SUSTAINABLE NEW PRODUCT INTRODUCTION
SUCCESS AND THE IMPACT OF MARKETING
MEASURES AND PRODUCT CATEGORY
CHARACTERISTICS
by
Katharina Bergmann
SUSTAINABLE NEW PRODUCT INTRODUCTION SUCCESS
AND THE IMPACT OF MARKETING MEASURES AND PRODUCT
CATEGORY CHARACTERISTICS
Master Thesis, MSc Marketing Intelligence and Marketing Management University of Groningen, Faculty of Economics and Business
12
thJanuary 2015
Katharina Bergmann Student number: s2405121 Folkingestraat 15d 9711JS Groningen, NL tel.: +31 (0)647 - 606069 e-mail: k.bergmann@student.rug.nl First supervisor: Jenny van Doorn Second supervisor:Felix Eggers
_____________________________________________________________________________
Acknowledgements: First, I would like to express my gratitude towards Jenny van Doorn, who
had to patiently bear my persistent questioning and insecurity. She provided me with helpful comments and supporting guidance. Further, I thank Rick, Patrick & Rebekka for reading through my thesis and providing me with valuable feedback.
Abstract
Launching new products is a major strategy that ensures company survival and growth. Yet, large resource investments and high failure rates make managerial decision-making complex, particularly concerning what stimulates market acceptance. Companies are releasing an increasing amount of green products in response to a growing demand for sustainability, although empirical evidence concerning sustainable new product introduction success is missing. Considering marketing measures and product category characteristics, we explore whether sustainable new products are more successful than non-sustainable versions. Using a Multilevel Negative Binomial Model, we discover higher sales for regular new products. However, with price promotions and increasing degrees of innovativeness, sustainable new products’ sales numbers grow and overtake those of regular items. Moreover, new vice products are more successful, whether sustainable or not, while regular virtue products are more successful than sustainable virtue products. The findings are based on product introductions of the Dutch market and are used to provide practical and theoretical guidance for managers and scholars.
Keywords: New product success, new product sales, sustainable, sustainability, degree of
Table of contents MANAGEMENT SUMMARY ... a LIST OF TABLES... c LIST OF FIGURES ... c LIST OF ABBREVIATIONS ... d 1 INTRODUCTION ... 2 2 RESEARCH FRAMEWORK ... 6 2.1 Theoretical background ... 6
2.1.1 New product introduction success ... 6
2.1.2 Marketing measures & NPI success ... 8
2.1.3 Product category characteristics & NPI success ... 12
2.1.4 Sustainability ... 13
3 CONCEPTUAL FRAMEWORK ... 16
4 HYPOTHESES GENERATION ... 17
4.1 New product success and product sustainability ... 17
4.2 Marketing measures ... 19
4.2.1 Degree of product innovativeness ... 19
4.2.2 Price promotion ... 21
4.3 Product characteristics ... 22
4.3.1 NPI success in vice and virtue categories ... 22
5 RESEARCH METHODOLOGY ... 23 5.1 Data ... 23 5.1.2 Sample description ... 23 5.1.3 Measures ... 23 5.1.4 Descriptives ... 26 5.2 Analytical procedure ... 27 6 EMPIRICAL RESULTS ... 30 6.1 Model Fit ... 30 6.2 Results ... 32 6.3 Robustness checks ... 34 7 DISCUSSION ... 35 7.1 Managerial implications ... 42
7.2 Limitation & future research ... 44
MANAGEMENT SUMMARY
LIST OF TABLES
Table 1: Comparative overview of selected literature concerning degree of innovativeness Table 2: Comparative overview of selected literature concerning sustainability & green
products
Table 3: Overview of categories & (non-) sustainable new product introductions Table 4: Theoretical overview of control variables
Table 5: Variable operationalisation & descriptives
Table 6: Overview of multilevel negative binomial regression results Table 7: Overview of findings according to hypotheses
Table 8: Summary of managerial recommendations for marketing measures & product category characteristics
LIST OF FIGURES
Figure 1: Conceptual model including expected effects Figure 2: Distribution of unitsales
Figure 3: The effect of product innovativeness on (sustainable) new product introduction success Figure 4: The effect of price promotions on (sustainable) new product introductions
LIST OF ABBREVIATIONS
“Whatever we wish to be in the future depends on our present actions”
1
INTRODUCTION
The Fast Moving Consumer Goods (FMCG) industry is notorious for its volatility and strong competition, turning the launching of new products into a difficult undertaking. Yet new product introduction success is indispensable for firm survival as well as growth and can threaten long-term corporate success if unsuccessful (Pauwels, Silva-Risso, Srinivasan & Hanssens 2004; Cooper & Kleinschmidt 1987; Gielens & Steenkamp 2007). Considering that three out of four new products fail during the initial year of launch, companies must be knowledgeable about factors that impact new product success - also defined as market/consumer acceptance and as proxy for new product trial probability (e.g. Marketingweek 2014; Steenkamp & Gielens 2003; Evanschitzky, Eisend, Calantone & Jiang 2012).
2012 the European market alone grew by 6% (Schaack, Lernoud, Schlatter & Willer 2014). A large body of literature is attending the potential reasons for this growth including consumer attitudes, purchase behaviour and willingness to pay, although findings are inconsistent: contrasting the positive consumer associations scholars discover on dimensions such as taste, safety and health, others encounter negative associations concerning e.g. price premiums, product performance and taste, too (Aertsens, Mondelaers, Verbeke & van Huylenbroeck 2009; Van Doorn & Verhoef 2011; Luchs, Naylor, Irwin & Raghunathan 2010; Verhoef 2005).
Launching sustainable products seems a worthwhile strategy not only because of the apparent market potential, but also related to the larger profit margins of sustainable items (Bezawada & Pauwels 2013). However, market shares of sustainable products remain low relative to regular products (Schaack et al. 2014). As potential reasons research cites an attitude-behaviour gap on the one hand, where consumers do not engage in sustainable consumption despite their positive attitudes, as well as consumers’ perceived risk towards newly introduced products on the other (Prothero et al. 2011; Cooper & Kleinschmidt 1987; Larceneux, Benoit-Moreau & Renaudin 2012; Schiffman 1972; Steenkamp & Gielens 2003). This ambiguity poses the question whether firms should invest into the launching of sustainable NPIs. Does making a product sustainable increase the odds of new product trial? Bearing in mind both the importance and low rate of new product success, this is an essential topic to be investigated and has, to the best of our knowledge, not yet been done despite large public interest.
(2003) for example show how product novelty, promotion as well as a hedonic product character1 stimulate new product trial within the FMCG industry. Product novelty has been
extensively discussed, yet findings diverge depending on industry and context (e.g. Ali 2000; Goldenberg, Lehmann & Mazursky 2001). With regards to promotions, price promotions are consistently found impactful on consumer purchase behaviour, especially for higher priced items (e.g. Gupta 1988; Neslin 2002; Pauwels et al. 2004). Scholars address product characteristics using the distinction of vice and virtue product categories2 and their significant influence on
purchase behaviour (e.g. Steenkamp & Gielens 2003; Wertenbroch 1998). Only few studies include sustainability in this realm, in spite of rising consumer interest and media attention (Prothero et al. 2011; Mick 2008). Several investigate how price promotion impacts sustainable products, but with conflicting results (e.g. Ngobo 2011; Bezawada & Pauwels 2013). Another study finds that consumers are less willing to pay for organic vice products, yet does not specifically target new product success for FMCGs in combination with sustainability, marketing elements and product category characteristics (van Doorn & Verhoef 2011). Further, most research concentrates on new product success for durables (e.g. Stremersch & Tellis 2004), only Steenkamp and Gielens (2003) research new product trial for FMCGs. We therefore use Steenkamp and Gielen’s (2003) study as starting point to fill the identified literature gap on sustainable versus regular NPI success and expand their research by including general NPI success factors related to marketing elements and product category characteristics as identified by academics in our proposed framework (figure 1). First, we investigate the impact of (1)
marketing measures on new product success, where we examine new products’ degree of
innovativeness and the effect of price promotions. Second, we include (2) product category
1
This includes products that are bought on impulse
2
characteristics, where new products are rated according to the relative concepts of vice and
virtue. Our study differs from that of Steenkamp and Gielens (2003) in that we use new product unit sales in a given period as proxy for new product success to see whether regular or sustainable NPIs are more successful, while Steenkamp and Gielens (2003) investigate general factors influencing general new product trial probability. The main objective is to contribute to the body of literature dealing with (sustainable) new product success within the Dutch FMCG industry. Based on that we derive the following research questions:
1. Are sustainable new product introductions more successful than regular new product introductions in terms of sales?
2. What effect do marketing measures and product characteristics have on sustainable new product introduction success?
The contribution of this paper is beneficial for both corporations as well as academics and those interested in new product performance within the (green) FMCG market. Following previous studies, we enrich marketing literature on sustainable new product trial probability, the related impact of marketing measures (e.g. Bezawada & Pauwels 2013; Neslin 2002) and product characteristics (e.g. van Doorn & Verhoef 2011). Because Bass and Wind (1995) point out the importance of empirical generalisations, we investigate this with large, time-framed product-level sales data of the Dutch FMCG market, which includes 884 newly introduced items. For the analysis we use STATA and a Multilevel Negative Binomial Model that accounts for the characteristics of hierarchically structured and overdispersed sales data.
presented using a conceptual model (figure 1) that forms the foundation of this investigation. Consequently, we give motivations and derive hypotheses for each factor. Second, we introduce the employed research method in more detail, before we present and discuss the study’s results. Based on those, we derive practical advice in the section of managerial implications. Lastly, we present limitations, future research opportunities and a concluding section.
2
RESEARCH FRAMEWORK
In the following we list and discuss existing literature concerning NPI success, its theoretical background within the FMCG environment together with the aspect of sustainability. After, we elaborate on (1) marketing measures (i.e. degree of product innovativeness and price promotion) and (2) product category characteristics (i.e. vice and virtue categories). The resulting conceptual model is derived from the identified research gaps.
2.1 Theoretical background
2.1.1 New product introduction success
Various research streams deal with NPI success and what the potential reasons for the high failure rates are. Considerable attention has been paid to the notion of market acceptance and how consumers’ risk-averseness and natural uncertainty hampers new product trial (e.g. Gielens & Steenkamp 2007; Holak & Lehmann 1990; Pauwels et al. 2004; Krishnan & Zhu 2006; Steenkamp & Gielens 2003). Every NPI lacks substantial information on their product performance, so consumers assume potential negative consequences when considering new product trial such as economic loss or physical danger, and thus associate NPIs with risk (Seth & Ram 1987).
Shimp 1982; Bettman 1973 & 1975; Cooper & Kleinschmidt 1987; Gielens & Steenkamp 2007; Ross 1975; Schiffman 1972; Steenkamp & Gielens 2003). Early research of Fishbein and Ajzen (1975) for example links consumers’ perceived risk to the eventual NPI adoption in their theory of reasoned action. Theoretically, this model describes the formation of behavioural activities where intention, attitude formation and beliefs play an important role in this order. New product trial as a behavioural activity is thus initiated through an intention that emerges after attitude formation. Attitudes in turn are created through a person’s beliefs regarding consequences a certain action might have. If consumers belief that a novel product performs dissatisfactory, a negative attitude in form of perceived risk arises and negatively impacts the formation of an intention - consumers become reluctant to new product trial. In other words, this illustrates how any uncertainty feeding a negative belief concludes in a decreasing willingness to try. The same also holds for positive beliefs (Holak & Lehmann 1990; Fishbein & Ajzen 1975).
certain consumer characteristics (e.g. innovative and younger consumers are more likely to accept new products) or brand strength (e.g. larger consumer acceptance with stronger brands). Most studies exploring market success factors for e.g. durable goods (e.g. Stremersch & Tellis 2004) and solely Steenkamp and Gielens (2003 & 2007) refer to NPIs and NPI success within the FMCG industry. Nevertheless, the authors only investigate regular NPIs and make no distinction with sustainable NPI versions to see whether there is a significant difference in terms of unit sales. Also, they only indirectly deal with product category characteristics by investigating impulse buying. In sum, literature lacks research on NPI success concerning sustainable new products within the FMCG industry, including the marketing elements and product category characteristics labelled as influential by scholars.
2.1.2 Marketing measures & NPI success
Based on research related to marketing measures and their impact on new product trial and success, we review literature concerning two sub-items: the degree of product
innovativeness (e.g. Steenkamp & Gielens 2003; Salavou & Avlonitis 2008) and price promotion
(e.g. Bezawada & Pauwels, 2013) in this block.
Degree of innovativeness
high, but so are levels of complexity (Steenkamp & Gielens 2003). Complexity is related to lacking product information and difficulties in understanding new products - if it is too high consumers become uncertain and perceive new products as risky, which hampers new product trial probability (Angelmaar 1990; Brentani 2001; Mukherjee & Hoyer 2001; Szymanski, Kroff & Troy 2007). New products of low innovativeness on the other hand have little relative advantage over other products and are thus less difficult to understand, which is accompanied by both lower uncertainty and perceived risk (Kleinschmidt & Cooper 1991; Steenkamp & Gielens 2003). Hence, the degree of innovativeness is a crucial factor that can influence consumers’ propensity to new product trial and ultimately NPI success (Pauwels et al. 2004). Scholars agree on the conceptual definition of innovativeness and its influence on NPI (e.g. Kleinschmidt & Cooper 1991), but disagree on which level of innovativeness is most impactful on new product success. Table 1 depicts a summary of relevant literature on innovativeness and their findings.
Table 1
Comparative overview of selected literature concerning degree of innovativeness Study Industry Main Results
Ali (2000) High-tech Product innovativeness moderates the relationship between product development time & new product performance: taking a long time to launch a product at low innovation degree has a lower probability of meeting sales revenue goals, as opposed to launching a highly innovative product with long development time (indirect influence)
Goldenberg et al. (2001)
Books Non-linear relationship between product novelty & market success: high & low product novelty promises low market success, while moderate product novelty promises high market success (inverted u-shape)
Kleinschmidt & Cooper (1991)
Industrial Products
Non-linear relationship between product innovativeness & commercial success: high & low degree of innovativeness are more successful than in-between innovativeness (u-shape)
Steenkamp & Gielens (2003)
FMCG Non-linear relationship between product novelty & new product trial: high & low product novelty both lead to high new product trial rates, while medium product novelty leads to low new product trial rates (u-shape) Salavou & Avlonitis
(2008)
Danneels &
Kleinschmidt (2001)
Industrial Projects
Positive relationship between product innovativeness & firm performance
Evanschitzky et al. (2012)
Across industries
Non-significant relationship between product innovativeness & new product success
Firth & Narayanan (1996)
Product Manufacturing
Positive relationship between innovativeness in the market & firm-level performance
Yap & Souder (1999) High-Tech Negative relationship between innovativeness & new product success
The majority finds a non-linear relationship between product innovativeness and outcomes such as market/commercial/product success or product trial probability but vary on the outcome, where some find a u-shaped relation while others find the opposite (inverted u-shape relation). This can be explained by the different industries and degrees of innovativeness that were investigated: e.g. Goldenberg et al. (2001) find an inverted u-shape within the durables industries, while Steenkamp and Gielens (2003) find the opposite within the FMCG industry. While the non-durables industry, or FMCGs, is characterised by innovativeness ranging from low to incremental, the durables industry includes the whole spectrum ranging from low to radical.
Price promotion
The impact of price promotion has been studied extensively and its short-term effectiveness on sales, purchase rates and category demands proven empirically and consistently (e.g. Nijs, Dekimpe, Steenkamps & Hanssens 2001). Blattberg, Briesch and Fox (1995) define price promotions as temporary price reductions offered to customers and literature describes it as impactful competitive tool to attract consumers, because price promotions reduce purchase barriers related to high prices (e.g. Arora & Henderson 2007; Gupta 1988; van Heerde, Leeflang & Wittink 2004; Nijs et al. 2001; Pauwels et al. 2004). Additionally, Arora and Henderson (2007) write how temporary price reductions can jeopardise a product’s image and consumers’ product associations, implying that price promotions and product image are interrelated. This can be explained by the price-quality heuristic, where a high price implies good quality and vice versa (e.g. Cronley, Posavac, Meyer, Kardes & Kellaris 2005).
In 2000, Glaser and Thompson investigate to what extent sustainable product sales are influenced by price promotions. The authors find that the marketing tool stimulates sales of sustainable products by lowering their price premiums and thus confirm previous price promotion research (e.g. Nijs et al. 2001). Subsequent research by Ngobo (2011) supports this by demonstrating that price promotions significantly increase organic product sales. The author, too, argues with sustainable products’ price premiums along with the notion of price-quality associations. Contrastingly, Bezawada and Pauwels (2013) find that price promotions do not significantly impact organic product sales and imply that most consumers are not price sensitive towards organic goods.
2.1.3 Product category characteristics & NPI success
Literature shows the differential impact product category characteristics have on product success. Particularly the product category type has been investigated with a major focus on vice and virtue categories (e.g. van Doorn & Verhoef 2011; Hui, Bradlow & Fader 2009).
Vice and virtue
Wertenbroch (1998) defines vice and virtue products as relative to one another where vice products are associated with wants and can create immediate feelings of joy. Also, vice products are related to feelings of guilt before or after purchase and consumption due to potential long-term consequences such as weight gain after chocolate consumption (Khan & Dhar 2006; Okada 2005; Mishra & Mishra 2011; Milkman, Rogers & Bazerman 2010). Literature depicts guilt as critical emotion in consumption and as trigger for purchase behaviour, e.g. impulse or bulimic buying (Dahl, Honea & Manchanda 2003). Virtue products on the other hand represent the healthy option and are labelled as shoulds. They are less appealing for consumption, yet do not have negative long-term consequences like vice products (Khan & Dhar 2006; Milkman, Rogers & Bazerman 2010; Mishra & Mishra 2011; Okada 2005; Wertenbroch 1998).
consumers’ low quality perceptions for these product types (Luchs, Naylor, Irwin & Raghunathan 2010).
In sum, the cited studies indirectly investigate sustainability and product categories as of vice and virtue, yet either do not account for sustainable products as such or do not differentiate between existing products and NPIs or vice versa (van Doorn & Verhoef 2011; Steenkamp & Gielens 2003).
2.1.4 Sustainability
Media and consumer attention on problematic environmental and societal topics is growing and causes consumer demand for sustainability to increase, which is reflected in an increasing amount of green products entering the FMCG market (Environmental Leader 2009; Food Marketing Institute 2009; Schaack et al. 2014). In 2012 the European organic market was worth 20.9 billion € and grew by 6% in that year, with Germany alone noting a growth rate of 6%. Fresh products, including dairy and milk items, vegetables and fruit are dominating retail sales, which amounted to 7.04 billion € in 2012 (Schaack et al. 2014; Willer, Lernoud & Schlatter 2014; Vaclavik 2012). Further growth along with demand is expected based on an increasing interest in sustainability and products (e.g. Larceneux et al. 2012; Prothero et al 2011), setting of a large stream of research of which a selection is depicted in table 2.
Table 2
Comparative overview of selected literature concerning sustainability & green products
Research focus Authors Dependent
Variable: New
product success Consumer-side:
Profiling green consumers Diamantopoulos, Schlegelmilch, Sinkovics & Bohlen (2003)
Jain & Kaur (2006)
D’Souza, Taghian, Lamb & Peretiatko (2007) No
General underlying motivations to choose for green product versions
Aertsens et al. (2009) Bourn & Prescott (2002)
De Pelsmacker, Driesen & Rayp (2005) Fotopoulos & Krystallis (2002) Freestone & McGoldrick (2008) Gleim, Smith, Andres & Cronin (2013) Kihlberg, Risvik (2007)
Marian et al. (2014)
Thøgersen & Ölander (2006) Vermeir & Verbeke (2006)
Yiridoe, Bonti-Ankomah & Martin (2005)
Consumer- side*:
Consumer associations concerning sustainable products
Aertsens et al. (2009) Baker et al. (2005) Ehrich & Irwin (2005) Luchs et al. (2010)
Lindner, Fliessbach, Trautner, Elger & Weber (2010)
Paul & Jyoti (2012) Prothero et al. (2011)
No
Consumers’ willingness to pay for sustainable products
Van Doorn & Verhoef (2001) Janssen & Hamm (2012)
Van Loo, Caputo, Nayga, Meullenet & Ricke (2011)
No
Company-side:
Importance of green/sustainable firm actions
Kotler (2011)
Latacz-Lohmann & Foster (1997) Peattie (2001)
No
Green claims and brand attitude on new product introductions
Olsen, Slotegraaf & Chandukala (2014) Yes * attitude-behaviour gap
We identify two major research foci, namely the consumer-side and the company-side. The consumer-side focus includes inter alia the profiling of customer purchasing sustainable or general underlying motivations to do so (e.g. Aertsens et al. 2009). Research concerning the latter diverges especially on the dimensions of taste and quality: Kihlberg and Risvik (2007) e.g. depict that the majority of consumers purchasing organic food belief these to be better tasting than regular food products, while Bourn and Prescott (2002) report the opposite. Health and protection of the environment are further major consumer motivations to consider sustainable consumption (e.g. Bourn & Prescott 2002; Fotopoulos & Krystallis 2002).
Consequently, literature investigates this attitude-behaviour gap by exploring consumer associations concerning sustainable items and presents conflicting findings. Some discover negative associations towards specific sustainable product types on aspects of price premiums, taste, product appearance as well as conservation, availability and performance for other cases (e.g. Luchs et al. 2010; Marian et al. 2014; Gleim et al. 2013; Vermeir & Verbeke 2006), while others find positive associations for specific sustainable product types and their labels on dimensions such as taste, safety and health as well as trust (Aertsens et al. 2009; Bourn & Prescott 2002; Janssen & Hamm 2012). Baker et al. (2005) show positive perceptions for quality dimensions (Paul & Jyoti 2012), while van Doorn and Verhoef (2011) uncover consumer perceptions concerning sustainable vice products to be qualitatively less convincing. Lindner et al. (2010) again, reveal positive attitudes concerning organic labelling (Alvensleben & Bruhn 2001) and are supported by Bauer et al. (2013), who find that organic labels lead consumers to associate the product itself with dimensions that they originally only relate to organic labels, including healthfulness, food safety and environmental friendliness. The authors relate this to the theory of Rokeach’s value and halo effects (1968). Literature relates these inconclusive results and diverging associations to sustainable products’ credence character and the accompanied informational asymmetry. Here consumers are unable to verify the product’s attributes before or after usage, which creates uncertainty (Darby & Karni 1973; Srinivasan & Till 2002). Lastly, the topic of willingness to pay involves sustainable goods and their price premiums and findings reveal how consumers’ willingness to pay varies with consumers’ quality perceptions of virtue and vice products (van Doorn & Verhoef 2011).
Table 2 demonstrates that a major focus rests on the consumer-side, while investigations on the company-side are comparably scarce. Olsen et al. (2014) represent the minority in inspecting NPIs, yet with a focus on brand attitudes. Precise success factors such as sales are missing, which leads to the question how sustainability and NPI success are related and whether sustainable NPIs sell better than regular NPIs or vice versa.
3
CONCEPTUAL FRAMEWORK
The literature review illustrates that researchers so far have not considered sustainable new products and their success potential within the FMCG industry. Therefore, the present paper investigates whether sustainable or regular NPIs are more successful, where NPI success is a proxy for new product unit sales in given time period initiated through new product trial (Rogers 1995).
while Appendix A gives a comparison of selected literature concerning NPI success and the variables we choose to investigate.
Figure 1
Conceptual model including expected effects
4
HYPOTHESES GENERATION
4.1 New product success and product sustainability
Consumers are risk-averse and face NPIs with uncertainty because informational asymmetries make it difficult to accurately infer new product performance. New product trial is thus perceived to be risky, which ultimately hampers NPI success (e.g. Rogers 1995; Steenkamp & Gielens 2003).
FMCG environment. Extrinsic cues provide consumers the opportunity to infer product performance by acting as mental shortcuts that compensate information asymmetries and hence stimulate new product trial probability (Cronley et al. 2005; Mukherjee & Hoyer 2001; Olsen 1972; Wansink 1989). Bearden and Shimp (1982) are the first to expose the impact extrinsic cues in form of product warranties have on mitigating perceived consumer risk towards product adoption. Later research by Valceschini and Nicolar (1995) reveals that product labels are utilised for inferring intrinsic product attributes, specifically product quality - with the help of labels as extrinsic cues consumers transfer credence characteristics into objectified product information to support their assessment of quality. Decision-making is simplified this way and product trial probability for NPIs positively impacted (Janssen & Hamm 2012; Larceneux, Benoit-Moreau & Renaudin 2012; Giannakas 2002).
Additionally, the high presence of certified sustainability labels on the market increases their perceptual fluency, i.e. the ease of information processing. Literature relates this to familiarity, liking and consequently believability (Golden 2010; Winkielman, Schwarz, Fazendeiro & Reber 2003; Kardes 2002). In turn, this can be linked to a reduction of uncertainty based on consumers’ subconscious familiarity with certified sustainability labels. Encountering them on new products triggers this feeling of familiarity, which again is transferred to the unknown product (Fennis & Stroebe 2010; Rokeach 1968) - another way in which labels function as heuristics that reduce uncertainty and enlarge new product trial probability.
hand miss this extrinsic cue, presenting consumers with larger informational asymmetries and difficulties on inferring product performance, which leads to higher risk perceptions and lower product trials. This counters the wealth of arguments against sustainable product success, yet because of our NPI focus, risk reduction and the low involvement environment of the FMCG industry, we derive the following hypothesis:
H1: Sustainable product introductions are more successful in terms of unit sales than
non-sustainable product introductions.
4.2 Marketing measures
4.2.1 Degree of product innovativeness
Research testifies the impact product innovativeness has on market and new product performance. Especially within the highly competitive FMCG industry, innovativeness can be a crucial advantage (e.g. Steenkamp & Gielens 2003; McKinsey 2014). This present study focuses on the range between low and incremental innovativeness, which is typical for this industry (Gielens & Steenkamp 2003).
in the sustainable NPI’s categorisation as straight me-too product, which Brentani (2001) defines as a good with a small relative advantage over other products. Academics relate me-too products with lower market performance (Robinson & Fornell 1985; Robinson 1990), which suggests lower success probabilities for sustainable me-too products’.
Contrary to Steenkamp and Gielens’ (2003) non-linear effect, we assume that a higher the degree of innovation leads consumers to conclude that additional efforts were put into the differentiation of the sustainable new product on top of making it green. The NPI is consequently more than a straight me-too product in that it not only has a significant relative advantage over already existing products but is sustainable, too. We expect the label and a higher degree of innovativeness to reinforce each other. This notion is fortified by the finding of Mukherjee and Hoyer (2001), who investigate the differing effect the adding of features on low- and high complexity products has on consumer evaluation and sales. The authors find a positive effect in relation to low complexity products – of which the latter is comparable to the low involvement products as found in the FMCG industry. Because the adding of features complies with our focus of incremental innovativeness, we posit that a higher degree of innovativeness positively influences new sustainable product sales:
H2: The degree of product innovativeness has a positive effect on new sustainable product
4.2.2 Price promotion
Price promotion is known as impactful tool to stimulate product sales and demand (e.g. Neslin 2002; Nijs et al. 2001), yet findings diverge concerning price elasticities and sustainable product sales (e.g. Glaser & Thompson 2000; Kiesel & Villas-Boas 2007). Ngobo (2011) e.g. shows that price promotions significantly impact organic product sales. Subsequently, Bezawada and Pauwels (2013) question this finding and observe that price promotion does not significantly impact organic product sales, which poses the question whether price promotions are an effective sales enhancer in the scenario of sustainable NPIs.
Because the present paper focuses on NPIs and success, we assume that temporary low prices encourage consumers to try the to them still unknown sustainable NPIs, which contrasts the finding of Bezawada & Pauwels (2013). We thus ascribe price promotions a positive effect on new sustainable product success and base this on their risk mitigation potential through lowering sustainable NPIs’ otherwise high premium prices (e.g. Aertsens et al. 2009; Fotopoulos & Krystallis 2002; Glaser & Thompson 2000; Thøgersen & Ölander 2002; Van Loo et al. 2010; Yiridoe et al. 2005; Verhoef 2005; Wier et al. 2003). This line of reasoning directs attention to risk perceptions as inhibitor to new product trial and price promotions as factors lessening these perceptions through the lowering of prices. Also, when including the concept of price as quality indicator in low involvement situations (Olson 1977), price promotions on sustainable NPIs make a perceived expensive, good quality product more affordable for consumers, further stimulating product trial. The following hypothesis is derived:
H3: Price promotions have a positive impact on unit sales of sustainable new product
4.3 Product characteristics
4.3.1 NPI success in vice and virtue categories
Consumers’ associations towards sustainability labels relate to e.g. a product’s fit of healthiness with the virtue category’s characteristic of wholesomeness, health and planned action as opposed to hedonic urges (e.g. Okada 2005; Mishra & Mishra 2011; Kivetz & Simonson 2002). Sustainability labels used on virtue NPIs are hence expected to be beneficial in inducing new product trial, because they reinforce the NPIs’ intrinsic values and positive associations (e.g. healthiness) and consequently reduce perceived risk concerning expected product performance. We expect sustainable NPIs to be more successful within the virtue category than within the vice category and derive the following hypothesis:
H4: Sustainable new product introductions are more successful in terms of unit sales in virtue
than in vice categories.
5
RESEARCH METHODOLOGY
5.1 Data
5.1.2 Sample description
Data were collected within the Dutch GfK household panel and contain information on 884 NPIs that were introduced in the Netherlands for a period of 24 months within 2008 – 2010 and 14 categories, ranging from a variety of food products to detergents (table 3). We focus on the first 12 months, since literature shows that within the FMCG industry these are critical for NPI success or failure (Ernst, Young & AC Nielsen 2000). The data set includes 58 sustainable NPIs, of which the largest share belongs to the category of chocolate bars and the smallest to the category of sugar, coffee and ketchup (table 3).
5.1.3 Measures
The dependent variable new product success is represented by unitsales, which is defined as the number of Stock Keeping Units (SKU) sold monthly over the course of our observation period. The independent variable is a dummy variable indicating whether a newly introduced product has a sustainability label or not (1 = sustainable, 0 = not sustainable). Sustainable here includes fair trade and organic labels. Fair trade labels on products ensure socially fair treatment by aiming “at improving the livelihood of (...) farmers and plantation and factory workers in developing countries” (Oosterveer et al. 2014; Voedingswaar 2014). Organic products are those with the Dutch eko label, which stands for pesticide free production methods (Voedingswaar 2014).
Marketing Measures
Experts from the GfK rated the innovation degree of every NPI on a continuous scale ranging from 1 – 7 for four items (table 5; Appendix B), which were then averaged to attain a single value. Expert ratings are a widespread tool in other recent research (e.g. Steenkamp & Gielens 2007, Goldenberg et al. 2001, Pauwels et al. 2004). Price promotion is inferred from the data and is represented by a dummy variable indicating whether price promotion took place within a calendar month or not (1 = promotion took place in month x; 0 = promotion did not take place in month x).
Product Category Characteristics
a category to have a virtue character, while a low shouldscore indicates a vice category. An overview of the product categories is shown in table 3.
Table 3
Overview of categories & (non-) sustainable new product introductions Category Total number of NPIs Number of sustainable NPIs Average unitsales Average degree of innovativeness Average number of price promotions Shouldscore* Baby food 104 6 5.03 3.188 .26 .58 Buttermilk 3 2 49.89 6.75 .39 2.59 Chocolate bars 53 19 13.35 3.41 .3 -4.42 Coffee 48 1 18.89 4.38 .34 -2.32 Drinking yoghurt 66 2 75.76 2.144 .55 .98 Packaged ice cream 35 6 2.907 13.7 .4 -4.98 Ketchup 24 1 21.40 2.75 0.31 -1.73 Liquid cleaner 67 3 3.940 3.40 0.55 2.56 Liquid detergent 163 2 13.63 3.715 0.56 2.51 Milk 19 4 109.20 2.375 0.2 2.59 Ready meal mix/sauces 99 3 41.66 4.21 0.62 -2.14 Soup 126 4 33.22 3.7381 0.54 .7 Sugar 14 1 52.11 4 0.13 -2.9 Yoghurt 63 4 46.47 2.456 0.65 .98 Total 884 58 - - - -
* relates to vice & virtue categorisation; high shouldscore = virtue category (Milkman et al. 2010)
Control Variables
Across our framework we include additional variables to control for effects that potentially influence NPI success such as time and advertising factors. Table 4 depicts an overview of all used control variables and sources for theoretical reasoning.
Table 4
Theoretical overview of control variables
Category Source Control variable
Time Gatignon, Weitz & Bantsal (1990) Month Introdate Category, price & van Heerde, Gijsenberg, Dekimpe & Introprice competition Steenkamp (2013) Categoryprice
Balachandra & Friar (1997) Categorycompetition Hoyer, MacInnis & Pieters (2013) Category sustainable (Lag-) Sales Cooper & Kleinschmidt (1987) LagUnitsales
Categorysales Purchasefrequency Advertising Van Heerde et al. (2013) (Lag-)Ownadvertising
5.1.4 Descriptives
The outcome variable unitsales is characterised by a left-skewed distribution with an overproportion of zeros. At a mean of 28.87, the zeros make up 11.9% of the total count, whereas the leftover outcomes are distributed between 0 - 7%. An independent sample t-test shows that the difference between sold units of non-sustainable and sustainable NPIs is significant over the course of the observation period (F (882) = 4.241; p < .05): on average 356 and 208 units of non-sustainable and sustainable NPIs have been sold, respectively. An overview of the descriptive statistics of all used variables can be found in table 5.
The correlation matrix (Appendix D) shows that unitsales negatively correlates with the two price variables introduction price and category price. This implies that unitsales are higher at lower introduction prices or lower average category prices. Furthermore, unitsales negatively correlates with category advertising but also with innovation. In other words, unitsales appear to be lower for higher advertised categories and NPIs with a higher degree of innovativeness. Despite a few exceptions, the correlations are below .42: the number of already existing products within a category (category competition) and sales per category (category sales) at .52, as well as
introprice and category price at .73, then shouldscore and the category price at -.58. The latter
implies, that in this case virtue products are higher priced than vice products.
Table 5
Variable operationalisation & descriptives
Variable Measurement Mean SD
Main Variables:
Unitsales Sold units (count) 28.87 56.89 Sustainable Organic/fair trade (yes/no)* - -
Marketing Measures:
Promotion Monthly price promotion (yes/no)** - - Innovation Degree of innovativeness
(averaged value taken from 4 items measured on 7-point-Likert scale)
Product Characteristics:
Shouldscore Score on category’s virtue character (first 7-point-Likert scale for want & should score respectively; then want score - should score = shouldscore)
0.11 2.29
Control variables Specification Mean SD
Month Month after introduction (dummy variable for 12 months)
0.7 – 0.8 0.269 - 0.276 Introdate Introduction month after 2008 20.10 9.68
Introprice Price at introduction per unit 0.64 0.76 Categoryprice Average category price 48.18 0.41 Categorycompetition Number of already existing products
within category
287.53 97.46 Category sustainable Number of already existing sustainable
products within category
6.04 6.075 Lagunitsales Sold units of previous month 26.59 55.26 Categorysales Sales per category 975460.18 854880.49 Purchasefrequency Average purchase frequency per category 23.91 21.13 (Lag-)Ownadvertising Own advertising (of previous month) 175596.78 405750.58 (Lag-)Categoryadvertising Category advertising (of previous month) 1782590.76 1564845.12 Distribution Weighted product distribution across
channels
0.85 0.17 *93.4% are non-sustainable NPIs
** 52.3% no price promotion took place
5.2 Analytical procedure
Hierarchical Linear Model
accounts for separate variation effects within and between-groups, where within-group variation is thus the individual-level effect, while between-group variation takes place across groups and levels (Snijders & Bosker 2012).
Because the variability within product categories (i.e. within-group variability) is assumed to be similar as compared to between product categories (i.e. between-group variability), we assume random intercepts and thus apply the Random Intercept Model, which is one of two types of random effect models in the HLM application.3 Here between-group
variability is expressed through an intercept that varies across groups. The intercept depends on the group mean and can take heterogeneous values, while the regression coefficients are homogeneous, i.e. held constant across groups4 (Heck & Thomas 2009; Snijders & Bosker
2012). The fact that observations within groups vary together is what violates an OLS’ assumption of independent error terms (Snijder & Bosker 2012).
Negative Binomial Model
To effectively answer our research questions, another aspect has to be considered: our dependent variable unitsales is based on count data, which means it is discrete, ordered and non-negative (Cameron & Trivedi 1986). Figure 2 shows its non-normal left skewed distribution, with a high proportion of zeros meaning that the mean and the variance are unequal. The Negative Binomial Regression accounts for such overdispersion (Cameron & Trivedi 1996).
3
The other method is called Random Coefficient Model. A combination of both Random Coefficient and Random Intercept Model is possible, too (Snijders & Bosker 2012)
4
Figure 2
Distribution of unitsales
The Negative Binomial Regression is part of the regression family and based on a negative binomial distribution. The model extends the Poisson distribution via a separate parameter, which allows for extra variance and in this manner for overdispersion (Cameron & Trivedi 1986; Moghimbeigi, Eshraghian, Mohammad & Mcardle 2008). To test the assumption of overdispersion, we use the Fisher Index Ratio of overdispersion, which uses the ratio of the variance to the mean (Kokonendji, Mizere & Balakrishnan 2008). If it is > 1 then the data is overdispersed: in our case the dependent variable unitsales has a variance of 326.434 and a mean of 28.87. The resulting ratio of 11.31 verifies overdispersion. We thus use a Multilevel Negative Binomial Model to measure monthly unitsalesi of SKUsj in product categoryk, where the
The following equations are derived:
Unitsales λijk = ß0jk + ß1Sustainableijk + ß2Shouldscoreijk + ß3Innovationijk + ß4Promotionijk + ß5Shouldscore*Sustainableijk +
ß6Innovation*Sustainableijk + ß7Promotion*Sustainableijk + ß8 Distributionijk + ß9Categorysustainableijk + ß10LagUnitsalesijk + ß11Categorysalesijk + ß12Purchasefrequencyijk + ß13Categorypriceijk + ß14Intropriceijk + ß15Ownadvertisementijk + ß16LagOwnadvertisementijk + ß17Categoryadvertisementijk + ß18LagCategoryadvertisementijk + ß19Introdateijk + ß20month02ijk + … +ß31month12ijk + Rijk,
(1)
ß0jk = δ00k + U0jk, (2)
δ00k = γ000 + V00k, (3)
where the letter i depicts level one dependency, while j and k stand for level two and three dependency. The variables Rijk,U0jk, V00k represent the error terms for level one, two and three
respectively. These allow the intercepts to vary across the three levels. Level one (1) includes the independent variable whether an NPI is sustainable, the utilised interactions and the respective main effects as well as additional control variables. The fixed coefficients are represented by the
ßs. We account for fixed time effects at level one and include the dummies month02 – month12
(see table 5 for a complete overview and specification of all utilised variables). As the equations suggest, we have three nested levels where level one (1) refers to monthly unitsales, level two (2) to the SKUs as such, while level three (3) represents the product categories for the respective SKUs.
6 EMPIRICAL RESULTS
6.1 Model Fit
is significant (χ2 (31) = 5129.29; p < .01), which shows that at least one parameter in our model is different from zero. The likelihood ratio test on whether a Multilevel Model is more appropriate as opposed to a model not accounting for multiple levels is also significant (χ2 (3) =
4282.77; p < .01). In other words, our Multilevel Negative Binomial Model fits our data better
than a regular Negative Binomial Model not accounting for hierarchical data structures (Snijders & Bosker 2012). Another test to verify the appropriateness of a Multilevel Model and the utilised groupings is the calculation of its intraclass correlations, which measure the dependency between variables for the utilised Random Intercept Model. Specifically, it shows whether variables within a group have correlating residuals. For calculation the following formula is used:
𝐼𝑛𝑡𝑟𝑎𝑐𝑙𝑎𝑠𝑠 𝑐𝑜𝑟𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛 = (𝜎𝑢)²
(𝜎𝑢)²+(𝜎𝑒)²=
𝑠𝑑(_ 𝑐𝑜𝑛𝑠)²
𝑠𝑑(_ 𝑐𝑜𝑛𝑠)²+sd(residual)², i.e. for level two the intraclass
correlation coefficient lies at 0.0255 and for level 3 at 0.2356. Both intraclass correlations are
between 0 and 1, which means that the groupings are useful and that there is enough variance to explain at the individual level, i.e. not every SKU is the same (Garson 2013; Snijders & Bosker 2012).
To assess whether the final model has more predictive power compared to the null model, we use the likelihood test statistic of both models as comparison. In this case we have calculated as follows: χ2 (31)= -2 (LLnull - LLfull) = 3320.7467. Our full model predicts significantly better
than the null model (χ2 (31) = 3320.746; p < .05), because it is larger than the critical value of 46.194. Also the Akaike Information Criteria (AIC) and the Bayesian Information Criteria (BIC) indicate a better model fit because both decrease in value from the null model to the full model: AICfull = 75681.09, BICfull = 75935.44; AICnull = 78939.83, BICnull = 78968.9 (Blattberg, Neslin
5Level two Intraclass Correlation = .0732531²/(.0732531² + .4577807²) = .02496644 6
Level three Intraclass Correlation = .3948769²/(.015714²+ .8143333²) = .235048444
7
χ2 (31)=-2log (Lnull/Lfull) = -2 (LLnull - LLfull) = -2 (-39465.913 - (-37805.54)) = 3320.746
& Kim 2008; Snijders & Bosker 2012; Schwarz 1978). The parameter lnalpha accounting for the overdispersion of our data is significant, too (ß= -.756, p < .01) and indicates that the Negative Binomial Model is a better fit than a Poisson Model, which does not account for overdispersion. Exponentiating lnalpha8 gives a result larger than 0 (e -0.756 = 0.47), which once again proves the
overdispersion of our data.
6.2 Results
STATA undertakes the analysis using the Maximum Likelihood Estimation to estimate components and automatically tests and accounts for heteroscedasticity during this process. The outcomes show that sustainability labels have a significant negative effect on new product success (ß = -.851, p < .01), which is not in line with the positive effect as expected: H1 is rejected. This negative effect also holds in a Main Effects-Only Model (ß = -.190 p < .01). The degree of innovativeness has a significant positive impact on new sustainable product sales (ß =
.091, p < .05), which confirms our expectations: H2 is accepted. However, in the Main
Effects-Only Model the same variable turns out insignificant (ß = -.022, p = .068). Price promotion
positively impacts new sustainable product sales (ß = .153, p < .05), which conforms to our theoretical reasoning and lets us accept H3. In the Main Effects-Only Model price promotion has a positive significant effect, too (ß = .903, p < .01). The categorisation of sustainable NPIs according to the shouldscore has a significant negative effect on new sustainable product sales (ß
= -.049, p < .05), which is contrary to our expectations: H4 is rejected. The same significant
negative effect is shown in the Main Effect-Only Model (ß = -.292, p < .01).
Concerning the control variables (see Appendix C), distribution (ß = 2.23, p < .01) has a significant positive impact on new product success, while introprice (ß = -.179, p < .01) has a
8
significant negative effect. Additionally, lagunitsales (ß = .004, p < .01) and category sales (ß =
-.000, p < .01) have a significant positive and negative effect on unitsales respectively, while purchase frequency (ß = -.009, p < .01) as well as the dummies for month03, 04, 06 (month03: ß = -.126, p < .01; month04: ß = -.105, p < .05; month06: ß = -.086, p < .05) all significantly
impact new product success negatively. Lastly, introdate (ß = .006, p < .05) has a positive impact on new product success. Notably, the control variables dealing with advertising, i.e.
category advertising, Ownadvertising and LagOwnadvertising are insignificant, while LagCategory advertising (ß = -.000, p < .01) has a significant negative impact. Similar effects
Table 6
Overview of multilevel negative binomial regression results
Parameter Coefficient Std. Error
Main effects: Sustainable -.851** .161 Innovation -.042* .0178 Shouldscore .0834** .0144 Promotion .889** .020 Interactions: Sustaintable*innovation .091* .039 Sustainable*promotion .153* .075 Sustainable*shouldscore -.049* .02 Control variables: Categorysustainable -.000 .003 LagUnitsales .004** .000 Categorysales -.000** .000 Distribution 2.23** .171 Purchasefrequency -.009** .001 Category price -.047 .066 Introprice -.179** .029 LagCategoryadvertising -.000** .000 Categoryadvertising .000 .000 LagOwnadvertising -.000 .000 Own advertising .000 .000 Category competition -.000 .000 Introdate .006* .002 Month02 -.037 .036 Month03 -.126** .036 Month04 -.105** .037 Month05 -.022 .037 Month06 -.086* .037 Month07 .02 .036 Month08 -.039 .037 Month09 .066 .038 Month10 -.050 .037 Month11 -.042 .037 Month12 -.017 .036 Constant: .066 .153 /lnalpha -.756* .019 Keycat: var(_cons) .073 006 keycat>id: var(_cons) .395 .016
Model fit statistics: LLfull
-37805.544 χ2 (31)= 5129.29** LLnull -39465.913 χ2 (0) = - Number of observations: 10584 **significant at p < .01 *significant at p < .05 6.3 Robustness checks
Anderson, Babin & William 2009). The full model shows higher VIFs ranging from 1.13 to 8.54, which we explain by the included interaction effects.
Another robustness test includes the estimation of an OLS, ignoring both the dependent variable’s count character, and the hierarchical data structure. Whether a new product was sustainable or not (sustainable) still has a significant negative impact on new product success (ß
= -.219, p < .05). The innovation-interaction is significant, yet with a positive impact on new
product success (ß = .159, p < .01). The shouldscore-interaction remains negative and significant (ß = -.069, p < .01), while the price promotion-interaction turns insignificant. The insignificant effects are likely to be caused by ignoring the nested data and count character of the dependent variable.
Lastly, we replace the dependent variable unitsales with a continuous variable namely
sales9. The predictor sustainable (ß= -.069, p < .05), the innovation- (ß = .05, p < .05), price
promotion- (ß = -.034, p < .05) and shouldscore-interaction (ß = -.035, p < .01) all stay
significant. The variable sales is measured in Euros instead of sold units and can therefore bias outcomes when comparing sustainable and regular NPIs, as regular products typically do not comprise the premium pricing sustainable items do.
7
DISCUSSION
Enterprises increasingly follow the growing demand for green products by allocating resources into the development of environmentally and socially sustainable new products. However, the high failure rate of NPIs and the low actual market share of sustainable products poses the question whether launching sustainable NPIs leads to more success as compared to regular NPIs. This knowledge gap inspired this paper’s topic, where we propose a framework
9
that considers marketing measures (i.e. price promotion and the degree of innovation), and product category characteristics (i.e. whether a product is classified as vice or virtue product). From that the following research questions are derived:
1. Are sustainable new product introductions more successful than non-sustainable new product introductions?
2. What effect do marketing measures and product characteristics have on sustainable new product introduction success?
To be able to answer these questions, we use data from the GfK panel, which includes 884 NPIs within 14 categories of which 6.56% are sustainable NPIs. We estimate a Multilevel Negative Binomial Model to gain theoretical and practical insights. All investigated factors have a significant impact on NPI success, although the effect counters our expectations in specific cases. Table 7 again summarises the hypotheses and the expected along with the actual effects.
Table 7
Overview of findings according to hypotheses
Hypothesis Variable Expected
effect
Actual effect
H1: Sustainable product introductions are more successful in terms of new product sales than non-sustainable product introductions.
Sustainable (yes/no)**
+ -
Marketing measures
H2: The degree of product innovativeness has a positive effect on sustainable new product introduction success.
Degree of innovation*
+ +
H3: Price promotions have a positive impact on unit sales of new sustainable product introductions.
Price promotion (yes/no)*
+ +
Product Category Characteristics
H4: Sustainable new product introductions are more successful in terms of unit sales in virtue than in vice categories
Shouldscore* + -
*significant at p > .05 **significant at p > .01
general are stronger than the risk reducing properties of sustainability labels on NPIs. Literature discusses these negative associations with diverging findings and predominantly refers to the dimensions of price, quality and taste (e.g. Bezawada & Pauwels 2013; Bourn & Prescott 2002; van Doorn & Verhoef 2011; Gleim et al. 2013; Kihlberg & Risvik 2007; Luchs et al. 2010). Premium pricing is mentioned as a strong argument for consumers to abstain from purchasing sustainable (e.g. Gleim et al. 2013), while scholars classify the dimensions of quality and taste as priority factors in purchase decisions concerning food and green items (Vermeir & Verbeke 2006; Newman et al. 2014; Tsiotsou 2006). Considering that our utilised dataset includes a large number of food products, the negative quality and taste associations in combination with the importance of these dimensions for food and green items alongside their premium pricing, might explain our finding. Relating this to the first research question, it can be said that sustainable NPIs are not more successful than regular NPIs.
Figure 3
The effect of product innovativeness on (sustainable) new product introduction success
Figure 3 depicts how innovativeness affects sustainable compared to regular NPI unit sales and demonstrates its significant effect on sustainable NPI success. The higher the degree of innovativeness, the more successful sustainable NPIs are, even exceeding regular NPIs’ sales at a low degree of innovativeness. This is an interesting finding, because according to our first research question and H1, regular NPIs are more successful than sustainable NPIs. This demonstrates the influential role the combination of sustainability labels with product novelty has on sustainable NPI success – it reinforces FMCG unit sales and lessens the effect of negative consumer associations on sustainable products. Regular NPIs are not significantly impacted by this factor as of our study, which we suspect is related to our focus on low to incremental innovativeness and the highly competitive FMCG industry, where major product advantage is necessary for consumers to opt for a product – without the sustainability label NPIs seem to vanish among the masses.
consumers use premium prices as indicator for intrinsic product attributes, e.g. a sustainable NPI’s higher quality (Olson 1977). Premium prices can be a major reason inhibiting new product trial (Gleim et al. 2013), yet together with a temporary price reduction they work as stimulant by making usually expensive products more affordable. Price promotions lower trial barriers and tempt consumers into purchase by lessening the risk that they perceive with higher priced products (e.g. Gleim et al. 2013; Cronley et al. 2005), confirming Ngobo’s (2011) line of reasoning who also mentions the price-quality heuristic. The negative associations that appear to be impactful for H1 seem to dissolve in the background when product prices are temporarily lowered in a similar way as for H2 (i.e. the degree of innovativeness). However, at regular prices these prevail again, which confirms researching dealing with negative associations and sustainable products (e.g. Gleim et al. 2013; Luchs et al. 2010; Vermeir & Verbeke 2006).
Figure 4
The effect of price promotions on (sustainable) new product introductions
important marketing tool for both NPI types, although the effect itself is visibly stronger for sustainable NPIs.
Third, we find that sustainable virtue NPIs (i.e. those with a higher shouldscore) are less successful in terms of unit sales than sustainable vice NPIs (i.e. those with a lower shouldscore), which contrasts our expectations. Figure 5 depicts a similar effect for regular vice and virtue NPIs and also shows that virtue NPIs in general are less successful – whether sustainable or not. Interestingly, sustainable new vice products are almost as successful as regular new vice products yet again less successful than virtue NPIs.
where cleaning power is a main differentiating feature. Additionally, sustainable products’ price premiums reinforce these negative associations and make consumers more reluctant to try.
Figure 5
The effect of a category’s shouldscore on (sustainable) new product introductions
A noteworthy aspect is that the in this paper utilised vice and virtue categorisation is according to Milkman et al. (2010), who have based their measurement on consumer assessment. We posit that this can also be biased as to what extent consumers perceive categories to be more of a should or want category, possibly influencing our research outcome. Specifically, the determination of non-food items as vice or virtue products depends on perception and can thus bias outcomes.
al. 2004). Another interesting finding is the negative influence of the average purchase frequency per category on NPI’s success, which means that NPI success is lower in categories with a high average purchase frequency. We relate this to consumers’ low involvement and their tendency to engage in straight rebuy of familiar items (e.g. milk) without considering any alternatives or further information search (e.g. Wind & Mahajan 1988). Lastly, the number of already existing sustainable products within a category positively affects NPI success, which we link to consumers’ probability of noticing changes within a category. The low market shares suggest sustainable products to be underrepresented within categories, so the more sustainable NPIs exist the more likely it becomes that consumers notice and purchase them through e.g. enlarged in-store visibility (e.g. Aertsens 2009; Steenkamp & Gielens 2003).
Regarding the second research question, we conclude that marketing measures and product characteristics are crucial factors for both regular and sustainable NPI success.
7.1 Managerial implications
NPI, the higher sustainable NPI success, exceeding that of regular NPIs. Therefore, we advise managers to launch sustainable NPIs at higher degrees of innovativeness.
Moreover, our findings illustrate the positive effect of price promotions on both regular and sustainable NPI success. Managers should thus consider using temporary price reductions to ensure high initial sales. Should managers decide against price promotions due to their short-term effects, we advise launching regular NPIs because these are more successful in this scenario.
Then, we find that new products of the vice category are more successful than those of the virtue category, no matter whether products are sustainable or not. Managers consequently are advised to launch vice NPIs. Should managers still decide to launch new virtue products, chances for success are larger with regular instead of sustainable NPIs. On top of that, we suggest managers to critically reflect over the date on which to launch its NPIs, as this has potential positively effects on unit sales, and to launch in a category that already includes a number of sustainable products, since this enlarges visibility and chances for success. Table 8 summarises the managerial recommendations.
Table 8
Summary of managerial recommendations for marketing measures & product category characteristics NPI success factors Recommendations for improving NPI success
Marketing Elements:
Degree of Innovativeness
Low Introduce a sustainable NPI version High Introduce a sustainable NPI version Price Promotion
Yes Introduce a regular/ sustainable NPI version* No Introduce a regular NPI version
Product Category Characteristics:
Vice Introduce a regular/sustainable NPI version* Virtue Introduce a regular NPI version
