Do multichannel shoppers buy and spend more? : the effect of product category, demographic and psychographic characteristics

(1)

University of Amsterdam

Do multichannel shoppers buy and spend more?

The effect of product category, demographic

and psychographic characteristics.

Laure van Ravensberg (10431411) Supervisor: dr. Umut Konuş

(2)

Statement of Originality

This document is written by Student Laure van Ravensberg 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.

(3)

Acknowledgement

This thesis is my last and final step towards achieving my master’s degree in Business

Administration and specialization in Marketing at the University of Amsterdam. I have had a great affinity with multichannel shopping from the moment dr. Umut Konuş gave a lecture about this subject. He is a true expert in his field and therefore I was very happy to find out that he was going to be my supervisor.

I would like to give thank my supervisor dr. Umut Konuş for supporting me in writing my thesis in every possible way he could! It has been a great pleasure to work with him. I was very inspired by his great expertise in digital marketing and he was always able to transfer his enthusiasm to me. This made writing a master thesis very pleasurable.

Furthermore, I would also like to thank my family and friends for supporting me and having faith in me. I would in particular like to thank dr. Jorik J. Reimerink for giving me feedback and great support all the time!

Hopefully, you will enjoy reading this master thesis and it will gain your interest in multichannel shopping.

Kind regards,

Laure van Ravensberg 24th_{of June 2016}

(4)

Abstract

The shopping environment has drastically changed since the emergence of internet and mobile smartphones. Previous research took multichannel shoppers as a whole and did not make a distinction between several multichannel shoppers. Multichannel shopping includes the information search and purchasing phase. This research will assess the difference in customer expenditures, i.e. in purchase frequency and euros, of multichannel versus single channel shoppers in offline, online and mobile channels. The possible moderating effect of product category, psychographic and demographic characteristics will be determined. An online survey will be conducted. Marketers can benefit from this research by making decisions for their channel strategy based on the results of this research. They can also optimize a certain channel in order to meet the customer preferences in terms of service orientation, variety-seeking and convenience orientation.

(5)

1. Introduction

The shopping channel environment has changed a lot over the past decades. With the emergence of internet and mobile phones, new possibilities for marketers to respond and get in touch with their customers arose with it. The approach of marketers to connect with their customers through online and mobile channels differs a lot from the approach that comes along with offline stores, i.e. brick-and-mortar stores. Over the past years, the popularity of online shopping (on a PC or laptop) has kept growing (NU.nl, 2015). The mobile shopping channel (on a smartphone or tablet) also keeps gaining popularity amongst shoppers (NU.nl, 2014). Stores who started off as a brick-and-mortar – i.e. offline – store, are now increasingly selling their products online as well. Some main warehouses seem to be following this trend, e.g. HEMA and the Bijenkorf. Both Bijenkorf and HEMA seem to be getting less and less customers in their brick-and-mortar stores, while their customers from the online shops are growing rapidly (Keuning, 2015).

A channel is defined as a customer contact point or a medium through which the firm and customer interact (Neslin, et al., 2006). This only includes two-way communications – such as mobile channels - and excludes one-way communications – such as television without further calls for interaction. A new feature that comes along with online and mobile channels is that people can decide when, where, what and how they want to shop. They can, for example, look at items without an employee present. However, this can result in both a positive and negative influence depending on the customer’s preferences. Also, an employee has the ability to persuade a customer into buying a product, which inherently can increase sales. Then again, it must be taken into account that employees do come with costs to hire them.

(7)

University of Amsterdam 2 Mobile channels have the advantage of being more flexible due to their availability - i.e. wireless network, easily portable - while online channels have a better overview of products due to their bigger screen. Depending on their preference, some people are really devoted to shop at a single channel while others prefer shopping using multiple channels, i.e. multichannel shopping. Shopping includes three different phases: pre-purchase, i.e. searching for information about the product or service on the internet, purchase, i.e. actually buying the product or service, and post-purchasing activities, i.e. the after-sales services. Multichannel shopping was defined bij Neslin et al. (2006) as shopping across two or more channels in the shopping process, including pre-purchase (i.e. information search), purchase and post-purchasing activities.

The studies on whether multichannel shoppers actually do spend more, have contradictory outcomes (Deloitte, 2014; Kushwaha & Shankar, 2013; Bilgicer, Jedidi, Lehmann, & Neslin, 2015). Even though there is some evidence that multichannel shoppers spend more, various measures for monetary value have been used. Therefore, it is hard to compare one to another. This research will determine expenditures in terms of purchase amount, i.e. how many customers spend in euros (EUR), as well as in purchase frequency, i.e. how many times they purchase a product. When talking about whether multichannel shoppers spend more, we should wonder whether this includes all three types of multichannel shoppers. Multichannel shoppers are customers that 1) search for information on several channel and purchase the product on one channel; 2) search for information on one channel and purchase the product on several channels or 3) search for information on several channels and purchase the product on several channels.

Little research has been done regarding different types of expenditures, thus more insights are needed. Some studies used the purchase- or sales volume as a measure to

(8)

University of Amsterdam 3 decide wether a company should add another channel or not (Inman, Shankar, & Ferraro, 2004; Ansari, Mela, & Neslin, 2008), while others look at the profit (Neslin, et al., 2006; Cheng, Tsao, Tsai, & Tu, 2007; Venkatesan, Kumar, & Ravishanker, 2007), sales (Deleersnyder, Geyskens, Gielens, & Dekimpe, 2002) or sales revenue (Avery, Steenburg, Deighton, & Caravella, 2012).

A commonly used measure to determine monetary value is purchase volume. But does volume really correlate with the actual expenditures of customers? The purchase- or sales volume can mean many different things. We should question if a firm can decide on the basis of purchase- or sales volume whether they should add another channel or not. An increase in volume can indicate that people shop more frequent, but not necessarily in higher quantities. What if people tend to buy cheaper products in larger volumes online, because it is heavy and now they do not have the disadvantage of lifting it to their home, but it gets homedelivered instead? Or if people only buy the cheap light products in large quantities in the store, because that is just more convenient since they do not have to wait for the delivery man to deliver the package and they can carry it themselves.

Profit is another oftenly used measure to determine monetary value. But can we fully assign single or multichannel usage to the overall profitability of a company? Therefore, this research can add value by looking at the actual expenditures of customers. If the spendings of the single channel shoppers versus multichannel shoppers are mapped, we can assess whether there is an actual difference between these two and whether a firm should add another channel or not. In our opinion, both profitability and purchase-/sales volume are not sufficiently good measures to determine whether this argument is generalizable to all multichannelshoppers. This study aims to give better insights at the different channels and the connection it has with the usage and expenditures of customers in order to adapt

(9)

University of Amsterdam 4 their channel strategy to it. There might be a ‘golden’ combination of certain channels that will work better than others.

If we assume that multichannel shoppers do spend more, there could be variables influencing this relationship? For example, could the relationship depend on product category, psychographic, or demographic characteristics? When multichannel shoppers search for information on multiple channels, but purchase a product on a single channel, is it because they are loyal to a specific channel for the purchase phase? Or do they prefer to touch the product or gain information by talking to an employee prior to purchasing it? And when multichannel shoppers search for information on one channel, but buy on several, does this indicate that they purchase the product on the channel that sells it the cheapest? Or do they find it more convenient to search online and buy the product on the channel that is most convenient for them, i.e. they will go to this store anyways or ordering online due to their busy schedule. The articles that did study similar relationships were mostly done using existing databases. Therefore, the generalizability of the outcomes were limited. To answer these questions, more research should be done to see if different factors are influencing the expenditures of multichannel shoppers.

In sum, this study tries to answer if multichannel shoppers, i.e. information search and/or purchase, actually spend more, while taking into account purchase amount and frequency. Also, it seeks to determine if product categories, demographic and psychographic characteristics moderate the relationship of single versus multichannel shoppers and their expenditures. Since the shopping environment is dynamic, it is important to keep track of possible changes in the channel shopping behavior of customers. The mobile and online channels are relatively new to the shopping environment. The outcome

(10)

University of Amsterdam 5 of this research will help marketers to gain insights in whether a company should change their single channel environment to a multichannel environment et vice versa. They can possibly base their channel strategy decisions on the outcome of this study and relate it to their own customers. This might help them to use as an indication whether they should expand or decrease their channels and thus consequently be more profitable.

The focus in this research lays on the pre-purchase phase - in which customers search for information - as well as the purchasing phase. Post-purchasing activities are beyond the scope of this research. It will also try to find out whether product category, demographic and psychographic characteristics influence whether multichannel shoppers actually spend more than single channel shoppers.

This study will start with reviewing the existing literature on channel usage and different expenditures. It will also review literature on the moderating effects of product category, demographic and psychographic characteristics. The research gap and question will be discussed prior to hypothesis, managerial and theoretical contribution. Next, the conceptual framework is visualized followed by the research design, results, conclusion, discussion and finally, the reference list.

(11)

2. Literature Review

In this literature review, we elaborate upon the theoretical background of the key concepts of this research. First, multichannel shopping will be briefly discussed. Next, insights in the different expenditures of multichannel channel shoppers will be provided. The available literature on the moderating effect of product category, i.e. tangible versus intangible products, demographic variables, i.e. age, gender and education, and psychographic characteristics, i.e. service orientation, variety-seeking and convenience orientation, will be discussed. At the end of each paragraph, a hypothesis based on the discussed literature is formulated. Finally, with the help of the previously described theoretical framework, the research gap and question for the current study is stated as well as the theoretical and managerial contributions.

2.1. Multichannel Shopping

In general, three different phases can be identified in the normal shopping process, i.e. information search, purchase and after sales services (Neslin, et al., 2006). In the

information search phase, customers research products, compare prices with competitors, access a coupon/promotional code, check reviews about the product/retailer, check funding available prior to purchasing, try and/or touch the product etc. In the purchase phase, customers exchange goods, i.e. buy products with money or an equivalent of money. The after sales service phase, refers to the phase where customers can contact the firm in case they want to return or advice about the purchased product, e.g. when your computer is broken and you get technical support from the service department. In this research,

(12)

University of Amsterdam 7 are a result of the process consisting of search and purchase behavior. Therefore, we did not consider after-sales as a dimension in our analysis.

As stated before a channel is defined as a customer contact point or a medium through which the firm and customer interact (Neslin, et al., 2006). Multichannel shoppers are customers who shop across different channels within a category and/or within a firm (Kushwaha & Shankar, 2013). Thus, single channel shoppers are customers who shop across one channel within a category and/or within a firm. There are different types of

multichannel shoppers that can be distinguished; customers that: 1) search for information on more than one channel and purchase the product on one channel; 2) search for

information on one channel and purchase the product on several channels or 3) search for information on several channels and purchase the product on several channels. Therefore, a multichannel shopper can be a single channel shopper in purchasing a product, but still search for information on multiple channels et vice versa. Inherently, a single channel shopper refers to a customer who searches for information about the product and purchases in the same single channel. A simplified visualization is shown in Figure 3.

Figure 3. Multichannel Shopping Phases (Neslin, et al., 2006)

Information search Purchase

Shopping phase: Offline Online Mobile Offline Online Mobile Channels:

(13)

University of Amsterdam 8 With the emergence of internet, a new era for shoppers arises. Previously, shoppers were only able to shop products in an offline brick-and-mortar store, and now shopping online via a laptop/pc and later also buying via mobile applications are added. Little

knowledge about the effect of this new mobile channel is available and thus more research needs to be done. In 2015, 46% of all customers purchased products with their mobile globally, which is 6% more than in 2014 (PwC, 2016). Also, in the United States 29.7% of all ecommerce was done by mobile in 2015 over 24.6% in 2014 (internetRETAILER, 2015). Recent research shows that adoption of mobile shopping as customers can be assigned to the higher convenience since customers appreciate the ability to shop whenever and wherever they want (Deloitte, 2015).

2.2. Drivers and Antecedents of Multichannel Shopping

Previous research has been done on several moderating effects and an overview of differences in the extending literature can be seen in table 1. This overview shows that most articles studied online and offline channels most frequently and studied the mobile channel less frequent since it was recently introduced and is a relatively new field of research. Also, despite the rapid growth of mobile channels usage, most firms still have an offline channel and more and more are getting an online channel (CBS, 2015). From the multichannel shoppers who purchased clothing, “53% still preferred to make these purchases in store, while 25% preferred to do it online via PC, and 6% each online by tablet and online via phone” (PwC, 2016). Of the multichannel shoppers who purchased consumer electronics, 52% preferred to do it in store and 40% online, primarily via computer (32%). Only 4% liked to do so by mobile phone” (PwC, 2016). This means there are still differences per product

(14)

University of Amsterdam 9 category, which influence the expenditures of multichannel shoppers. This all taken

together makes it more interesting to study the upcoming mobile channel since little is known about this channel.

Little research has been done after which effects psychographics have on the expenditures of single and multichannel shoppers. Heitz-Spahn (2013) did look at the need for flexibility, convenience-, price comparison- and enjoyment orientation customers, but did not look at which effect they had at their expenditures, nor at their channel usage.

Various articles show that multichannel shoppers spend more compared to single channel shoppers, but most articles look at purchase-/sales volume, sales revenue, profit or other sorts of monetary values. Since the effect of multichannel shopping on expenditures is possibly moderated by other factors and differences between these shoppers. It is

important to gain a further understanding of the possible variables influencing this effect. If we have a better insight which antecedents influence the expenditures of single channel versus multichannel shoppers, marketers can take this into account while developing a marketing (channel) strategy.

(15)

Table 1. Prior research overview of multichannel literature (chronological order)

Authors (year) Multichannel usea _Empirical/

Theoretical

Antecedents Monetary value

Demo graphics Psycho graphics Multiple categories Purchase frequency

Deleersnyder et al. (2002) Online, Offline Empirical - - - - Sales

Inman, Shankar, & Ferraro (2004) Offline Empirical √ - √ √ Purchase

Volume

Kumar & Venkatesan (2005) Online, Offline,

Other

Empirical - - √ √ Sales Revenue

Thomas & Sullivan (2005) Online, Offline,

Other

Empirical √ - √ - Expenditures

Neslin, et al. (2006) Online, Offline,

Other

Theoretical √ √

(behavior)

- - Profit

Cheng, Tsao, Tsai, & Tu (2007) Online, Offline Empirical - - - - Profit

Venkatesan, Kumar, & Ravishanker (2007)

Online, Offline Empirical √ - - √ Profit

Ansari, Mela, & Neslin (2008) Online, Other Empirical √ - - - Sales Volume

Avery et al. (2012) Online, Offline,

Other Empirical √ √ (consumer shopping goal) - - Sales Revenue

Hsiao, Yen, & Li (2012) Online, Offline,

Mobile, Other

Empirical √ √ - - -

Heitz-Spahn (2013) Online, Offline,

Mobile, Other

Empirical √ √

(_variousb)

√ √ -

Kushwaha & Shankar (2013) Online, Offline,

Mobile, Other

Empirical √ - √ √ Monetary Value

Campo & Breugelmans (2015) Online, Offline Empirical - - √ -

This research: Mobile, Online,

Offline

Empirical √ √ √ √ Expenditures

a_{Online in this context means online non-mobile PC/Laptop}

(16)

Authors (year) Product category Industry

Deleersnyder et al. (2002) - Newspaper industry

Inman, Shankar, & Ferraro (2004)

Low vs. high differentiation, low vs. high purchase frequency

Automotive/Beauty care/Candy Drug/Cigarettes/Cleaning products/Cosmetics/Feminine hygiene/Bulk foods/Gifts/ Magazines or newspapers/Miscellaneous household items/Paper goods/Pet foods/Photo supplies/Snacks/Grocery/Soap

Retail formats (i.e. grocery stores, drugstores, mass merchandisers,

club stores, and convenience stores)

Kumar & Venkatesan (2005) Aerospace / Consumer Packaged Goods / Education / Financial Services / Government / Manufacturing /

Technology / Travel

Technology industry

(High technology hard- and software)

Thomas & Sullivan (2005) Not specified Retailer (not specified, United States)

Neslin, et al. (2006) - Not firm nor industry specific

Cheng, Tsao, Tsai, & Tu (2007) - Financial service sector (Taiwan)

Venkatesan, Kumar, & Ravishanker (2007)

- Apparel industry

Ansari, Mela, & Neslin (2008) - Consumer durable and apparel products in

mature categories

Avery et al. (2012) - High-end apparel, accessories, and home

furnishings

Hsiao, Yen, & Li (2012) - Not firm nor industry specific

Heitz-Spahn (2013) Non-food product categories (furniture, appliances, electronics, house linen, music-video-books, apparel and

accessories)

Non-food industry

Kushwaha & Shankar (2013) Utilitarian vs Hedonic 22 subcategories

Not firm nor industry specific

Campo & Breugelmans (2015) Different grocery categories Groceries (European chain)

This research: Different product categories Not firm nor industry specific

(17)

2.3. Expenditures

The newer channels, i.e. online and mobile, offer new possibilities for buying products at the lowest cost, but also buying products that previously were not available to the customer. Therefore, new patterns of customer expenditures can be identified. According to the Cambridge English Dictionary (2016), expenditures refers to the total amount of money that a government or person spends. As discussed in the introduction, expenditures in this research refers to the purchase amount, i.e. how many customers spend in euros (EUR), as well as purchase frequency, i.e. how many times – not to be mistaken for how many - they purchase a product.

In order to maximize product sales, firms who are now merely selling their products on one single channel, might question whether they should be adding another, i.e. offline, online and/or mobile, channel. According to Kollmann, Kuckertz and Kayser (2012), it is important for these firms to investigate whether this will increase or decrease the expenditures of their customers. One of the possible dangers lays with the possible cannibalization of having multiple channels over a single channel. For example, there could possibly be a cannibalization of the offline channel over the online channel due to the differences in service and customers’ service orientation. Nevertheless, we might also assume that adding an online channel comes with its own privileges, e.g. the ability and freedom of customers to shop products without people watching over their shoulders.

A recent longitudinal research found that multichannel shoppers are only more profitable during the first three years (Bilgicer, Jedidi, Lehmann, & Neslin, 2015). The data used for this research, was obtained from a major US catalog company. Therefore, it should be takin into account that this affects the external validity. Especially since another

(18)

University of Amsterdam 13 longitudinal study had the opposite outcome and provided evidence that multichannel shoppers are associated with higher customer profitability (Venkatesan, Kumar, & Ravishanker, 2007).

Many other studies found evidence that customers who are shopping on multiple channels are more profitable. Even though not all studies used the same measure, the outcome of several other studies was that multichannel shoppers appear to be associated with customer expenditures (Deleersnyder, Geyskens, Gielens, & Dekimpe, 2002; Inman, Shankar, & Ferraro, 2004; Kumar & Venkatesan, 2005; Thomas & Sullivan, 2005; Neslin, et al., 2006; Cheng, Tsao, Tsai, & Tu, 2007; Venkatesan, Kumar, & Ravishanker, 2007; Avery, Steenburg, Deighton, & Caravella, 2012; Kushwaha & Shankar, 2013). Since the shopping environment is very dynamic, it must be taken into account that articles can become obsolete. Therefore, it is necessary to keep getting new insights to see if the previously found evidence still holds up. Also, the Deleersnyder et al. (2002) article is an old study that used data from a firm that works within the newspaper industry and might have become obsolete. Since the newspapers might have a specific type of buyers and the reasoning behind buying a newspaper online, offline and mobile might be different than the reasoning of buying other products in different channels, the generalizability of this article to other products might be questioned. The same is applicable for Kumar & Venkatesan (2005) with data from a high technology hardware and software manufacturer and Campo and Breugelmans (2015) with data from a European grocery chain. Heitz-Spahn (2013) and Kushwaha and Shankar (2013) kept their data broader by focussing on data within the non-food industry respectively using data that is nor firm nor industry specific.

Inherently, a single channel environment can be considered as an unprofitable business strategy (Hsiao, Yen, & Li, 2012). According to research of Deloitte, multichannel

(19)

University of Amsterdam 14 shoppers spend more than single channel shoppers. (Deloitte, 2014; Thomas & Sullivan, 2005). Their expenditures are 20 to 30% higher, on average, than single-channel shoppers do (Neslin, et al., 2006). Thomas and Sullivan (2005) also found that single channel shoppers buy more items, but have half the expenditures in total. This is evidence that purchase- or sales volume is not a good measure for marketers to rely on and base their channel strategy on.

2.4. Factors Affecting Customer Spending in a Multichannel Environment

Several factors influence the relationship of multi versus single channel usage on

expenditures. Amongst them are demographic characteristics, as this relationship might differ per gender, age or education level. Psychographic characteristics also affect this relationship, amongst them are service orientation, variety-seeking and convenience

orientation. Previous research has shown that there is a different effect per product

category. Due to time and financial restraints, this research will focus on the difference in tangibility for ‘Holiday/Event Tickets’, ‘Consumer Electronics’ and ‘Clothing/Apparel’.

There are also different factors influencing the expenditures of multichannel customers. Behavioral factors are known to have an effect, i.e. past purchases, past

shopping behavior. For example, Bilgicer et al. (2015) found that past purchase incidence is positively associated with customer expenditures. This suggests that more frequent buyers spend more. In order to do research after these behavioral factors, actual data must be obtained of customers’ shopping behavior. This research only has survey data to work with and is therefore beyond the scope of this research.

(20)

University of Amsterdam 15 2.4.1. Demographic Characteristics

Demographic characteristics are features of individuals such as age, sex and education. These factors can determine whether there are certain differences amongst groups. This research will assess if any differences can be found in channel usage and expenditures based on demographic characteristics.

Age

Previous research shows that generation ‘Baby Boomers’ (born between 1945 and 1958) are more loyal to a certain retailer than Generation Y (born between 1977 and 1989), who are less loyal and mostly interested in

the product they want to buy (Parment, 2013). Even though the mobile phone use of ‘Baby Boomers’ is high, they have a limited use and understanding of functions beyond simple voice calls and SMS (McLeod, 2009). Since this is an older article,

‘Baby Boomers’ might attain greater knowledge over the years, but overall we see a decrease in new technology use the older people get (Williams & Page, 2011). One might assume that, due to the better understanding and use of new technology, that the multichannel shoppers will mainly be younger customers. The relationship between age and expenditures is inverted-U-shaped (CBS, 2014; Foster, C.A., 2015; CBS, 2015). This means that really young (<25) and really old people (65+) will be spending the least. As you can see

(21)

University of Amsterdam 16 in Figure 1, the biggest spenders will be customers with the age of 25 to 45 (figure 1) (CBS, 2015).

Gender

Compared to females, males have a shorter channel adoption duration which does not have a relationship with income (Venkatesan,

Kumar, & Ravishanker, 2007). Males tend to be spending more online than females (figure 2) (CBS, 2015). Males also tend to buy more expensive products, i.e. electronics, software, hardware, online than females.

Education

Not much research has been done about the moderating effect of education on customer expenditures between multichannel and single channel shoppers. It seems that people who have a higher education level, have a higher probability of multichannel shopping (Strebel, Erdem, & Swait, 2004). Since people with a higher education are expected to adapt faster to different situations, i.e. shopping channels, they are expected to be spending more than lower educated customers (Kushwaha & Shankar, 2013).

2.4.2. Psychographic Characteristics

Psychographic characteristics relate to opinions, interests, and emotions of customers (Cambridge English Dictionary, 2016). Psychographics have been shown to influence the

(22)

University of Amsterdam 17 behavior of multichannel and single channel shoppers (Neslin, et al., 2006; Avery, Steenburg, Deighton, & Caravella, 2012; Hsiao, Yen, & Li, 2012; Heitz-Spahn, 2013; Ailawadi, Neslin, & Gedenk, 2001). The outcomes of research on psychographic characteristics will give greater insights in the motives of spending more while shopping in a multichannel or single channel environment.

Service Orientation

Service orientation isthe degree to which the shopper finds it important that the firm is helpful, thoughtful, considerate, and cooperative (Hogan, Hogan, & Busch, 1984). Customers that value service orientation, prefer to get expert advice about the products (Holzwarth, Janiszewski, & Neumann, 2006). Expert advice about products is one of the main features of an offline channel which can help customers to make a purchase decision. Even though research shows that replacing an expert with an avatar, i.e. animated graphic representation of a sales agent, has a positive impact on consumers’ attitude towards the product and purchase intention, it pales in comparison to a real life sales agent. Therefore, this indicates that offline channels will generate customers to spend more over online and mobile channels.

The possibility for customer to experience, i.e. touch and try, a product previous to purchasing it, is another unique feature of offline channels (Gupta, Su, & Walter, 2004a; Gupta & Walter, 2004b). Due to the uncertainty with the use of an online channel, customers with higher service orientation, might have a lower propensity to shop at an online or mobile channel since it comes with a higher risk (Kollmann, Kuckertz, & Kayser, 2012). Although, nowadays firms give their customers the option to buy their products online or mobile more and more, try it at home and return it for free if it does not live up to

(23)

University of Amsterdam 18 their expectations. Of course, this increases the firm’s service orientation on the one hand, but the action a customer has to take to return a product decreases the service orientation, i.e. wrapping the product, walking to the post delivery office and posting the package. Previous studies are too narrow to be used as valid evidence to state that service orientation has a negative influence on the expenditures of multichannel shoppers.

Variety-Seeking

Variety seeking is “the choice of an alternative in order to experience diversity or variety in consumption over time” (American Marketing Association, 2016). Consumers seek for variety, because they have a desired stimulation level that they want to satisfy (Kwon & Jain, 2009). Especially when the psychological arousal level is lower than their desired level, which can be changed, e.g. using multiple channels (Donthu & Garcia, 1999). This need for satisfying a certain desired stimulation level can be explained by the theory of optimal stimulation level (OSL). This theory assumes that the optimal stimulation level differs per individual and has the tendency to keep the stimulation at the preferred level (Raju, 1980; McAlister & Pessemier, 1982; Baumgartner & Steenkamp, 1992). In case that the stimulation is not at an optimal level, customers will seek for variety or novelty in order to bring the stimulation to the desired level.

Convenience Orientation

Convenience can be described in terms of saving time and physical and mental effort (Mitchell, 1998). Store accessibility, i.e. location, opening hours, can contribute to this concept (Corby, 1994). Customers tend to weigh the costs, e.g. perceived search effort, and benefits, e.g. perceived convenient browsing, while shopping in different channels. Time

(24)

University of Amsterdam 19 restrictions and efforts are features that are associated with offline channels (Bhatnagar, Misra, & Rao, 2000; Rohm & Swaminathan, 2004). During the pre-purchase phase, i.e. information search, online channels come with less effort since it is easy to browse the internet to look for information and compare product and price information (Gupta, Su, & Walter, An, 2004a; Gupta & Walter, 2004b; Kollmann, Kuckertz, & Kayser, 2012). Online and mobile channels consume less time than offline channels since customers have do not have to visit the store while shopping online or mobile. The accessability of online and mobile stores is overall more convenient since online and mobile channels are always accessible and offline channels are restricted to certain opening hours.

2.4.3. Product Categories

Product categories are distinctions made between products. These distinctions can be made on the basis of particular features of a product. Previous research shows that different categories can lead to different consumer behaviors (Inman, Shankar, & Ferraro, 2004; Kumar & Venkatesan, 2005; Heitz-Spahn, 2013; Kushwaha & Shankar, 2013; Campo & Breugelmans, 2015), which also leads to different expenditures (Thomas & Sullivan, 2005; Kamakura & Du, 2012). An overview can be seen in table 2. Therefore, it is important to decide how to distinguish and categorize the different products alongside which any possible difference, i.e. of the effect of channel usage on different customer expenditures, can be assessed.

Konuş et al. (2008) make a distinction between: Books, Mortgage, Electronics, Holidays, Clothing, Computers and Insurance. This article will add ‘Event Tickets’ as a category since this has been a growing online purchase (CBS, 2014). Konuş et al. (2008)

(25)

University of Amsterdam 20 found differences between different categories, therefore this might also be the case in this research. The differences between these categories, which influence the customer’s shopping behavior, are expected to vary on the basis of several different product characteristics. Balasubramanian, Raghunathan and Mahajan (2005) suggest that the product type can affect the customer’s shopping goals which will lead to a difference in perceived attractiveness of a particular channel. This research will assess the differences between tangible versus intangible products.

Intangible versus Tangible Products

An intangible product is a product that cannot be touched, e.g. online tickets, holidays, mortgage, software, whereas a tangible product is a product that can be touched, e.g. clothing, books, and furniture. Previous literature discusses these products as so-called “low touch”, i.e. intangible, and “high touch”, i.e. tangible, product types (Lynch, Kent, & Srinivasan, 2001; Konuş, Verhoef, & Neslin, 2008). Customers who want to purchase a high touch – or tangible - product tend to have a preference towards the offline channels, while customers who want to purchase a low touch – or intangible - product tend to have a preference towards the online channels due to the speed of shopping online (Lynch, Kent, & Srinivasan, 2001). In line with this, Lee and Kim (2008) found that customers shopping for intangible products are shopping online more often compared to customers shopping for tangible products (Lee & Kim, 2008).

On the contrary, other studies say that it is not adequate to expect differences of channel usage based on the tangibility of a product (Bock, Lee, Kuan, & Kim, 2012; Zeng & Reinartz, 2003). For example, Zeng and Reinartz (2003) found that within the tangible product category, books are sold very often via an online channel, while beauty products

(26)

University of Amsterdam 21 and furniture were not. Although, this article can be outdated and this might have changed over time.

2.5. Contributions

2.5.1. Theoretical Contribution

This research adds to the extending literature, contributing to the research gap summarized in table 1. Because the introduction of mobile and online channels was fairly recent, little research has been conducted comparing the mobile, online and offline channel shoppers and their expenditures together. Knowledge on this subject is therefore far from complete and new research might add to our knowledge of these new channels. Since the mobile channel is an upcoming and evolving channel, the usage of this channels as well as customers’ behavior changes a lot. Inherently, the shopping behavior of customers via online and offline channels changes as well. Thus, it is important to renew existing literature so it will not become obsolete.

Most articles study reasons for multichannel shoppers to spend more.However, this research will go more in-depth in which moderating variables might influence this. When discussing the monetary value of customers’ purchases, previous research mostly assessed different measures than customer expenditures.

Previous research (Bilgicer, Jedidi, Lehmann, & Neslin, 2015) did longitudinal studies of multichannel profitability. This study showed a decline of expenditures after 2 to 3 years following the introduction of multichannel shopping. It is therefore important to study multichannel shopping in the current time frame – years after the introduction – to assess if this decline is real or multichannel shopping is still a positive influence on expenditure.

(27)

University of Amsterdam 22 A lot of articles base their conclusions on multichannel shopping behavior based on quantitative data from big data sets. Thereby, the wishes and psychographics of the actual customers are not taken into account. Also, research is usually limited to a dataset from a specific company, which makes it harder to generalize the outcome of that research. This research will provide insight in how the customers’ preferences might influence their shopping behavior and expenditure patterns in multichannel shopping. Previous literature saw multichannel shopping as a whole, while this research will make a distinction between three different types of multichannel shoppers, who either search for information on multiple channels and/or purchase products on multiple channels. This research will add the effect of different product categories, demographic and psychographic variables on the multichannel and expenditures relationship. Therefore, value will be added to previous research and its findings.

2.5.2. Managerial Contribution

With the upcoming of online and mobile channels, marketers might consider to expand their reach through these new channels. Organizing and setting up an online channel will imply a significant investment. The possible benefit from expanding to these new channels will differ between companies and their type of product and customers. It is therefore very important to be able to predict if this multichannel approach will be beneficial for a company. If the moderating influence of product category, demographic and psychographic characteristics are better known, a tailored prediction of benefits per company might be easier. If the product category that their firm sells products in, show significant positive or negative differences in the expenditures of single versus multichannel shoppers, it is very

(28)

University of Amsterdam 23 important for them to take this into account in developing a channel strategy. The rationale applies while looking at the demographics, i.e. age, gender and education, and psychographics, i.e. service orientation, variety-seeking and convenience orientation, of a firm’s target group. When the influence of these variables is known better, the possible benefits can be predicted better according to their target group. With the outcomes of this research, marketers will be able to weigh the costs and benefits of their current channel strategy and possibly adjust it to better fit the needs of their target group.

If service orientation affects the expenditures of multichannel shoppers positively, marketers should focus on improving the service of mainly the mobile and online channels over the offline channel. Especially since mobile and online channels come with lower fixed costs; the products can be stored in cheaper locations since the ‘visual store’ is on a laptop or mobile. Also, online and mobile store need less employees who inherently bring extra costs since they need to get paid and insured.

When variety-seeking affects the expenditures of multichannel shoppers positively, marketers should try to come up with new ways of increasing the level of variety. This can be done by having various channels, but also by making sure to have several brands or the newest products within a certain product category. These are several ways of increasing the level of variety for customers which might increase customer expenditures.

When convenience orientation positively affects the expenditures of multichannel shoppers, marketers should try to increase the store accessibility by making sure that it is easy for customers to shop the products, by having a clear website, app or store, having long opening hours and easily reachable store for an offline channel, having a return policies for mobile and online channels that makes it easier for customers to return a product so that they find it convenient to try the product while ordering it online without associating it

(29)

University of Amsterdam 24 with a financial risk due to time and convenience restraints, i.e. finding it too much effort or do not have enough time to return the ordered product.

When multichannel customers’ buy and spend more on tangible products compared to intangible products, marketers selling tangible products should make sure they have multiple channels in order to maximize the amount of money customers spend and the amount of times a customer buys products.

2.6. Gaps and Research Question

Research gaps

Quite extensive research has been done about multichannel shoppers. The researchers of leading articles take into account some sort of monetary value (see table 1). The monetary value described in previous studies differs throughout the articles, i.e. sales, purchase volume, sales volume, sales revenue, profit, monetary value, expenditures. The study by Thomas & Sullivan (2005) is the only one that focuses on the expenditures of multichannel shoppers. However, this article lacks description of the psychographic and demographic characteristics. Little research has been done on psychographic characteristics and what has been done on psychographic characteristics, differs a lot in the studied variables. Therefore, there is a need to expand the knowledge on several moderators, i.e. product categories, demographic and psychographic characteristics, that influence the relationship of channel usage on expenditures.

Most studies use a big dataset from a company which makes it harder to generalize the outcome of the research. In contrary, this study will conduct a survey where the

(30)

University of Amsterdam 25 customers will be asked what they spend throughout different categories, firms and channels. Therefore, this research is trying to fill the existing research gap.

Research Question

This research is trying to answer the questions:

Do multichannel shoppers buy and spend more? What is the effect of product category, demographic and psychographic characteristics?

(31)

3. Conceptual Framework

This chapter will visualize the hypothesized relationships in a conceptual framework, and will briefly explain why these hypothesis will be studied. A summary table of all hypothesis can be found in Appendix A.

Figure 4. Conceptual Framework

Figure 4 provides a visualization of the conceptual framework of this research. The main purpose of this study is to determine whether there is a difference in expenditures between single versus multichannel shoppers. This research will look at offline, i.e. brick-and-mortar stores, online, i.e. PC or laptop, and mobile channels. It aims to find out if single or

multichannel shoppers have different expenditures while using different channels or combinations of channels. As explained in the literature review, the expenditures of multichannel shoppers and single channel shoppers are expected to differ. Various

Single versus Multichannel Shoppers Customer Expenditures

Psychographics: - Service Orientation (H3) - Variety-Seeking (H4) - Convenience Orientation (H5) Product Category:

- Tangible versus intangible (H6)

Phase _{Average Order per Purchase}

Purchase Frequency

Total Expenditures (in euros) (H1a,b)

Offline - Store

Information Search Purchase

Channel

Online - PC/Laptop

Offline - Store

Online - PC/Laptop

Mobile - Apps (no browser)

*

*variable not analyzed

Mobile - Apps (no browser)

Demographics:

- Age (H2a)

- Gender (male) (H2b) - Education (H2c)

(32)

University of Amsterdam 27 psychographic characteristics will influence this relationship, as well as different product categories.

3.1. Expenditures

There were contradictory studies showing both that customer profitability is associated with multichannel shoppers as well as that this could be a temporal increase of customer

profitability, observed by Bilgicer et al. (2015). Since the data that was used for this research comes from a major catalog company based in the United States, it might be restricted regarding the external validity and the generalizability towards other product categories. Based on this, an improvement can be done in this study by using data that is not only firm or industry specific.

Despite the use of different measures among various studies, the outcomes appeared to be leading towards the same direction regarding the profitability of

multichannel shoppers versus single channel shoppers. Sales volume, purchase volume, sales revenue, profit and expenditures all seem to be greater for multichannel shoppers than for single channel shoppers. Therefore, we expect to find that multichannel shoppers spend more than single channel shoppers in our study as well.

H1a: Customers using multiple search channels spend more per transaction and purchase

more frequently compared to customers using a single search channel.

H1b: Customers using multiple purchase channels spend more per transaction and purchase

more frequently compared to customers using a single purchase channel.

H1c: Customers using multiple search and purchase channels spend more per transaction

(33)

3.2. Demographics

Age

People who are older appear to be more loyal and less interested in the product they are going to purchase (McLeod, 2009). This indicates they prefer to stick to the well-known area instead of exploring new ones and might therefore spend less in a multichannel

environment. On average, older people also seem to have a lower understanding of new technology and will therefore be less likely to adopt other channels, i.e. less likely to be multichannel shoppers (Williams & Page, 2011).

H2a: The positive effect of multichannel shopping on total expenditures is less pronounced

when age increases.

Gender

As discussed in our literature review, males have a shorter channel adoption duration and therefore it is more likely that they will adopt newer channel, i.e. online and mobile channels (Venkatesan, Kumar, & Ravishanker, 2007). This might increase their likelihood to spend also spend more on these channels since they will feel more comfortable using these channels. Since males tend to spend more online than other females, this will increase the likelihood. It is hypothesized that the relationship of expenditure difference from multichannel shoppers compared to single channel shoppers will be stronger for male sex.

H2b: The positive effect of multichannel shopping on total expenditures is more pronounced

(34)

Education

There is little evidence stating that customers with a higher education level, have a higher probability of multichannel shopping (Strebel, Erdem, & Swait, 2004). Kushwaha and Shankar (2013) state that the increased spending of higher educated multichannel shoppers can be assigned to their possibility to adapt faster to different situations. It is therefore well possible this research will get similar results.

H2c: The positive effect of multichannel shopping on total expenditures is more pronounced

when the customer has a higher education level.

3.3. Psychographics

Service orientation

Offline channel appear to be having the advantage of having the direct ability for customers to touch and try the products. This increases the likelihood of customers to buy a product offline. An online avatar might help the online and mobile channels as an alternative for a real person, but there seems to be a gap between the effectiveness between an avatar and a real person (Holzwarth, Janiszewski, & Neumann, 2006). Kollmann et al. (2012) found that ‘the higher the customers’ service orientation, the lower their propensity to seek information through the online channel first and the lower the propensity to place an order via the online channel’ (p.188).

Even though the service of online and mobile channels is improving due to convenient return policies which allows customers to experience the products better, offline channels are still expected to outdo online and mobile channels. Due to the overall larger screens which inherently comes along with better sight of the products, we expect online

(35)

University of Amsterdam 30 channels to slightly generate higher expenditures than mobile channels. Even though we expect a large gap between the offline versus online and mobile channel. Therefore, it is hypothesized that service orientation will have a negative effect on the relationship of expenditure difference from multichannel shoppers compared to single channel shoppers.

H3: The positive effect of multichannel shopping on total expenditures is less pronounced

when the customer has a higher service orientation.

Variety-seeking

Little research has been done on the influence of variety seeking on expenditures of multichannel versus single channel shoppers. Variety-seeking comes along with a certain level of satisfaction of stimulation (Kwon & Jain, 2009). When this desired level is not reached, customers will do efforts to change the stimulation level to the desired level by seeking variety according to the Optimum Stimulation Level (OSL) (Raju, 1980; McAlister & Pessemier, 1982; Baumgartner & Steenkamp, 1992). This novelty or variety can be found in shopping in different channels, i.e. offline, online and mobile. Due to the variety multichannel shoppers have while shopping across different channels, this might have a positive influence on the customer expenditures. Therefore, it is most likely that multichannel shoppers have a high variety seeking inclination (Donthu & Garcia, 1999). It is hypothesized that variety-seeking positively affect the relationship of expenditure difference from multichannel shoppers compared to single channel shoppers.

H4: The positive effect of multichannel shopping on total expenditures is more pronounced

(36)

Convenience orientation

Since convenience orientation is focused on store accessibility, saving time and physical and mental effort, it is most likely that the use of online and mobile channels will score higher on these features than offline channels. Mainly because the online and mobile channels are 24/7 accessible and it is easier to implement features in the app or website that helps you filter the products and customers do not have to leave their home to buy the products. Since the amount of choices is greater in online and mobile channels compared to an offline channel, this might have a bad influence on customers’ mental effort. Since websites nowadays still have more possibilities to be designed to decrease the customers’ mental efforts, online channels will probably be experienced as more convenient over mobile channels. It is hypothesized that a high level of customers’ convenience orientation will positively affect the expenditures of multichannel shoppers.

when the customer has a higher convenience orientation.

3.4. Product Categories

Intangible versus tangible

There are contradictory studies about whether the tangibility of a product has an impact on the expenditures of single versus multichannel shoppers. Some studies did not find an effect (Zeng & Reinartz, 2003; Bock, Lee, Kuan, & Kim, 2012), while other studies did. Lynch et al. (2001) found that tangible products have a higher need to be touched, which can more easily be assessed via an offline channel. Inherently, intangible products do not have this

(37)

University of Amsterdam 32 need and therefore they might as well be bought via an online or mobile channel (Lee & Kim, 2008). Of course it must be taken into account that there is an upcoming trend towards making it more convenient for the customers to order and return products in order to touch and try them. Since it still takes effort to do return the products and it is less convenient since you cannot directly order and return the entire collection, the offline channel is still the most convenient for tangible – or high touch – products. Therefore, we expect that customers will more easily use multiple channels to buy intangible products.

(38)

4. Research Design

For this research, an online questionnaire was distributed. Therefore, a quantitative approach is needed. In order to assess the expenditure of multichannel shoppers, the difference between product categories and demographic and psychographic characteristics, a survey was conducted to gather cross-sectional data. All respondents received the same online questionnaire. The questionnaire was available in Dutch and English. A pilot study was done to make sure that the questionnaire was clear and attractive. Also, the pilot-study was done to make sure that all questions could be answered within the given time slot.

4.1. Population Sample

The population-sample used for this research was taken from online, mobile and offline shoppers aged eighteen and older. The questionnaire was made with Qualtrics and was spread via Internet while using a convenience sample. Due to financial and time restricted reasons, the questionnaire was spread via social media, i.e. Facebook, email, peers, friends and family.

4.2. Measures

The respondents were asked to answer which channels, i.e. offline/online/mobile, they used to search for information about and – separately - purchased products in the past six months. Total product expenditures and purchase frequency were asked separately for three different product categories: holiday/event tickets, consumer electronics and

clothing/apparel. Since the main research question of this study is finding out whether single or multichannel shoppers spend more, the questionnaire started with questions on

(39)

University of Amsterdam 34 the amount of money the respondents spend on which channels. First, the three definitions of online, offline and mobile channels were given to increase the internal validity.

To measure the convenience orientation, variety-seeking and service orientation, the commonly used questionnaires from Kolmann et al. (2012) and Baumgartner et al. (1992) were used. A more detailed description is found in appendix B.

Convenience orientation was measured using the validated CONVOR scale with a

Cronbach’s alpha of .735. (Montoya-Weiss, Voss, & Grewal, 2003; Chiang, Zhang, & Zhou, 2006; Kollmann, Kuckertz, & Kayser, 2012). The 3 items used was measured on a 7-point Likert scale. Service orientation was measured using the validated SERVOR scale with a Cronbach’s alpha of .735. The 3 items used was measured on a 7-point Likert scale

(Montoya-Weiss, Voss, & Grewal, 2003; Chiang, Zhang, & Zhou, 2006; Kollmann, Kuckertz, & Kayser, 2012). Variety-seeking was measured using the validated Exploratory Acquisition of Products (EAP) scale with a Cronbach’s alpha of .85 (Baumgartner & Steenkamp, 1992). The 4 items was measured on a 7-point Likert scale.

The questionnaire ended with asking the respondents about their demographics, i.e. age, gender and education. Education was divided in 7 levels, depicted in table 3 and

clarified in the next paragraph. Since these questions were confidential, respondents were asked at the end of the survey in order to increase the response rate (Gravetter & Forzano, 2015; Saunders & Lewis, 2012). All responses were processed anonymously. An overview of all variables, their measures and measure levels can be seen in table 3.

(40)

Table 3. Summary table measures and measure levels of variables

Variable Measure Level

Demographics Age Numeric Ratio

Gender Female = 0, Male = 1 Nominal (dummy variable) Education Less than high school (1);

High School / GED (2); Intermediate Vocational College Degree (MBO) (3); Bachelor’s Degree (HBO, University of Applied Sciences) (4); Bachelor’s Degree (WO, University) (5); Master’s Degree (6); Doctoral Degree (7)

Ordinal

Student No Student = 0, Student = 1 Nominal (dummy variable)

Psychographics Service Orientation 7-points Likert Scale Interval/Ratio Variety-seeking 7-points Likert Scale Interval/Ratio Convenience Orientation 7-points Likert Scale Interval/Ratio

Expenditures Purchase Frequency Numeric Ratio

Purchase Amount Numeric Ratio

Product Categories

Holiday/Flight Tickets No = 0; Yes = 1 Nominal (dummy variable) Consumer Electronics No = 0; Yes = 1 Nominal (dummy variable) Clothing/Apparel No = 0; Yes = 1 Nominal (dummy variable)

Recoding Variables

The variable education was recoded to ‘Low’ (Less than high school, High School, Intermediate Vocational College Degree (MBO), Bachelor’s Degree (HBO), University of Applied Sciences, Bachelor’s Degree (WO, University)) = 0 and ‘High’ (Master’s Degree, Doctoral Degree) = 1. Two items from the scale Variety-Seeking, were counter-indicative,

(41)

University of Amsterdam 36 meaning that when the respondents scored high on these two items, they should score low on the scale. Therefore, these items were recoded.

Multi Search Channel Shoppers were recoded as a dummy variable with multi search

channel shoppers (searching for information about a product or service on multiple channels) = 1, and single search channel shoppers (searching for information about a product on a single channel) = 0. Multi Purchase Channel Shoppers were recoded that multi purchase channel shoppers (purchasing a product or service on multiple channels) = 1, and single purchase channel shoppers (purchasing a product or service on a single channel) = 0. To analyze the moderating effect, the dummy variable Multichannel shopping was recoded if respondents were ‘multichannel shoppers’ (searching for information on multiple channels and purchasing a product on a single channel, or searching for information on a single channel and purchasing a product on multiple channels, or searching for information and purchasing a product or service on multiple channels) = 1, or ‘single channel shoppers’ (searching for information and purchasing a product on a single channel) = 0. This variable was multiplied with the moderating variable in order to analyze the moderating effect.

4.3. Analyses

The data were analysed with SPSS (Version 22.0.0.0, IBM Corporation). Since the purchase frequency and spending variables were asked separately for the three product categories, i.e. tickets, consumer electronics and clothing, the data were analysed separately based on these three product categories. The main effect of being a multichannel shoppers on customer expenditures, was also analysed per category. Customer expenditures in euros were analysed with the use of a multivariate linear regression analysis. For customer

(42)

University of Amsterdam 37 expenditures in purchase frequency, a Poisson regression was performed. The next section is a more thorough explanation of the use and advantage of the Poisson Regression Analysis for this incremental variable. In total twelve models were constructed and analysed. A separate analysis were done per 3 (holiday/event tickets, consumer electronics,

clothing/apparel) x 2 (purchase frequency, spending in euros) x 2 (main effect, moderating effects). An overview can be seen in table 4. Additional analyses were conducted to see if the interaction between the information search and purchase of a product affects customer expenditures in purchase frequency and euros. In these analyses multichannel shoppers are considered customers who both search for information and purchase on multiple channels.

For each separate three product categories a multivariate linear regression analysis on the main effect total expenditures were performed, as well as a multivariate linear regression analysis with moderators included as independent variables. Next to this a Poisson regression analyses for each three product categories was performed to assess purchase frequency with the main effect only and a model with the moderators included as well.

Table 4. 12 Analyses Models

Customer Expenditures in Purchase Frequency Customer Expenditures in Euro’s Holiday/Event Ticket 1. Main Effect 2. Moderating Effects 3. Main Effect 4. Moderating Effects Consumer Electronics 5. Main Effect 6. Moderating Effects 7. Main Effect 8. Moderating Effects Clothing/Apparel 9. Main Effect 10. Moderating Effects 11. Main Effect 12. Moderating Effects

(43)

University of Amsterdam 38 Since purchase frequency is in ‘amount of times’ and therefore an incremental variable, it would not be correct to do a simple linear regression analysis. A Poisson regression is more suitable instead, since this is used when the dependent variable contains count data. The downside of doing a simple linear regression is that it will use a fixed amount of addition per unit change in the predictor without taking into account the circumstances, but a Poisson regression analyzes this while using percentages, i.e. per unit change in the predictor x% increase of y. Therefore, it is better to do a Poisson regression analysis when customer expenditures in amount of times is analyzed.

In order to be able to do a Poisson regression analysis, the data has to meet five criteria points. First of all, it is important to use when the dependent variable contains count data. Second, one or more independent variables are needed that contain nominal, ordinal or continuous data. Third, an independence of observations is needed. Fourth, the

distribution of counts follow a Poisson distribution. And fifth, the mean and variance of the model are identical. This dataset meets the first four criteria, but the mean and variance of the model appear not to be identical. Nevertheless, a Poisson regression analysis will be done because it is the most proper way to do so in this case and more correct than doing a simple linear regression.

In a Poisson regression, the beta coefficient, Wald Chi-Square value and p-value will be interpreted. The beta coefficient and p-value need the same interpretation as in a linear regression analysis. When the p-value is significant, the effect of the variable on the

predictor, i.e. purchase frequency, can be done. The beta coefficient indicates that when the customers purchase frequency went up by 1, the variable went up by the beta coefficient. Lastly, the Wald Chi-Square is testing the difference between the log of expected counts of

(44)

University of Amsterdam 39 purchase frequency on the independent variables. Wald Chi-Square is the beta coefficient divided by the standard error and multiplied by two, i.e. (B/SE)*2.

Do multichannel shoppers buy and spend more? : the effect of product category, demographic and psychographic characteristics