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 non-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’
responses. A widely used personal trait in sales promotion literature is deal-proneness, or rather promotion-proneness (Crespo-Almendros & Del Barrio-García, 2016). This variable can be measured through quite explicit questions such as: “I have a favorite brand, but most of the time I buy a brand that’s on special offer (Crespo-Almendros & Del Barrio-García, 2016).”
Another potential source of valuable insights on personal characteristics in risk reduction research could be to include whether someone is a Maximizer or ‘Satisficer’ (Schwartz, Ward, Monterosso, Lyubomirsky, White & Lehman, 2002). Satisficers (non-maximizers), a wordplay of satisfy and suffice, pursue a good enough option instead of the best option. Maximizers are people who want the best option possible. Therefore, they want to have the feeling they have considered all options. Maximizers tend to depend more on social comparison and self-esteem, hence social and psychological risk, than satisficers. Therefore, which risk perception(s) is most predominant for someone might differ depending on whether they are a maximizer of a satisficer. As a consequence, the most suitable sales promotion type for risk-relieving strategies may vary for these characteristics.