, there is a non-significant
positive relationship between condition and green brand equity, b = 0.12, 95% CI [-0.24,
0.49], t = 0.66, p = .508. The moderation effect only occurs for high levels of sustainability
self-identity, i.e., one standard deviation above the mean, b = 0.65, 95% CI [0.08, 1.22], t =
2.27, p = .025. When sustainability self-identity is high, people in the sustainability condition
rate the green brand equity of H&M significantly higher (M = 3.85), than people in the
regular condition (M = 3.20), p = .025. Figure 3 shows the interaction between sustainability
Fig. 3: Relationship between condition and green brand equity, moderated by sustainability
self-identity.
The Johnson-Neyman method reveals that the moderation holds also true for very low
levels of sustainability self-identity. When sustainability self-identity is between 1 and 2.81
(on a scale from 1 to 7) it has a significant negative moderating influence on the relationship
between condition and green brand equity (at SSI value level 1: b = 1.44, 95% CI [2.59,
-0.29], t = -2.47, p = .015; at SSI value level 2.81: b = -0.65, 95% CI [-1.31, 0.00], t = -1.97, p
= .050). The negative interaction indicates that people with a very low sustainability
self-identity rate the green brand equity of H&M in the sustainability condition lower than in the
regular condition. This is not depicted by the figure because the very low levels of
sustainability self-identity were far below the mean and the low levels depicted in the figure
start one standard deviation below the mean of sustainability self-identity, at 3.38 which is
already too high according to the findings of the Johnson-Neyman method.
High levels of sustainability self-identity have a significant positive moderating effect
on the relationship between condition and green brand equity. The significant moderation
starts at the value level of 5.25, b = 0.42, 95% CI [0.00, 0.84], t = 1.97, p = .050. The positive
interaction means that for high levels of sustainability self-identity (starting point: 5.25)
H&M’s green brand equity is rated significantly higher in the sustainability condition than in
2.94 3.20 2.54 3.85 0 1 2 3 4 5
GBE at low levels of SSI GBE at high levels of SSI
G re en B ra nd E qu it y
the regular condition. None of the other relationships became significant when prior brand
attitude was controlled for.
Due to the insignificant moderation effect on brand equity, willingness to pay a price
premium, and brand attitude, Hypothesis 2 can only be partially accepted.
Three-way interaction between condition, prior brand attitude and sustainability self-identity
Fig. 4: Process Model 3 by Hayes (2012)
Hypothesis 3 proposed a three-way interaction between condition, prior brand
attitude, and sustainability self-identity on brand attitude. Namely, hypothesis 3.1 proposed
that this would only lead to an attitude change if a person has a negative prior brand attitude,
a high sustainability self-identity, and is exposed to the sustainability stimulus. This was
investigated with Process Model 3 by Hayes (2012), ΔR2 = .77, F(7, 179) = 101.14, p < .001, however, no three-way interaction between condition, prior brand attitude, and sustainability
self-identity on brand attitude was found, p = .582. Prior brand attitude significantly
influences brand attitude measured after presentation of the stimulus, p < .001, as does
condition, p = .037, and sustainability self-identity, p = .003. Also a two-way interaction
between prior brand attitude and sustainability self-identity proofed to be marginally
Coefficient Standard Error t p
Constant 3.54 .06 62.04 <.001
Condition -.24 .11 -2.10 .037
Prior brand attitude (Att1) .87 .04 21.42 <.001 Sustainability self-identity (SSI) -.13 .04 -2.98 .003 Cond x Att1 -.01 .08 -.13 .900 Cond x SSI -.12 .09 -1.33 .184 Att1 x SSI -.03 .02 -1.68 .095
Cond x Att1 x SSI -.02 .04 -.55 .582
Table 2: Model overview. Effect on brand attitude measured after exposure to the stimulus.
Condition has a negative significant effect on brand attitude, b = -.24, p = .037,
namely, when the model goes from the regular condition (coded as 0) to the sustainability
condition (coded as 1) brand attitude is reduced. This confirms again the main effect of
condition on brand attitude that was found when prior brand attitude was controlled for. Prior
brand attitude has a positive effect on brand attitude, b = .87, p < .001, which is of course
logical since it is a repeated measure. Sustainability self-identity has a negative effect on
brand attitude, b = -.13, p = .003, meaning that the higher a person ranks on sustainability
self-identity the lower his or her brand attitude towards H&M is.
The interaction between prior brand attitude and sustainability self-identity negatively
influences brand attitude at a marginal significant level, b = -.03, p = .095. As shown in
figure 4 above, sustainability self-identity moderates the controlled influence of prior brand
negative influence on the relationship between prior brand attitude and brand attitude, i.e., the
higher the sustainability self-identity of a person the less positive the effect of prior brand
attitude on brand attitude. So when two people have the same prior brand attitude about
H&M, but one has a high sustainability self-identity and the other one a low sustainability
self-identity, then the one with the high sustainability self-identity will show a less positive
attitude after exposure to either of the stimuli than the person with the low sustainability
self-identity.
Figure 5 depicts the proposed but non-significant three-way interaction between
condition, prior brand attitude, and sustainability self-identity. The values of brand attitude in
the regular and in the sustainability condition are compared within different levels of prior
brand attitude (Att1) and sustainability self-identity (SSI). The slopes in figure 5 are not
significant; however, they hint the negative effect of sustainability self-identity. High levels
of sustainability self-identity always lead to the least positive brand attitude within their prior
brand attitude categories (high, low). This effect becomes more pronounced in the
sustainability condition than in the regular condition. Nevertheless, the three-way interaction
of condition, prior brand attitude, and sustainability self-identity is not significant, p = .582.
Fig. 5: Proposed (non-significant) three-way interaction between condition, prior brand
attitude, and sustainability self-identity on brand attitude.
5.12 4.86 2.38 2.28 5.05 4.42 2.26 1.95 0 1 2 3 4 5 6
Att2 at low levels of SSI and high levels of Att1
Att2 at high levels of SSI and Att1
Att2 at low levels of SSI & Att1
Att2 at high levels of SSI and low levels of Att1
B ra nd At tit ud e (At t2 )
The figure also clearly shows that the hypothesis has to be rejected, the condition that
was hypothesized to produce an attitude change (key: Att2 at high levels of SSI and low
levels of Att1) has actually the lowest levels of brand attitude (Att2) and there is also no
attitude change since the differences between the slopes in this three-way interaction model
are not significant.
Thus, hypothesis 3 has to be rejected: there was no attitude change based on an
interaction of condition, prior brand attitude, and sustainability self-identity, however an
interaction of prior brand attitude and sustainability self-identity was marginally significant.
Discussion
Results overview
Fig. 6: Overview over all significant and marginally significant effects.
Figure 6 displays an overview of the main results. A positive main effect of condition
on brand equity (via brand attitude) and willingness to pay a price premium was found, as
well as a negative main effect of condition on brand attitude. There was also a significant
proposed three-way interaction between condition prior brand attitude and sustainability
self-identity did not hold true, however an interaction between prior brand attitude and
sustainability self-identity was found.
Differences in effects on brand attitude and brand equity
The main effect of condition on willingness to pay a price premium is in line with the
expectations. Respondents in the sustainability condition were willing to pay 16 percent more
for sustainably produced apparel; this also is in line with the fair trade likers (17 percent
more) de Pelsmacker et al. (2005) identified. The way willingness to pay a price premium
was measured made sure that participants in the sustainability condition only saw a piece of
apparel framed as from the sustainable collection and participants in the regular condition
saw the same piece but framed as from a regular condition. Since there was no direct
comparison between a sustainably produced piece of apparel and a regular one within one
group of participants it can be excluded that the price is just set higher through the
comparison. The participants in the sustainability group were willing to pay a significantly
higher price for the piece of sustainably produced apparel than the participants in the regular
condition for the regular apparel. Therefore it can be concluded that people think that it is
worth paying more for sustainably produced apparel.
That condition had a positive indirect effect (via brand attitude) on brand equity is
also in line with the expectations. The sustainability branding strategy significantly increased
the value of the brand. However, it was quite surprising that condition had a negative effect
on brand attitude, meaning, that when people were exposed to the sustainability branding
strategy they reported a significantly lower attitude towards H&M than when people were
exposed to the regular branding strategy. Some comments that participants gave about the
study might bring clarity to that issue: “is it true that H&M plans to become more fair trade?”
comitmeent to envrotment or etc they still capitalist product that on the way they creat,
distribution, accumulation give bad impact on the third country or human being. For me us
the shirt do not use the brand” (sic), and “For me, environmental friendliness and fair trade
are two separate things. Even if H&M uses organic cotton, the conditions for their workers in
Bangladesh etc. do not improve from that.” In summary, the comments can be interpreted in
a way that they criticize the sustainability branding strategy as not credible and as
green-washing.
This interpretation is supported by various studies. Carrete, Castaño, Centeno, and
González (2012), for instance, report in their study that their respondents perceive
sustainability claims as deceptive and that companies only use them out of self-interest to be
able to ask for higher prices. Crane (2000) explored the “green backlash” with case studies
and interviews with managers and he confirmed the assumptions regarding credibility. Due to
the widespread use of (often unproven) green claims consumers have become very suspicious
towards it. This would explain why brand attitude is lower in the sustainability condition.
People do not believe the sustainability branding strategy.
This is also in line with the way how Hartmann et al. (2005) explain attitude
formation towards sustainability branding under the Elaboration Likelihood Model (ELM). It
could be assumed that participants processed the information more deeply because of how the
information, i.e., the stimuli, was presented in the study. After the instructions to carefully
read the information because questions are going to be asked later about it the participants
only saw the information on their screen. This is not like in a real life setting when an
advertisement has to compete with other stimuli for attention, thus it seems likely that the
participants processed the given information more deeply than they usually would. On this
route of deeper processing, however, sustainability branding can only lead to a positive
Cacioppo, 1986). Apparently, as the comments indicate, this was not the case, thus brand
attitude was lower in the sustainability condition.
This is of course a problem for established brands who want to change their image.
They have their image often for years and are thus known for certain things, in H&M’s case
this could be cheap fashion. Information that differs from this perception, like the given
information about H&M’s sustainable production processes, might therefore be regarded as
not credible. This assumption is further supported by the negative direct effect of
sustainability self-identity on brand attitude and also by the negative interaction effect of
sustainability and prior brand attitude on brand attitude, both in the three-way interaction
model. The higher the sustainability self-identity of the participants the less positive was the
influence of prior brand attitude, meaning that people who are concerned with sustainability
had a significantly lower attitude towards H&M. In sum, it seems problematic for an
established brand like H&M to rebrand itself as sustainable because it might be too far from
its brand core, i.e., cheap fashion.
The negative effect on brand attitude and the positive effect on brand equity of
condition might seem contradicting at a first glance. However brand equity is measured in
comparison with any other brand (e.g. “It makes sense to buy H&M instead of any other
brand, even if they are the same”). Of course H&M invites comparison with other fashion
giants like Zara for instance. Participants in the sustainability condition might not have been
totally convinced by H&M’s sustainable actions, however, in comparison with other brands
like Zara, H&M could be perceived as a brand which is at least doing something. These are,
of course, just assumptions.
For future research it would be interesting to look for one thing at the credibility of
big established companies, and for another thing to measure brand equity not in comparison
could be also interesting to use a big brand for which sustainability branding would seem
more believable. A brand, for example, which does not produce in third-world countries but
in Europe.
Sustainability self-identity and prior brand attitude as moderators
Sustainability self-identity moderates the relationship between condition and green
brand equity. People with a high sustainability self-identity did rate the green brand equity of
H&M higher when they were exposed to the sustainability branding strategy than when
exposed to the regular branding strategy. This was in line with the expectations. It was
measured similarly to brand equity in comparison with any other brand. Nevertheless, it is
surprising that sustainability self-identity did not moderate the relationship of condition and
any of the other consumer response measures. However, green brand equity is the only
response measure that is really concerned about the “green factor”. Therefore, it seems
logical that people who are concerned with the environment (high sustainability self-identity)
value a brand higher (in comparison with any other brand) that is concerned with the
environment (sustainability vs. regular condition).
Brand attitude, brand equity, and willingness to pay a price premium are all not
moderated by sustainability self-identity. For these three there is only a significant main
effect of condition, i.e., the branding strategy has an effect (positive (and indirect via brand
attitude) for brand equity and willingness to pay a price premium, negative for brand
attitude), however this effect occurs regardless of the sustainability self-identity levels of the
consumer. Generally speaking, consumers are willing to pay a significantly higher price for a
sustainably branded product, consumers value a brand higher in comparison to any other
brand if it engages in sustainability branding and consumers report a slightly more negative
sustainability self-identity of a consumer does not play a role for variables that are not
directly related to sustainability.
When the three-way interaction was tested a not-proposed two-way interaction of
prior brand attitude and sustainability self-identity was found. Thus, prior brand attitude does
not moderate the relationship of condition on brand attitude (the three-way interaction and the
interaction between condition and prior brand attitude were non-significant). The direct effect
of prior brand attitude on brand attitude, however, is moderated by sustainability self-identity.
In contrast to the moderating role sustainability self-identity plays on green brand equity, the
interaction with prior brand attitude is negative, i.e., sustainability self-identity makes the
positive effect of prior brand attitude on brand attitude less positive. As already explained
above, this could be due to credibility issues of already established brands. In other words,
when the (high) sustainability self-identity of a person is taken into account his attitude
towards H&M is reduced. Sustainability and H&M seem not to go well with each other.
Practical implications
This study was conducted to shed light on the workings of sustainability branding for
established brands. The sustainability branding strategy led to a higher willingness to pay a
price premium, a higher brand equity and (for people with high sustainability self-identity) to
a higher green brand equity, however, antithetical to the expectations the sustainability
branding strategy led to a lessened brand attitude compared to a regular branding strategy. As
indicated above, this could be due to credibility issues. An established brand has not a blank
slate like a new brand, and might be therefore less credible. Compared with other brands, the
sustainability branding increases the value of the brand; but when focusing on a sustainability
strategy the information given should be credible and also be perceived as credible in order to
2000). Additionally, the strategy and consumer response should be monitored closely to be
able to make changes during the campaign.
The higher willingness to pay a price premium for a sustainably branded product is
good news for the industry. People are willing to pay more for sustainably produced
products, and this demand could in turn magnify sustainable production processes and
increase working conditions.
Limitations
One limitation of the study is the made assumptions based on some comments of
participants, namely that sustainability branding has a negative effect on brand attitude due to
credibility issues of the established brand. It of course makes sense, but the credibility of the
brand was never measured in this study and thus, this is just an assumption that requires
investigation in further studies.
Another limitation is the way brand equity and green brand equity were measured.
They were measured in comparison with other brands. The consideration of “any other
brand” might have been the reason for the diverging results of brand equity and brand
attitude. This also should be investigated in future studies.
A major limitation is the way prior brand attitude was measured. Due to the time
constraints of this study is was not possible to actually measure it prior to the main study. It
was only possible to measure it prior to the exposure of the stimulus. Of course it was mixed
with demographics and other attitude filler questions but it cannot be completely excluded
that the participants’ answers were distorted because they were primed with what to expect.
One last limitation is due to the experimental design of the study. No long term
effects could be measured. All measured effects are based on one exposure. Therefore it
cannot be concluded that after multiple exposures to a sustainability branding strategy all
Conclusion
This study investigated how sustainability branding influences the consumer response
towards an established brand and also took the personality trait of sustainability self-identity
into account. While research in this field was thus far driven by studies using fictional brands
or just industries without focus on specific brands (e.g., Lee et al., 2010; Kalafatis et al.,
1999; Lee, 2008) this study focused on a well-established brand. This changes the focal point
completely because in this scenario did not only the sustainability branding have an effect but
also the attitudes the consumers already held towards the brand. This study furthermore
confirmed that a sustainability branding strategy should be used carefully because if the
brand and the strategy together are not seen as believable image damage in form of lowered
attitudes towards the brand can result. Lastly, this study showed that the personality trait of
sustainability self-identity is a factor that should not be neglected. Even if it only moderated
the relationship between branding strategy and green brand equity it had, however, also an
effect on brand attitude and is therefore an important factor for the success or failure of a
Literature
Akerlof, G. A., & Dickens, W. T. (1982). The economic consequences of cognitive
dissonance. The American Economic Review, 72(3), 307-319.
Anderson, J. R. (1983). A spreading activation theory of memory. Journal of verbal learning
and verbal behavior, 22(3), 261-295.
Arvola, A., Vassallo, M., Dean, M., Lampila, P., Saba, A., Lähteenmäki, L., & Shepherd, R.
(2008). Predicting intentions to purchase organic food: The role of affective and
moral attitudes in the Theory of Planned Behaviour. Appetite,50(2), 443-454.
Carrete, L., Castaño, R., Felix, R., Centeno, E., & González, E. (2012). Green consumer
behavior in an emerging economy: confusion, credibility, and compatibility. Journal
of Consumer Marketing, 29(7), 470-481.
Chen, Y. S. (2010). The drivers of green brand equity: green brand image, green satisfaction,
and green trust. Journal of Business Ethics, 93(2), 307-319.
Crane, A. (2000). Facing the backlash: green marketing and strategic reorientation in the
1990s. Journal of Strategic Marketing, 8(3), 277-296.
De Pelsmacker, P., Driesen, L., & Rayp, G. (2005). Do consumers care about ethics?
Willingness to pay for fair‐trade coffee. Journal of consumer affairs, 39(2), 363-385. Franzen, G. (2009). Brand portfolio and brand architecture strategies, SWOCC book series
49. Amsterdam: SWOCC.
Gardner, M. P. (1985). Does attitude toward the ad affect brand attitude under a brand
evaluation set?. Journal of Marketing Research, 22(2), 192-198.
Gilg, A., Barr, S., & Ford, N. (2005). Green consumption or sustainable lifestyles?
Identifying the sustainable consumer. Futures, 37(6), 481-504.
H&M (2013) Conscious Actions. Sustainability Report 2013. Retrieved on January 20, 2015
