Going Green for a healthy Planet or a
guaranteed Return on Investment?
The effect of signaling warranty and different types of benefit in
marketing communications on willingness to pay for green services
25th of June 2014, Amsterdam
Bachelor Thesis Business Studies
Pieter van Exter
Supervisor: Bram Kuijken
2
Abstract
Increasing the energy efficiency of residential buildings is not only beneficial for the environment, it is increasingly regarded as an investment opportunity. From a green marketing perspective it is important to understand which offer appeals most to whom. For this study, a Vickrey auction was set-up, measuring real willingness to pay for a personal energy efficiency advice and establishing the impact of an added financial and environmental benefit. The effect of warranty signaling was tested as well. The level of environmental involvement and warranty signaling appear to have a significant and strong positive effect on the willingness to pay. However, willingness to pay was not significantly affected by the type of benefit that was shown to consumers, regardless of their level of environmental involvement
Keywords: green marketing, marketing communications, environmental involvement,
3
Contents
Abstract ... 2 Acknowledgements ... 4 1. Introduction ... 5 2. Theoretical framework ... 8 2.1 Diffusion of innovations ... 82.2 Consumer decision model ... 9
2.3 Perceived risk ... 10
2.4 Warranty signaling ... 11
2.5 Consumer response towards green products ... 11
2.6 Consumer response towards green advertisements ... 12
2.7 Conclusion ... 13
3. Conceptual Framework ... 14
3.1 Benefit signaling ... 14
3.2 Level of environmental involvement ... 14
3.3 Personal benefit signaling ... 15
3.4 Common benefit signaling ... 16
3.5 Perceived risk and warranty signaling ... 17
4. Methodology... 18 4.1 Research design ... 18 4.2 Research sample ... 20 4.3 Procedure ... 21 4.4 Measures ... 21 5. Results ... 24 5.1 Descriptive statistics ... 24 5.2 Reliability ... 26 5.3. Correlations ... 27 5.4 Hypotheses testing ... 27 6. Discussion ... 32 6.1 Findings ... 32 6.2 Theoretical contributions ... 35 6.3 Managerial implications ... 35
6.4 Limitations and recommendations for future research ... 35
7. Conclusion ... 37
Bibliography ... 38
Appendix A: Six Treatments ... 43
4
Acknowledgements
With this thesis I am completing my bachelor Future Planet Studies with a major in Business Studies at the University of Amsterdam. I have always been interested in the apparent contradiction between the environment and economy. In my opinion these two domains are actually strongly related and dependent on each other. I am pleased that I found more supportive evidence for this during my research.
I would like to thank Gijs van Wijk and Marjan Minnesma from Urgenda for their support and inspiration. Organizations such as Urgenda give me hope that we can mitigate climate change before the most severe predictions by the IPCC become reality and make me proud to be involved with.
I am also very thankful for all the support and feedback I received from my supervisor Bram Kuijken. The quick responses and ideas from him made writing my thesis much more enjoyable. I really belief in the strength of Veylinx especially because of the dedicated people behind it, working often till after midnight.
5
1. Introduction
According to the latest assessment report by the Intergovernmental Panel on Climate Change (IPCC) released in 2014, global warming is becoming an emerging problem in the 21th century. Due to the intensive use of fossil fuels by men, severe consequences are expected: heavy floods, risk of food shortages, rise in mortality due to extreme temperatures and loss of many species and important ecosystems (IPCC, 2014). Even the government of the United States, which used to be skeptical about the human influence on global warming, released recently a Climate Action Report. Secretary of State John F. Kerry spoke of the ‘(..) Most urgent and profoundly
complex challenges we face (..) we cannot afford today to delay the reckoning with climate change’ (CAR, 2014 p. 2).
Commercial and residential buildings in developed countries are responsible for approximately 40% of the total energy consumption (EIA, 2013). Initiatives to reduce this in existing buildings, to decrease carbon-dioxide emissions and to become less dependent on fossil fuels, receive increasing attention over the last years from policy makers, NGO’s, multinationals and consumers. A growing number of individuals and companies have applied energy saving measures and installed applications, such as solar panels and wind turbines, that actually generate renewable energy.
Reducing fossil energy consumption is not only beneficial for the climate, it can be regarded as a lucrative financial investment as well, due to increasing energy prices and decreasing prices for new energy efficiency improvements. The average energy bill has risen three times faster than the Consumer Price Index in the Netherlands over the last years (Compendium voor de leefomgeving, 2014), while the price per watt of solar power has become hundred times cheaper in 35 years (BNEF, 2013). The payback period for Dutch consumers was in 2013 between eight and ten years, while they last at least 25 years (ECN, 2013).
Increasing energy efficiency of buildings is becoming a new industry and potentially billions of dollars could be saved annually by property owners (Koopmans & Te Velde, 2001; IEA, 2009; Hirst & Brown, 1990; Backlund et al., 2012; Copenhagen Economics, 2012; McKinsey, 2009). An iconic example of an energy saving investment is the Empire State Building in New York City which was completely assessed and renovated in 2009, reducing the energy demand with 38%, resulting in more than 4 million dollars annual savings.
Thus, there is both great environmental and economic relevancy to develop the energy efficiency industry. This research focusses on the effectiveness of highlighting the financial
6
benefits or environmental benefits in marketing communication messages, promoting energy efficiency renovation to house owners.
Reviewing the existing literature, there seems to be an increasing demand for environmental friendly or green products in general (Grimmer & Woolley, 2012). However, the reaction of consumers towards green products appears to be dependent on personal variables and the consumers’ level of awareness of environmental problems. Consumers who are more concerned about the environment will, on average, valuate green products higher , compared to lesser involved consumers (D’Souza & Taghian, 2005; Grimmer & Woolley, 2012). Grimmer & Woolley (2012) found that consumers with low environmental involvement favored products, promising a personal (financial) benefit . However this was not confirmed in other studies. Some found more moderate correlations between environmental involvement and motivation for green products, while some other studies even suggest an inverse relation. Many articles argue that the literature on responses towards green advertisements is still limited (D’Souza & Taghian, 2005; Grimmer & Woolley, 2012; Do Paco & Reis, 2012).
Another individual characteristic which could have influence on the appeal of energy saving measures, is the innovativeness of the consumer. According to the diffusion of innovations theory (Rogers, 2003), new innovations are firstly adopted by a small atypical group of consumers who are usually higher educated, more involved and more resistant to risk (Rogers, 2003). Warranty could take away some of the customers uncertainty. According to the signaling theory, a warranty is a signal that can contribute to a higher perceived quality and reliability (Murthy & Blischke, 2006; Murthy & Djamaludin, 2001). Warranty can be seen as an important marketing tool for new products, since it has a significant positive influence on consumer product choice (Murthy & djamaludin, 2001).
In order to retrieve additional knowledge about effective ways to promote innovative and green products and services, the relation between the level of environmental involvement and the attitude towards personal (financial) or more altruistic (environmental) benefits, is studied, as well as the added value of warranties in marketing communications of green services. Therefore the main research question is: What is the effect of signaling warranty and different
type of benefits in marketing communications on the willingness to pay for green services?
In order to obtain an answer to the research question, an empirical experiment is set-up to measure Willingness to Pay (WTP). The research is conducted via an online Vickrey auction site ‘veylinx’, which is a research project of the Business School of the University of Amsterdam.
7
First, the existing literature will be reviewed on green marketing, the level of environmental involvement and the effect of consumers level of innovativeness on willingness to pay. Based on the retrieved knowledge, a conceptual framework is developed with a number of hypotheses. After the framework is set-up, the research design and methodology are described after which the results are presented. At the end of this paper the research will be reflected upon in the discussion and a general conclusion will close this research.
8
2. Theoretical framework
In the theoretical framework, literature is evaluated on new and innovative products and services, the consumer response towards innovative products and the effect of warranty signaling. Thereafter, previous studies are reviewed looking at the relation between level of environmental involvement and attitude towards green products and advertisements.
2.1 Diffusion of innovations
The possibilities of installing solar panels and applying energy saving measures in such a way that a building becomes self-sufficient (in terms of energy) is still relatively unknown in the Netherlands. Only a few experiments with zero energy renovations of houses have been conducted so far as part of government programs. Although exact numbers are lacking, probably less than hundred houses in the Netherlands are now completely energy neutral. Previous researchers found some general theories about the introduction and adoption of new innovative products. One of the most important ones is the diffusion of innovations theory. This theory was formulated by Rogers in 1962. According to Rogers, new products will undergo a certain adoption process. The first adopters appear to have different characteristics compared to their followers. Rogers distinguishes five different consumer groups based on the time it takes them to adopt an innovation.
Figure 1 five distinct consumer groups based on adoption rate (Rogers, 2003)
According to Rogers (2003), an innovation will first be adopted by a relatively small number of consumers, who are referred to as innovators and early adopters. The majority, divided into an early majority and a late majority, follows. The majority consists of approximately 68% of the consumers. The laggards are the latest group of consumers to adopt a new innovation. Early adopters are relatively higher educated, have greater empathy, enjoy a higher status and income, can deal better with uncertainty and risks and are more involved with latest trends in technology (Rogers, 2003).
9
Diffusion of innovations theory has been applied on many different product types over the last decades. An example is the adoption of solar panels. In the Untited Kingdom a significant difference was found between the early adopters and the early majority. Compared to the first group, the early majority perceived solar panels to a greater extent as unattractive, unaffordable and insufficiently subsidized. The payback time of the investment has to be reduced in order to convince the early majority (Faiers & Neame, 2006).
2.2 Consumer decision model
Not only did Rogers distinguishes five different consumer groups and their rate of adoption, he also developed a consumer decision model. This contains five succeeding stages and describes the process of adopting or rejecting the innovation. The first stage is knowledge, this is where the consumer comes in contact with the product for the first time. It is followed by persuasion: the retailer tries to convince the consumer, advertising the benefits of the innovation Thirdly is the decision moment, when the consumer decides to purchase the product or service after weighing all the benefits. After the decision, the product is implemented. The final stage is called the confirmation: the product is evaluated and reviewed.
This decision model has been used as a starting point in many researches of marketing related topics. The adapted decision model by Wilson and Dowlatabadi (2007) (figure 2) makes clear that the knowledge phase is affected by the innovativeness of the adopter. The more a consumer has the characteristics of an early adopter, the more positive he or she will react towards the offered product or service (Wilson & Dowlatabadi, 2007). A second influencing factor is the perceived attractiveness of the innovation.
10
Figure 2 traditional consumer decision model adapted by Wilson and Dowlatabadi (2007)
The attractiveness of the innovation is determined by five perceived attributes, according to Rogers (2003): - relative advantage - compatibility - complexity - trialability - observability
According to Rogers (2003), relative advantage and compatibility are found to be the most important attributes. Völlink et al. (2002), found that for consumers the perceived relative
advantage of energy conservation measures is in some cases the first and most critical attribute
that is reviewed by consumers. If the relative advantage is perceived as low, they stop the evaluation and reject the conservation measure without considering other attributes.
2.3 Perceived risk
According to Labay and Kinnear (1981), a sixth attribute should be added to determine the attractiveness of an innovation. This attribute is the perceived risk that comes with adopting.
Perceived risk would be inversely related to adoption. The level of perceived risk can have a
11
The concept of risk is divided into five sub-types of risk described by Murthy and Blischke (2006): - Financial risk - Performance risk - Physical risk - Social risk - Psychological risk
Of these five sub-types financial risk, the risk of losing money, and performance risk, the risk of malfunctions, are found to be the most important types of risk regarding purchase decisions. For most consumers, the perceived risk is inversely related to the willingness to try a new product (Murthy & Blischke, 2006). The perceived risk of services is found to be higher than the perceived risk of products. As a service is perceived by consumers as more intangible and less clear what the service actually comprehends and what not (Murray & Schlachter, 1990).
Labay and Kinnear (1981) reviewed the adoption of solar panels in the 1980’s and found that adopters perceive a solar energy system lower in financial and social risk compared to non-adopters.
2.4 Warranty signaling
According to signaling theory, warranties can be signals addressing both product quality and product reliability. Thus, perceived risk can be decreased with risk mitigating signals such as warranties (Murthy & Blischke, 2006; Murthy & Djamaludin, 2001; Havlena & Desarbo, 1991). Warranty signaling can be an important marketing tool, since it has a significant positive influence on consumer product choice (Murthy & djamaludin, 2001). An example of warranty contributing as a marketing tool is Hyundai which quadrupled its sales in the United states after offering a 7-year warranty (Murthy & Blischke, 2006).
For innovative products, an outstanding warranty does reduce the perceived financial risk, however it appears to be incapable of reducing the perceived performance risk (Shimp & Bearden, 1982).
2.5 Consumer response towards green products
Consumers are willing to pay higher prices for environmentally friendly or green products (Haytko & Matulich, 2008). This applies also for house owners who are willing to pay a higher price for energy conservation measures, whether they had previous experience with green products or not (Chau & Chung, 2010). Another research on the willingness to pay for
12
improvements in environmental performances of residential buildings, found that the marginal willingness to pay to reduce the energy demand by 50% is approximately 18.825 dollar, which is about 2.4 times the actual price it would cost to achieve this reduction (Park et al., 2012). A note should be made here, since marginal willingness to pay is not the same as the real willingness to pay. Since it is an hypothetical question without actual consequences, marginal willingness to pay is usually higher (Voelckner, 2006).
A positive effect on the environment is not the only reason for house owners to implement energy saving measures. Financial benefits are often regarded as much more important. Initial costs and the amount of savings, were for more respondents more important factors than environmental protection (Zhao et al., 2012).
2.6 Consumer response towards green advertisements
Consumers can be informed about the benefits of an energy neutral renovation of their house through all sorts of different media: word-of-mouth, news articles or commercials and advertisements. Mentioning either financial benefits or environmental benefits in advertisements, appear to appeal to different consumer groups. However the literature on responses towards green advertisements is limited (D’Souza & taghian, 2005; Grimmer & Woolley ,2012; Do Paco & Reis, 2012) and appear to have contradictory outcomes.
A famous experiment conducted by Schuhwerk and Lefkoff-Hagius (1995) did look at the effect of different appeals in advertisements of green products and the concern about the environment of the consumer. Two different advertisements of a green product were shown to the respondents; one with a green appeal and the other with a financial non-green appeal. The type of appeal did not matter to respondents who were more concerned about the environment. However, consumers with low environmental concerns did valuate the advertisements with the green appeal just as much as the group with high concerns about the environment. The group that was little concerned about the environment did value the advertisement with a cost-saving appeal significantly lower, since their general perception is that green products are more expensive and that therefore this advertisement is not credible (Schuhwerk & Lefkoff-Hagius, 1995).
A similar research was conducted by Grimmer and Woolley in 2012, who studied the effect of signaling personal (financial) benefits or environmental benefits and the personal level of environmental involvement. The results showed, a personal benefit added to a green product does not increase the purchase intention, compared to adding an environmental message,
13
looking at the entire research sample. However, the environmental message did increase the purchase intention of consumers with a high environmental affect, while respondents with low environmental affect were more intended to purchase the product with a personal benefit (Grimmer & Woolley, 2012). The first outcome is in line with the experiment by Schuhwerk and Lefkoff-Hagius (1995). However, the fact that less involved participants reacted more positively towards the personal benefit during the experiment by Grimmer and Woolley (2012), is contradictory to its outcome. Research by Tucker et al. (2012) found that consumers with high environmental involvement are more sensitive to green advertisements, while consumers with low environmental involvement had a stronger disregard towards green advertisement according to D’Souza and Taghian (2005).
2.7 Conclusion
Previous research showed that new and innovative products are first adopted in general by a specific group which is relatively higher educated, more involved with development of new technologies and less vulnerable to risk. Besides the personal characteristics, the perceived attributes of the innovation are also of importance during the decision process. Relative advantage and compatibility are found to be the most important attributes, however research on solar energy adoption concluded that perceived risk should be added as a sixth attribute, which is inversely related to adoption. Perceived risk could be mitigated by warranty signaling. Just as there are specific responses towards innovative products, the consumer characteristics play an important role in purchase behavior towards environmentally friendly products. Although there is a limited amount of research conducted, there seems to be a relationship between the level of environmental involvement and the attitude towards green advertisements and purchase behavior and the attitude towards green advertisements. Although the literature is not unambiguous, consumers with low environmental involvement tend to be more susceptive to environmental messages.
14
3. Conceptual Framework
Based on the theoretical framework, the conceptual framework is defined in this chapter. It is the fundament for the research. First, hypotheses about the effect of benefit signaling and level of environmental involvement on WTP are formulated. Secondly, the expectations about the relations between the level of environmental involvement and the willingness to pay for common benefit and personal benefit are stated. Secondly, hypotheses are presented regarding innovativeness of the consumer, its effect on perceived risk and ultimately on the willingness to pay.
3.1 Benefit signaling
Consumers with more knowledge of the product can better determine the real value of it (Bei & Widdows, 1999). The actual value of the auctioned service in this experiment will be unknown to all respondents, since it is not available on the market and it is unlikely that respondents have much to refer to in their own environment. The provided description of the service is also limited. Therefore it is expected that when extra text is added such as a benefit, financial of environmental, the value of the service is perceived higher (Bei & Widdows, 1999). An experiment by Grimmer and Woolley (2012) looked at the differences in consumer reaction between three different advertisements of a water bottle. Environmental benefit, personal benefit or no benefit was signaled in the advertisements. No significant difference was found between any of the advertisements. In the discussion, the ordinariness of the promoted product was addressed as one of the reasons for lack of effect. It was suggested that an effect might be found for more high-involved products (Grimmer & Woolley, 2012). The following first hypotheses is formulated:
H1: Mentioning any benefit in an advertisement has a positive effect on willingness to pay
3.2 Level of environmental involvement
In general consumers with a high level of involvement with the environment were found to differ in several aspects from consumers who are less involved . For instance, consumers with
BENEFIT SIGNALING
WILLINGNESS TO PAY
15
great concerns about the environment are on average higher-educated and have a higher income (Schwepker & Cornwell, 1991; Straughan & Roberts, 1999). This higher socio-economic status could result in a higher WTP and therefore H2 is formulated as follows:
H2:Respondents with a high level of environmental involvement are willing to pay more for an environmentally friendly service compared to respondents with a low level of involvement
3.3 Personal benefit signaling
Consumers expect that environmentally friendly products are sold at premium prices and a trade-off is made between price and environmental impact (Peattie, 2001). Consumers with high level of environmental involvement are more likely to pay the premium, since they believe in the importance of green products. Consumers with less environmental concerns are less likely to pay for reduced environmental impact, since they are not as much aware of environmental issues. With increasing fossil energy prices and decreasing renewable energy prices, this trade-off is disappearing. Investing in energy efficiency and solar panels is more and more considered to be a financial investment. It is expected that less involved consumers are willing to pay more for a service that is to their personal benefit, rather than to a common/general benefit.
H3a: Respondents with a low level of environmental involvement are willing to pay more for a service if a personal benefit is mentioned compared to a general benefit or no benefit at all. H3b: Respondents with low level of environmental involvement are willing to pay more for a service if a personal benefit is mentioned compared to respondents with a high level of environmental involvement. LEVEL OF ENVIRONMEN-TAL INVOLVEMENT WILLINGNESS TO PAY
+
16
3.4 Common benefit signaling
Some consumers which are concerned about the human impact on the environment have an altruistic believe that they have to reduce their personal impact on the environment as much as they can. Research has shown that these consumers are prepared to pay a premium for environmentally certified products (Vlosky et al., 1999). Others found that in advertisements, an environmental benefit (common benefit) appeals more to consumers with a high level of environmental involvement (Grimmer & Woolley, 2012; Tucker et al., 2012).
H4a: Respondents with a high level of environmental involvement are willing to pay more for a service where a common benefit is mentioned compared to a personal benefit or no benefit at all.
H4b: Respondents with a high level of environmental involvement are willing to pay more for an offered service where a common benefit is mentioned compared to respondents with low level of environmental involvement.
COMMON BENEFIT SIGNALING WILLINGNESS TO PAY HIGH LEVEL OF ENVIRONMENTAL INVOLVEMENT
+
PERSONAL BENEFIT SIGNALING LOW LEVEL OF ENVIRONMENTAL INVOLVEMENT WILLINGNESS TO PAY+
17
3.5 Perceived risk and warranty signaling
The literature reviewed in the theoretical framework showed an inverse relation between the perceived risk of a product or service and the actual adoption (Yeung & Morris, 2006; Murthy & Blischke, 2006). Perceived risk is a subjective concept and previous research showed that the level of perceived risk is partly explained by the innovativeness of the consumer. According to Rogers (2003), early adopters are in general higher educated, have higher incomes and are more resistant to risk compared to late adopters. Another perceived risk mitigating method could be guaranteeing the quality of the offered service. Adding a warranty system to a product can have a significant positive effect on purchase behavior of consumers (Murthy & djamaludin, 2001). This warranty signaling is a marketing tool which has been successfully used in the automobile industry for instance (Murthy & Blischke, 2006).
H5a: Warranty signaling has a positive influence on WTP. H5b: Warranty signaling has a negative effect on perceived risk.
H5c: Innovativeness of the consumers has a negative effect on perceived risk.
H5d: Respondents that consider the perceived risk of the offered service as low are willing to pay more compared to respondents with a high perceived risk
WILLINGNESS TO PAY INNOVATIVENESS OF THE CONSUMER WARRANTY SIGNALING
-
PERCEIVED RISK-
-
18
4. Methodology
In this chapter the research design, research sample and the process are described and justified. subsequently, measurements scales are selected and explained.
4.1 Research design
4.1.1 Vickrey Auction
There are multiple methods of measuring WTP such as contingent valuation, conjoint analysis, auctions, lotteries and revealed preference data. The contingent valuation and conjoint analysis method measure a hypothetical WTP, since the participants do not actually have to pay for the product during the experiment and thus there are no financial consequences for them (Balisteri et al., 2001; Voelckner, 2006). It was found that respondents are willing to pay more for a product or service when they only have to fill in an amount, instead of actually paying for it. Thus, experiments measuring hypothetical WTP overstate the amount consumers are willing to pay up to three times compared to real WTP (List & Gallet, 2001). A proved method to measure real WTP, is a Vickrey auction. This type of auction is invented by William Vickrey in 1961 and is often referred to as second-price-sealed-bid-auction. Participants place an anonymous bet on the offered product. At the end, the highest bidder wins but pays the second-highest price. This system of paying the second highest price will lead to the most truthful bids, since bidding strategies other than offering the real amount of maximum WTP turn out to be disadvantageous. Vickrey auctions have been widely used for experiments for three properties: the auction is weakly demand revealing, the rules are relatively easy to explain and bidders set a market-clearing price (Lusk & Shogren, 2007).
4.1.2 Auctioned service
The service that will be offered at the auction is a tailor-made advice on how your house could become completely energy neutral. An energy neutral house is defined as a building which generates its own building-related and non-building-related energy demand in or around the house via renewable energy sources such as wind, solar power or biomass (Marszal et al., 2011). First, the condition of the building is assessed as well as the personal energy use. Based on the gathered data, a number of measures are proposed in a report, with the expected investment costs and savings based on calculations from a specialized software program. The advice will be provided by in2ecobuildings, which has more than 25 years of experience with energy efficiency advice, especially in the non-residential sectors. In2ecobuildings is now focusing more on the residential market and is offering a warranty, when the advice is actually executed.
19
The costs of the advice is around 1200 euro’s. This is financed by stichting Urgenda, a Dutch NGO which aims to accelerate the transition to a sustainable society and economy. Urgenda is currently experimenting with energy neutral renovations and has set the ambition to renovate the entire existing building stock in the Netherlands within 20 years. Urgenda is cooperating with this research, since the results are potentially relevant for the NGO. It could give insight in the attitude of different types of consumers towards different types of benefit and warranty signaling. Urgenda could adapt its marketing strategy based on the outcome.
4.1.3 Different Treatments
In order to test the hypotheses, different treatments were prepared. The energy neutral advice will be auctioned with a financial argument, an environmental argument or no argument. And additionally three treatments contain warranty signaling and three do not.
The six different treatments that will be shown:
- Warranty, no financial argument, no environmental argument (WNA) - Warranty, financial argument, no environmental argument (WFA) - Warranty, no financial argument, environmental argument (WEA)
- No warranty, no financial argument, no environmental argument (NWNA) - No warranty, financial argument, no environmental argument (NWFA) - No warranty, no financial argument, environmental argument (NWEA)
The financial argument is described as:
When your house is energy neutral, you don’t longer have an energy bill.
The environmental argument is described as:
20
Figure 3 One of the six treatments. This example contains a financial argument and warranty (WFA)
All six treatments are enclosed as Appendix A
4.2 Research sample
Usually, only subscribers of veylinx.nl are invited for the auction. This group consists of approximately 3400 respondents and is a reasonably fair representation of the Dutch population, regarding age, gender and education. Based upon experience between 600 and 900 members of veylinx usually take part in the auction.If the offered product is not of any interest to the bidder, he or she can offer 0 euro’s. The percentage of zero-bidders is very volatile and ranges from 20 up to 50%. When art is auctioned, a high percentage of zero-bidders is usually found, whereas small consumer electronics receive more bids in general. Since the offered service is rather specific and only applies to people who are owner of a house, the auction was sent to 600 members of veylinx who declared to be home-owners in previous auctions. In order to get a larger research sample and a potentially interesting new group, a customer database of Urgenda was invited to the auction as well. 3700 customers, who have ordered solar panels in the past via Urgenda, were sent an email with a brief explanation and a link to the auction.
21
4.3 Procedure
After the experiment has been set-up, members of Veylinx.nl and participants of ‘wij willen zon’ received an invitation with a direct link to the auction. They all received a personal link so that their email address was already filled in into the registration form. This step reduced the effort respondents had to make in order to participate. The emails were sent via mailchimp which is an online newsletter tool. The participants of ‘wij willen zon’ have received emails via mailchimp before. Unlike the members of Veylinx.nl, the newsletter receivers are unfamiliar with the auction procedure. Therefore, the procedure of the auction is explained, emphasizing that the highest bidder is obliged to actually purchase the product paying the second highest bid. There is no further information given about the auctioned service, to prevent information asymmetry between both groups.
When the respondent clicks on a direct link to the auction, one of the six different descriptions (see research design) of the service is shown to him. After reading the description, the respondent can fill in the amount he is willing to pay.. After the ‘place bid’ button is clicked on, a box will appear which asks for confirmation of the bid. After the bid is confirmed, a new page is loaded with a few multiple choice scales. These scales have to be filled in, in order to take actual part in the auction. All auctions have a set expiration time and after one has expired, a winner, who is obliged to purchase the offered service, is selected. Whereas the Veylinx.nl members are used to the fact that auctions expire within one day, the ‘wij willen zon’ group is not. Therefore, they receive the invitation to the auction one day earlier, in order to have more time to participate.
4.4 Measures
Four different variables will be measured during the research in order to test the hypotheses. Three of four variables will be gathered via the questionnaire after the bid is placed. The number of questions that can be placed after the auction is limited in order to get a maximum response rate. Scales will be derived from existing literature and have to be both validated and comprehensive.
4.4.1 Height of bid
The most important dependent variable for this research is the bid amount. This reflects the amount a respondent is willing to pay for the offered advice. The bid is entered in euro’s, up to whole cents. In case a respondent is not interested in the advice, it is possible to place a zero-bid.
22 4.4.2 Innovativeness of the respondent
The innovativeness of consumers has immediate significant relevance to consumer behavior. Although this concept has been researched extensively, no consensus has been reached on the measurement method nor the definition of innovativeness (Roehrich, 2004; Hauser et al., 2006). Roehrich (2004) describes the most important operational measurements of innovativeness in a meta-research. A limitation of most of the measurements is the low or non -predictive validity of actual product consumption (Roehrich, 2004). A measure developed by Roehrich in 1995 consists of six scales. This method appears to have a low to average predictive validity (r between 0 and .30), which is higher than most other measurement methods (Roehrich, 2004).
- I am more interested in buying new than known products - I like to buy new and different products
- New products excite me
- I am usually among the first to try new products - I know more than others on latest new products - I try new products before my friends and neighbors
The first and last questions were left out in order to present a compact measurement method and to increase the response rate.
4.4.3 Level of environmental involvement
A consensus on how a consumers level of environmental involvement should be measured, has not been reached yet, though environmental involvement has been measured in many different papers. A research conducted by Grimmer and Woolley (2012) measured environmental involvement with three times eight multiple choice questions for the level of cognitive, affective and behavioral involvement. Other experiments used four 5-point Likert scales to measure involvement with the environment (Schuhwerk & Lefkoff-Hagius, 1995; D’Souza & Taghian, 2005), or five different scales to measure environmental attitude and behavior (Tucker et al., 2012).
For this experiment, the scale by Schuhwerk & Lefkoff-Hagius (1995) will be applied since it has been used in a research by D’Souza & Taghian (2005) as well, while other reviewed measurement methods were not adopted in other experiments.
23
Another advantage of the Schuhwerk & Lefkoff-Hagius measurement is the fact that it consists of only four comprehensive 5-point Likert scales:
- I am concerned about the environment
- I am willing to take sacrifices to protect the environment - The quality of the environment affects the quality of my life - My actions impact on the environment
The third question was removed for this research since the Dutch translation did not reflect the original meaning and intention well.
4.4.4 Level of perceived risk
Just as environmental involvement, perceived risk is usually measured via multiple statements in the existing literature. Perceived risk contains many different aspects of which financial risk and performance risk (the risk of malfunctioning) are considered as the most significant according to Murthy and Blischke (2006). Havlena and Desarbo (1991) conducted a research on the measurement of perceived risk and found that financial risk, performance risk and safety were the three most important concerns and gave the best reflection of the overall perceived risk. Again, for this research the number of questions that can be asked is limited and one overarching measurement has to be used to measure perceived risk. Therefore the following statement will be shown to the respondents and a 5-point Likert scale is added from totally ‘agree’ up to ‘totally disagree’:
- I perceive this service as riskful
These 8 scales in total are translated in Dutch with most care to make sure that the translated scales are exactly the same as the original scales.
24
5. Results
In this section, the results of the experiment will be displayed. First some descriptive statistics of the research sample and the data are presented. A test of normality follows in order to determine whether parametric or non-parametric tests have to be used. Then a reliability test is conducted of the Likert scales in order to test the internal cohesion of the scales. Before the hypotheses are actually tested, the correlations will be reviewed.
5.1 Descriptive statistics
5.1.1 Sample characteristics
In total 649 respondents actually filled in an amount they were willing to pay for the service. 41 respondents did not completely fill in the Likert scales after the actual auction. 3 other respondents were removed from the data sample, since they filled in that their year of birth was 1903, 1909 and 2000 respectively, which is highly unlikely. Therefore, the final sample consists of 605 respondents. Of the 649 respondents, 469 were male (72,3%) and 180 were female (27,7%). The youngest participant was 19 and the oldest had the age of 80. The mean was 52 years old and the median 53.
5.1.2 Data characteristics
In total 308 (46,8%) of the 649 respondents placed a zero-bid. This relatively high percentage was expected since it is a rather specific service. Previous auctions via Veylinx with more specific products or services had comparable percentages of zero-bids. The WTP ranged from 0 euro’s up to 1600 euro’s.
N Range Minimum Maximum Mean Std.
Deviation Bid amount 646 160000 0 160000 2697,55 8200,066 PR 622 4 1 5 3,74 1,110 EI1 624 4 1 5 2,06 ,978 EI2 622 4 1 5 2,14 ,939 EI3 624 4 1 5 1,93 ,990 CI1 624 4 1 5 2,67 1,012 CI2 621 4 1 5 2,33 ,983 CI3 624 4 1 5 3,26 1,124 CI4 623 4 1 5 3,08 1,141
Looking at the six different treatments, no significant difference was found between any of them. Although rather large differences were found between the means of the six treatments, there is a big variance, which explains why no significant results were found.
25
Treatments Frequency Percentage zero-bids
Minimum Maximum Mean Std. Deviation 1. NWNA 102 49% 0 17500 2210,25 3788,536 2. NWFA 125 48,8% 0 36000 2194,81 5707,01 3. NWEA 113 44,2% 0 20100 1777,42 3196,574 4. WNA 103 44,7% 0 35000 3112,84 6553,288 5. WFA 98 48% 0 160000 4134,74 16995,88 6. WEA 105 45,7% 0 45000 3010,91 6167,443 Normality test
The main dependent variable, the bid amount, is highly skewed. A normality test was executed in order to test whether the data are normally distributed. Both, Kolmogorov-Smirnov- and Shapiro-Wilk-tests showed a p-level smaller than 0.001 and therefore the bid amount is not normally distributed and non-parametric tests have to be used in order to test the hypotheses. The natural log (ln) of the bid amount is not normally distributed either, not even when only the top 50% percentile is considered.
Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
bid_amount ,371 646 ,000 ,304 646 ,000
a. Lilliefors Significance Correction
0 1000 2000 3000 4000 5000 6000 7000 8000 0 1000 2000 3000 4000 5000 6000 7000 8000
1. NWNA 2. NWFA 3. NWEA 4. WNA 5. WFA 6. WEA
Bid Amount Treatments
26
5.2 Reliability
In order to test whether the Likert scales consistently reflect environmental involvement and consumer innovativeness respectively, a reliability analysis is conducted measuring Cronbach’s alpha α. This analysis technique is found to be the most common measure of scale reliability (Fields, 2009, p. 674).
Environmental involvement:
The Cronbach’s alpha α of three scales, measuring environmental involvement was found to be 0.830 (N=618). In the table below it is found that eliminating the third question could slightly increase the Cronbach’s alpha α, but since combining the three questions is significant (α > 0.7), it was chosen not to.
Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Corr. Squared Multiple Corr. Cronbach's Alpha if Item Deleted EI1 4,08 2,929 ,727 ,561 ,726 EI2 3,99 3,021 ,730 ,563 ,724 EI3 4,20 3,154 ,613 ,375 ,840 Consumer innovativeness:
The Cronbach’s alpha α of four scales measuring consumer innovativeness was 0.885 (N=617), which is well above 0.7. The table below shows that all four questions have a positive influence on the Cronbach’s alpha α, therefore all question are included.
Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Corr. Squared Multiple Corr. Cronbach's Alpha if Item Deleted CI1 8,66 8,197 ,747 ,589 ,853 CI2 8,99 8,416 ,724 ,556 ,862 CI3 8,06 7,411 ,787 ,644 ,837 CI4 8,24 7,503 ,746 ,598 ,854
Both measures were proven to be consistent and the three scales measuring environmental involvement are merged onto one average of the three scales ranging between 1 and 5. The same merging is done for the four consumer innovativeness scales: 1 means a high level and 5 means a low level.
27
5.3. Correlations
In advance of testing the hypotheses, several correlations will be looked into. The entire correlation table is enclosed as Appendix B. Since, the data is normally distributed a parametric correlation was chosen. Kendall’s tau b (τ) was found to be the most accurate non-parametric correlation of two available methods (Fields, 2009). WTP was significantly positively related to age of the respondent, τ = .063, p < .05. The gender of the respondents is also correlated with WTP, τ = -.132, p < .01. Male respondents were willing to pay significantly more for the offered advice. Two other significant relationship were found between the level of environmental involvement and WTP (τ = -.139), and between level of education and WTP (τ = .138), (both ps < .01). Considering the entire research sample, no other variables such as perceived risk or consumer innovativeness were significantly related to WTP. However, a significant relationship was found between perceived risk and consumer innovativeness, τ = -.063, p < .05. Level of environmental involvement has a stronger negative correlation with perceived risk, compared to consumer innovativeness, τ = -.112, p < .01.
5.4 Hypotheses testing
First the entire research sample will be used to test a hypothesis. Then, the same tests will be done for the 50% highest bidders. This split is made in order to review the respondents that did not place a zero-bid at the auction.
5.4.1 Benefit signaling
The first hypotheses (H1) states that respondents are willing to pay more for a service with a benefit mentioned in the advertisement compared to no benefit. Since the WTP is non-normally distributed, a Mann-Whitney tests is executed. Looking at the entire database (N=646), the WTP for advertisements without a benefit (Mdn = 100) did not significantly differ from advertisements with a benefit (Mdn = 100) mentioned.
When looking at the top 50% percentile of bids (N=319) no significant difference was found in WTP for advertisements without a benefit (Mdn = 2500) compared to treatments with benefit (Mdn = 3000). Therefore, H1 can be rejected.
5.4.2 Environmental involvement and WTP
In order to measure H2, firstly 4 groups were created based on the level of environmental involvement which has been split up in quartiles. A Kruskall-Wallis test was conducted and it showed that height of bid was significantly affected by level of environmental involvement,
28
respondents in the first quartile (N=172) were assigned with a high level of environmental involvement (between 1 and 1.33), the second (N=206) and third quartile (N=125) were determined as medium-high and medium-low level of involvement, while the fourth quartile (N=115) was assigned with a low level of environmental involvement (between 2,66 and 5). Respondents with a high level of environmental involvement (Mdn = 1000) were significantly willing to pay more for the offered service than respondents with a low level of environmental involvement (Mdn = 0), U = 7240.5, z = -4,055, p < 0.001, r = -.239.
Looking at the top 50% of bids, no significant difference was found between the first (N=81) and second quartile (N=74) (U = 2948.5 , r = -.014), nor between the first and third quartile (N=79), (U = 2669, r = -.14). Only the fourth quartile (N=71) showed a significantly lower WTP compared to the first at P<.0167, (U = 1958, r =-0.28). Therefore H2 can be accepted; respondents with a high level of environmental involvement are significantly willing to pay more for the offered advice compared to respondents with a low level of involvement.
5.4.3 Financial benefit signaling
In order to test whether respondents with a low level of environmental involvement are willing to pay more for an offered service with a personal benefit, compared to a common benefit or no benefit (H3a), the fourth quartile (N=115) is further examined with respondents with a relatively low level of environmental involvement. A Mann-Whitney test was conducted and no significant difference was found for the low environmental involved group between financial benefit (Mdn. = 100) and other mentioned benefits (Mdn = 100), U = 44947, z = -1.038, ns, r = -.041. Looking at the third quartile, consumers with medium-low level of environmental involvement, WTP was significantly higher for the financial benefit, H(1), 4.936, p < 0.05. Since this effect was only found for the third quartile and not for the fourth, hypothesis H3a is still rejected.
29
Respondents with a high level of environmental involvement were willing to pay significantly more for an advice with a personal (financial) benefit mentioned compared to respondents with a low level of environmental involvement, U = 7240.5, z = -4.055, p < 0.001, r = -.239. Thus
H3b, which stated actually the opposite is rejected.
5.4.4 Environmental benefit signaling
Considering the graph on the next page, it seems that the mean bid amount for respondents with a high level of environmental involvement was actually lower when an environmental benefit is mentioned, while it seems to be the opposite for respondents with low environmental involvement. However the difference in bid for respondents with high level of environmental involvement appears not to be significant, U = 3259.5, z = -.417, ns, r = -.032. Neither was a significant difference found for the low environmental involved group, U= 1299.5, z = -1.069,
ns, r = -.10. Thus, H4a is rejected.
0 1000 2000 3000 4000 5000 6000
Low Environmental Involvement High Environmental Involvement
Mean bid amount for financial benefit signaling
30
A significant difference was found conducting a Kruskall-Wallis test between level of environmental involvement and WTP for financial and no benefits signaled, H(3) = 16.314, p
< .01. However there was no significant difference found between environmental involvement
and WTP for the respondents who were shown an environmental benefit, H(3) = 3.358, ns. Mann-Whitney tests were executed to retrieve a better insight. WTP did not differ between high level of environmental involvement or medium-high level of environmental involvement (U = 1964, r = -.072), nor between high level and medium-low level of environmental involvement (U = 1060.5, r = -.11). Even between high and low level of involvement group there was no significant difference found (U = 918, r = -.18), while a clear significant difference (p < 0.001) was found between high environmental involvement and groups with a lower level of environmental involvement when financial or no benefit was shown. H4b was rejected, since no significant difference was found in WTP between level of environmental involvement when an environmental benefit is shown.
5.4.5 Warranty signaling and perceived risk
An Analysis of Variance (ANOVA) showed a significant difference in WTP for treatments with and without warranty, F = 4.352, p < .05. A regression analysis showed that for the 50% highest bids, signaling warranty results in a 40% increase of WTP (p < .01). Non-parametric tests were conducted as well. Looking at the entire research sample, no significant difference was found between treatments with warranty (Mn. = 3405.14) and without warranty (Mn. = 2060.72) although the means seem to differ highly, U = 49460, z = -1.141, ns, r = -.045. However when only the upper 50% bids are regarded, and thus all zero-bids are eliminated, a significant (p <
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Low Environmental Involvement High Environmental Involvement
Mean bid amount for environmental benefit signaling
31
0.05) and positive effect was found between warranty signaling and WTP, U = 10560.5, z = -2.593, r = -.145. The regression analysis found that warranty signaling has a positive influence on WTP and H5a can be accepted. An additional non-parametric test found that H5a can only be accepted when the lowest 50% bids were left out.
Perceived risk appears not to be a mediator for the effect of warranty signaling. No significant relation was found between perceived risk and warranty signaling, U = 45745, z = -1.186, ns, r = -.048, nor is a difference found in perceived risk between any of the six treatments, H(5) = 6.006, ns. H5b is rejected, warranty signaling does not appear to have a significant effect on perceived risk.
To test whether innovativeness of the consumer has a negative effect on perceived risk (H4c), the respondents were divided into four groups based on quartiles, since WTP is not normally distributed. The first quartile (N=151) was considered as high level of innovativeness whereas the respondents in the fourth quartile (N=123) were assigned to have a low level of innovativeness No significant difference was found in perceived risk between the most innovative and most less-innovative group, U = 8461.5, z = -1.129, r = -.068. H5c is therefore rejected.
It is noticeable that only 22 of the 622 perceived the offered service as very risky, and 59 as rather risky. The majority perceived the service as not really risky (N=169) or not risky at all (N=198). However no significant difference was found in WTP between several levels of perceived risk, H(4), 6.015, ns. Neither was a significant difference found when selective samples were regarded, such as respondents with a high education level, top 50% bids, top 25% bids and gender. H5d is rejected, since no relation was found between perceived risk and WTP.
32
6. Discussion
In this section the results will be discussed and compared with existing literature and managerial implications of the outcome will be summed up. Finally, the issue of limitations of this research will be addressed and some suggestions for future research will be given.
6.1 Findings
Benefit signaling
No significant difference was found in benefit signaling. Respondents did not value treatments with a benefit (financial or environmental) higher than treatments without a benefit. This contradicts with the findings by Bei and Widdows (1999), which found that extra information such as a benefit would result in a higher perceived value, but this result is partially in line with research by Grimmer and Woolley (2012). They did not find a significant result, measuring purchase intention of consumers after being exposed to different water bottle advertisements. Grimmer and Woolley mentioned the possibility in their discussion that for high-involved products there would be an effect for benefit signaling. However, for this personal advice no difference was found either.
There are several possible explanations for the outcome. First of all, it is possible that respondents found benefits implicitly mentioned in the service description. An energy neutral house, could well be seen as good for the environment without being explicitly mentioned as such. In a previous experiment it was found that highly environmentally involved consumers came up with more arguments to buy the green product compared to lower involved consumers, even when no benefit was added (Schuhwerk & Lefkoff-Hagius, 1995). Unfortunately there were no scales included in the questionnaire afterwards in order to measure whether benefit signaling was noticed at all.
Another explanation might be that respondents were not convinced by the mentioned benefits and required more information before valuating it accordingly. Research into the purchase intention for renewable energy, demonstrated that consumers require much more detailed information in order to perceive a significant utilitarian benefit (Hartmann & Apaolaza-Ibanez, 2012). The lack of more concrete figures about both the financial as well as the environmental benefit in the treatment, could have decreased the perceived extra value of the benefit.
33
did not notice the benefit. However, this last explanation is more unlikely since the benefits were highlighted with a big bold font just below the title of the advertisement.
Level of environmental involvement
Level of environmental involvement has a very strong and significant positive effect on WTP, regardless of the mentioned benefit. As was found in literature, consumers with a high level of environmental involvement tend to differ from non-environmentally involved consumers regarding socio-economic variables. Consumers with high concern are on average higher-educated and have a higher income (Schwepker & Cornwell, 1991; Straughan & Roberts, 1999). A higher income could be an explanation for the higher WTP of respondents. A second explanation could be that the offered advice was perceived as beneficial for the environment, regardless of the mentioned benefits, a benefit which was then perceived as more valuable by environmentally involved consumers. The experiment by Schuhwerk and Lefkoff-Hagius (1995) found that more environmentally involved consumers came up with significantly less counterarguments of buying a product, regardless of the appeal mentioned.
Consumers with a low-level of environmental involvement were not willing to pay more when a financial benefit was mentioned, compared to environmental or no benefit. This result is in line with research by D’Souza and Taghian (2005), but in contradiction with findings by Schuhwerk and Lefkoff-Hagius (1995) and Grimmer & Woolley (2012). Schuhwerk and Lefkoff-Hagius found that non-involved respondents were actually willing to pay more for a product when environmental benefit was shown.
Consumers who were more concerned about the environment, were prepared to pay more for the service. However, mentioning benefit for the environment did not increase the bid, as was expected. This result is in line with the results by Schuhwerk and Lefkoff-Hagius (1995). They did not found a difference in valuation by highly environmentally-involved consumers. But it is in contradiction with results by Grimmer and Woolley (2012) who did find that highly-involved consumers value advertisements with green benefits mentioned higher than low-involved consumers. These contradicting findings can also be explained by the fact that for this experiment real willingness to pay is measured instead of hypothetical willingness to pay. Real WTP is a much more accurate reflection of actual behavior compared to hypothetical WTP (Balistreri et al., 2001; Voelckner, 2006). Although no significant difference was found, the mean WTP for the financial benefit was higher compared to environmental benefit. A recent study by Green and Peloza (2014) found that consumers valuate the financial appeal for green
34
products more, compared to environmental benefit, when the ad is communicated via private media such as advertising.
Warranty signaling
The effect of warranty signaling has a great effect on WTP. Regression analysis showed a 40% increase for the top 50 percent bidders. This result confirms findings from previous research that warranty signaling can be a powerful marketing tool (Murthy & Blischke, 2006; Murthy & djamaludin, 2001). However, no relation was found between warranty signaling and the level of perceived risk, which is in contrast with previous findings (Murthy & Blischke, 2006; Murthy & Djamaludin, 2001; Havlena & Desarbo, 1991). Neither has this experiment found that perceived risk has a negative effect on WTP. One explanation could be that the scale of perceived risk did not actually measure perceived risk properly. Perhaps more scales regarding risk perception should have been added in order to collect better data. Another explanation might be that perceived risk has no effect on willingness to pay for auctions compared to fixed-price products or services. When consumers can fill in an amount themselves, it could well be that respondents determine their WTP after they have assessed their risk perception, which will ultimately have a mitigating effect on risk perception of the product once the amount was placed. This could be a plausible explanation why no relation was found between perceived risk and willingness to pay.
Consumer innovativeness
No link was found between consumer innovativeness and perceived risk, nor between innovativeness and willingness to pay. Literature suggests that innovative consumers would be more likely to adopt a new product or service and would be less receptive to risk (Rogers, 2003). However, not even a hint was found for this suggestion. A first explanation could be that the offered service was not perceived as new or innovative. Another possibility is that the wrong measurement method was used. Unfortunately, previous studies have not agreed upon a uniform method for measuring consumer innovativeness (Roehrich, 2004). For this experiment the most compact and actual demand revealing measuring method was used. A third explanation might be that the scales do not reflect the behavior for this specific offered service. Consumers who are very much into the latest gadgets, might show different behavior when it comes to renovation or energy efficiency improvements. One respondent wrote an email in which he argued that he did not care for the latest innovations, but bought his first solar panels in 1970’s.
35
6.2 Theoretical contributions
This study generated several findings on the effect of warranty and benefit signaling on WTP as well as personal characteristics on WTP. Some of these results are in line with previous experiments while other findings appear to contradict. Previously, in other industries such as the automobile industry, it was found that warranty signaling has a very strong positive effect on WTP.
Another contribution to the theory is the strong relation between environmental involvement and willingness to pay. The top 25% environmentally involved respondents were willing to pay almost 4 times as much compared to the lowest 25%.
Benefit signaling does not appear to have an effect for this specific offered object.
6.3 Managerial implications
Looking at the results from a marketing perspective it was found that adding a warranty to the offered advice, would result in higher results. Of course, the extra expenses of adding a warranty regulation have to be weighed against the extra generated income.
A second managerial implication would be to focus on environmentally involved consumers first. These consumers will be much easier to persuade and are willing to pay more for the offered advice. For example, advertisements could best be placed in more environmentally involved magazines such as ‘Milieudefensie’ or ‘De wereld van morgen’ magazine, or commercials before or after specific radio shows such as ‘BNR duurzaam’ and ‘Vroege vogels’.
6.4 Limitations and recommendations for future research
Although, this research was conducted with great care, there are a number of limitations of this study. First, a higher number of respondents would improve the quality of this research. A larger research sample could provide more results, especially when zooming in at selections of the research sample. Perhaps, benefit signaling would have an effect for specific groups, since we found some hints already. A larger sample could also result in more serious bids. There were not only many zero-bids, a high number of respondents also placed a bid below 10 euro, which is less than 1% of the market-value. The high value and effort could have been more emphasized in the description of the service in order to get higher bids, although it is mentioned that an expert will visit and asses the house and personal energy consumption.
A second limitation of the research was the limited time and information that was granted to make a decision. For a high-value service as this, consumers might find it necessary to retrieve additional information, think it through and discuss it with other members of the
36
household, before making a purchase decision.
A third point of improvement regards the consumer innovativeness scale Perhaps the statements should have been more specified towards energy efficiency measures, renewable energy or renovation. An extra scale could be added to determine the extent into which the participants consider the offered service to be innovative and new.
A last point of improvement would be to provide more detailed information about the benefit. For instance, the average return on investment could be mentioned.
Future research should be conducted to retrieve better understanding of the effect of benefit signaling and willingness to pay, since this research did not came up with a decisive answer. It would be interesting to experiment with both environmental and financial benefits combined in one advertisement. Another important topic for future research could be to examine whether perceived risk plays a role when a respondent can fill in the amount he or she is willing to pay. A third and last suggestion for future research would be to conduct the experiment with more information about the service as well as providing more time to the participants in order to think through the amount they are willing to pay.
37
7. Conclusion
The main purpose of this research was to examine whether different marketing communication messages have an effect on WTP and whether this was affected by warranty signaling, consumer’s level of environmental involvement and level of innovativeness. In our experiment, mentioning a financial benefit or an environmental benefit appeared not to have any influence on the willingness to pay. The level of environmental involvement however, was strongly related to WTP. Highly environmentally involved consumers were prepared to pay significantly more for the offered service.
Adding a warranty to the advertisement has a strong positive effect as well, although this has no influence on perceived risk. Consumer innovativeness does not appear to have influence on perceived risk or willingness to pay either. This research has given a contribution to existing literature by measuring real willingness to pay and it gave new insights by studying warranty signaling for a green service.
Although green marketing is becoming increasingly important , the scientific knowledge of consumer response towards marketing communications is still rather limited and should be explored more in the future in order to improve the theoretical knowledge and in the end also the tools for marketeers in the field.
38
Bibliography
Balistreri, E., McClelland, G., Poe, G., & Schulze, W. (2001). Can hypothetical questions reveal true values? A laboratory comparison of dichotomous choice and open-ended contingent values with auction values. Environmental and Resource Economics, 18(3), 275-292.
BEI, L., & Widdows, R. (1999). Product knowledge and product involvement as moderators of the effects of information on purchase decisions: a case study using the perfect information frontier approach. Journal of Consumer Affairs, 33(1), 165-186.
BNEF. (2013). More than 100x cheaper in 35 years. Retrieved 07/05, 2014, Retrieved from http://www.treehugger.com/renewable-energy/graph-shows-why-solar-power-will-take-over-world.html
Chau, C., Tse, M., & Chung, K. (2010). A choice experiment to estimate the effect of green experience on preferences and willingness-to-pay for green building attributes. Building
and Environment, 45(11), 2553-2561.
D’Souza, C., & Taghian, M. (2005). Green advertising effects on attitude and choice of advertising themes. Asia Pacific Journal of Marketing and Logistics, 17(3), 51-66.
do Paço, Arminda Maria Finisterra, & Reis, R. (2012). Factors affecting skepticism toward green advertising. Journal of Advertising, 41(4), 147-155.
ECN. (2013). Energietrends 2013, vier gevolgen van de groei van hernieuwbaar voor het
energiesysteem. ( No. 2).Energie Centrum Nederland.
Faiers, A., & Neame, C. (2006). Consumer attitudes towards domestic solar power systems.
Energy Policy, 34(14), 1797-1806.
Field, A. (2009). Discovering statistics using SPSS Sage publications.
Granade, H. C., Creyts, J., Derkach, A., Farese, P., Nyquist, S., & Ostrowski, K. (2009).
Unlocking Energy Efficiency in the U.S. Economy. ().Mckinsey Global Energy and
