How does forced migration to mobile shopping apps affect intentions/willingness to buy : the moderating effect of impulsiveness, innovativeness and mobile shopping orientation

(1)

0

How does forced migration to mobile shopping apps affect

intentions/willingness to buy: The moderating effect of impulsiveness,

innovativeness and mobile shopping orientation.

Course: Master’s Thesis Marketing

Program: MSc Business Administration - Marketing Track

Student: Danny Dekker

Student Number: 10534296 Thesis Supervisor: Umut Konuş

(2)

1 STATEMENT OF ORIGINALITY

This document is written by Danny Dekker 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)

2

Abstract

Consumers are increasingly using their mobile phones for various reasons, including shopping. Some mobile shopping apps have been downloaded several million times. With this new channel companies will want to know how to migrate consumers to this channel in the best way however there has been no research on migration to this specific channel. Therefore, this research aims to find the effects of rewarded and forced migration and try three possible variables that might moderate the negative effects of forced migration. To research this, respondents were asked to do a survey where they were presented with the three types of migration across three product categories; clothing, electronics and flight tickets. They were asked how innovative, impulsive and experienced with mobile shopping they are. Rewarded migration was found to have a positive effect on willingness to buy, valence and channel choice. Forced migration was found to have a negative effect as expected. This research however could not find any significant moderating effects of impulsiveness, innovation and mobile orientation.

(6)

5

1. Introduction

1.1 Introduction

By the end of January 2018, the Amazon shopping app has been downloaded over a hundred million times and the Walmart shopping app between ten and fifty million times through the Google Play Store (Google Play Store, data retrieved January 28 2018). Some Dutch

companies also have shopping apps. Bol.com, a Dutch version of Amazon and Albert Heijn, a Dutch grocer, both have been downloaded between one and five million times. Some companies, such as Rivièra Maison, are thinking about or did eliminate their websites, completely or just the mobile one, in favour of such shopping apps. An example is Rivièra Maison who for a while did not have a proper mobile webshop and instead wanted visitors to download their app. Others might follow or start pushing their shopping apps in favour of their webshops.

Companies are making these shopping apps because they believe the apps have certain benefits over desktop websites and mobile websites, such as the benefits that Mobify gives. Mobify, a shopping app developer gives several benefits on their website (Cyr, 2015). They state that overall mobile apps can be better designed and are more convenient due to them being independent instead of within a browser.

When companies make such a mobile shopping app, they will have think of how to change their multichannel marketing strategy to include this new channel. Most studies in the multichannel marketing field find that having more channels is better for the company. Examples are Blattberg, Kim, Neslin (2008) and Kumar and Venkatesan (2005). They found that multichannel customers have a more positive view of the company and are more

(7)

6 However, companies will want at least a part their customers to go to their newly developed channel, the shopping app. Companies thus have several options on how to handle their new shopping app. They can have the mobile shopping app be equal to the other channels and let the customers completely voluntarily migrate to the new channel if they want to. The

companies can however also try forcing the customers to the new channel, which can be done in multiple ways. They could completely remove previous channels, such as Rivièra Maison tried. This was looked at by a few studies and they found some success. Examples of this are Chu, Chintagunta, and Vilcassim (2007) who found Dell made the right decision when they removed their retail channel. And another example is Konuş, Neslin and Verhoef (2014) that found a small increase in profits for a company after they removed their catalogue. These examples of completely forced migration are however specific and not necessarily from one retail channel, such as a normal website, to another, the mobile shopping app. Also, a possible consequence with forced migrations is reactance. When the freedom of a person is threatened they will try to regain their freedom by opposing or resisting whatever causes the threat (Brehm & Brehm, 2013).

To create less reactance with their customers, companies might try other ways to force migration. They could punish customers if they keep using older channels, for example companies could have increased prices for these channels. Or companies could reward their customers for using the new channel, for example by giving a discount for using that channel. Trampe, Konuş and Verhoef (2014) looked into the amount of reactance caused by the

various types of migration. They found that completely forced migration and punishing customers for using older channels cause similar levels of reactance, while rewarding customers causes less reactance.

(8)

7 There has been some research on the effects of customers migrating to mobile from other channels, however there has been little research about the effects when people start using mobile shopping apps specifically. Studies have found that this migration to mobile shopping increases spending. Wang, Malthouse and Krishnamurthi (2015) found that

customers who migrated to mobile buy more often and in the case of previously low-spenders the order sizes also go up. This is an interesting result, as Ansari, Mela and Neslin (2008) found migration from offline channels to online made customers less loyal to a company. They propose this effect comes from the lowered switching costs and less interaction with employees.

Thus, there has been research in the fields of mobile shopping and the effects of channel migration. However, there has been no research on how to get customers to actually migrate to mobile shopping. And, it is not known if the migration literature would directly translate to migration towards mobile shopping apps, as there is evidence that migration effects depend on the context. Right now, companies spend large amounts of money on mobile shopping apps without knowing how to get their customers to use it and if it would actually benefit the company, and if it does by how much. If the mobile shopping apps are much better for companies, forcing their customers to use them might make sense. If more research is done on this topic, the managers could more effectively choose to make a mobile shopping app or not. And if the reasons for customers to migrate or not are studied,

companies can use their marketing more efficiently to get customers to migrate to these shopping apps. For example, Wang, Malthouse and Krishnamurthi (2015) found that habitual products were purchased the most through mobile, maybe an airline would not benefit from an app at all. Or maybe they will because likely most of their customers already buy online so they will be more willing to go to a mobile app.

(9)

8 Therefore, the research question will be:

How does forced migration to mobile shopping apps affect willingness to purchase and what is the moderating effect of attitudes, demographics and online shopping orientation?

Although mobile shopping apps are becoming more commonplace, no research has been done on this subject. Therefore, this research will make a start by looking at what might affect customer’s reaction to migration to mobile shopping apps. Hopefully this will create a starting point for other researchers to further develop this field.

1.2 Research structure

To fully understand the theory behind channel migration and mobile shopping, first mobile shopping will be quickly discussed. After that multichannel strategy and channel migration will be discussed, after which the forced migration to mobile shopping apps and its effects on online shopping will be discussed. After that other possible factors will be

discussed, such as demographics. After these topics are discussed in chapter 2, the conceptual framework will be discussed in chapter 3 and how it will be researched in chapter 4.

(10)

9

2. Literature review

2.1 Multichannel Marketing

A channel is defined as a “customer contact point, or a medium through which the firm and the customer interact” by Neslin et al. (2006, p. 96). This means that channels are contact points where customers not only get communicated to by a company but where the customers themselves can also communicate or buy. There are three visions that can be used to give an explanation for a multichannel strategy, it is more efficient, easier to segment or better for customer satisfaction (Neslin & Shankar 2009).

There has been little research on which is the view that should be used when implementing a multichannel strategy. What is known however is that customers who buy from multiple channels have been found to have a greater trust and lower perceived risk company, at least in a business-to-business setting (Kumar, & Venkatesan, 2005). This is further supported by (Blattberg, Kim, & Neslin, 2008) who found that firms with a multichannel strategy will get more customers and that they will generate more revenue and more profit.

Kushwaha and Shankar (2013) however found that this is not true for all product categories. They found there to be differences between hedonic and utilitarian products and high-risk and low-risk. For large mass-merchandise and hedonic products multichannel customers are most valuable, while for utilitarian products any single channel customer is valuable. And specific channels for the combinations of those two categories and low-risk and high-risk.

(11)

10 There has been plenty of research on variables that influence channel choice.

Examples are Balasubramanian, Raghunathan and Mahajan (2005) who looked at the goals of the purchase, such as if it is thrift, affirmation of expertise or if the purchase has a symbolic meaning. There has also been research on the effect of demographic and psychographic characteristics such as price consciousness, loyalty and time pressure (Konuş et al., 2008) and other characteristics such as privacy concerns and enjoyment (Verhoef, Neslin, & Vroomen, 2005).

Besides the benefits mentioned earlier there are also risks involved with the use of multiple channels. There is a risk of channel cannibalization (Pauwels, & Neslin. 2015; Avery, et al. 2012; Pauwels et al. 2011), though the exact results in what channel addition hurts what other channels differs per study. This might be explained by the difference in product category. There is also risk of channel conflict if the company is unwilling or unable to reduce this conflict (Vinhas & Anderson, 2005).

2.2 Channel Migration

Because of the aforementioned risks and to increase revenue and profit, many firms are stimulating customers to use certain channels, either all customers or segmented, and some even remove channels to force customers to use the other channels (Trampe, Konuş & Verhoef. 2014; Neslin & Shankar. 2009; Konuş, Neslin & Verhoef. 2014).

Despite this migration becoming increasingly used by companies, there has been little research about the consequences of this. And there has also been little research on a possible difference in these consequences depending on the type of migration. A distinction between these migrations can made. Migration can be forced, voluntary, rewarded or punished (Trampe, Konuş & Verhoef. 2014).

(12)

11 They found that rewarding customers if they migrate to the new channel, there will be less reactance than when punished or forced, and that forcing consumers to use the new channel will even cause reactance in the customers already using the new channel, also shown by Moon and Frei (2000).

There are also several consumer characteristics moderating the impact of forced migration. Konuş, Neslin and Verhoef (2014) looked at a company that eliminated their search catalogue and found three factors moderating the impact on purchase utility. First, they found that customers with a high baseline firm preference are less affected from the catalogue elimination.

They found that the customers who are most negatively affected are those who used the purchase channel that went along with the search channel, the telephone. And lastly, they found that receiving marketing communications diminishes the negative impact of the elimination. Li, Konuş, Langerak and Weggeman (2017) looked at the adoption of a new online channel and what the customer already used for competitors. They found that existing customers of the focal firm are more likely to buy through the existing catalogue channel than new customers. However, they found that customers who already adopted the online channel from a competitor are more likely to adopt the focal firm’s online channel. Those who do then adopt the online channel are more likely to have an increased share of wallet at the focal firm. This effect could be explained by the previous experience they have with an online channel, as found by Bigné, Ruiz and Sanz (2007). They found that mobile affinity and previous experience positively influences future mobile commerce intention.

This is further supported by Lu and Su (2009) who found mobile skillfulness to play an important factor in intention to engage in mobile shopping. In their study skillfulness increases enjoyment, usefulness and reduces anxiety, which increase customers intention to engage in mobile shopping.

(13)

12 If people have former mobile experience it could be proposed that they are more skillful. Agrebi and Jallais (2015) also found that perceived ease of use, usefulness and enjoyment increase the intention to use the mobile channel for shopping. They find that this effect is caused by an increase in satisfaction. However, this is only the case for customers who have already used the mobile channel at least once. Non-purchasers were found to only have an increased intention to use if the perceived usefulness was high. This further suggests that previous mobile experience is an important factor that decides that if customers will use a company’s new mobile shopping app.

Despite the literature suggesting that multichannel customers spend more and that people experience reactance when there is forced migration, there is proof of it actually improving profits of companies. Examples are Dell, which would have made less profit if they kept their retail channel (Chu, Chintagunta, and Vilcassim, 2007) and an anonymous company used in Konuş, Neslin and Verhoef’s (2014) research had increased profit after they removed their catalogue, albeit small.

For the latter the extra profit came from the savings on the catalogues, however for some companies it might be a bigger difference as they could possibly close a call center or support staff. They also found the impact of the channel elimination to be for a longer term, in their case 28 months.

2.3 Mobile and Mobile App Shopping

because shopping apps are still relatively new there has been little research about this subject. Therefore, instead of talking about shopping apps specifically, mobile commerce in general will be discussed in this part of the literature review. Mobile commerce can include shopping apps; however, no one has made a real distinction when researching it, putting both groups together.

(14)

13 The new wave of channel migration might be that to shopping apps. In one research 18% to 38% of the respondents use their mobile phone to purchase different types of products. Meanwhile 29% to 73% of the respondents use their computer to do these

purchases (Holmes, Byrne, & Rowley. 2013). The percentage of purchases done through the mobile channel is however likely to increase. Because of continuously increasing capabilities of smartphones and lowering costs of data, customers are using their smartphones to be online more (Persaud & Azhar, 2012). Companies can reach customers increasingly more through mobile means thanks to this.

They can reach them not just through websites but also through apps. Companies are already doing this, both for purely marketing, such as Maybelline with their Make-up coach app, but also with shopping apps such as Amazon. Wang, Malthouse and Krishnamurthi (2015) found that customers who start using their mobile phone for shopping have increased order rates, and in case of low-spenders not just order rates but also order sizes. They found that

customers tend to buy habitual products through mobile shopping, proposing that is due to its convenience.

Some companies have abandoned their mobile website and instead use it mainly to promote their app shop which has a better experience in an effort to increase profits. This could go in line with Wang, Malthouse and Krishnamurthi (2015) who propose that the increased spending through mobile shopping is due to convenience. An app is more convenient for the customer, you only need one click and you are at their store, instead of going through a browser. There is however no research the companies can rely on to see if they should fully focus on their apps and force their customers to these apps. There is only limited research on m-commerce in general, of which most is about the technological adoption model and specifically the involvement of perceived usefulness.

(15)

14 And as stated by Zhang, Zhu and Liu (2012) who did a meta-analysis on mobile commerce adoption, it is very specific for a country and setting or old. An example of this is Chong, Chan and Ooi (2012) who looked at a prediction model for mobile commerce adoption. They however only looked at China and Malaysia. They found that social influence was the biggest predictor of adoption and that demographics mattered in Malaysia but not China. Surprisingly age had a positive effect, with older Malaysians being more likely to adopt mobile commerce. This is contradictory to Bigné, Ruiz and Sanz (2007) who found that younger Spanish people are more likely to use their mobile have a more positive attitude towards mobile commerce and increased mobile affinity.

2.4 Forced channel migration to Mobile shopping apps

So, it has been shown that there is an increasing interest in mobile shopping and it is expected to increase further (Holmes, Byrne, & Rowley. 2013; Ajax Persaud, Irfan Azha. 2012). And it has been shown that it is likely to be beneficial to companies if their customers use their mobile devices to shop (Wang, Malthouse, & Krishnamurthi. 2015).

One of the main issues that companies will face however is actually getting their customers to use these mobile devices. The only determinant found to increase intention to use the mobile channel to purchase in the case of non-purchasers is usefulness (Agrebi, & Jallais. 2015). Only after they have become mobile users will more determinants, such as ease of use and satisfaction, come into effect. Thus, according to Agrebi and Jallais (2015) companies should focus on marketing the advantages of the mobile channel over the computer.

Little research however has been done on finding other possible strategies companies could use to get their customers to use the mobile channel. Instead of marketing the

(16)

15 Instead of telling the customers the advantages and hoping that they believe these messages, it could be better to make them use the mobile channel. Once they have done this they can directly see the advantages. This is especially the case with shopping apps, as they can be better designed and more convenient than a mobile website (Cyr, 2015). This should further increase the satisfaction and perceived usefulness, making the customers more likely to keep purchasing through this channel (Agrebi & Jallais, 2015; Wang, Malthouse, &

Krishnamurthi, 2015; Bigné, Ruiz, & Sanz. 2007).

The issue with this approach however is the reactances it can cause (Trampe, Konuş & Verhoef. 2014; Moon & Frei. 2000). Customers might feel that their freedom is threatened if they are forced to use a company’s mobile shopping app. If this happens they might not make use of the promotion, or even worse stop buying from the company altogether and start buying from a competitor. As Trampe, Konuş and Verhoef (2014) found, rewarding

customers for making the switch reduces the reactance significantly. They also found that customers with higher loyalty experience less reactance. This research will look at forced migration and several customer traits to see if this is a viable strategy for companies to get their customers to use the more profitable mobile channel.

As has been discussed there are various things that can influence the intention to use the mobile channel. For this study three will be looked at. As it is at the core of a lot of other studies, mobile orientation will be looked at. If customers are already experienced with mobile shopping they are more likely to adopt another company’s mobile channel (Li, Konuş, Langerak & Weggeman.2017) and have a higher intention to use the mobile to make future purchases (Bigné, Ruiz & Sanz. 2007). As most research so far has been focussed on perceived usefulness, impulsiveness will be looked at as another possible moderator. This because impulsiveness increased the perceived usefulness of online stores (Cass & Fenech. 2003).

(17)

16 The last possible moderator that will be looked at is innovativeness, as high innovativeness is found in at least particular categories to make customers try new channels (Cho & workman. 2011).

It is likely that even when done properly, forcing customers to use a mobile shopping app will cause a part of the customers to stop purchasing from a company due to experiencing reactance. However, the increased purchase quantities (Wang, Malthouse & Krishnamurthi. 2015) should make up for that.

2.5 Theoretical Contribution

This research will contribute to the research fields of both mobile shopping apps and forced migration. There has been a lot of research on the adoption of mobile shopping, but most of them have been focussed on the perceived usefulness of mobile shopping in general, ignoring other possible customer traits and missing out on the newest channel, the mobile shopping app. It contributes to the field of forced migration by looking at it in combination with a different channel and by looking at possible moderating variables that have not been looked at before. Hopefully future research will further expand on this research and look at these moderators further, or at different ones.

(18)

17 Table 1. Summary of prior research.

Topic Findings Study

Multichannel Marketing

Customers who buy from multiple channels typically generate more revenue and profit.

Blattberg, Kim, & Neslin (2008)

The benefits of multichannel marketing are not true for all product categories.

Kushwaha and Shankar (2013)

There are also downsides to having multiple channels. Examples are possible channel cannibalization and channel conflicts.

Pauwels & Neslin (2015) Avery et al. (2012) Vinhas & Anderson (2005)

Channel Migration

To remove these downsides and to focus customers on certain channels that might be more profitable, some companies try migrating their customers.

Trampe, Konuş & Verhoef. (2014) Neslin & Shankar (2009)

Konus, Neslin, & Verhoef (2014)

Migration Strategies

Channel migration can be voluntary, rewarded, punished and forced.

However, if customers are punished or forced to migrate they experience reactance.

Trampe, Konuş & Verhoef. (2014)

Migration Determinants

Customers who have prior experience with a new are more likely to use it if a company starts using it.

Li, Konuş, Langerak and Weggeman (2017)

Bigné, Ruiz, & Sanz (2007) Migration

Determinants

Perceived usefulness increases the intention to use a new channel, specifically the mobile channel.

Agrebi & Jallais (2015) Wang, Malthouse, & Krishnamurthi (2015) Bigné, Ruiz and Sanz (2007) Migration

Determinants

The effects of variables such as age and social influence differ per country.

Bigné, Ruiz, & Sanz (2007) Chong, Chan, & Ooi (2012)

Mobile Shopping

An increasing percentage of customers is using mobile devices to purchase

products

Holmes, Byrne, & Rowley (2013) Persaud & Azhar (2012)

Mobile Shopping

Customers that start using mobile devices for purchasing spend more.

Wang, Malthouse, & Krishnamurthi (2015)

(19)

18

2.6 Managerial Contribution

As has been stated before, some companies are already trying to migrate their customer base to their mobile shopping app. However currently they have little guidance from the scientific field in how they should do this. It is possible that forcing customers to use it will cause a lot of reactance, or it might be surprisingly little. This could differ per customer and per product category. This study tries to find a best option and what could affect it. If managers can learn about this they can more effectively migrate their customers to a more convenient and profitable channel or decide not to if their customers tend to have traits that cause them to have more reactance to forced migration. Instead then they can opt to go for rewarded migration, or even none at all.

(20)

19

3. Conceptual Framework

and Hypothesis

3.1 Research Framework

Fig. 1. Conceptual framework.

For this study a conceptual framework has been developed to test various possible traits that could be influencing the relationship between the type of migration and various reactions to this migration. This model suggests that rewarded migration however should have a positive impact on willingness to buy, channel choice and valence. Forced migration is however expected to have a negative impact on willingness to buy and valence. Channel choice is not included for forced migration due to a mistake made in the study.

Type of migration Forced Willingness to Buy Channel Choice

Valence

H1a; H1b; H2 Impulsiveness Innov Mobile Orientation Control Variables Demographics Mobile Phone OS H3a H4a H3b H4b H3c H4c

(21)

20 It is expected that impulsiveness, innovation and mobile orientation each have a moderating effect on these relationships for forced migration. Thus, if someone is impulsive, innovative or already spend more on mobile, they will be less impacted by forced migration. Lastly the analyses will also account for several control variables; age, nationality, gender and mobile operating system.

3.2 The effects of rewarded migration

When looking at the effect of different types of migrations, Trampe, Konuş and Verhoef (2014) found that even when people are rewarded for using the new channel, consumers still had negative reactance towards this migration. This is in line with Brehm and Brehm (2013) who talk about people’s reactions when they feel they are coerced or being persuaded to change their behaviour. However as stated by Trampe, Konuş and Verhoef in their research in their experiment consumers went from a higher cost and service level channel to a lower cost and service level channel. This study however looks at migration from a company’s website to a mobile shopping app. The level of service would remain equal or improve as mobile apps can have a better suited interface. Also, reactance is reduced if the reward is congruent with the consumers effort (Kivetz, 2005). This study has a different reward scenario from Trampe, Konuş and Verhoef (2014). This scenario has a discount for both the website and mobile shopping app, however the mobile shopping app has more.

So, everyone gets to enjoy a discount but people who are willing to try the app get more discount. This should negate the effect of people reacting badly to the migration. This should reduce reactance, making people more likely to use the new channel and feel better about the promotion and as there is an increased discount increase the willingness to buy.

(22)

21 This increase while having a possibility to still use either channel is also in line with the extant literature showing that customers who use multiple channels buy more overall (Blattberg, Kim, Neslin (2008); Kumar and Venkatesan (2005). Therefore, the hypotheses regarding rewarded migration are as following:

H1a: Rewarded migration has a significant positive effect on willingness to buy and valence

H1b: Rewarded migration has a significant positive effect on consumers choosing a mobile app instead of a website.

3.3 The effects of forced migration

The forced migration scenario in this study is different from Trampe, Konuş and Verhoef (2014). In their forced scenario they truly forced the consumers to the new channel as the previous one got removed. They also had a scenario where the previous channel still existed however consumers got punished to still use that, which had the biggest negative impact out of all scenarios. Companies want their consumers to use their mobile shopping app, but they want to keep other channels open as well due to the higher profitability of multi-channel consumers. Therefore, in this study consumers will not be forced due to removal, but forced due to a sale only being for the app. This is what Trampe, Konuş and Verhoef had as their rewarded scenario. In their case this scenario had a positive impact on reactance, meaning people were annoyed and possible making them act against what the company wants. In accordance to the reactance theory, this should lead to a reduction in willingness to buy as people would feel annoyed by the discount for just the mobile shopping app.

(23)

22 Therefore, the hypothesis regarding forced migration is as following:

H2: Forced migration has a significant negative effect on willingness to buy and valence.

3.4 The effects of impulsiveness

There has been little research on forced channel migration and none on forced migration to mobile shopping apps, thus this study will look at three traits that could possibly have an effect to make a start in this field. The first of these is impulsiveness. Impulse buying can occur when a consumer feels a sudden and strong urge to buy an item. Cass and Fenech (2003) found that impulsiveness increased consumers’ perceived usefulness and perceived ease of use of online stores. This is due to websites providing customers verbal and graphical cues that can trigger this impulsiveness. Impulsiveness is also increased and triggered by discounts (Chu, Shen, Liao, 2009). In the case of this study’s forced migration, this could have an effect. As there is a discount when you use the mobile shopping app, impulsive buyers might get triggered by that, caring less that they are restricted to only using the app for the discount. Impulsive buyers are also more activated by discounts. Therefore, it can be hypothesized that:

H3a: The negative effect of forced migration on willingness to buy is moderated by impulsiveness, so that this relationship is weaker for higher levels of impulsiveness.

H4a: The negative effect of forced migration on valence is moderated by impulsiveness, so that this relationship is weaker for higher levels of impulsiveness.

(24)

23

3.5 The effects of innovativeness

The second variable that will be discussed as possible moderating variable is innovativeness. Cho and Workman (2011) found that consumers high in fashion innovativeness prefer to shop from non-brick and mortar stores, choosing for the newer channels. Park and Jun (2002) also supported this, finding that innovativeness was associated with an increase in online shopping.

Innovative people are also more willing to explore new products and methods to buy products and actively recommend them to others (Lee, Son, 2017). Lu (2014) also found that

innovativeness has a direct positive impact on perceived ease of using mobile commerce. This means that innovative people are likely more willing to try out a company’s mobile shopping app and react less negatively to a promotion for such an app. Therefore, it is hypothesized that:

H3b: The negative effect of forced migration on willingness to buy is moderated by innovativeness, so that this relationship is weaker for higher levels of innovativeness.

H4b: The negative effect of forced migration on valence is moderated by

innovativeness, so that this relationship is weaker for higher levels of mobile orientation.

3.6 The effects of Mobile Orientation

The last variable that will be discussed as a possible moderating variable for forced migration is mobile orientation, or how experienced consumers already are with mobile shopping. Consumers that already have affinity with their mobile phone and that have previous experience with mobile shopping are more likely to buy something on their phone in the future (Bigné, Ruiz and Sanz, 2007). This could be explained by multiple factors.

(25)

24 Schoenbachler and Gordon (2002) hypothesize that the more familiar the consumer is with the internet, the less risky they perceive the risk of shopping from multiple channels. This could extend towards being familiar with mobile shopping as well. Lu and Su (2009) found an important indirect factor in intention to engage in mobile shopping to be mobile skillfulness, which can lead from being experienced with it. This increase in intent to engage in mobile commerce might reduce the effect of forced migration on reactance, due to

consumer already planning to use the mobile anyway. This is supported by Trampe, Konuş and Verhoef (2014) who found reactance in response to migration to be lower for customers that are already using the new channel. Thus, it is hypothesized that:

H3c: The negative effect of forced migration on willingness to buy is moderated by mobile orientation, so that this relationship is weaker for higher levels of mobile orientation.

H4c: The negative effect of forced migration on valence is moderated by by mobile orientation, so that this relationship is weaker for higher levels of mobile orientation.

4. Research Plan/Design

4.1 Method

To test the model and its hypotheses an experiment is conducted through a cross-sectional online survey. Every respondent will have to fill in 23 questions related to the hypotheses. The questions are a combination of Likert-scale questions for all variables except mobile orientation and the control variables, which are also asked through multiple choice questions.

(26)

25 This study would benefit from using actual sales data from a company who is trying or

successfully forced their customers to their mobile shopping app. This study however has no access to such data.

4.2 Experimental Design and Survey structure

This study has a repeated measures design. This is done so that fewer respondents are needed for accurate results (Girden, 1992). This design can cause bias or fatigue due to being asked similar questions multiple times. To counter these disadvantages all scenarios are in random order and the questions concerning the dependent variables are limited to one per variable.

Because of the repeated measures design the survey will include all types of migration and product category, assigned randomly to each other. This means that each respondent will get a scenario related to voluntary, rewarded and forced migration and they will see each of these tied randomly to either an clothing, electronics, or flight tickets retailer. To minimize the potential bias and to ensure as many people as possible can participate in this study, no actual retailers are used. After a short introduction to the survey the respondents first be shown a short message that they should assume it is from their favourite store in that category. Also included in the message is that they will be shown three different sales and that they are ongoing in both physical and online stores unless mentioned otherwise, thus for the forced migration scenario. After these scenarios are shown and the respondents answered the questions relating to them, the moderating variables are tested. And lastly the controlling variables are asked.

To ensure there were no mistakes and everyone can fill in the survey without issues multiple pre-tests were done on different layouts of this survey (van Teijlingen & Hundley, 2001). Each time three to five family members and friends filled in the survey and gave their feedback on layout, comprehensibility and time required.

(27)

26 Together with the people who did the pre-tests, the survey was translated into Dutch so that both Dutch and foreign people can fill it in and so that the Dutch version is as close as possible to the English without it being confusing.

4.3 Variables

In this section all variables will be discussed, which ones they exactly are, what type of questions will be asked and how they can be answered.

4.3.1. Independent Variables

How the independent variables will be measured can be seen in figure 2. Every respondent will be shown voluntary, rewarded and forced migration examples together with a randomly assigned product category.

Figure 2. Survey Design

Type of migration Example of scenario

Voluntary Your preferred clothing retailer has a sales promotion ongoing where every product is discounted by 10%.

Rewarded Your preferred electronics retailer has a sales promotion ongoing where every product is discounted by 10% and if you buy from them through their mobile shopping app you will get an additional 5% discount for a total discount of 15%.

Forced Your preferred flight tickets retailer has a sales promotion ongoing where every product is discounted by 10%, however this sale is only valid if you buy through their mobile shopping app.

(28)

27

4.3.2 Dependent Variables

The dependent variables are intention to purchase, willingness to buy, and reaction. Due to the nature of this study each variable will only consist of one question to minimize the fatigue. This is because each of these variables will be measured three times, thus each additional question for these variables would mean three more questions to be answered.

Intention to purchase will be measured with the question “How likely are you to purchase something from this retailer during this promotion?”. To this question the

respondents answer on a Likert scale of 1 to 7 “Extremely unlikely” to “Extremely Likely”. Channel choice will be measured with the question “Which of these would you buy

through?” and respondents can choose between “their website” and “their mobile app”. The reaction will be measured with the question “How do you feel about this sale?”. This question is also answered on a Likert scale of 1 to 7 but ranged from “Dislike a great deal” to “Like a great deal”.

4.3.3 Moderating Variables

The moderating variables are mobile orientation and the shopping attitudes, impulsiveness and innovativeness. Impulsiveness is measured with a four item scale made by Ridgway, Kukar-Kinney and Monroe (2008). It has a Cronbach’s Alpha of 0.8 and an example is “I consider myself an impulse purchaser”. Innovativeness is measured with a five item scale made by McKnight, Choudhury and Kacmar (2002). It has an alpha of 0.89 and an example is “I like to explore new Websites.”.

Respondents are asked how much they agree with these statements and both of these variables are answered on a Likert scale of 1 to 7. These range from “Strongly disagree” to “Strongly agree”.

(29)

28 Mobile orientation is measured by asking two questions regarding their online

spending. These are “Roughly how much of your total spending in the past half year has been done online?” and “Roughly how much of your online spending in the past year has been done through a mobile phone or tablet?”. These questions have range answers instead of exact answers. 0% and 100% are individual ranges so that these individuals can be clearly identified.

4.3.4 Controlling Variables

The controlling variables are age, gender, nationality and the type of operating system that the respondents have on their mobile or phone. These variables are measured with the following questions:

● What is your age? ● What is your gender? ● What is your nationality?

● What operating system do you have on your mobile phone/tablet?

Age ranges from “under 18” then go “18-24” up to “65+”. For nationality the predefined answers are the nationalities that are most expected by the researcher with the option to fill in “other” to not upset respondents, however afterwards they will be combined into two groups “Dutch” and “Non-Dutch”.

This study will also control for the type of operating system on the customers’ phones. This is done because Taylor and Levin (2014) have found that Apple iOS users are more receptive to retail mobile apps than Android users.

(30)

29 As Android has a significantly bigger market share than Apple iOS, with Android having 88 percent market share in 2016 (“Strategy Analytics”,2016) this could blur the results.

The options for operating system are iOS, and Android and others. The market share for other phone operating systems is very small and due to the nature of Android the people with those operating systems should act similar. These were chosen by looking at the market shares that each mobile OS has (Mobile Operating System Market Share Worldwide, march 2018). All scales are in the Appendix.

4.4 Data Collection

The survey for this study will be made and distributed on the online survey software “Qualtrics”. The population for this study can be any consumer across the world as long as they have a smartphone. However due to the type of study and the researcher it will likely be mostly Dutch consumers. Also due to the constraints it will be done using a non-probability convenience sample. Respondents will be contacted through social media and personal messages. Also, survey sharing websites will be used. It is preferable to get as many

respondents as possible during the data collection period of roughly three weeks, however the minimum amount should be between 200 and 300 respondents to ensure the analyses can be done properly.

(31)

30

4.5 Data Analysis

To test the hypotheses the statistical analysis program SPSS is used. For every hypothesis except for H1b the dependent variable is a scale variable and can be tested through linear regressions. H1b however has a binary dependent variable, either the consumer goes through the website or the mobile shopping app. Therefore, to test H1b a binary logistic regression is done. So that the different effects of the moderating variables can be compared to each other in case they are significant, all three interaction terms will be in the same analysis instead of being tested one by one.

5. Results

5.1 Descriptive Statistics

The survey was filled in by 186 respondents. Before anything was done with the data, first the descriptives were checked to see if there were any issues with the data. Nothing wrong was found so all respondents will be used for the analyses. Table 2 gives an overview of the means, standard deviations, reliability and the correlations between each variable.

However, before doing any other descriptives or analyses the frequencies of the variables “Nationality” and “OS” are checked to see if it is fine to converge them into easier to use variables. This means that “Nationality” will become Dutch and non-Dutch and “OS” will become Androids and others, and iOS. Out of 186 respondents, 96 were Dutch and 90 non-Dutch. And out of all respondents, only 4.3% has an OS other than Android (52.2%) or iOS (43.5%). With these numbers it is safe to do the proposed changes to the data.

(32)

31 After these checks, the variables “Innovativeness” and “Impulsiveness” were tested for their reliability, due to them being multi-item variables. To check the reliability,

Cronbach’s Alpha was checked. Both Impulsiveness (α = .877) and Innovativeness (α = .817) are good. The Cronbach’s Alpha of Innovativeness can be increased to .824 by removing “In general I am not interested in trying out new websites”, however this increase is very small and the Cronbach’s Alpha is already high, so it is kept.

After this, the Pearson correlations between the variables were examined. Due to the within subjects design of this study and the nine different scenarios the general mean of Willingness to Buy, Channel Choice and Valence were used. This was done because

otherwise the correlation matrix would have too many different variables and become too big to easily extract information from. These correlations can be seen in table 2.

(33)

32 The first strong correlation that is significant is between nationality and age (r = -.546, p < .01). Due to the way the sampling was done, the non-Dutch respondents (M = 25.889) are on average younger than the Dutch respondents (M = 40.750). This can cause skewing for all correlations between variables and nationality and age, where it is not necessarily that

specific variable that causes it. However, both are only controlling variables, so it will not have a main effect on the hypotheses.

Besides the more obvious correlations, between valence and willingness to buy (r = .726, p < .01) and between online and mobile spending (r = .416, p < .01), the strongest correlation is between willingness to buy and innovativeness (r = .393, p < .01). This gives some initial support for the hypothesized moderation of innovativeness on willing to buy. Also, important to note is the correlation between channel choice and willingness to buy (r = .353, p < .01). Due to the nature of the survey and this correlation matrix it can be said for certain till proper regressions are done, however it can give support to this study by showing that people who would choose the mobile app more are more willing to buy. This does however have some connotations to it which will be discussed in the discussion section. Lastly it shows correlations between all dependent variables and moderators except for a few exceptions. Impulsiveness appears to not correlate with channel choice and valence, and online spending appears to not correlate with channel choice. The latter should however not matter as an effect between the two is not hypothesized.

5.2 Hypotheses Testing

To conduct this study two types of regressions are used. For most hypotheses linear regression is used, however those involving channel choice have to be done with binary logistic regression due to channel choice having only two outcomes. Every hypothesis is tested for each category, meaning all hypotheses require 3 analyses.

(34)

33 First hypothesis H1 and H2, the direct effects of the type of migration on the

dependent variables is tested.

H1a: Rewarded migration has a significant positive effect on willingness to buy and valence so that willingness to buy and valence are higher when there is rewarded migration than if there is voluntary or forced migration.

H1b: Rewarded migration has a significant positive effect on channel choice so that people are more likely to choose the mobile app when there is rewarded migration than if there is voluntary migration.

To test H1a standard linear regression was performed to measure the effect of rewarded migration on the dependent variables. For H2b the only case of binary logistic regression in this study had to be used. Due to a mistake in the survey channel choice was only measured for voluntary and rewarded migration. Due to this the sample size is lower (clothing: 125, electronics: 123, flight: 124) and this can not be calculated for forced

migration. For channel choice Nagelkerke R square is used. Table 4 shows that for the most part the hypotheses H1a and H1b can be accepted (results be seen in table 3). For the category clothing all three relations with rewarded migration are positive and significant, however the relationship with channel choice is the only one with p < 0.01. For electronics rewarded migration is positive and significant for both willingness to buy (Beta: 0.266, p < .05) and channel choice (Exp(B): 28.547, p < .01), however it is insignificant for valence. For the category flight tickets, it is the other way around, channel choice (Exp(B): 12.363, p < .01) and valence (Beta: .196, p < .01) are significant, but willingness to buy is not.

(35)

34 Table 3: Regression results Hypothesis 1

H2: Forced migration has a significant negative effect on willingness to buy and valence so that willingness to buy and valence are lower when there is forced migration than if there is voluntary or rewarded migration.

For H2 and any subsequent hypotheses linear regression is performed (results be seen in table 4). Unexpectedly forced migration does not appear to have a significant effect on willingness to buy in case of the categories clothing and flight tickets. It does however have a significant negative effect for electronics (Beta: -.239, p < .01). There are however significant negative effects on valence for all three, clothing (Beta: .145, p < .05), electronics (Beta: -.145, p < 0.05) and flight tickets (Beta: -.272, p < 0.01).

(36)

35 Table 4: Regression results for Hypothesis 2

H3a: The negative effect of forced migration on willingness to buy is moderated by impulsiveness, so that this effect is weaker for higher levels of impulsiveness.

H3b: The negative effect of forced migration on willingness to buy is moderated by innovativeness, so that this effect is weaker for higher levels of innovativeness.

H3c: The negative effect of forced migration on willingness to buy is moderated by mobile orientation, so that this effect is weaker for higher levels of mobile orientation.

For these three hypotheses unexpectedly only one interaction effect was found to be significant (results be seen in table 5). This is the moderating effect of mobile spending on the relationship between forced migration and willingness to buy for electronics, meaning H3c is supported in the case of electronics but not the other categories. This moderating effect is however small (Beta: .199, p < .05) and the direct effect of forced migration on willingness to buy is not significant anymore for electronics.

(37)

36 The only direct effect of forced migration that is still significant is that of forced migration on willingness to buy in the case of clothing (Beta: -.580, p < .05). In this case the direct effect of innovativeness is also significant (Beta: .194, p < .05). Innovativeness also has a direct significant effect in the case of electronics (Beta: .185, p < .05) and flight tickets (Beta: .225, p < .05). Lastly there’s a significant effect of impulsiveness on willingness to buy in the case of electronics (Beta: .204, p < .05).

That there are no other significant effects can however indicate there is indeed an effect of the interaction terms, as besides the cases of innovativeness with clothing and mobile orientation with electronics, the significant effects found in H2 are non-significant now. However, the interaction terms themselves are not significant so it can not be said with certainty that it truly is thanks to a moderating effect. This means that H3a and H3b are not supported for all product categories. Though interestingly, the interaction term with

impulsiveness is almost significant at the <0.05 level in the case of electronics (Beta: -0.326, p: 0.074). However, it is a negative effect, so that would mean that if someone is impulsive and they are forced to the mobile shopping app they would be even less willing to buy. H3c is supported for electronics but not supported for clothing and flight tickets.

(38)

37 Table 5: Regression results for Hypothesis

a _{0 = male, 1 = female;}b _{0 = Dutch, 1 = Non-Dutch;}c _{0 = Android and others, 1 = iOS}

Clothing Electronics Flight Tickets

Dependent Variable Willingness to Buy Willingness to Buy Willingness to Buy

Coefficient Beta p-value Coefficient Beta p-value Coefficient Beta p-value Direct Effects Constant 3.360 3.058 2.600 Forced -2.171 -.0580 0.033 -0.855 -0.249 0.301 -0.681 -0.175 0.499 Impulsiveness 0.045 0.039 0.658 0.230 0.213 0.018 0.096 0.079 0.398 Innovativeness 0.284 0.194 0.027 0.250 0.185 0.033 0.344 0.225 0.012 Mobile Orientation 0.007 0.106 0.217 0.007 0.101 0.248 0.002 0.029 0.747 Interaction Effects ForcedxImpuls 0.050 0.051 0.785 -0.287 -0.326 0.074 -0.223 -0.230 0.228 ForcedxInnov 0.389 0.421 0.092 0.165 0.197 0.402 0.240 0.263 0.315 ForcedxMobile 0.002 0.022 0.830 0.019 0.199 0.049 0.10 0.096 0.362 Control Variables Age -0.007 -0.058 0.504 -0.003 -0.028 0.736 -0.007 -0.053 0.544 Gendera _0.145 _0.041 _0.578 _-0.191 _-0.059 _0.416 _-0.046 _-0.012 _0.867 Nationalityb _-0.066 _-0.019 _0.823 _0.172 _0.053 _0.509 _0.630 _0.171 _0.043 OSc 0.418 0.257 0.107 -0.081 -0.025 0.728 0.090 0.024 0.738 R2 _0.184 _0.256 _0.177

(39)

38 H4a: The negative effect of forced migration on valence is moderated by

impulsiveness, so that this relationship is weaker for higher levels of impulsiveness. H4b: The negative effect of forced migration on valence is moderated by innovativeness, so that this effect is weaker for higher levels of mobile orientation.

H4c: The negative effect of forced migration on valence is moderated by mobile orientation, so that this effect is weaker for higher levels of mobile orientation.

Similar to the results of hypothesis 3, the results for hypothesis 4 are also unexpected (results can be seen in table 6). There are even less significant effects found for these than for

hypothesis 3. No significant direct effect of forced migration is found in these regressions. This despite the direct effect of forced migration on valence being significant for every category if there are no interaction terms included. This means there might be some effect, however as the interaction term is not significant nothing can be said with certainty. This means H4a. H4b and H4c are all rejected for all product categories.

(40)

39 Table 6: Regression results for Hypothesis 4

Clothing Electronics Flight Tickets

Dependent Variable Valence Valence Valence

Coefficient Beta p-value Coefficient Beta p-value Coefficient Beta p-value Direct Effects Constant 3.917 3.733 4.438 Forced -.1383 -0.445 0.108 -0.131 -0.044 0.864 -1.010 -0.341 0.175 Impulsiveness 0.018 0.018 0.837 0.135 0.145 0.134 0.015 0.016 0.861 Innovativeness 0.171 0.140 0.118 0.149 0.127 0.171 0.154 0.133 0.123 Mobile Orientation 0.008 0.136 0.122 0.005 0.097 0.301 0.002 0.038 0.663 Interaction Effects ForcedxImpuls 0.114 0.138 0.469 -0.221 -0.291 0.138 -0.151 -0.204 0.271 ForcedxInnov 0.129 0.169 0.508 0.061 0.083 0.740 0.164 0.236 0.354 ForcedxMobile 0.003 0.031 0.763 0.012 0.149 0.167 0.004 0.049 0.636 Control Variables Age -0.001 -0.069 0.945 -0.004 -0.037 0.675 -0.012 -0.114 0.180 Gendera 0.043 0.015 0.846 -0.123 -0.044 0.572 0.104 0.037 0.608 Nationalityb 0.188 0.064 0.451 0.219 0.078 0.366 0.493 0.177 0.032 OSc 0.370 0.126 0.092 0.217 0.077 0.316 0.193 0.069 0.332 R2 0.144 0.142 0.221

(41)

40 Table 7: Summary results of Hypotheses

Hypothesis Description Clothing Electronics Flight

Tickets H1a Rewarded migration has a positive effect on

willingness to buy and valence.

Supported Supported Partly Supported *

H1b Rewarded migration has a positive effect on channel choice.

Supported Supported Supported

H2 Forced migration has a negative effect on

willingness to buy and valence.

Partly Supported ** Supported Partly Supported **

H3a Impulsiveness has a moderating effect on the effect of forced migration on willingness to buy. Not supported Not supported Not supported

H3b Innovativeness has a moderating effect on the relationship between forced migration and willingness to buy. Not supported Not supported Not supported

H3c Mobile orientation has a moderating effect on the effect of forced migration on willingness to buy. Not supported Supported Not supported H4a

Impulsiveness has a moderating effect on the effect of forced migration on valence.

Not supported Not supported Not supported

H4b Innovativeness has a moderating effect on the effect of forced migration on valence.

Not supported Not supported Not supported H4c Mobile orientation has a moderating effect on

the effect of forced migration on valence.

Not supported Not supported Not supported *Relationship with channel choice supported, willingness to buy not supported.

(42)

41

6. Discussion and Conclusion

6.1 Discussion and Theoretical implications

Having more channels to reach the consumers with is a goal companies should go for as consumers who use multiple channels are more profitable and look more positively at the company (Blattberg, Kim, Neslin (2008); Kumar and Venkatesan (2005). A relatively new channel is the mobile shopping app, which some companies are already using and have been downloaded millions of times, such as Amazon. Getting consumers to use this channel is likely profitable as both purchase frequency and size go up when consumers migrate to mobile (Wang, Malthouse, & Krishnamurthi, 2015). However, the most effective type of migration to this new channel has been and still is unclear. Therefore, this study tried to find if it might be an appropriate strategy to force consumers to the mobile shopping app. The main previous research that this has been based on is Trampe, Konuş & Verhoef (2014) who looked at the effect of different types of migration on reactance.

Instead of looking at the effect of migration reactance this research however looked at the effect on several behaviours towards the migration and different consumer traits that might have an effect on this relation. It was found that rewarding consumers with an extra discount if they use the mobile shopping app does indeed increase the willingness to buy, valence and the likelihood consumers would use the mobile shopping app. Only in the case of flight tickets it did not increase willingness to buy, which was an expected possibility. A possible explanation for this is given by Wang, Malthouse and Krishnamurthi (2015). They found that consumers who do already use their mobile phone to make purchases mostly do so with habitual products.

(43)

42 Electronics are not habitual products either however people have more experience buying these already. Flight tickets however are rarely bought, with most consumers buying them yearly or even less frequently.

Interestingly forced migration did indeed have a direct negative effect on valence towards the sales promotion in all categories, however it only had a negative effect on willingness to buy in the case of electronics. A possible explanation to explain this is how people respond when asked how willing they are to buy something together with the scenario. People often say they are more willing to buy or intent on purchasing something than they truly are (Jamieson, & Bass, 1989). The willingness to buy flight tickets was already the lowest out of the three product categories so possibly respondents do not want to go below a certain point when they respond to such a question. Also, in the presented scenarios the discount is only 10%, or 15% if they use the mobile shopping app in the rewarded case. Flight tickets are often bought when there are bigger discounts than this, or through auction websites.

Unexpectedly none of the other expected effects were found except for the

moderating effect of mobile experience in the case of electronics. The moderating effects that were expected to be found did not exist in any case. Impulsiveness was expected to help alleviate the negative effect of forced migration on valence and willingness to buy as consumers were expected to feel an impulse to still use the discount despite having to use a mobile shopping app. And it was also found that impulsiveness made consumers perceive online stores more useful and easy to use (Cass, & Fenech, 2003), which were expected to translate to mobile shopping apps as well, again alleviating the effect of forced migration. Instead, the opposite was found and it might be the case that customers who score high on impulsiveness react stronger on forced migration than non-impulsive customers.

(44)

43 This might be explained by affective-impulsive people having stronger reactance (Steindl, Jonas, Sittenthaler, Traut-Mattausch, & Greenberg, 2015). These people react more quickly on their emotions than reflecting on changes. It is possible that there is a relationship between people who are more likely to react impulsively on emotions and people who are more

impulsive with purchases. This could also explain the almost significant effect found of the moderating effect of mobile spending in case of electronics, where impulsive consumers are less likely to buy when they are being forced to a mobile shopping app.

The degree of how innovative consumers feel also had no moderating effect on the negative effects of forced migration. Innovativeness was found to have a direct effect on willingness to buy, however it does not appear to have a moderating effect. No clear reason could be found in previous research. It appears that innovative people are willing to try new ways, however they do not want to be forced to do so. They maybe want to try it on their own accord.

And for mobile experience as well, only moderating effect was found for the product category electronics. Intuitively this is surprising as it can be expected that people who already use a platform would be fine to use it for discounts. Trampe, Konuş & Verhoef (2014) found that customers who did already use the new channel in their study still had reactance under two scenarios. They still had reactance when they were forced to change channel because the previous one was removed or if they were punished if they wanted to keep using the previous channel. The forced scenario in this study however is more alike to their rewarded scenario and under that scenario already experienced consumers had less reactance. One possible explanation might be the difference in channel, they investigated a customer service channel while this research investigates purchase channels.

(45)

44

6.3 Managerial Implications

Companies are increasingly using a mobile shopping app as a new channel to get customers to purchase more and to collect more data on these customers. Some companies are already putting their focus on their app instead of their website without knowing the implications of this. This study tried to help these companies understand more about possibilities for this migration however much is still unclear. It is shown that forced migration will have an effect on valence, meaning that their consumers will react negatively to such migration. However, it is unclear if this actually will result in reduced revenues and if people will refuse to use the app or not. It is shown however that it is more likely that rewarding customers for trying the app increases their willingness to buy and has a positive reaction. Thus, it is recommended for companies to weigh the benefits of the lower risk of losing consumers by giving this extra reward or if they truly want their entire customer base on the mobile app which might result in a loss of some consumers.

Sadly, this research did not make help answer the question if there is a way to get people to be more willing to migrate or if there are consumer traits that can be targeted for a migration to mobile shopping apps. Hopefully future research will be done to help companies find answers to these questions so that they can more safely increase their mobile shopping app usage.

6.4 Limitations and Suggestions for Further Research

This research provides a starting point for research into migration to mobile shopping apps, however it also has some limitations. The first limitation is the measurement of willingness to buy. Jamieson and Bass (1989) found that respondents typically far overestimate their future behaviour regarding purchasing products.

(46)

45 Considering the nature of this research it is difficult to measure the consumers’ behaviour more accurately, as is the case with many similar studies. Results would be more accurate if company data was used. The issue with using actual company data however is multifold. Few companies are currently focussing on migrating towards a mobile shopping app, and when they are they are still likely still in the early stages. And as this study has shown there is support for there to be differences in reactions to both rewarded and forced migration and product categories, thus a full study using actual company data would need data from different kinds of companies for the most accurate results.

Another limitation is the lack of channel choice for the scenarios involving forced migration. Due to a mistake while making the survey this variable is missing, despite it being valuable data. It could have shown us if people were still willing to actually try the mobile shopping app if the discount was only for that channel, or if people would keep buying through the website. This should be included in future studies, unless these studies looks at a true forced migration where previous channels are removed.

This also relates to the third limitation and suggestion for future research. Trampe, Konuş and Verhoef (2014) had their forced migration involving the removal of the previous channel and their rewarded migration had a reward for using the new channel. This study however had the forced migration being like their reward migration while the rewarded migration had a discount, albeit smaller, for the website as well. Websites and mobile shopping apps use many similar resources so it could make sense to keep both which can make this study’s forced migration look more realistic. However, it might also prove for some companies to be better to get rid of their website. To account for this, future studies could include more scenarios.

(47)

46 Thus, you would have the voluntary scenario, a rewarded scenario where consumers get discounts on both channels but a bigger one for the app, a rewarded scenario where

consumers only get a discount for the app, and lastly a forced migration where the website is removed.

Lastly as this study found almost significant moderations, future studies should look at different variables that might change the effect of forced migration. An example could be that loyalty could reduce the negative effect. Or loyalty could possibly increase the negative effect as they feel accustomed to the current channel that they do not want to move to another.

How does forced migration to mobile shopping apps affect intentions/willingness to buy : the moderating effect of impulsiveness, innovativeness and mobile shopping orientation