• No results found

creating an attitude towards Webrooming

N/A
N/A
Protected

Academic year: 2021

Share "creating an attitude towards Webrooming"

Copied!
44
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

~ Master’s Thesis ~

An Evaluation of Modern Shopper

Characteristics as a source for creating an attitude towards Webrooming.

Written by: Kamran Mahmudov – S2751194 Supervisor: Berger, J.

Faculty of Economic and Business Marketing Programme

Date: 17 June 2019

Abstract

Webrooming behaviour is one of the most common cross-channel shopping behaviours. Thus, this study aims at finding a better understanding of the behaviour and its antecedents. More specifically, on top of the perceptions and beliefs, which can influence the attitude towards webrooming, I examine the effect of individual characteristics, namely uncertainty avoidance and future orientation. Firstly, the research is based on the theory of planned behaviour, but instead of utilising all three predictors of the behaviour, this study mainly focuses on the attitude and how individual characteristics can play a role in shaping them. The findings of eight hypothesised statements illustrate the link between an individual characteristic and formation of perceptions, which in turn leads to a specific attitude. Moreover, the results stress the importance of modern shopper characteristics when studying cross-channel shopping behaviour.

Keywords: webrooming, attitude, intention, perceived benefits, perceived risks, online channel, offline channel, uncertainty avoidance and future orientation.

(2)

Table of Contents

Abstract ... 1

Introduction ... 3

Theoretical Framework ... 4

Research Methodology ...10

Field Research Design. ... 10

Measures... 11

Statistical Procedure. ...15

Results ... 15

Measurement Model. ...15

Structural Model. ... 18

Discussion and Conclusion ... 20

Scientific and Managerial Relevance. ... 22

Limitations and Future Research. ... 23

The References ... 24

Appendix 1: Measures. ... 30

Appendix 2: R-squared and Adjusted R-squared values. ... 31

(3)

Introduction

Currently, habits of shoppers are rapidly changing, which creates pressure to clarify an idea of modern shoppers and their behaviour. Every shopper has a unique behaviour, which is derived from intention and attitude (Ajzen, 1991). The Internet and technology have empowered consumers beyond the level of convenience offered by marketers and businesses. The term savvy is related to those consumers who are much better informed, capable and connected (Macdonald & Uncle, 2007). These shoppers know exactly what they are purchasing and from where. Furthermore, the consumer journey is becoming more extensive, involving many different touchpoints, regardless whether it is online or offline. The different channels are used to suit consumers’ needs, and every shopping occasion can be comprised of different journeys. Therefore, new concepts, such as showrooming and webrooming, started becoming popular among Marketing researchers (Arora & Sahney,2018). Showrooming is referred to as the process whereby consumers initially visit the physical store to gather information, but make the purchase decision through online channels (Flavian et al., 2016). Whereas, webrooming, also known as reverse showrooming, refers to the situation where consumers gather information online prior to the visit to a physical store (Flavian et al., 2016). The phenomenon of webrooming is becoming more and more prevalent, even though it is not completely novel, as marketers refer to it as cross-channel free riding behaviour. According to the Forrester Research (2014), which was conducted to predict cross channel free riding behaviour in Europe between 2013 and 2018, sales generated from webrooming outweigh online sales by 500 per cent. PwC’s Annual Global Total Retail Consumer Survey (2015), also confirms that 70 per cent of shoppers around the world have collected information online prior to the purchase decision done through an offline channel. Thus, this phenomenon raises a challenge for marketers to understand how webrooming shapes the customer journey. Moreover, as customers are becoming more well-informed and savvier, understanding their behaviour becomes even more complicated. Therefore, researchers try to study the antecedents of customer behaviour more in depth.

So far literature has proposed utilising the theory of planned behaviour, which uses attitude as one of the primary antecedents of behaviour (Ajzen, 1991). More specifically, authors utilise perceived benefits and risks as antecedents of attitudes toward certain parts of the customer journey (i.e. online search, offline purchase), which in turn predicts the attitude towards webrooming (Arora & Sahney, 2017; Arora

& Sahney, 2018). Nevertheless, there is a lack in the literature in regards to understanding whether the characteristics of a modern shopper shape the extent and magnitude of their perception formation.

Therefore, this research proposes to find whether characteristics such as future orientation and uncertainty avoidance, which can vary at the individual level, will lead to creating a different level of attitude towards webrooming. Consequently, will these characteristics explain the behaviour of a modern shopper engaging in webrooming.

The research paper is structured as follows. Firstly, we dive into the literature and form a theoretical framework, which in turn acts as a basis for the conceptual model. Then, research

(4)

methodology section explains the manner in which the field study was performed. In this paper, the millennial shoppers (i.e. individuals born between 1981 and 2001) were chosen as the sample for modern shopper. Further reasons for this choice are discussed in the research methodology chapter. This is followed by depicting findings and discussing the theoretical importance of the constructs and relationships between them. The paper finishes with the summary of all the findings, limitations and propositions for the further research.

Theoretical Framework

The term ‘webrooming’ started to get greater attention recently and most of the authors research this topic more in-depth to understand, not only the behaviour, but also intention and attitude towards it (Arora & Sahney,2018; Arora & Sahney, 2017; Nesar & Sabir , 2016; Flavian, Gurrea & Orus, 2016).

Studies have shown that the main reason customers use multiple channels during the single purchase process (Verhoef et al., 2015; Frasquet et al., 2015), is to maximise benefits and minimise costs of shopping (Verhoef et al., 2007). Furthermore, even though most of the literature follows almost the same base conceptual model, each of them adds certain new variables, which can potentially influence the attitude towards webrooming. In the research by Flavian et al. (2016), authors argue that consumers engage in the process of reverse showrooming to utilise the Internet reviews to make a better choice, whereas Verhoef et al. (2007) highlights that it is risks that push customers to visit a physical store before making the final purchase decision after collecting information online. The latter research is coherent with findings by Chiu et al. (2011) and Chou et al. (2016). The lack of product diagnosticity and need for touch are other reasons which acts as a push to create favourable webrooming attitude (Flavian et al., 2016; Reid et al., 2016). Due to the choice overload, shoppers also use webrooming in order to narrow down the consideration set and learn more technical information, before visiting a physical store to close the sale (Wolny & Charoensuksai, 2014).

According to the Theory of Planned Behaviour (Ajzen, 1991), before behaviour occurs people form an attitude, which has an impact on the intention to execute the behaviour. The intention to act is an immediate antecedent of the behaviour; hence, if the intention to webroom is positive, the favourable behaviour can be almost guaranteed (Arora & Sahney, 2018). The gap between intention and behaviour will be narrowed in the case when intention is positive and favourable, as this type of intention does not require additional evaluation of possible outcomes. Nevertheless, whether consumers will have the willingness to continue and proceed with webrooming behaviour depends on post-purchase satisfaction (Flavian, Gurrea & Orus, 2016). Post-purchase choice confidence is viewed as a crucial outcome of the reverse showrooming, which illustrates that shopper has put a sufficient amount of effort to arrive at the best decision (Flavian, Gurrea & Orus, 2016). Whereas intention occurs prior to behaviour (Ajzen, 1991), attitude is one of the most important antecedents of the intention.

(5)

Attitude towards the behaviour is defined as the extent to which evaluation of the behaviour in question is favourable or unfavourable (Ajzen, 1991). The more favourable the attitude, the stronger should be the intention to act on certain behaviour. Even though, the theory of planned behaviour also describes social norms and perceived behavioural control as antecedents of the intention, the relative importance of each of these variables can vary across different situations and behaviours. Therefore, in some situations only attitude could be a significant predictor of intention, while in others it might be the combination of attitude with one of other two variables or all three of them (Ajzen, 1991). However, whereas perceived behavioural control’s and social norms’ influence alters based on the situation, the attitude is always present as a predictor of intention and behaviour subsequently. According to the expectancy-value model of attitudes, depicted in work by Fishbein and Ajzen (1975), it is beliefs about the attitude object that develop people’s attitude. Generally, people form beliefs by linking them to certain attributes. Specifically, each of the belief associates the behaviour either to a potential outcome or to another attribute, such as costs of performing certain behaviour (Fishbein & Ajzen, 1975).

Therefore, in a case when the outcome of the behaviour is mostly desirable, people form a favourable attitude. Following the topic, the attitude object is webrooming, and when people form a positive and favourable attitude towards it then the intention to engage in the behaviour of webrooming will be greater, which leads to the first hypothesis:

H1: Positive attitude about webrooming behaviour will have a positive and direct influence on the intention of webrooming behaviour.

Even though it can be evident from previous researches that positive attitude will lead to favourable intention towards webrooming behaviour, it is a necessary step before diving into the analysis of antecedents of attitude. Based on the definition of the webrooming (Flavian et al., 2016), the attitude towards webrooming might consist of two different attitudes formed by beliefs, namely the beliefs about online and offline channels of the customer journey. Starting with the integration of both channels, the attitude towards webrooming will be more significant if people actually perceive multi-channel search as beneficial. According to Neslin et al. (2006), the cross-channel shopping is a highly involving behaviour, because when consumers are highly involved with a product, their information needs are more extensive than usual, which creates a motivation to engage in extensive information search prior to the purchase (Eagly & Chaiken, 1993). Certain reasons of the strong motivation to put effort, time, and energy in multichannel search could be due to the desire to choose the best available option (Puccinelli et al., 2009), overload of information relative to the limited cognitive capacity (Walsh &

Mitchell, 2010), and inability to physically interact with the product or other people (i.e. such as other customers or service personnel) in the online channel (Mitchell, 1999). Despite the fact that digital revolution has not altered the stages consumers go through in the shopping journey (i.e. search and evaluation), the path of the journey has changed (Kruh, 2017; Stein & Ramaseshan, 2016). Whereas the traditional path to purchase is linear, currently due to the availability of a myriad of both online and offline channels at every stage the path has become more like a cycle, where consumers move back and

(6)

forth between stages while utilising both the Internet and physical stores (Kruh, 2017). Consequently, as consumers started to use various channels at different stages to maximise the satisfactory outcome, the terms of webrooming and showrooming emerged in marketing literature. Therefore, the consumer will create a positive attitude towards webrooming, in order to satisfy these needs through the offline channel after building confidence through informational search via the online channel.

H2: Perceived Benefits of Cross-Channel Search will have a direct and positive effect on attitude towards webrooming.

Nevertheless, the online channel part of the cross-channel search behaviour should be analysed more thoroughly, because people engaging in the webrooming behaviour search online, but when it comes to the stage of the purchase decision, those customers turn to physical stores. Consequently, the attitude towards the webrooming behaviour will be favourable if there is a distrust in the online channel, which can be identified through comparing benefits of the search and risks of purchase in the online context. Distrust can be defined as negative expectations towards one’s conduct (Lewicki et al., 1998), which can be regarded towards online channels as a lack of personal interface, which complicates the process of building confidence in regards to the seller and product (Harridge-March, 2004). The level of the distrust among webroomers, which can range from full distrust to full trust, might encompass two parts, namely benefits of the search and risks of the purchase online, which can explain the reason of engaging in the webrooming process. Therefore, in order to understand the elaborative definition of the distrust, one should first look into two separate parts (i.e. benefits and risks).

According to the literature, there are several reasons why customers pursue online search and collect information before purchasing in the physical store (Arora & Sahney, 2018; Flavian et al., 2016;

Noble et al., 2005; Balasubramanian et al., 2005). Specifically, studies identify three main benefits of the online search, namely search costs, electronic word of mouth (e-WOM) and convenience. In the study by Noble et al. (2005), the authors mention that the Internet is one of the most important and strongest channels used in the search stage of the consumer journey. The vast amount of different websites and platforms gives an opportunity to consumers to collect and compare different sorts of information, such as attributes of the product, quality measures and price. Moreover, one can access a wide range of information without a need to contribute much time, effort and energy, which is more efficient compared to visiting brick and mortar stores (Noble et al., 2005; Jepsen, 2007). Consequently, the search costs have considerably decreased since the introduction of the online channels.

Along with reduced search costs, consumers usually prefer to use the Internet to collect information about the product due to the perceived benefits of the electronic word of mouth (e-WOM), which is usually referred to as the online reviews (Cui et al., 2012). The advent of the Internet has provided opportunities to consumers to gather unbiased product information based on the experience of other consumers and to offer their advices via electronic word of mouth (Hennig-Thurau et al., 2004).

According to Strauss (2000), e-WOM communication can be defined as any negative or positive statement about a product or a firm made by former, actual or potential customers, and available to the

(7)

audience via the Internet. This communication can occur in different ways, such as discussion forums, news groups or web-based platforms (Hennig-Thurau et al., 2004). The information available online in a form of the reviews can significantly affect consumers’ decision-making process (Zhang et al., 2014).

Such information assists consumers to understand product features and functionality, to compare possible alternative and hence raise confidence in making the better choice (Arora & Sahney, 2018).

Finally, the third benefit of the online search is the convenience of gathering information online, which is straightforward as consumers can access a vast amount of websites and online channels through the computer, tablet or a mobile phone whenever and wherever it is convenient. To summarise, the ability to access and compare information provided by both the store website and e-WOM at anytime, anywhere reduces the search cost and hence creates a positive belief about online search. Therefore, when consumers have strong perceptions about the benefits of online search, his/her confidence about the Internet shopping process will be improved in a positive way.

H3: Perceived Benefits of Online Search will have a negative effect on Online Distrust.

As there are benefits of the Internet, there are also disadvantages. As mentioned above, consumers might hold strong beliefs about online search, but still choose to visit brick and mortar due to risks of online purchase. Based on the literature review, there are two main risks associated with the online channel, which withholds customers from proceeding with purchasing via the Internet, namely financial and performance risks. Generally, perceived risk is defined as “the likelihood that purchase of the item will result in general dissatisfaction of the consumer” (Pires et al., 2004). While purchasing online, customers tend to feel insecure about the misuse of credit card information, deceptive activities online or just money loss (Chiu et al., 2011; Wu & Wang, 2005; Forsythe & Shi, 2003), which eventually drives them to use a physical store. On the other hand, the lack of product diagnosticity leads to the performance risk, as consumers might be worried about unsuitability of the product or quality (Chou et al., 2016; Forsythe & Shi, 2003). Moreover, due to the vast amount of information and limited cognitive capacity, consumers usually find it challenging to avoid uncertainty and hence increase confidence about a chosen product. All these risk factors can be one of the main aspects which explains the drive of consumers avoiding online channels and preferring offline ones during the purchase decision. Therefore, the higher risks reduce the confidence of a shopper and pushes him/her away from making a final decision in the online channel and hence increases distrust.

H4: Perceived risks of online purchasing have a positive effect on online distrust.

Distrust acts as a mediator between the perception of online shopping and attitude towards webrooming. It has been confirmed by several authors that the Internet is considered to be the strongest channel for information search (Widing & Talarzyk, 1993; Peterson & Merino, 2003; Arora & Sahney, 2018). Nevertheless, after making a choice and putting a product in the virtual basket, customers face the risks discussed above. These risks (i.e. risks of online purchase) outweigh risks of offline purchase and hence pushes consumers to engage in the webrooming behaviour. Therefore, distrust in online shopping will be an important determinant of the attitude of webrooming, which will be determined by

(8)

the perception of the search and risk of purchase via an online website. In most of the scenarios, the risks of purchasing online outweigh the perceived benefits, hence increasing the distrust. Consequently, distrust of the online shopping will determine the attitude towards preferring to purchase through physical stores after gathering the information online.

H5: The Distrust in Online Shopping has a positive and direct effect on the attitude towards webrooming.

As mentioned above, all the risks associated with the online channel increase the level of uncertainty, which pushes customers to prefer buying though the offline channel. Specifically, according to the uncertainty reduction theory, feeling uncertain has a strong motivational effect (Lee, 2001;

Stafford and Grimes, 2012; Flavian et al., 2016). Not being able to physically interact with either the product or other people (i.e. other customers or service personnel), information overload, and willingness to make the best choice increase the amount of uncertainty. On the other hand, a satisfactory experience through webrooming behaviour results in the choice confidence, which can be defined as the extent to which consumers feel that the cross-channel performance combination has allowed him/her to acquire the best choice (Bettman et al., 1998; Flavian et al., 2016). Therefore, by combining the two theories, one can conclude that by using online channel as an information gathering source a consumer’s uncertainty level is decreasing, which automatically increases the mental state of certainty, hence confidence. Nevertheless, multiple studies have recognised that humans have an inherent level of uncertainty in them, which can vary on aggregated, individual level (House R.J., 2004).

GLOBE project, which is built on multiple cross-cultural and organisational studies, has developed a scale to measure dimensions such as uncertainty avoidance, assertiveness, gender differentiation, performance orientation, humane orientation, in-group collectivism, institutional collectivism, power distance, and future orientation. All the dimensions are described both at the organisational and individual level. The GLOBE survey is unique due to the focus on the values and practices, which clarifies what are the internal values of people along the variables and what are the subsequent practices (House, 2004). Most of these values are implicit to humans and are established based on the background and society of one’s habitat. Even though the study is looking at the organisational and cultural values, the survey can be used to measure those variables at the individual level. Out of all the dimensions, uncertainty avoidance (UA) and future orientation (FO) seem to be the most related to the topic of webrooming as both of the dimensions are related to the perception of benefits and risks.

Information seeking is associated with uncertainty avoidance, as people tend to require feedback in uncertain situations, especially if feedback has high instrumental value (House, 2004). Therefore, when certain consumers have a higher uncertainty avoidance level, their perception of risks of online purchase and the level of distrust will be larger compared to the consumers with lower levels of uncertainty avoidance. Those people could also be referred to as risk-averse.

(9)

H6: The level of Uncertainty Avoidance will positively influence the perceived risks of online purchase.

H7: The level of Uncertainty Avoidance will positively affect the perceived benefits of cross channel searches.

Moreover, one additional factor is inherent to the characteristics of a human, which can predict whether a consumer will engage in the webrooming behaviour, namely future orientation. According to the GLOBE study (2004), future orientation can be defined as the level to which an individual exhibits future-oriented behaviour, such as investing in future planning. This behaviour is usually observed in people who strive to reduce the level of uncertainty and hence increase confidence. According to House (2004), uncertainty avoidance is positively correlated with Future Orientation. Aspirations for future orientation might mean that most uncertainties can be managed through planning and it can be reduced via better information and knowledge (House, 2004). Therefore, being future oriented might convey putting effort, time and energy during the search process in order to increase the probability of a successful outcome. However, the magnitude of the correlation is not clarified, which questions whether the relationship between two dimensions will interfere with the analysis, hence further multicollinearity testing is required. Webrooming strengthens control over the purchasing process and reduces information asymmetry, which allows consumers to exhibit a high degree of confidence when arriving at the final stage (i.e. purchase) (Flavian et al., 2016). Therefore, as an additional antecedent, the degree of future orientation will influence the perception of the benefits of cross channel shopping and thus affect the attitude towards webrooming.

H8: The degree of Future Orientation will have a positive and direct effect on the perceived benefits of cross channel search.

Figure 1 illustrates an overview of the research proposal. Considering the combination of various literatures, this research examines the effects of what is known as an important influencer of the attitude toward webrooming. Furthermore, as an additional part, which is limited in prior research, this paper analyses inherent human factors to determine whether the modern shoppers who have a positive

(10)

attitude toward webrooming share similar characteristics, which in turn predicts the degree of the perceptions.

Research Methodology

Prior research shows that there is a need for conducting an experiment at the disaggregated, individual level to better explain the relationship towards attitude and intention. Therefore, considering it, this research uses questionnaires to calculate and quantify the variables. The objective of the current analysis lies within the relationship between the two characteristics, namely uncertainty avoidance and future orientation, and attitude towards webrooming. The relationship is not directly observable, but instead goes through several additional latent variables, such as perceived benefits and risks of online shopping and that of cross-channel search.

The structure of this chapter is as following. Firstly, the design of the field and data collection will be discussed, which is followed by measures and finalised with an examination of the statistical procedure.

Field Research Design.

The first step of designing the field study is developing a questionnaire. The questionnaire gathers information about beliefs which are denoted as perceived benefits and risks in the conceptual model, and also about inherent values, which can predict certain beliefs and attitudes. Values underlie people’s attitudes, which later predict their behaviour (Schwartz, 2012). An individual will evaluate behaviour, belief or an attitude favourably in the case if they promote the achievement of valued goals, whereas in case of threatening the attainment of these goals, an individual will rate above mentioned variables negatively (Schwartz, 2012). Beliefs, on the other hand, can be defined as ideas about the extent of truthfulness of things related in particular way (Schwartz, 2012). For instance, a customer might have a belief that online shopping carries more risks than benefits, which can vary based on the level of certainty or confidence. One should not confuse values measured in this paper with traits, because people exhibiting a certain trait may not value the corresponding goal. For example, a customer might behave cautiously yet not value uncertainty avoidance. Therefore, traits illustrate what people are like instead of describing what they value as more important.

The latent variables described in the conceptual model cannot be directly observed, and hence the data was obtained in order to overcome this problem. Since the analysis involves 8 different variables, in order to find the relationship between them a qualitative research methodology was used.

Moreover, utilising a qualitative method allows researchers to acquire insights in a relatively quicker time span (Blumberg, Cooper & Schindler, 2014). Furthermore, the qualitative research methodology involves nonprobability sampling, more specifically this study uses convenience sampling, where any readily available individual is selected as a participant of a study. Due to the low budget and short span

(11)

of time the convenience sampling was the best option to choose for the current analysis (Blumberg, Cooper & Schindler, 2014). Consequently, the data was obtained from 115 respondents between age of 19 to 35. Moreover, to support diversity, the data was distributed among people from several countries.

This way, one can analyse whether the characteristics and relationship between measures are stable regardless of the background of an individual. As discussed above, due to the expansion of the communications technologies and global marketing strategies, the consumer behaviour and consumption pattern illustrate increasing global convergence (Mirty & Smith, 2009). Furthermore, based on the Global Online Consumer Report by KPMG (2017), millennials (i.e. people in age range of 18 to 37) are more likely to research a product by visiting a physical store and make the final choice of the channel based on attributes, such as price. This might be due to the lower disposable income or being relatively savvier when it comes to shopping in general (Kruh, 2017).

Measures.

This study incorporates several existing scales from previous research, and some of those are modified to match the current study. Two of the variables, namely uncertainty avoidance and Future Orientation, were developed and discussed in the Study of GLOBE (2004), which has based the creation of its dimensions on various earlier researches, such as Hofstede’s Cultural Dimensions Theory and Schwartz’s Theory of Values. Therefore, it incorporates different studies and each dimension of GLOBE reflects individual differences and then aggregates the results in order to present it on a societal level.

Therefore, even though GLOBE original questionnaire cannot be used for measuring individual level data, the dimensions can act as a basis for the creation of a new survey. Another reason for basing the new questionnaire on the GLOBE Study (2004), is due to the measurement technique, as this study incorporates both practices and values to find the accurate measures for predicting the behaviour.

Whereas the development of the abovementioned measures was based on the cross-cultural literature, the variables related to the perceived benefits, risks and distrust were developed in the context of interactive marketing and online shopping (Porto & Okada, 2018; Forsythe, Liu, Shannon & Gardner, 2006). This study utilises both first and second order constructs, while the former is a latent factor that has several observed variables called indicators, the latter construct consists of other facets (i.e. first- order), which also serve as indicators (Matsuno, Mentzer & Özsomer, 2002; Vinzi et al., 2010).

Furthermore, the current analysis distinguishes indicators relating to reflective and formative scales. In the reflective model, indicators represent a set of items that reflect the latent variable, thus differences in the value of the indicators are determined by the change in the value of the corresponding latent variable (Hair et al., 2011). Whereas in the formative model the opposite effect is illustrated; hence, discrepancies in the value of the latent variable are determined by the changes in the value of the indicators. Furthermore, formative indicators measure different facets of the latent variable, which explains the reason of formative indicators not being related to each other. On the other hand, in case of the reflective model, high correlation between indicators of the same latent variable is expected (Hair et

(12)

al., 2011; Berger, 2015; Garson, 2016). The measurement scales, their format, and sources are presented in appendix 1. In order to measure the indicators a Likert scale was utilised almost for the entire survey, ranging from “strongly agree” (1) to “strongly disagree” (7), with the exception for the fourth question measuring uncertainty avoidance (House, 2004), which was calibrated between “has a lot to be grateful for” (1) and “is missing out” (7).

To analyse the uncertainty avoidance and future orientation, one should look into not only values but also practices. Therefore, using GLOBE dimensions (2004) as a basis for the questionnaire will assist in establishing the degree of these variables at the individual level. Even though initially GLOBE scales were used to explain differences between societies, this research uses it as a basis for the creation of the survey to measure differences at the individual level. Thus, instead of explaining and describing societal culture, the new survey (appendix 1) focuses on finding implicit values and practices on two dimensions (i.e. uncertainty avoidance and future orientation). According to House et al. (2004), uncertainty avoidance has certain characteristics, namely fear of failure, less risk taking, high anxiety levels and more resistance to change. At the individual level of analysis, uncertainty avoidance describes the tendency to perceive ambiguous situations as risky and threatening (Hirsch et al., 2016). As customers are presented with vast amount of information online, they might exhibit values such as preference for clear instructions and requirements and lower tolerance to compromise. In the webrooming context, individuals engage in an online research of a product and approve the chosen object by visiting the store, which minimises potential risks and costs associated with the shopping and maximises confidence. Therefore, most people exhibit the abovementioned characteristics might have higher attitude towards webrooming.

Future Orientation dimension is important for measuring the values of a modern shopper, because people have a tendency to be oriented towards the past, the present and/or the future (House et al., 2004). Future orientation plays a role in developing adaptive capacities and skills by encouraging behaviours of learning and saving. Therefore, delay of gratification is positively correlated with future orientation (Davids & Falkof, 1975). Delay of gratification can be defined as the ability to avoid immediate reward for the sake of later greater reward (Nadler, 1975). In the shopping context, once a customer has established a need for a certain product, he/she might indeed delay the gratification in order to extensively search for information online. However, even if an individual will buy a product immediately through one channel, or use multichannel search, the end product of possession will not be a larger reward. Thus, in this context delay of gratification might suggest that customers involved in webrooming spend more time searching for the product to advance the confidence and reach more successful outcome. Hence, to these shoppers larger reward is greater confidence in the final outcome.

The general analysis focuses on determining the attitude towards webrooming, based on personal differences in the dimensions of uncertainty avoidance and future orientation. Furthermore, in case there is evidence of consistent results, one might imply that there are certain values inherent to the modern shopper, due to the vast amount of the contradicting information online and the need to achieve

(13)

a successful outcome. The basis for creation of the questionnaire to measure Future Orientation and Uncertainty was taken from the existing GLOBE study (House et al., 2004). Uncertainty Avoidance (UA) originally has nine items, four related to practices and five to values (House et al., 2004). The current analysis uses only five items for measuring UA, which could be interpreted at the individual level. Furthermore, Future Orientation (FO) also consists of nine items (i.e. five questions related to practices and four to values), and this research has incorporated only four. The choice was mainly based on the interpretation of questions at the individual level, compared to analysis of the societal values and practices.

The measure of the perceived benefits of online search, which is part of online shopping, has been used widely in the previous literature. Here, the analysis draws attention to the non-store patronage behaviour. According to Forsythe et al. (2006), there are three dimensions related to the perceived benefits, namely convenience, product selection and ease of shopping. Convenience can be defined as the ability to access an online store at any time from almost any location (Forsythe et al., 2006). On the other hand, ease of utilising online channels is linked to the avoidance of emotional and physical hassle of visiting store. Product selection is related to the variety and quantity of the products available online, which surpasses the brick-and-mortar’s portfolio. Moreover, the internet provides a vast amount of product information to support the decision-making process of a customer (Forsythe et al., 2006).

Consequently, based on the previous research by Forsythe et al. (2006), the six-item construct was established to measure different facets of the perceived benefits of online search (i.e. formative scale).

Perceived risks of the online purchase can be defined as the subjective perception of possible losses from online shopping (Forsythe et al., 2006; Arora & Sahney, 2017). This measure can also be divided into three subparts: financial risks; product related risks and time/convenience risks (Forsythe et al., 2006). Compared to the former two types of risks, the latter describes the cost of time searching for an appropriate website and information. However, as discussed above most of the customers find it less time costly to use the internet channel, compared to other alternatives. Therefore, in order to avoid theoretical and statistical correlation this research uses only two types of risks, namely financial and product related. The four-item scale was created to measure perceived risks of the online purchase, two of which are related to the financial risks and two related to the product related risks, which is based on the research by Forsythe et al. (2006). The financial risks encompass the information a customer needs to provide in order to proceed through the checkout, which involves bank card and personal information.

Whereas, product related risks would include the quality and delivery.

Previous research has inserted the measure of the cross-channel behaviour into the webrooming literature as it encompasses the motives of, for example, searching product information online before carrying out the purchase through the offline channel (Gerritsen et al., 2014; Zang, 2012). Two pre- current behaviours can be considered as parts of the cross-channel behaviour, namely the simultaneous search for product information, and product and price comparison (Porto & Okada, 2018). Searching information via various channels is considered to be an important consumer experience to improve the

(14)

outcome of purchase (Shin, 2009). Even though an individual capacity restrains the ability to efficiently compare all the possible price and product options, the capability of being able to view various products through various web sites enhances customer’s confidence in choosing ‘the best option’ and thus improves customer satisfaction. The survey of the current analysis utilises three items related to simultaneous information search, price comparison and product comparison. Based on the research by Porto and Okada (2018), the three-item scale was created to measure both product/price comparison and simultaneous information search facets of the construct.

Even though the previous literature discusses and incorporates the online distrust into the analysis (Arora & Sahney, 2018), appropriate scales for the dimension were not available in the scope of the research and hence were generated based on the theory and previous measures. Distrust of online shopping is usually described as the consumer’s tendency of inhibitions towards online purchasing, which can be based on the beliefs about potential risks and benefits (Kim & Ahmad, 2013; Moody et al., 2014). Therefore, based on the definition and description of distrust the three items were created to measure the extent of the reliability of online shopping, and since all three of the items measure the same part of the construct this variable is considered to be reflective.

The last two variables of the conceptual model measure the attitude and intention towards webrooming. Attitude refers to an overall evaluation of webrooming, which can vary on favourability (i.e. positive/favourable or negative/unfavourable) (Ajzen, 1991). According to Huang (2005), the attitude has a utilitarian aspect, which measures the level of safety, riskiness, usefulness and reliability.

Thus, this research encompasses three items to measure attitude on the basis of perception of its riskiness, reliability and effectiveness. Even though based on the literature attitude can be divided into subparts, all of them are related to the utilitarian part of shopping and thus are viewed as a reflective model. Whereas attitudes measure evaluations, intentions represent the conscious plan of an individual to exert effort to perform the behaviour (Eagly & Chaiken, 1993), which looks into the certainty of the motivation and interest of carrying out the behaviour of webrooming. Hence, as intention is not the focus of the current research, based on the research by Spears and Singh (2004) two items were utilised with slightly differing wording, measuring the intention towards webrooming, which makes this construct reflective.

To summarise, there are eight variables, some of which take a role of both dependent and independent variable. For instance, in case of the sixth hypothesis uncertainty avoidance is an independent variable, which predicts the value of the Perceived Benefits of Online Search (i.e.

dependent variable), whereas in the third hypothesis Perceived Benefits of Online Search becomes an independent variable that supposedly affects Attitude towards Webrooming (i.e. dependent variable).

Variables of Perceived Benefits of Online Search and Perceived Risks of Online Purchase take formative form, while others are represented in the reflective scale model.

(15)

Statistical Procedure.

The survey will be collected based on the Likert scale, which quantifies the data in a range from 1 to 7, and thus eases the process of the analysis. This analysis includes path model of several variables taking form of both dependent and independent measures, with both formative and reflective constructs.

Consequently, the gathered data will be analysed with Partial Least Squares, a variance-based Structural Equation Model (i.e. PLS-SEM) to find the relationship between given variables. Since indicators are measuring different latent variables in this analysis, PLS-SEM is an appropriate technique to use because it combines both factor analysis and multiple regression analysis (Berger, 2015; Hair, Ringle & Sarstedt, 2011). According to Chin (1998), the minimum sample size for the analysis in the PLS-SEM should be based on the following two conditions: either equal to the number of indicators of the largest formative construct times ten, or ten times the number of variables in the longest structural path to a dependent variable. In this research the sample size of 130 respondents exceeds the suggested size.

Results

PLS-SEM path model’s analysis formally includes two set of linear equations, the measurement model, also known as outer model, and the structural model, also called the inner model (Hair et al., 2010; Berger, 2015). The measurement model estimates the relations between each latent variable and its observed indicators, whereas the structural model specifies the relationships between constructs.

SmartPLS software was utilised in the current analysis and PLS results were assessed after 16 iterations.

Measurement Model.

The measurement model should meet minimum requirements of reliability and validity before proceeding to the structural model. These requirements can be assessed by its convergent and discriminant validity. Whereas convergent validity measures whether factor is unidimensional, discriminant validity examines whether each latent variable statistically represent theoretically different concepts (Hair et al., 2010).

Table 2.1: Convergent validity and construct reliability of reflective scales.

Constructs: Cronbach’s alpha: Composite reliability:

Uncertainty Avoidance (UA) 0.775 0.817

Future Orientation (FO) 0.738 0.827

Online Distrust (OD) 0.815 0.890

Attitude (ATT) 0.417 0.720

Intention (INT) 1 1

Measures such as composite reliability and Cronbach’s alpha are appropriate tests for the convergent validity only in reflective models, since the formative scales are not expected to relate to

(16)

each other (Hair et al., 2011). Based on the results of the Cronbach’s alpha, all indicators except for ATT illustrate high reliability, as value is greater than 0.70. On the other hand, composite reliability includes factor loadings of each indicator and is preferred to Cronbach’s alpha as a test of convergent validity in the PLS-based analysis (Götz, Liehr-Gobbers & Krafft, 2010; Berger, 2015; Henseler et al., 2009). All the coefficients of composite reliability are greater than 0.7, which illustrates acceptable convergent validity. For the formative scales, convergent validity can be assessed by the significance (i.e. p-values) and magnitude of outer weights and the variance inflation factors of those weights. The VIF coefficient values should be low because formative construct indicators are expected to estimate different facets of the latent variable. Consequently, Garson (2016) illustrates that in the case of an indicator having a theoretically significant dimension and high loading, then it should be retained regardless of a non-significant weight. To assess outer weights’ significance bootstrapping technique was applied.

Table 2.2: Convergent validity of formative indicators.

Indicators: Indicator

weights:

p-value: VIF:

Perceived Benefits of the Online Search (PBOS)

PBOS1 0.691 0.000 1.439

PBOS2 -0.109 0.585 1.978

PBOS3 0.461 0.019 2.025

PBOS4 0.120 0.478 1.410

PBOS5 -0.126 0.342 1.228

PBOS6 0.119 0.464 1.117

Perceived Risks of the Online Purchase (PROP)

PROP1 0.085 0.626 1.419

PROP2 0.519 0.000 1.695

PROP3 -0.134 0.333 1.073

PROP4 0.605 0.003 1.414

Perceived Benefits of the Cross-Channel Search (PBCCS)

PBCCS1 0.144 0.564 1.220

PBCCS2 0.392 0.096 1.297

PBCCS3 0.712 0.003 1.208

Based on the table 2.2, only the p-values of five indicators (i.e. PBOS1; PBOS3; PROP2; PROP4 and PBCCS3) are lower than the threshold value of 0.05. After a thorough analysis of those indicators with high p-value it was decided to retain the most of the indicators, except for PBOS5 and PROP3 since their outer loadings also show insignificance. The PBOS5 item is created to measure the product

(17)

selection facet of the construct, same as PBOS1 and PBOS6. Specifically, according to Forsythe et al.

(2006), one of the dimensions of the product selection is named as following – “can get good product information online”, which measures the favourability and perceived usefulness of online information search. However, the literature does not necessarily discuss what good product information mean, thus leading to complications in the interpretation of the scale item. Consequently, the structure of the question might lead to contradicting subjective interpretations, hence explaining insignificant results.

Therefore, deleting the item will not lead to any complications in the theoretical framework.

Complaisant with the PBOS5, PROP3 can be removed because same as PROP4 it measures the product related risk of the online purchase. Therefore, removing PROP3 will not change the meaning of the factor significantly. Additionally, PROP3 is related to the delivery of the product, and due to the rapid globalisation and technological innovations the delivery is considered to be less problematic and risky amongst millennials (Kruh, 2017). Furthermore, the reason for retaining other scales is due to the fact that they are theoretically significant for the latent variable and without them the factor might not yield meaningful information. Despite the non-significance within the indicators, the VIF coefficient is lower than threshold value of 3.3, which illustrates that indicators measure different facets of the same latent variable. Therefore, even though indicators are not significant they adequately capture the theoretical dimensions.

Discriminant Validity utilises average variance extracted (AVE) scores by the Fornell-Lacker criterion. For each latent variable, the square roots of AVE should be greater than correlation coefficients of those LVs with other constructs. Table 2.3 illustrates square roots of AVE scores in a diagonal, and correlation coefficients below it.

Table 2.3: Discriminant Validity

ATT FO INT OD PBCCS PBOS PROP UA

ATT 0.684

FO 0.298 0.791

INT 0.345 0.286 1.000

OD 0.482 -0.035 -0.001 0.854

PBCCS 0.371 0.269 0.259 0.138 0.752

PBOS 0.123 0.579 0.483 -0.223 0.301 0.653

PROP 0.266 0.049 -0.015 0.617 0.140 -0.117 0.692

UA 0.405 0.712 0.374 0.011 0.324 0.572 -0.043 0.628

As can be seen from the table above, the square root of each construct’s average variance extracted is larger than any other correlation coefficient, in absolute value term. Thus, there is a discriminant validity (Fornell & Larcker, 1981).

(18)

Based on the tests performed and reported above, the measurement model of this analysis is deemed of sufficient quality to proceed with the structural model.

Structural Model.

The structural or inner model consists of the relationship between latent variables. Table 2.4 presents the structural path coefficients, which can be defined as weights connecting latent constructs representing hypothesized direct effects.

Table 2.4: Standardised path coefficients.

Independent variable à

Dependent variable Beta: Sample

Mean: SD: T statistic: p-value:

UA à PROP -0.043 -0.048 0.146 0.292 0.769

UA à PBOS 0.572 0.596 0.050 11.376 <0.001***

PROP à OD 0.599 0.610 0.059 10.097 <0.001***

PBOS à OD -0.153 -0.141 0.085 1.799 0.075*

OD à ATT 0.439 0.440 0.084 5.241 <0.001***

FO à PBCCS 0.269 0.301 0.096 2.804 0.005***

PBCCS à ATT 0.311 0.325 0.084 3.696 <0.001***

ATT à INT 0.345 0.343 0.082 4.197 <0.001***

Note: ***! ≤ 0.01;∗∗ ! ≤ 0.05;∗ ! ≤ 0.10.

Since the data in the analysis is standardised, path coefficients take value in a range of 0 to 1.

The significance of the loadings is measured by using bootstrapping technique. Every direct relationship has significance, except for the effect of uncertainty avoidance on perceived risks of the online purchase.

Due to the non-significance, the model can call for respecifying without that specific relationship.

However, after deleting the path from UA to PROP, the path coefficients and significance levels of other relationships do not change. Consequently, as UA is a crucial part of the theoretical framework and its absence does not lead to any further significant changes in the results, the decision was to retain the factor in the path model.

The goodness of fit of the structural model can be checked by utilising the explained variance (i.e. R-squared and Adjusted R-squared), multicollinearity test and f-square effect. Tables 2.5 and appendix 2 depicts the collinearity VIFs scores together with f-square values and the explained variance score, respectively.

The explained variance or the coefficient of determination can be applied only to the endogenous variables, which explains the absence of two variables, namely FO and UA. R-squared is a statistical measure depicting the percentage of the variance for a dependent variable explained by the inputs or independent variable(s) in a model (Vinzi et al., 2010), whereas adjusted R-squared takes into consideration the number of predictors in the model (Harel, 2009; Garson, 2016). All the R-squared

(19)

Table 2.5: Collinearity statistics (VIFs) and f-square effect size.

f-square:

VIFs: coefficient: p-value:

UA à PROP 1.000 0.002 0.957

UA à PBOS 1.000 0.485 0.001***

PROP à OD 1.014 0.593 0.002***

PBOS à OD 1.014 0.039 0.369

OD à ATT 1.019 0.280 0.057*

FO à PBCCS 1.000 0.078 0.214

PBCCS à ATT 1.019 0.141 0.144

ATT à INT 1.000 0.135 0.076*

Note: ***! ≤ 0.01;∗∗ ! ≤ 0.05;∗ ! ≤ 0.10.

values for each endogenous construct in a model can be viewed in the appendix 2. Latent constructs of Attitude, Online Distrust and Perceived Benefits of the Online Search show moderate variation of approximately 33 - 40%, whereas Intention, Perceived Benefits of the Cross-Channel Search and Perceived Risks of the Online Purchase illustrates weak results – only 12%; 7.2% and 0.2%, respectively. Based on the adjusted coefficients of determination, the strongest equation is that of Online Distrust, where 39.4% of the variation is explained by the model.

The VIF scores, depicted in Table 2.5, are below the threshold of 3.3, which implies that there are no collinearity issues in the structural model between latent constructs. Furthermore, the effect size (i.e. f-square) indicates the influence of the latent exogenous factor on the latent endogenous one. Values of 0.35, 0.15 and 0.02 demonstrate the high, medium and small effect sizes, respectively (Cohen, 1988;

Chin, 1998b). Consequently, as shown in Table 2.5 out of four significant results, two have high effect size (i.e. UA à PBOS; PROP à OD), one has medium and the last relation has small effect size (i.e.

OD à ATT and ATT à INT, respectively).

To test the hypotheses, we have to take the estimated direct effects (i.e. path coefficients) into consideration. Table 2.6 demonstrates the coefficients of every single relationship in the path model.

Table 2.6: Path coefficients

Variables: Beta: Hypothesis: Effect:

1. ATT à INT 0.345*** (0.083) H1 (pos. effect) sup.

2. PBCCS à ATT 0.311*** (0.084) H2 (pos. effect) sup.

3. PBOS à OD -0.153* (0.086) H3 (pos. effect) sup.

4. PROP à OD 0.599*** (0.058) H4 (neg. effect) sup.

5. OD à ATT 0.439*** (0.085) H5 (pos. effect) sup.

6. UA à PROP -0.043 (0.145) H6 (pos. effect) rej.

(20)

Continued: Path coefficients

7. UA à PBOS 0.572*** (0.050) H7 (pos. effect) sup.

8. FO à PBCCS 0.269*** (0.095) H8 (pos. effect) sup.

Note: ***! ≤ 0.01;∗∗ ! ≤ 0.05;∗ ! ≤ 0.10. Values in the parentheses are standard deviations.

Hypothesis 1 predicted that in case of a positive attitude towards webrooming, the intention would also be positive among the respondents. The effect of ATT on INT is positive and significant at 1% significance level (p<0.01) with magnitude of 0.345. Thus, H1 is supported. The direct effect of PBCCS on ATT is positive and significant at 1% significance level (beta=0.311; p<0.01), which supports the hypothesis 2. Furthermore, hypothesis 3 and 4 predicted that while perceived benefits of the online search would decrease the distrust, the risks of the online purchase would have an opposite effect. In total both of the variable (i.e. PROP and PBOS) will determine the level of the distrust an individual has. Both of the hypotheses 3 and 4 were supported at 10% and 1% significance level, respectively. As the Distrust level can change according to the extent of the perceived benefits and risks, the trust issues with the Internet were predicted to have a positive effect on the attitude towards webrooming, since the definition of webrooming postulates that individuals search online but keep the purchase decision to the physical store. Following that, the hypothesis 5 was supported at 1%

significance level (beta=0.439; p<0.01).

The last three hypothesis are related to the individual characteristics of individual who establish the attitude towards webrooming, whether negative or positive. The relationship between uncertainty avoidance and perceived benefits of the online search is positive and significant at 1% significance level (beta=0.572; p<0.01), supporting H7. However, based on the results of the analysis, uncertainty avoidance does not have an effect on the perceived risks of the online search, rejecting H6. The final hypothesis took into account the future orientation of an individual and its influence on the general level of perceived benefits of engaging in cross-channel search, which is a crucial part of webrooming.

Consequently, the effect of future orientation on perceived benefits of cross-channel search is positive and significant at 1% significance level with magnitude of 0.269.

The discussion chapter of this research will further interpret and discuss the findings and reason the observed effects.

Discussion and Conclusion

The 5th chapter of this research is organised as follows. Firstly, the results of the analysis are highlighted and discussed, reflecting the findings from Table 2.6. Then, the research question is revised and aims of the study are discussed. Finally, the chapter is concluded with limitations and suggestions for the further research.

(21)

As expected the level uncertainty avoidance predicts the increased level of perceived benefits of information search online. According to Table 2.6, if the level of uncertainty avoidance is increased by one unit, the perceived benefits of the online search will increase by 0.572. Individuals with high levels of uncertainty avoidance are also known to have a characteristic of information seeking (House, 2004). Therefore, in order to increase the confidence of the final choice of a product, an individual will highly rely on information gathered online. Moreover, as was observed from the previous literature, perceived benefits of the online channel include convenience, large and varied product selection and comfort. Therefore, the cost of product search and selection has been decreased with the introduction of an online channel (Forsythe et al., 2006), which implies that uncertain people will be more eager to look up the information online to select the most satisfying option possible.

Nevertheless, the effect of uncertainty avoidance on the perceived risks of online purchase were not significant. Even though it might seem indisputable that the level of uncertainty will increase the level of perceived risks (House, 2004), the current research has not considered that the majority of millennials (i.e. respondents) have Internet experience (Kruh, 2017). Consumer’s Internet experience increases the likelihood of online purchasing (Bhatnagar et al., 2000; Kim et al., 2000). Furthermore, the analysis by George (2002) shows that the frequency of the online purchases increases with the level of Internet experience. Therefore, as the experience of a customer increases, the level of perceived risks might decrease, since individuals are aware of the potential risks and the ways of avoiding them.

Following the uncertainty avoidance, the second chosen characteristic for the research is future orientation. Future orientation has a positive and direct effect on perceived benefits of the cross-channel search with a magnitude of 0.269. Online shopping involves vast amounts of information, which includes abundant product selection with different quality and price criteria. Thus, online shopping discourages unplanned purchases (Frambach et al., 2007) and instead pushes customers to define the goal in terms of the product selection (Chatterjee, 2009). Most of the uncertainties can be reduced by planning, which is crucial characteristic of a future oriented person (House, 2004). Consequently, the future orientation characteristic of a customer, will encourage him/her to engage in cross-channel search.

After an individual forms certain beliefs about an online channel, the issue of trust on online shopping is examined. According to Luarn and Lin (2005), distrust is one of the highest barriers stopping consumers from online purchasing. The effect of perceived risks on distrust can be clarified by reviewing the definition. Distrust can be defined as “confident negative expectation regarding another’s conduct”

(Lewicki et al., 1998), hence in case of negative beliefs outweigh positive ones the distrust is enforced.

Nevertheless, the perceived benefits were predicted to reduce the distrust due to the increasing confidence. The relationship is supported by the analysis, despite the fact that it is significant only at 10% significance level. Even though the online information search might increase the confidence with the choice, the potential risks create feelings of fear and worry as consumers do not have an opportunity to inspect a product of choice (McKnight et al., 2004; Arora & Sahney, 2018). Hence, people’s perception of risks has a stronger effect on the e-distrust, compared to benefits of the search and since

(22)

webrooming implies the avoidance of the online purchase e-distrust was predicted to positively enhance the attitude towards webrooming. This hypothesis is fully supported and when the distrust in online channel increases the attitude towards webrooming will increase with a magnitude of 0.439. The finding of a strong and positive connection between e-distrust and attitude is in line with several previous researchers (McKnight et al., 2004; Moody et al., 2014; McNeish, 2015; Arora & Sahney, 2018).

As noted earlier, consumers engage in cross-channel searches due to the great amount of provided information and selection. Another reason for utilising a cross-channel search is time convenience. Ordering a product online requires delivery and in the case of dissatisfaction it might require a return, whereas visiting store gives an advantage to inspect a product without even purchasing it (Chatterjee, 2010). Among the listed benefits of cross-channel shopping are experience and socialisation (Mitchell, 1999). Consequently, as predicted an individual will build a positive attitude towards webrooming if he/she believes that the use of cross-channel shopping is beneficial.

Finally, based on the research, existing literature and field study, due to the overload of information and product selection people are trying to make the best choice with high confidence and minimum effort. The information overload and limited capacity causes people to question all information provided in the Internet, while the strive to achieve a goal urges an individual to become more future oriented. As a consequence, these two characteristics are becoming crucial to the majority of modern shoppers. They might not have a direct effect on webrooming, but instead they develop beliefs about potential benefits and risks, which in turn creates a need to utilise webrooming technique.

Scientific and Managerial Relevance.

There are several implications of this study. The following discussion firstly describes scientific relevance, which is followed by the managerial implications.

First, compared to other researches related to webrooming, this conceptual model includes antecedents for the perceived benefits and risks. Therefore, by using specific characteristics, such as uncertainty avoidance (UA) and future orientation (FO) this study expands the research by exploring new variables affecting the attitude towards webrooming. The findings illustrate that these two variables (i.e. UA and FO) indeed have a significant effect on perceptions, beliefs and values. Moreover, this study utilises GLOBE measures for individual level measurement, justifying that the survey can be adjusted and used to measure uncertainty avoidance and future orientation on the individual level.

Secondly, the variable of an Online Distrust acts as a stabiliser of perceived benefits against perceived risks. Based on the results of this study and previous literature, e-distrust plays a crucial role in predicting the attitude towards webrooming. Nevertheless, the variable itself can increase or decrease based on the extent and amount of beliefs about benefits and risks of online shopping. Despite living in the century of Internet, the online distrust still exists amongst modern shoppers, which explains the reason for individuals preferring to utilise webrooming.

(23)

Thirdly, besides exploring additional antecedents for the conceptual model, the current study also can be viewed as confirmatory to a certain extent. The relationship between attitude, perceptions on benefits and risks of facets of online shopping and online trust level has been discussed in the previous literature and has been confirmed through this study.

Finally, the study determined that the amount of product and complementary information is rapidly increasing, while the cognitive capacity of an individual cannot process each and every piece of information. Therefore, it becomes more difficult for consumers to make a confident choice, which pushes them to utilise every possible channel and acquire the best possible option of the needed product.

Thus, instead of focusing on the survival of one specific channel, along with implementation of the omni-channel strategy companies should focus on delivering transparency and convenience via each and every channel, in order to increase trust. Increasing trust will assist in decreasing the level of perceived risks among individuals scoring high on uncertainty avoidance. There are several suggestions regarding how companies can reduce the level of perceived risks. For instance, by facilitating and easing the process of contact with employees of the firm, and/or allowing consumers to control how the personal data is used. Moreover, firms should focus on social responsibility and try to be transparent and honest with ongoing firm issues.

Limitations and Future Research.

Although this study provides significant insights on webrooming attitude, there are several limitations that need to be addressed, which provides a scope for the future research.

Firstly, this study empirically investigated how individual characteristics can indirectly influence attitudes and intentions towards webrooming. However, it does not involve a time-series analysis, which might omit to capture the long-term relations and expose results to the risk of reflecting short-term effects. Thus, future research may analyse more modern shopper characteristics, and perform a longitudinal study, where firstly the most salient shopper characteristics are discovered, followed by an analysis of their relationship with webrooming. A broader perspective is needed to understand how human beliefs and characteristics shape the attitude towards certain shopping channel.

Secondly, even though the majority of researchers have already incorporated the online distrust in the research of webrooming (Moody et al., 2014; McNeish, 2015; Arora & Sahney, 2017), there is a lack of understanding as to what shapes online distrust at the individual level. This research looks at the balance effect of benefits and risks related to online shopping which in turn determines the value of online distrust. Therefore, it is suggested that further research thoroughly analyses the variable of online distrust, in order to find the antecedents.

Thirdly, the advent of mobile technology in the shopping context is expected to substantially influence the webrooming intention (Kim & Hahn, 2015), but it has not been considered in this research and hence can be stressed by future research.

(24)

Furthermore, brands are an important factor in purchase decision making process for certain people, but not all (Park & John, 2010). Certain people buy products to enhance and/or signal the self (Park & John, 2010). Nevertheless, certain customers might have preference for several brands and the abundance of different complementary and substitutional brands with thousands of stimuli make it more complicated for a customer to make a well-informed choice. Therefore, during the research it was discovered that in order to make a well-informed decision, individuals participated in webrooming, showrooming and, more interestingly, sometimes they engaged in omni-channel shopping. Omni- channel shopping might imply that a customer goes back and forth from offline to online channel and constantly engages in a search of the best option possible. For instance, a consumer might gather information online, then visit offline channel due to the need to inspect the physical aspect of a product and lastly return to the online channel to compare the prices with physical store. After comparing and contrasting all the gathered information the customer makes a decision based on the prioritised criteria (e.g. price, quality or instant gratification). Therefore, further research should include this omni perspective into the research of webrooming and showrooming.

The References

Ajzen, I. (1991). The theory of planned behaviour. Organisational Behavior and Human Decision Processes, 50(2), 179-211.

Ajzen, I. & Fishbein, M. (1980). Understanding Attitudes and Predicting Social Behaviour, Prentice Hall, Englewood Cliffs, NJ.

Arora, S. & Sahney, S. (2017). Webrooming behaviour: a conceptual framework. International Journal of Retail and Distribution Management, 45(⅞), 762-781.

Arora, S. & Sahney, S. (2018). Consumer’s Webrooming conduct: an explanation using the theory of planned behaviour. Asia Pacific Journal of Marketing and Logistics, 30(4), 1040-1063.

Balasubramanian, S., Raghunathan, R. and Mahajan, V. (2005). Consumer in a multichannel

environment: product utility, and channel choice. Journal of Interactive Marketing, 19(2), 12- 30.

Bettman, J.R, Luce, M.F., & Payne, J.W. (1998). Constructive consumer choice processes. Journal of Consumer Research, 25(3), 187-217.

Berger, J. (2015). Essays on the governance of buyer-supplier relationships. Groningen: University of Groningen, SOM research school.

Bhatnagar, A., Misra, S. & Rao, H.R. (2000). On risk, convenience and Internet shopping behaviour. Communications of the ACM, 43(11), 98-105.

Blumberg, B., Cooper, D., & Schindler, P. (2014). Business research methods (Fourth ed.). Berkshire:

McGraw-Hill Education.

Referenties

GERELATEERDE DOCUMENTEN

To regularize the sparse attraction-forces calculated from the image data, the deformation is described using a B-spline grid.. Edge attraction registration algorithms usually

(2003) Problem-based learning Critical thinking; Communication; Leadership Helped in developing critical and creative thinking, decision-making, communication and leadership

(2b) Verondersteld wordt dat de mate van symptomen op de somatische depressiedimensie het laagste zal zijn voor de veilige hechtingsstijl, hoger voor de

Naast dat het executief functioneren als geheel meegenomen wordt in dit onderzoek, wordt er ook gekeken naar verschillende executieve vaardigheden en hoe deze in verband

The objective of the company project was to study the applicability of the developed tools and methods during the ENABLE S3 initiative for the Verification and Validation

In this paper, our main contribution is that we present combinations of measurements for error modeling that can be used to estimate the quality of arbitrary GNSS receivers

On the other hand, poly- 2 reduced the fire growth rate more strongly than poly-1 in EP: this resulted from the better thermal barrier properties of EP-poly-2 ’s char, which

Aspects examined include: the type of grants provided by government; the nominal value of grants; the number of beneficiaries receiving assistants from grants;