University of Amsterdam Thesis
The Moderating Effect of Sales Promotion Type on Perceived Value, Perceived Risk and Purchase Intention
Supervisor: drs. F. Slisser
Study: Executive Programme in Management Studies
Track: Digital Business
Author: Thomas de Groot
Student Number: 12589683
EBEC Number: 20201101031121
Date of Submission: 29-01-2021
STATEMENT OF ORIGINALITY
This document is written by Thomas de Groot who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Signature
CONTENT
ABSTRACT ... 5
1. INTRODUCTION ...6
2. LITERATURE REVIEW ... 8
2.1. PERCEIVEDVALUEANDRISK ... 8
2.1.1. UTILITARIAN AND HEDONIC VALUE ... 9
2.1.2. RISK DIMENSIONS ... 9
2.2. SALESPROMOTIONS ... 12
2.2.1. SALES PROMOTION TYPES ... 13
2.2.2. EFFECTS OF SALES PROMOTION TYPES ... 14
2.2.3. SALES PROMOTION EFFECTS ON VALUE PERCEPTIONS ... 15
2.2.4. SALES PROMOTION EFFECTS ON RISK PERCEPTIONS ... 16
2.3. HYPOTHESES...17
2.4. CONCEPTUALFRAMEWORK ... 20
3. METHOD ... 20
3.1. RESEARCHDESIGN ... 21
3.1.1. EXPLORATORY STAGE ... 21
3.1.2. EXPERIMENTAL STAGE ... 22
3.2. DATASAMPLE ... 23
3.3. MEASURES ... 23
Purchase Intention ... 23
Value Perceptions ... 23
Risk Dimensions ... 24
Control Variables ... 24
4. RESULTS ... 25
4.1. DATAPREPARATION ... 25
4.2. EXPLORATORYANALYSIS ... 27
4.3. MAINANALYSIS ... 29
5. DISCUSSION ... 32
6. CONCLUSION ... 38
REFERENCES ...41
APPENDICES ... 45
A. SCALES ... 45
B. SURVEY ... 47
C. HOTELOFFERS ... 51
D. STATISTICS ... 52
D1. TESTS FOR LINEARITY ... 52
D2. TEST FOR HOMOSCEDASTICITY ... 56
D3. TESTS FOR NORMALITY ... 57
ABSTRACT
This paper addresses the effects of sales promotion types on perceived value, perceived risk, purchase intentions, and how sales promotion types moderate the relationships between these constructs. To investigate this, an experiment with a between-subject design was conducted.
The participants were divided into three groups through random assignment (no sales promotion, monetary sales promotion and non-monetary sales promotion), wherein the presence of sales promotion (type) was manipulated. As predicted, an increase in hedonic value perceptions of a promoted service (hotel offer) increases consumers’ purchase intentions. An increase in overall risk, financial risk and psychological risk decreases consumers’ purchase intentions. The results of this experiment do not indicate moderating effects of sales promotion type on the relationships between value perceptions, risk perceptions and purchase intentions.
However, this research shows that the perceptions of financial risk, overall risk and time risk are the least for monetary sales promotions. This is diametrical to prior research wherein non- monetary sales promotions evoked less financial risk perceptions than monetary sales promotions. Additionally, the presented research elaborates on managerial implications and suggestions for future research.
Keywords: sales promotion types, perceived risk, perceived value, risk-reliever.
1. INTRODUCTION
In 2019, global expenditures on advertisement have grown by 4% to over 560 billion U.S. dollars across all channels (Statista, 2020). As a tool of advertisement, sales promotions are crucial for many companies (Chandon et al., 2000; Zhang & Wedel, 2009). A large part of marketing budgets is spent on sales promotions by both national and small businesses, especially companies who are active in consumer-packaged goods and other consumer products (Blattberg & Briesch, 2012). The main goal of sales promotions is to drive consumer purchases (Blattberg & Briesch, 2012; Chandon, Wansink & Laurent, 2000).
There are two main consumers’ value perceptions of a product or service that increase purchase intentions: hedonic value and utilitarian value (Dhar & Wertenbroch, 2003; Voss, Sprangenberg & Grohmann, 2003). Additionally, risk perceptions of a product or service play an important role in influencing purchase intentions (Case, 2002; Kim, Kim & Leong, 2005).
Whereas high value perceptions increase purchase intentions of a consumer, high risk perceptions decrease the purchase intentions. The risk dimensions are overall risk, performance risk, financial risk, social risk, psychological risk, physical risk and time risk (Kaplan, 1974;
Kim et al., 2005; Stone & Grønhaug, 1993). Furthermore, research indicates that there is a moderating effect of sales promotion type on the relationships with perceived value (Chandon et al., 2000) and financial risk perceptions (Santini, Sampaio, Perin, Espartel & Ladeira, 2015) as independent variables and purchase intention as the dependent variable.
However, there is a clear gap in the literature. Research has paid little attention to the moderating effect of sales promotion types (monetary vs. non-monetary) on the relationship
between all of the aforementioned risk perceptions and purchase intention. This raises the following research question:
What is the moderating effect of sales promotion type on the relationship between perceived value, perceived risk and purchase intention?
The presented research builds on the remarks about future research of several authors (Cases, 2002; Kim et al., 2005; Santini et al., 2015). Kim et al. (2005) have found direct negative effects of perceived risk dimensions on purchase intention but did not investigate how to reduce the levels of perceived risk. Cases (2002) have found that sales promotions are an effective risk- reliever for several perceived risk dimensions related to buying a product on the internet.
Respondents in Cases’ (2002) study assigned risk-relievers, including sales promotions, to risk dimensions that they considered having the most effect on. This study does not measure the moderating effect of sales promotions on the relationship between perceived risk and purchase intentions. Moreover, this research does not consider different sales promotion types.
This is similar to the research of Mitchell & Greatorex (1993), wherein respondents appoint sales promotions as an appropriate risk-reliever but do not specify the sales promotion type.
Santini et al. (2015) explored the moderating effects of sales promotion types on the relationship between perceived financial risk and purchase intention. Moreover, they propose to analyze other risk dimensions since their research only addressed financial risk.
The presented research contributes to the literature by measuring the effects of sales promotion type (monetary vs. non-monetary) on the perceived risk dimensions. On a managerial level, this research contributes to perceived risk-relieving strategies. Based upon the results of the presented research, managers know which sales promotion type aligns best
with specific risk dimensions. It becomes clear which risk dimensions managers should avoid using sales promotions for.
2. LITERATURE REVIEW
To answer the aforementioned research question, several factors have to be studied in dept.
Firstly, the literature on perceived value and risk of a product is investigated. To understand their relationship with purchase intention (Kim et al., 2005; Voss et al., 2003), it is important to define these consumer perceptions first. Secondly, the definition and effects of sales promotions are explored. Sales promotion types are differentiated by the corresponding presented value and perceived benefits of the sales promotion (Chandon et al., 2000; Crespo- Almendros & Barrio-García, 2016). Furthermore, the direct and moderating effects of sales promotion types on perceived value, risk, and their relationship with purchase intention are discussed (Cases., 2002; Chandon et al., 2000; Santini et al., 2015).
2.1. PERCEIVED VALUE AND RISK
In the literature, three main consumer perceptions of a product or service are defined:
hedonic, utilitarian and risk. These perceptions influence consumer purchases and consumption behaviors (Kim et al., 2005; Voss et al., 2003). Additionally, perceived value and risk are related to consumer perceptions of shopping experience and channels (Babin, Darden & Griffin, 1994; Cases, 2002; Overby & Lee, 2006). The scope of this study includes the effects of perceived value and risk of a product. Therefore, perceived shopping value and risk are left out of this research.
2.1.1. UTILITARIAN AND HEDONIC VALUE
In early research, a consumer’s attitude towards a brand or product was seen as one- dimensional (Batra & Ahtola, 1991;Voss et al. (2003). Voss et al. (2003) explain that the value perception of a product is two-dimensional: hedonic and utilitarian. However, a product can be perceived as both utilitarian and hedonic (Okada, 2005). Therefore, when referring to a hedonic (utilitarian) product, the hedonic (utilitarian) value is relatively higher (Okada, 2005).
Utilitarian and hedonic perceptions of a product drive consumers’ choices for products and increase purchase intentions (Dhar & Wertenbroch, 2000; Voss et al., 2003). The utilitarian value is rather cognitive than emotional (Overby & Lee, 2006). Products perceived as utilitarian provide primarily instrumental and functional benefits (Dhar & Wertenbroch, 2000). For example, utilitarian features of a laptop are battery life and storage. The hedonic value is rather emotional than cognitive (Overby & Lee, 2006) and provides more fun and excitement (Dhar
& Wertenbroch, 2000). The hedonic value of a laptop consists of, for example, design and status. Voss et al. (2003) have established a valid and broadly accepted ten-item scale to measure consumers’ hedonic and utilitarian value perceptions of a product.
2.1.2. RISK DIMENSIONS
In 1960, Bauer urged researchers to consider risk perceptions in consumer behavior studies. This is defined as “that any action of a consumer will produce consequences which he cannot anticipate with anything approximating certainty, and some of which at least are likely to be unpleasant (Bauer, 1960).” Perceived risk is referred to as the risk from a consumer’s perspective, which negatively affects their attitude and/or purchase intention of a product (Bauer, 1960; Kaplan, Szybilla and Jacoby, 1974). Essentially, the level of perceived risk emerges from two factors: uncertainty and seriousness (Bauer, 1960). Uncertainty about the
product/service and seriousness of the consequences of making a mistake (Bauer, 1960, Derbaix, 1982).
Stone and Grønhaug (1993) describe how the definition of risk in marketing research differs from other studies, such as psychology and economics. In marketing research, the concept of risk involves a focus on negative consequences alone. Whereas other disciplines’
risk concept involves the possibility of both positive and negative outcomes. Other disciplines make a distinction between uncertainty and risk, wherein uncertainty means that no probabilities can be attached to the outcome (Stone & Grønhaug, 1993). The terms ‘uncertainty’
and ‘risk, in marketing research, are used almost inseparable because a consumer cannot take the true probabilities of risk into account (Stone & Grønhaug, 1993). The chance and impact of risk can be approximated by a consumer. Therefore, the level of perceived risk differs per person. Therefore, Stone and Grønhaug (1993) describe perceived risk as “subjective expectations of loss; the more certain of this loin, the greater the risk perceived by the individual.”
Different components that affect one’s risk perception explain why the ranking of risk dimension importance varies (Stone & Grønhaug, 1993). Next to the product or service, other factors form consumers’ risk perceptions in the purchasing decision: technology, vendor and consumer characteristics (Lim, 2003). For example, the internet as a technology and mode of shopping involves risk dimensions such as delivery, privacy, payment (Cases, 2002) and security risk (Kim et al., 2005). Moreover, familiarity with a brand and experience with a product influence the level of consumers’ risk perceptions (Stone & Grønhaug, 1993).
Extensive research identified one or more dimensions of perceived risk that have a significant impact on attitude towards a product and purchase intention. Kaplan et al. (1974) added social and performance risk and summarized it in their study of measuring five risk
dimensions in product/service purchase. The five dimensions are performance risk, financial risk, social risk, psychological risk, physical risk (Kaplan et al., 1974). Several researchers mention a sixth dimension, namely time risk (Dholakia, 1997; Kim et al., 2005). Risk dimensions form the overall perceived risk of a consumer in a purchasing situation. Stone and Grønhaug (1993) instruct to measure overall risk as a separate dimension. These risk dimensions explain a large part (77% - 88%) of the variance in the overall risk construct (Kim et al., 2005; Stone & Grønhaug, 1993). The risk dimensions related to the product are described in Table 1. The presented research refers to these six dimensions when mentioning ‘risk dimensions’ or ‘risk perceptions’.
Table 1. Product and service risk dimensions based on Kaplan et al. (1974) and Kim et al. (2005)
Perceived risk dimension Description
Financial risk
The expectation for monetary loss or not getting value for money spent.
Performance risk The expectation that the product will not perform as expected or provide the preferred benefits.
Physical risk
The expectation of physical threats (e.g., injury, sickness) cause by purchasing a product.
Psychological risk The expectation if failure in reflecting one’s self-image.
Social risk The expectation of negative opinions of reference groups.
Time risk The expectation of time loss or a waste of time caused by product failure.
Over the years, there have been different views on the composition of the perceived risk construct. Mitchell and Greatorex (1993), for example, state that performance risk does not exist. The authors conclude that the consequences of performance risk occur prior to any type of loss (i.e., risk). Conversely, many other studies recognize performance risk (Kaplan et al., 1974; Stone & Grønhaug, 1993; Kim et al., 2005). In contrast to Kaplan et al. (1974), who claim that the risk dimensions are independent, Stone and Grønhaug (1993) explain how risk
dimensions are intercorrelated. The authors explain how the expectations of loss or discomfort (i.e., risk perceptions) are processed by one’s psyche. The results show that all risk dimensions are mediated by one’s psychological risk perception within the overall risk construct.
Researchers make a distinction between products and services (Kaplan et al., 1974;
Murray & Schlacter, 1990). Murray and Schlacter (1990) concluded that social, convenience, physical, psychological and overall perceived risk are significantly higher for services than products. Although the authors have found directional support, financial and performance risk perceptions for services were not significantly higher than for products. Conversely, Mitchell and Greatorex (1993) concluded that financial risk perceptions are significantly higher for products instead of services. The authors have found that perceived financial risk, of all risk dimensions, is the strongest risk dimension for both products and services.
2.2. SALES PROMOTIONS
According to Belch and Belch (2003), marketing communications consists of six elements which together form the ‘promotional mix’. These elements are advertising, direct marketing, interactive/internet marketing, sales promotion, publicity/public relations and personal selling. Advertising is used for different purposes, such as seeking for immediate response and developing awareness and/or brand image over a longer period of time (Belch &
Belch, 2003). According to Blattberg and Briesch (2012), a sales promotion is “an action- focused marketing event whose purpose is to have a direct impact on the behavior of the firm’s customer.” Sales promotions are temporary, include a call-to-action and should evoke an immediate response (Blattberg & Briesch, 2012). Although several elements of the promotional mix may contain a call-to-action, the temporary aspect elucidates the distinction between sales
promotions and the other elements of the promotional mix (Belch & Belch, 2003; Blattberg &
Briesch, 2012).
There are different constructs of sales promotion provider and receiver: retailers-to- consumer (retailer promotions), manufacturer-to-retailer (trade promotions) and manufacturer- to-consumer (consumer promotions) (Blattberg & Briesch, 2012). These constructs use similar sales promotion instruments (Table 2). Trade promotions are excluded because the presented research addresses the effects on consumers’ purchase intentions. Retailer promotions are used to convince consumers to purchase a product at a certain retailer. In the presented research, consumer sales promotions that stimulate purchases of a product/service are included.
2.2.1. SALES PROMOTION TYPES
Blattberg and Briesch (2012) have summarized traditional forms of sales promotions (Table 2). In the literature, these sales promotions are divided into two types: monetary or non- monetary (Chandon et al., 2000; Blattberg & Briesch, 2012). The division between monetary and non-monetary is made because consumers perceive them differently. Both types are perceived to have different benefits, which are further described in section 2.2.3. In alignment with Blattberg and Briesch (2012), Kumar, Rangachari, Jhingran and Mohan (1998) describe that traditional forms of sales promotions (Table 2) are used both offline and offline. However, Kumar et al. (1998) have found that ‘instant rebates’ can only be offered via the internet. Due to the internet’s interactive nature (Blattberg & Briesch, 2012), instant rebates can be offered to consumers who are on the website for a long time or returns to it often (Kumar et al., 1998).
Table 2. Sales promotions based on Blattberg and Briesch (2012) and Kumar et al. (1998)
Sales promotions Description
Price reduction Temporary price reduction.
Retailer or manufacturer coupon Coupons for products provided by retailers or manufacturers.
Free goods For example, buy X get Y free.
Sweepstakes The consumer has the chance of winning cash or other prizes.
Free Trial Free samples to encourage to purchase the product.
N-for For example, three for the price of one product.
Discount card A card that tracks consumers’ purchases. In return, the consumer receives discounted prices on some products.
(Instant) Rebates
Consumers send proof of purchase to the retailer or the manufacturer in order to receive a rebate. An Instant Rebate offer is offered to consumers who are on the website for a long time or returns to it often.
Bundled Promotion
Discounted prices for complementary products (e.g., toothbrush and toothpaste).
2.2.2. EFFECTS OF SALES PROMOTION TYPES
Gedenk and Neslin (2010) categorize the results of sales promotion campaigns on consumer purchases in the short-term and long-term. Short-term results are store switching, product switching (brand or category), new users, and purchase acceleration (consumption rate or stockpiling). Long-term results are stockpiling, product loyalty (brand or category) and store loyalty.
Over time, research has found positive effects of different sales promotion types on purchase intention (Chandon et al., 2000; Crespo-Almendros & Barrio-García, 2016; Santini et al., 2015; Zhang & Wedel, 2009). Monetary sales promotions are often found to be more effective for increasing purchase intention than non-monetary sales promotions (Chandon et al., 2000; Luk & Yip, 2008). However, monetary sales promotions have been criticized because it can result in more price sensitivity and destroy brand equity (Chandon et al., 2000; Luk &
Yip, 2008). Although a later study has found positive long-term effects of monetary sales
promotions (Mendez, Bendixen, Abratt, Yurova and O’Leary, 2015), the fear for price sensitivity and destroying brand equity urges companies to use non-monetary sales promotions as well (Chandon et al., 2000; Luk & Yip, 2008). In fact, Chandon et al. (2000) found that motivation for consumers to respond to sales promotions goes beyond just monetary savings.
This will be further discussed in section 2.2.3.
2.2.3. SALES PROMOTION EFFECTS ON VALUE PERCEPTIONS
Chandon et al. (2000) have found three utilitarian- and three hedonic consumer benefits of sales promotions. Utilitarian benefits are instrumental, functional and cognitive, whereas hedonic benefits are rather non-instrumental, experiential and affective. The three utilitarian benefits are saving money (saving benefit), the possibility to upgrade to a higher quality product (quality benefit) and reducing search and decision costs (convenience benefit). The hedonic benefits are the consumers’ self-perception of being smart shoppers (value expression benefit), the fulfillment of consumers’ need for information and exploration (exploration benefit), and fun to use (entertainment benefit). Non-monetary sales promotions are found to provide more hedonic benefits and, conversely, monetary sales promotions provide almost solely utilitarian benefits (Chandon et al., 2000).
Regarding the nature of products, Chandon et al. (2000) have found a phenomenon called “benefit congruency”. This means that a type of promotion (monetary or non-monetary) should be compatible with the type of product (hedonic or utilitarian) to positively affect purchase intention. Monetary sales promotions are more effective for utilitarian products because monetary sales promotions provide more utilitarian benefits. Conversely, non- monetary sales promotions are more effective for hedonic products because non-monetary sales
promotions provide more hedonic benefits. (Chandon et al., 2000). Later studies have found the same benefit congruency effects (Crespo-Almendros & Barrio-García, 2016; Santini et al., 2015).
2.2.4. SALES PROMOTION EFFECTS ON RISK PERCEPTIONS
In the literature, the relationship between perceived risk and sales promotion types has been discussed (Mitchell & Greatorex, 1993; Chandon et al., 2000; Cases, 2002; Luk & Yip;
2008; Lowe, 2010; Santini et al., 2015). Mitchell and Greatorex (1993) have investigated the usefulness of fourteen risk-relieving strategies. The authors have found that sales promotions are moderately useful as a risk-relieving strategy across multiple services and products. In an online shopping context, Cases (2002) found sales promotions to be a useful risk-reliever. More specifically, the results of Cases (2002) show that sales promotions are effective for performance, delivery and source risk. Performance risk being the more product-related risk dimension, whereas delivery and source risk are more related to shopping mode (Cases, 2002).
However, both Mitchell and Greatorex (1993) and Cases (2002) do not specify sales promotions types. Furthermore, Mitchell and Greatorex (1993) and Cases (2002) asked respondents to rank risk-relievers on usefulness. This is different from measuring sales promotions’ effects on the level of perceived risk and the relationship between risk and purchase intention.
Santini et al. (2015) describe that sales promotions negatively influence the relationship between financial risk and purchase intention. The authors found a moderating effect of sales promotion type. The financial risk perceptions are stronger for monetary sales promotions, whereas financial risk perceptions are weaker for non-monetary sales promotions (Santini et al., 2015). For perceived risk, Shimp and Bearden (1982) describe how non-monetary factors such as warranty are effective as risk relievers. Although a warranty is not a typical sales promotion, temporary non-monetary aspects can be an effective form of sales promotions. As
monetary sales promotions are rather cognitive than affective, this sales promotion type can result in consumer cautiousness (Chandon et al., 2000; Santini et al., 2015). Conversely, non- monetary sales promotions are more affective and, therefore, better suited for decreasing one’s risk perceptions. Moreover, non-monetary sales promotions can enhance consumers’
perceptions of product quality Chandon et al., 2000; Santini et al., 2015). Therefore, Lowe (2010) defines monetary and non-monetary sales promotions as respectively “reduced losses”
and “segregated gains” from the consumer perspective.
2.3. HYPOTHESES
Many studies have indicated that sales promotions have a positive effect on purchase intentions (Blattberg & Briesch, 2012; Chandon et al., 2000; Santini et al., 2015). Therefore, the following hypotheses is proposed:
H1a: sales promotions increase a consumer’s purchase intentions of the promoted product/service.
H1b: monetary sales promotions increase purchase intention more than non-monetary sales promotions.
The literature indicates positive effects of hedonic and utilitarian value perceptions of a product on consumer purchase intentions (Dhar & Wertenbroch, 2000; Voss et al., 2003).
Therefore, the following hypotheses is proposed:
H2a: the utilitarian value perception of a product/service has a positive effect on consumer purchase intention.
H2b: the hedonic value perception of a product/service has a positive effect on consumer purchase intention.
The risk dimensions overall, financial, performance, physical, psychological, social, and time risk are found to decrease purchase intention (Kaplan et al., 1974; Kim et al., 2005; Murray
& Schlacter, 1990). Therefore, the following hypotheses is proposed:
H3: an increase in the overall risk perception, and each risk dimension individually, decreases a consumer’s purchase intention.
H3a: an increase in the overall risk perception of a product/service decreases a consumer’s purchase intention.
H3b: an increase in the social risk perception of a product/service decreases a consumer’s purchase intention.
H3c: an increase in the time risk perception of a product/service decreases a consumer’s purchase intention.
H3d: an increase in the financial risk perception of a product/service decreases a consumer’s purchase intention.
H3e: an increase in the physical risk perception of a product/service decreases a consumer’s purchase intention.
H3f: an increase in the performance risk perception of a product/service decreases a consumer’s purchase intention.
H3g: an increase in the psychological risk perception of a product/service decreases a consumer’s purchase intention
H3h: financial risk is the most influential risk dimension, meaning high values of financial risk decrease purchase intention more than other risk dimensions.
Several studies have found the aforementioned phenomenon ‘benefit congruency’
(Chandon et al., 2000; Crespo-Almendros & Barrio-García, 2016; Santini et al., 2015). The following hypotheses are proposed:
H4: sales promotion type moderates the relationship between value perceptions of a product/service and purchase intention.
H4a: the utilitarian value perception, in its relationship with purchase intention, will be higher for monetary sales promotions than for no- and non-monetary sales promotions.
H4b: the hedonic value perception, in its relationship with purchase intention, will be higher for non-monetary sales promotions than for no- and monetary sales promotions.
Although respondents in the research of Michael and Greatorex (1993) and Cases (2002) chose sales promotions as potential risk relievers, Santini et al. (2015) have found negative effects of sales promotions on the relationship between financial risk and purchase intention.
However, non-monetary aspects of products and/or sales promotions are effective for relieving this negative effect (Chandon et al., 2000; Santini et al., 2015; Shimp & Bearden, 1982).
Therefore, the following hypothesis is proposed:
H5: monetary sales promotions have a negative moderating effect on the relationship between perceived risk dimensions and purchase intentions, whereas non-monetary sales
promotions have a weaker negative effect on the relationship between perceived risk dimensions and purchase intentions.
2.4. CONCEPTUAL FRAMEWORK
To answer the aforementioned research question and test the proposed hypotheses, a proper research design is needed (Figure 1). As explained in prior sections, the presented research is relevant due to the inclusion of other risk dimensions while being moderated by sales promotion types. Although many other factors can influence one’s purchase intention, this conceptual framework captures two important intrinsic variables that influence purchase intention: value and risk perceptions of the product (Stone & Grønhaug, 1993; Voss et al., 2003).
3. METHOD
In this section the research method, data collection and measures are described which are used to test the hypotheses.
Figure 1. Conceptual model
3.1. RESEARCH DESIGN
The research is based on quantitative methods. An experimental research design was used wherein the respondents were randomly assigned to three groups: a control group and two treatment groups. A between-subject design was used to measure the effect of sales promotion presence and moderating effect of sales promotion type.
The respondents were provided with a short introduction wherein respondents were asked to imagine themselves searching for a hotel for a night away within the Netherlands. It was assumed respondents were familiar with booking a night at a hotel or were, at least, capable of imagining themselves doing this. Further explanations were considered to be unnecessary since no exceptional terms and definitions were used in the survey questions. After this, 33 questions were used to measure value perceptions, risk perceptions and purchase intention. The order of these questions was randomized.
3.1.1. EXPLORATORY STAGE
In the exploratory stage, a suited product/service was searched for to evoke value and risk perceptions. A hotel was chosen as the promoted service, based on the findings of the studies of Mitchell & Greatorex (1993) and Voss et al. (2003). These studies have focused on, respectively, risk and value perceptions of products and services. Mitchell & Greatorex (1993) have found that hotels are perceived as one of the riskiest services. Although Santini et al.
(2015) caution for sales promotion research that involve too risky products/services, respondents in Mitchell & Greatorex’ (1993) study appoint sales promotions as an effective risk reliever for hotels. Voss et al. (2003), who have tested their scales for hedonic and utilitarian value on sixteen products/services, have found both hedonic and utilitarian value for hotels. The average price for one night at a hotel in the Netherlands is €104, according to NOS
(2018). Therefore, the hotel in the questionnaire was offered at this price. The hotel room photo used in this experiment was chosen based on the price per night (prior to COVID-19).
Secondly, the sales promotion types were chosen. A wide range of sales promotions has been tested in the literature. A price discount is seen as the purest form of monetary sales promotions (Blatt & Briesch, 2012; Chandon et al., 2000). For non-monetary sales promotions, sweepstakes are considered as one of the most effective sales promotions since it evokes the most hedonic benefits (Chandon et al., 2000; Santini et al., 2015). In the presented study, a price discount was used as a monetary sales promotion and a chance of winning a Michelin-star diner was used as non-monetary sales promotion. Based on secondary data from Groupon, the average price discount across 251 hotel offers was 35%. In a pre-test, a Michelin-star diner was picked as the best option out of three non-monetary sales promotions. Furthermore, the chosen sales promotions were tested for utilitarian and hedonic benefits. Monetary sales promotions should evoke utilitarian benefits whereas non-monetary sales promotions should evoke hedonic benefits (Chandon et al., 2000).
3.1.2. EXPERIMENTAL STAGE
In the experimental stage, all respondents were divided through random assignment.
Respondents were divided into three groups: the control group, monetary sales promotion and non-monetary sales promotion. The same hotel room was shown to each respondent.
Respondents in the control group were not exposed to any sales promotion. Respondents in one treatment group were exposed to a hotel offer with a price discount (monetary), whereas respondents from the other treatment group saw a hotel offer with a chance of winning a Michelin-star diner (non-monetary). While exposed to the hotel offer, the respondents answered questions to measure purchase intention, value- and risk perceptions.
3.2. DATA SAMPLE
Non-probability sampling methods (convenience, self-selection and snowball) were used to obtain as much respondents as possible. The survey, designed with Qualtrics, was spread via e-mail, LinkedIn and WhatsApp. To increase the response rate, the invitation message was personalized with the potential respondents’ names and respondents were asked to spread the survey among colleagues, fellow-students, friends and families (Healey, Baron &
Ilieva, 2002). The respondents (N = 149) did not receive an incentive for participating in- nor spreading the survey.
3.3. MEASURES
The items used for the questionnaire were selected from prior studies on value perceptions, perceived risk and purchase intention. In the following sections, these items are explained. Unless indicated differently, the items are measured on a 7-point Likert scale ranging from ‘strongly disagree’ to ‘ strongly agree’. All items and questions used for the questionnaire are listed in the Appendices (Appendix A).
Purchase Intention
To measure purchase intention, three items were retrieved from Barber, Kuo, Bishop and Goodman Jr’s study on purchase intention and willingness to pay (2012). An example of an item used to measure purchase intention: “I would consider purchasing a night at the hotel.”
Value Perceptions
The hedonic and utilitarian value of a product, perceived by a consumer, was measured with the ten-item scale of Voss et al. (2003). A 7-point Likert scale was used to measure these items (e.g., extremely ineffective – extremely effective). Each question was introduced with:
“How do you value a staying at this hotel? Please select a definition which describes your feeling the most?”
Risk Dimensions
The items used for measuring the risk perceptions were retrieved from Stone and Grønhaug’s (1993) research on the overall risk construct. These items were measured with a 7- point Likert scale. Stone and Grønhaug (1993) recommend customizing the 20 questions to be realistic and in alignment with the chosen product. For example, the physical risk for booking a night at a hotel had to be specified to the expectation of not getting a clean room. A broken leg caused by booking a night at a hotel is not a realistic physical risk. An example of a question used for overall risk in the presented research: “All things considered, I think I would be making a mistake if I book a night for this hotel.”
Control Variables
To enhance the research design, some factors were controlled for. Firstly, Santini et al., (2015) recommend aiming at a homogeneous sample much as possible to improve internal validity. Therefore, the respondents were asked for their age, gender and income. Secondly, research has found that brand familiarity and loyalty have a large impact on purchase intention and can be strong risk relievers (Chandon et al., 2000; Mitchell & Greatorex, 1993). Therefore, the hotel offer was not branded. Lastly, respondents were given a certain context in the introduction to avoid misconceptions. The respondents were asked to imagine themselves searching for a hotel prior to COVID-19. Due to the pandemic, hotel prices have dropped such that these are unrealistic in comparison with ‘normal’ circumstances.
4. RESULTS
4.1. DATA PREPARATION
In total, 219 responses were collected. The final sample consists of N = 149 participants after several cases were removed. Firstly, 52 responses were removed since they did not finish the survey. Secondly, 16 responses were excluded because of an incorrect answer to the attention check. Thirdly, 2 responses were removed due to detected outliers z > |3|. Furthermore, the completion time of the survey was checked. There were no respondents who finished the survey remarkably fast. The descriptive of the sample is presented in Table 3. The sample is rather young with a lower income. Furthermore, the sample consists of more females than males.
Table 3. Descriptive of Sample
Gender Percentage (%) Frequency
Male 38,9 58
Female 61,1 91
Age Percentage (%) Frequency
18 - 25 years 50,3 75
26 - 34 years 30,9 46
35 - 44 years 4,7 7
45 years or older 14,1 21
Net Income per month Percentage (%) Frequency
Less than €1500 37,6 56
€1500 - €2500 32,2 48
€2501 - €3500 16,1 24
€3501 - €4500 9,4 14
€4501 - €5500 2 3
More than €5500 2,7 4
The variables of risk, value and purchase intention were investigated for normality by an analysis of skewness and kurtosis. For all variables, the skewness and kurtosis levels were
between -1 and 1. Therefore, the data was assumed to be normally distributed and no variables had to be transformed.
Furthermore, the items of our scales were checked for reliability with Cronbach’s Alpha.
Value of a < 0.7 are considered as less reliable (Bland & Altman, 1997; Spajic, 2020). Purchase intention (a = .906), hedonic value (a = .860), financial risk (a = .776), performance risk (a = .701) and psychological risk (a = .862) showed to be reliable scales. The scales for utilitarian value (a = .687), overall risk (a = .623) showed to be slightly less reliable. Time risk (a = .544), physical risk (a = .524) and social risk (a = .484) showed to be unreliable. The variables are not excluded from the analysis. However, the insufficient Cronbach’s Alpha scores are taken into account in the discussion and conclusions. The reliability for the scales could not be enhanced. If items were deleted from these scales, Cronbach’s Alpha scores would be lower.
A variable ‘sales promotion type’ was created representing the groups respondents were assigned to: no sales promotion (N = 45), monetary sales promotion (N = 51) and non-monetary sales promotion (N = 53).
To get a first impression of the data, a Pearson correlation matrix of all variables is presented in Table 4. Although correlations do not indicate causal relationships, the correlation matrix allows us to have an impression of the extent of linear relationships between variables.
Utilitarian and hedonic values show moderate to high positive correlations with purchase intention at a significant level of p < 0.01. All risk dimensions, except physical risk, are significant (p < 0.01) negatively correlated with purchase intention. These correlations vary from low to moderate. Furthermore, all risk dimensions have moderate to high positive correlations with each other at a significant level of p < 0.01.
4.2. EXPLORATORY ANALYSIS
Before performing statistical analysis to test the hypotheses, the independent- and dependent variable(s) were analyzed for potential remarkable group differences. With age, income and sales promotion type as a factor, One-way ANOVA tests were performed with purchase intention, value- and risk perceptions as dependent variables. An Independent
Table 4. Correlation matrix with means, standard deviations and reliability scores.
Variables M SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1. Gender 1,61 0,49 -
2. Age 2,83 1,04 -,068 -
3. Income 2,14 1,23 -,223** ,456** -
4. Sales Promotion Type 1,05 1,35 -,083 ,123 ,195* -
5. Purchase intention 4,13 1,36 ,083 -,188* -,098 -,068 (.906) 6. Utilitarian value 4,65 0,74 -,048 -,171* ,006 ,086 ,400** (.687) 7. Hedonic value 3,9 1,01 ,226** -,293** -,244** ,019 ,686** ,278** (.860) 8. Overall risk 3,13 1,12 -,050 ,107 ,036 ,094 -,629** -,304** -,467** (.623) 9. Social risk 2,87 1,08 ,069 -,002 -,136 -,007 -,215** ,007 -,217** ,513** (.484) 10. Time risk 3,26 0,99 ,041 ,039 -,120 -,101 -,410** -,301** -,322** ,563** ,433** (.544) 11. Financial risk 3,96 1,29 ,031 ,032 -,081 -,086 -,616** -,364** -,384** ,664** ,283** ,391** (.776) 12. Physical risk 2,8 1,06 ,055 ,014 -,037 -,019 -,078 -,079 -,144 ,401** ,361** ,389** ,275** (.524) 13. Performance risk 3,76 1,17 ,045 ,072 -,006 ,071 -,409** -,128 -,361** ,604** ,372** ,452** ,478** ,469** (.701) 14. Psychological risk 2,73 1,21 ,081 ,120 -,009 -,056 -,531** -,219** -,459** ,626** ,434** ,647** ,453** ,527** ,586** (.862)
**. Correlation is significant at the 0.01 level (2-tailed)
*. Correlation is significant at the 0.05 level (2-tailed)
Samples T-test was used to analyze group differences between males and females with purchase intention, value- and risk perceptions as dependent variables. Levene’s statistic was checked for each test. None of these tests were significant, which means we assume homogeneity of variances.
There were found multiple significant group differences for hedonic value. Firstly, there was a significant difference between males and females. Male (M = 3,62; SD = 0,91) mean hedonic value was significantly lower (t(147) = -2,81; p < 0.01) than for female (M = 4,08; SD
= 1,03) Secondly, there was found a significant difference between age groups for hedonic value (F(3, 145) = 6,52; p < 0.001). Turkey post-hoc tests showed a significant difference (p <
0.01) between age group 18 – 25 years and the age groups 26 – 34 years and 45 years or older.
The hedonic value was higher for participants who were 18 – 25 years old. It must be stated that the group 45 years or older has much fewer observations than the two younger age groups (Table 3). Therefore, results might be different if the two older groups contained more observations. Thirdly, a significant difference between income groups was found for hedonic value (F(5, 143) = 2,40; p < 0.05). However, turkey post-hoc tests showed no significant difference between groups (p > 0.05).
The One-way ANOVA with sales promotion type as factor showed significant differences between groups for three variables: overall risk (F(2, 146) = 4,00; p < 0.05), time risk (F(2, 146) = 4,79; p < 0.05 ) and financial risk (F(2, 146) = 9,83; p < 0.001). Turkey post-hoc tests showed a significant difference for overall risk between the monetary- and non-monetary sales promotion (p < 0.05). Monetary sales promotion mean for overall risk was significantly lower than the non-monetary sales promotion. For time risk, monetary sales promotion mean was significantly lower than for no sales promotion (p < 0.01). For financial risk, the mean of monetary sales promotion is significantly different than no sales promotion (p < 0.001) and the
non-monetary sales promotion (p < 0.01). Monetary sales promotion mean for financial risk was lower than both no sales promotion and the non-monetary sales promotion.
4.3. MAIN ANALYSIS
To test H1, the One-way ANOVA results with sales promotion type as a factor were analyzed. There was no significant mean difference for purchase intention F(2, 146) = 2,33 (p
> 0.05) between sales promotion types. Therefore, H1a is rejected. Turkey post-hoc results showed directional support for purchase intention being the highest for the monetary sales promotion. However, it was not significant which means H1b is rejected.
In addition to prior tests, a Factorial ANOVA was performed to for check interaction effects between sales promotion type and our categorical control variables (age, income, gender) in their relationship with purchase intention. Firstly, there was a non-significant interaction effect between sales promotion type and gender in their relationship with purchase intention (F(2, 143) = 0,70; p > 0.05). There was a remarkable mean difference for purchase intention between male (M = 3,53; SD = 1,59) and female (M = 4,12; SD = 1,24) within the non-monetary sales promotion group. However, an Independent Samples T-test showed that this difference was not significant (t(51) = -1,54; p > 0.05). Secondly, there were non-significant interaction effects between sales promotion type and control variables age (F(6, 137) = 0,463; p > 0.05) and income (F(1, 131) = 1,80; p > 0.05).
H2 and H3, and corresponding sub-hypotheses, were tested with a hierarchical regression analysis. Before analyzing the results of the regression, the data was checked whether it satisfied all assumptions of regression. Firstly, the dependent variable purchase intention was continuous. Secondly, the independent variables utilitarian value, hedonic value and all seven risk dimensions were continuous. Therefore, the variables were suited for regression. The
variables were measured on a categorical Likert-scale, but were transformed into continuous variables as the items for each variable were combined by computing scale means. Thirdly, there was tested for the independence of residuals by assessing the Durbin-Watson statistic. As this was close to 2 (2,011), it is assumed there is no correlation between residuals. Fourthly, the data was checked for linearity by plotting studentized residuals and unstandardized predicted values in a scatter plot. Moreover, partial regression plots of purchase intention against the independent variables were inspected. Based on the scatterplots (Appendix D1), it is assumed that the relationships between the purchase intention and the independent variables are linear.
Fifthly, a scatterplot (standardized residuals vs. standardized predicted values) was used to check for homoscedasticity. Based on visual inspection (Appendix D2), homoscedasticity is assumed. Sixthly, the data was checked for multicollinearity among the independent variables by inspection of Pearson’s correlation coefficients and Tolerance/VIF values. There were no correlations stronger than 0.7 or -0.7 (Table 4) and none of the variables had a Tolerance value
< 0.1. Based on the observations of correlation coefficients and Tolerance/VIF values, we assume there is no multicollinearity among the independent variables. Lastly, a histogram and P-P Plot (Appendix D3) of the residuals were checked for normal distribution. Based on visual inspection, the data is assumed to be normally distributed.
Table 5. Hierarchical regression model of Purchase Intention.
R R² R² Change B SE β t
Step 1 0,2 ,04 ,02
Income ,01 ,16 0,01 ,09
Age -,24 ,12 -,18* -1,99
Gender ,19 ,23 ,07 ,82
Sales promotion type -,07 ,14 -,04 -,50
Step 2 ,84 ,71*** ,67***
Income ,02 ,06 ,02 ,29
Age ,01 ,07 ,01 ,17
Gender -,04 ,14 -,01 -,25
Sales promotion type -,18 ,08 -,11* -2,12
Utilitarian value ,18 ,10 ,10 1,78
Hedonic value ,56 ,08 ,41*** 6,81
Overall risk -,25 ,10 -,21*** -2,48
Social risk ,09 ,07 ,07 1,28
Time risk -,02 ,09 -,02 -,28
Financial risk -,29 ,07 -,28*** -4,17
Physical risk ,31 ,07 ,24*** 4,18
Performance risk ,02 ,08 ,02 ,30
Psychological risk -,26 ,09 -,23** -3,02
* is significant at the level p < 0.05, ** is significant at the level p < 0.01, *** is significant at the level p < 0.001.
The hierarchical regression (Table 5) was performed to test the ability of consumers’
value- and risk perceptions of a product to predict levels of purchase intention, after controlling for income, age, gender and sales promotion type.
In step 1, four predictors were entered: income, age, gender and sales promotion type.
This model was statistically not significant F(4, 144) = 1,58; p > 0.05. If the model was significant, it would explain 4,2% of the variance in purchase intention. In step 2, the value- and risk perception variables were entered in the model. This model was statistically significant F(13, 135) = 25; p < 0.001 and explained 71% of the variance in purchase intention. The addition of value- and risk perceptions explained 69% more variance (R² Change = 0.69; F(9, 135) = 34,97; p < 0.001).
In the final model, six of thirteen predictors were statistically significant. The Beta value for hedonic value is β = ,41; p < .001. When hedonic value perceptions increase for one, their purchase intention will increase by 0,41. Therefore, H2b is supported. Utilitarian value showed a positive non-significant effect (β = ,10; p > 0.05) on purchase intention. This means H2a is rejected.
Not all risk dimensions had a significant (negative) effect on purchase intention.
Therefore, H3 is rejected. As social- time- and performance risk were not significant, we reject H3b, H3c and H3f. Overall-, financial- and psychological risk had a significant negative
relationship with purchase intention. When overall risk (β = -,21; p < .001) increases for one, their purchase intention will decrease by 0,21. This means H3a is supported. Purchase intention will decrease by 0,28 when financial risk (β = -,28; p < .001) increases for one. Thus, H3d is supported. Furthermore, the effect (β = -,28) for financial risk is the strongest of all risk dimensions. Therefore, H3h is supported. When psychological risk (β = -,23; p < .01) increases for one, purchase intention decreases by 0,23. Thus, H3g is supported. Physical risk (β = ,24; p
< .001) has a significant positive Beta value in the regression model. This means when physical risk increases for one, purchase intention increases by 0,24. H3e is therefore rejected.
H4 and H5 included an expected moderating effect of sales promotion type on respective value- and risk perceptions. To test the hypotheses, a PROCESS moderation model (1) was performed (Hayes, 2012). For each independent variable, a moderation regression was performed with purchase intention as the dependent variable and sales promotion type as the moderation variable. The independent variables were utilitarian and hedonic value, and the risk perceptions; overall, social, time, financial, physical, performance and psychological. As the scales for value- and risk perceptions are numerical, these variables were standardized. The moderator, sales promotion type, did not need to be standardized as this variable is categorical.
After running all regressions, H4 and H5 are both rejected. There was no significant moderating effect of sales promotion type (p > 0.05). This means that high levels of our independent variables do not result in more or less purchase intention among participants who were shown a specific sales promotion type.
5. DISCUSSION
This experiment used theory-based constructs to measure purchase intention, value- and risk perceptions. The results of the Pearson correlation matrix show correlations between
dimensions within the value and risk constructs. This indicates construct validity for the measurement items. However, not all scales can be considered as reliable in this experiment since Cronbach’s Alpha for utilitarian value, overall-, time-, physical- and social risk is lower than a < 0.7. The positive correlations between risk dimensions are in alignment with the results of Stone and Grønhaug (1993). The positive correlation between the utilitarian and hedonic value is as Voss et al. (2003) have indicated in their study. As expected, value perceptions have positive correlations with purchase intention whereas risk perceptions have negative correlations with purchase intention.
The internal validity is increased by several factors. Firstly, the participants in this experiment are predominantly young (younger than 35 years) and approximately distributed equally across experiment groups. Homogeneity of the sample increases internal validity (Santini et al., 2015). Secondly, the responses are validated by the use of an attention check or i.e. Instructional Manipulation Check (Hauser, Ellsworth & Gonzalez, 2018). Thirdly, the experiment in this research contains a control group that represented no sales promotion (i.e.
no manipulation).
The inclusion of a control group in this experiment is in contrast with the study of Santini et al. (2015), wherein participants are divided into two groups: monetary sales promotion and non-monetary sales promotion. Santini et al. (2015) tested the presence of a sales promotion through a within-subject design. Participants filled in the set of questions without a sales promotion, whereafter the same set of questions were asked with a sales promotion (type). Our experiment tested the presence of a sales promotion through a between-subject design. It is possible that sales promotion (moderating) effects on value, risk and purchase intention are different for a between-subject design. Future research on sales promotion experiments could explore whether this is different and to what degree.
The experiment shows diverse results regarding the expected main effects of value- and risk perceptions on purchase intention. As expected, there are significant moderate to high positive correlations between value perceptions (utilitarian and hedonic) and purchase intention.
Furthermore, utilitarian and hedonic values have a significant moderate positive correlation.
This means that someone with high utilitarian perceptions of the hotel is likely to have high hedonic perceptions as well. All risk perceptions, except physical risk, have significant moderate to high negative correlations with purchase intention. Furthermore, all risk perceptions have moderate to high significant positive intercorrelations. Therefore, a participant with a high certain risk perception is likely to have high values for other risk perceptions and lower scores for purchase intention.
However, the regression model shows that only hedonic value, overall risk, financial risk, psychological and physical risk are significant predictors for one’s purchase intention. Only physical risk shows an unexpected direction of its effect on purchase intention. The results show that higher physical risk perceptions increase purchase intention. As the intensity of risk perceptions varies across products, services and people, it is not unexpected some risk dimensions did not significantly decrease purchase intention (Kim et al., 2005; Stone &
Grønhaug, 1993). Nevertheless, the significant positive effect of physical risk on purchase intention is unexpected. A possible explanation is that the scale of physical risk (a = .524) is not reliable in this research. Although non-significant, physical risk was a positive predictor for purchase intention in the study of Kim et al. (2005).
Furthermore, utilitarian and hedonic values are both positively correlated with purchase intention. The results of the regression show that only hedonic value predict one’s purchase intention at a significant level of p < 0.05. It is against expectations that utilitarian value is not a significant predictor for purchase intention. It was non-significant at the standard alpha level
p < 0.05. However, the p-value was lower than 0.1. Increasing sample size could result in a significant effect of utilitarian value on purchase intention (Wilkerson & Olson, 1997).
The results of our experiment do not show a moderating effect of sales promotion type on value perceptions, i.e. the benefit congruency phenomenon of Chandon et al. (2000). Benefit congruency in the study of Chandon et al (2000) was found for high-equity brands, thus not for low-equity brands. As our hotel offer was brandless, this might be an explanation for not finding benefit congruency. However, this is in contrast with Santini et al. (2015). They have found benefit congruency for a brandless promoted product. Opposite benefit congruency can occur as well (Kwok & Uncles, 2005), meaning monetary (non-monetary) sales promotions work best for products/services with hedonic (utilitarian) value perception.
Despite the fact that moderating effects of sales promotion types are not found, this research contributes to sales promotion literature with contradictive findings to prior studies. The One- way ANOVA tests with sales promotion type as a factor showed significant differences between groups for overall-, time- and financial risk. People who received the monetary sales promotion perceived the least financial risk. This is diametrical to the findings of Santini et al. (2015), wherein the monetary sales promotion evokes more financial risk perceptions than the non- monetary sales promotion. Santini et al. (2015) describe that monetary sales promotions erode quality perceptions, thus result in higher risk perceptions. It is possible that participants in our study perceived the monetary sales promotion for the hotel as a better value for money. A suggestion for future research is therefore to investigate which factors determine whether a monetary- or a non-monetary sales promotion evokes the least risk perceptions.
For marketing managers, this means they have to be cautious when implementing sales promotion techniques as part of risk-relieving strategies. With our contribution to sales promotion literature, it seems that it is not a foregone conclusion that a non-monetary sales
promotion will evoke less financial risk perceptions than a monetary sales promotion. For premiums (i.e. free goods) for example, promotion attractiveness positively affects consumers’
perceptions of a sales promotion (d’Astous & Landreville, 2002). A poor product-sales promotion fit could be seen as manipulative, but mentioning the value of the free good helps to reduce the perceptions of manipulation (d’Astous & Landreville, 2002). The same relationships might be applicable for other sales promotions, meaning the choice for a non-monetary or a monetary sales promotion as a risk reliever might be more complex.
This research comes with its limitations. Although the experiment did not take place in a fully controlled laboratory setting, this research can be characterized as a laboratory experiment wherein one variable is manipulated. As the participants received the online survey through e- mail and social media, we were not able to control for certain interferences. For example, we cannot check whether participants discussed the survey with others or compared hotel offers on the internet. Furthermore, the results from this experiment might differ from real consumer behavior. Rejected (confirmed) hypotheses in this study could be confirmed (rejected) in a semi-natural- or field experiment. Purchase intention and actual buying behavior do not only differ from each other, these variables are affected by different factors (Rimal, Flether, McWatters, 2004). For example, Rimal et al. (2004) conducted a lab- and field experiment comparison on purchase intention and actual buying behavior of beef. Results indicated that actual buying behavior is affected by package labels and appearance, whereas purchase intention is influenced by attitude and demographics. In short, actual consumer choices cannot always be predicted or inferred from purchase intentions (d’Astous & Landreville, 2003).
Therefore, we have to be cautious for generalizing our results to actual consumer behavior.
Regarding generalizability, two factors are a limitation for this research. Firstly, this research only focused on a single service (hotel). As the literature indicates (Mitchell &
Greatorex, 1993; Voss et al., 2003), value- and risk perceptions vary across products and services. The same research design should be replicated to investigate whether the effects are similar for other products and/or services. Secondly, the survey was spread among participants who are living in the Netherlands. Therefore, it is not certain to what extend similar results will be achieved in other countries and/or cultures. Research has found similar sales promotion effects between ethnic groups within the same country (Kwok & Uncles, 2005). Multi-country studies indicate differences between attitude towards sales promotions (Fam, Brito, Gadekar, Richard, Jargal & Liu, 2019) and sales promotion type (Huff & Alden, 2000).
Furthermore, the reader should bear in mind that this study measured the perceived value of the hotel offer with two dimensions. It is beyond the scope of this research to see what factors exactly made participants value or dislike the hotel offer. Several factors are decisive in consumers’ perceptions of a hotel: staff service, quality, room quality, general amenities, business services, value and security (Choi & Chu, 1998). In our experiment, it was a deliberate decision to exclude such information from the hotel offers. Mainly because these could interfere as risk relievers or affect purchase intention. Without these factors participants may find it more difficult to develop an attitude towards the hotel offer, thus answering more neutral in a survey.
Therefore, it would be interesting to investigate how different levels of these hotel perceptions (high/medium/low) influence the constructs in our experiment. As the importance of these hotel perceptions varies at price-levels (Choi & Chu, 1998), this should be taken into account in future research.
Lastly, the scope of this research does not grasp the personal characteristics of participants.
It would be insightful to include this in future sales promotion research within the same relationships as our study: purchase intention, value and risk perception. Including personal characteristics can potentially enhance the explanation of heterogeneity in participants’
