A study of consumer behaviour in the digital world - The power of alternatives

(1)

Bachelor Thesis

Bachelor's Program in International Marketing, 180 credits

A study of consumer behaviour in the digital world - The power of alternatives

Independent project in business administration, 15 credits

Halmstad 2021-05-25

Ludvig Mohlin, Eric Nilsson

(2)

(3)

Preface

This bachelor thesis in business economics aims to treat consumer behaviour in a digital context. The thesis was written during the authors last year of the International Marketing program at Halmstad University. The authors would like to thank Ulf Aagerup, who has provided the authors guidance and expertise during the process. The authors are also thankful to all respondents who participated in this thesis experiment. Hopefully findings of this thesis could enrich and influence companies and societal knowledge, which the authors are proud of. During the process of this report authors received a deeper knowledge within the area which they believe to be valuable in future marketing challenges.

(4)

Abstract

Title: A study of consumer behaviour in the digital world - The power of alternatives Date: 2021-05-25

Grade: Bachelor thesis in Business Economics Author: Ludvig Mohlin & Eric Nilsson

Supervisor: Ulf Aagerup Examiner: Svante Andersson

As COVID-19 has led to more desktop usage overall it could be said that the screen size has increased for the general consumer. Previous research shows that what device a consumer uses when browsing the internet also affects their behaviour and more particularly their decision making. Previous research regarding consumer behaviour also shows that consumers when exposed to different amounts of alternatives act differently when making a decision.

Connected to devices and their obvious distinction when it comes to screen size and therefore the amount of alternatives displayed to the consumer on the screen makes it interesting.

Most of the previous research regarding alternatives and consumer behaviour had not been done in a digital context and this thesis was aimed at studying how consumers react to

different amounts of alternatives in a digital context. The authors got interesting contradictory results from the two different AB-tested experimental questions indicating a more complex explanation to the causality of the research into alternatives and consumer behaviour. The author's main conclusion is that consumers' attitudes and then preferences towards products influenced by their internal search could have an explanation to why some products are more advantageous to be displayed as large assortments than others.

(5)

1. Introduction

This chapter aims at introducing, describing and elaborating the background for the selected area of study. Mainly discussed here are the disruption of the COVID-19 pandemic and the changes in consumer behaviour. The different aspects of this are being further discussed in the problem discussion and finally the purpose of the study is motivated.

1.1 Background

The year of 2020 was one of the most disrupting years in a long time. The pandemic

COVID-19 changed the global economy but also the way people act and behave. Due to the COVID-19 pandemic, its restrictions and lockdowns, people spend more of their time at home which have turned their attention and spare time to the only world at hand, the online world. The use and time spent on mobile devices, desktops and television has increased in the last decade and in connection with the lockdown due to COVID-19 the device usage of all kinds has increased rapidly (Bahkir, 2020).

The increase of time spent on digital devices due to COVID-19 and its restrictions could be compared with the usage pre lockdown. From Bahkir (2020) study who compared daily use of digital devices prelockdown to lockdown it’s shown that 94 percent of the participants replied to an increase of hours spent on digital devices during lockdown. From this, 50 percent logged an increase in average daily use of 5 hours.

It is essential to view this in a wider perspective, to see and understand the use of the internet pre- and during lockdown. In March, when the lockdown held people at home, the internet experienced an increase of 15-20 percent in a week. Studies show how the internet behavior at pre lockdown was allocated to working days, where the internet usage increases in the evenings compared to a weekend behavior where the usage starts at 9:00 AM and continues all day. Due to lockdown the weekday patterns are minimized and all days are like weekday patterns (Feldman et al., 2020).

(7)

Recent studies also show that proliferation and quarantine due to COVID-19 lead to an increase in time spent on digital screens. Even though quarantine is temporary, this development seems to continue. A shift in how we behave when it comes to digital usage may continue even after the pandemic (Wai et al., 2020).

1.2 Problem discussion

Consumer online behaviour is constantly changing and with the development of technology and several devices consumers now have the opportunity to choose devices. Desktop and mobile devices are the most popular devices where consumers search for information. Since the development in technology in the last decades the mobile device has brought the

opportunity for individuals to be anywhere and still be able to search for information. Mobile devices are more comfortable compared to desktops and, in general, people bring their mobile everywhere because it is a part of our identity and desktops, who’s harder to bring anywhere, are in general more frequently used for job purposes (Haan et al., 2018). This brings a new type of consumer behaviour. While consumers always have access to their mobile device, and it is the device they use on the go, behavior of time spent on devices online increases on mobile and decreases on desktop.

While technology in the recent years has improved drastically with use of the internet, (mobile, desktop and television), did the behavior of the individual follow. But since the start of the pandemic the location and therefore the devices people use seem to change.

According to a recent article in the New York Times Americans are turning away from their mobile devices for desktop and web usage instead, at least in some cases. Facebook has seen an increase for their app by 1.1 percent but 27 (!) percent increase for their website in the period 15th of January to 24th of March. In the meantime Youtube has had a decrease on their app by 4.5 percent and an increase on their website by 15.3 percent (Koeza & Popper, 2020). Thus, with the Koeza and Poppers (2020) article in mind it is not a longshot to think that there might be a change in online consumer presence due to COVID-19 and its

consequences. While people stay at home and work from home the use of desktops increases and since individuals are not on the go the use of mobile decreases. It is interesting to see the correlation between consumers' online behaviour and their real life behavior. There could be a chance that peoples’ new behavior, if they are home or on the go, at work etc., has an effect

(8)

on their choice of which device they use when conducting information search and browsing the internet.

With a new light on the selection of devices by consumers during the pandemic it is interesting to study how different situations may differ depending on what device the consumer uses. For instance, being able to view more or less alternatives on the screen because of the size of the screen when shopping on the internet could have an impact on the consumer's decision making. Research has been done before by Iyengar and Lepper (2020) on consumers' decision making when getting exposed to limited or extensive choices, but not so much in a digital context. Even though businesses could gather this information on their own in terms of what works best for them, the authors believe a deeper dive into the reasoning behind it is of interest.

1.3 Purpose

Because of the pandemic and its consequences of shifting the consumer behaviour in the course of what online device prefered by consumers the authors believe the distinction between the different online devices is of great importance. The greatest distinction from our perspective is the screen size which affects the alternatives shown. Studies on alternatives and consumer behaviour have been done before but not so much in a digital context. The authors of this thesis therefore have the intention to research how consumers make decisions based on the amount of alternatives they are exposed to in a digital shopping context.

(9)

2. Frame of reference

In this chapter previous literature and studies within the chosen area will be presented followed by selected theories to support and develop understanding for what the authors are going to examine.

The selected theories aim to give the reader a clearer insight into the study itself and a wider perspective of its importance.

2.1 When choice is demotivating

According to Jessup et al. (2009) economic theories propose that more alternatives are better and that people should prefer choosing from a wide range of options to find their most valued alternative. However recent studies show something called too-much-choice effect (a.k.a.

TMC). A study done by Iyengar and Lepper (2000) examined consumer behaviour when faced with different amounts of options, in this case different quantities of flavours of jams.

The study was conducted in order to see how consumers reacted when faced with an aisle of an extensive (24) different flavours compared to a limited (6) different flavored jams. The researchers’ two dependent measures of consumers’ motivation were their initial attraction and then their subsequent purchasing behaviour. The results showed that consumers who passed the extensive-selection display of jams got more attracted to the jams than the people who passed by the limited-choice condition. In other words the extensive-selection display won in terms of initial attractiveness. But when researchers looked at the other measure it turned out that consumers initially exposed to the limited-choice proved considerably more likely to make a purchase than the group of consumers exposed to the extensive-selection.

This is strengthened by a meta-analytic review made byScheibehenne et al. 2010 (gathered 50 studies, both published and unpublished) which shows that being exposed to an increased number of alternatives will decrease decision-making and is considered as “too many”

alternatives.

Greifeneder et al. (2010) also says that consumers prefer an increased amount of alternatives compared to decreased alternatives. In contrast they also report that increased number of alternatives correlated with decreased decisions.

In an online context, assortments are depicted differently compared to in store offline. The digital interface increases demand for attention from consumers due to the device and size of screen which needs a more narrow focus (Kahn, 2017).

(10)

2.2 Less is more

The conventional thought about the number of alternatives often relates to “more is better”. A larger assortment is more likely to satisfy consumers with a product that they like (Baumol &

Ide, 1956). Thereto a larger assortment also gives possibilities to satisfy unsure or searching consumers (Kahn & Lehman, 1991).

But, how come we don’t see unlimited assortment at any place? This is not a coincidence, in recent studies authors indicate that unlimited offerings affect consumers differently.

Wauxman’s (2004) presents a quote from the New York Times in his article pronounced by a salesman; “I was told never to show customers more than three pairs of shoes. If they saw more, they would not be able to decide on any of them.”

This assertion is strengthened by recent studies. Kuksov and Villas-Boas (2010) say that more alternatives in the assortment could lead to confusion among consumers to find the right product and may also irritate them. In addition, more alternatives also leads to decreased attraction to the product (Shang et al., 2010).Several experiments have also proven this intuition to be true. Iyengar and Lepper (2000) study how the number of chocolate bars exposed to a consumer could affect their decision-making. They investigated participants faced an extensive size of alternatives (30), compared to a limited size of alternatives (6).

Their findings indicated that the extensive assortment made it more complex in the decision-making process for participants compared to participants in the decision-making process of a limited assortment. This is in line with Boatwright and Nunes (2001) findings that less alternatives offered in a store resulted in greater sales. Again, Iyengar and Lepper (2000) investigated if there could be too few, too many or just the right number of

alternatives disposable. Participants exposed to 30 alternatives responded that 30 was seen as

“too many” compared to participants exposed to 6 alternatives who responded that 6 samples was “just right”. This finding was also significant. This is in line with Jessup et al. (2009) TMC- effect and Iyengar and Lepper (2000) choice overload. These effects are an indication

(11)

Conversely, consumers of non exciting preferences or consumers who are unfamiliar with the assortment have a more complex decision-making when faced with a larger size of

alternatives.

More shopping is frequently starting online and assortments are exposed via a digital context.

This underlies the importance of how the assortment are depicted. According to Kolesova and Singh (2018) an assumption should be made that products displayed in a digital context should imitate product offerings offline. Strengthened by Kahn (2017) who says that in an environmental context “less is more” and how small assortments online will be easier to process and be more appreciated among consumers.

2.3 Consumer behaviour & device usage

Relationship between consumer behaviour and size of assortment depends on preferences.

Spassova and Isen (2013) differ consumer preferences into liking. If consumers have positive preferences for the product a greater assortment size increases decision-making intentions in contrast to consumers with neutral preferences where a larger assortment makes

decision-making more complex. In contrast, neutral preferences on a small assortment size increases probability of making a decision where the difference on positive preferences on a small assortment size compared to a larger size is modest. Kuksov and Villas-Boas (2010) findings also indicate that consumer preferences are important. Their finding that mainstream consumers prefer a decreased base of alternatives and the more specific and extreme the consumer is, the greater the range of alternatives are to prefer.

Interesting combination into this is the consumer search phase, they talk about consumers who are willing to search prefer a decreased size of alternatives. Other way around, where consumers have low search frequency the alternatives offered should be larger. But, if the alternatives offered are too large consumers may find it too difficult to make a choice and completely stay out of the market. To find alternatives that match their preferences may take too much time of their searching phase. Vice versa with offering too few alternatives, may also lead to consumers staying out of the market. Because consumers feel that the alternatives offered are too far away from their preferences. Result estimates fewer alternatives best probability to consumers making a choice. In other words, consumers with low search frequency, impulse choice behaviour, prefer a greater size of alternative and a smaller size of alternatives offered is to prefer if the search frequency is greater. Studies have also identified inequalities among consumer behaviour on the various devices. One thing that distinguishes

(12)

device behaviour is time spent, it is shown that consumers who browse on desktops spend more time compared to mobile users, who seem to be more effective. Consumers who use mobile base their decisions on emotions and instinct while desktops users seem to be more rational and base their decisions on logic and thoughts. Thus, consumers are more likely to look and seek for information and find new products on their mobile and read and find out details about the product on their desktop (Jiang et al., 2018).

With mobile devices smaller screen consumers think they are more likely to miss out on something in the evaluation process and in the end miss out or purchase another product then they were supposed to. The small screen size also brings problems when the consumers are in the transaction phase (Haan et al., 2018). While consumers seem more comfortable with making rational decisions on desktops according to the bigger screen (Jiang et al., 2018).

2.4 Screen size and behaviour

In a digital context consumers are exposed to alternatives through different devices. These devices are dependent on various screen sizes. In a digital context, different consumers are exposed most through desktop and mobile devices. Search behaviour on mobile devices separates from search behaviour on desktops. This is mostly due to the difference in screen size. Mobile devices have smaller screen sizes which perform less search results. Studies about behaviour on various devices have been done,Tourangeau et al. (2017) made a study on inequalities among devices (desktop versus mobile) through conducting a web survey.

Results showed that participants using a desktop, in contrast to mobile, spent more time finishing the survey, performed longer answers on questions with own reflection alternatives and decreased response level on items and less coherent answers.Kim et al. (2015)

strengthened these findings where results showed that individuals search behaviour is different on various devices. Smaller screens entails more exertion for selection of the information with less eye-movements then on a wider screen.This seems to be in line with Jiang et al. (2018) estimation that mobile users behave more on intuition than desktop users

(13)

an important finding in Kim et al. (2015) study where results showed that participants needed extended time to obtain information on the smaller screen. Time was also examined in the Iyengarand Lepper (2000) experiment with chocolate. They investigated time on deciding if the participants were exposed to 6 chocolates or 30 chocolates. Results showed that

participants spent longer time to decide on which chocolate on the 30 chocolates compared to deciding on one of the 6 chocolates. This finding was also significant.In Söderlund et al.

(2019) study it says that screen size has an influence on the exposure of images to the receiver. A bigger screen also leads to a higher level of attention. The larger the screen the more attention is shown from the user (Söderlund & Sagfossen, 2015). Mobile devices are used by an increased number of individuals in various digital contexts (Tourangeau et al., 2017). Söderlund et al. (2019) says that screen size has an influence on behaviour. Desktops generated more positive answers among participants in contrast to mobile. But they also mention their findings as modest. Furthermore, Tourangeau et al. (2017) made findings that the use of mobile or desktop does not have an influence on behaviour.

Information search on mobile devices means decreased amount of information shown and more scrolling to reach more information. This could lead to negative effects and to miss out on some important information, therefore individuals choose to search on a desktop in special search events (Sweeney & Crestani, 2006). The study constructed bySöderlund et al. (2019) shows that when individuals search for information they prefer various devices on various occasions. Their result showed that mobile devices are not the device an individual chooses to search for new information, it's the device they use when they already have some

knowledge about the product, a habitual behaviour among the individual and they tend to buy brands or products they know about. On other occasions they use a desktop. This type of habitual behaviour on mobile devices also increases spendings in comparison to desktop and a more rational behaviour (Wang et al., 2015).

2.5 Consumer decision-making process

When trying to understand how consumers make decisions and how they search information the consumer decision-making process is a useful tool. The consumer decision-making process could be split up into five different steps as shown at Figure 1.1. (1) problem recognition, (2) information research, (3) evaluation of alternatives, (4) product choice and (5) post-purchase evaluation. In this study the second step, information search, is going to be used. Information search is the process where a consumer searches in their environment for

(14)

data appropriate to make a decision that is reasonable. This could be broken down into two different kinds of search: internal and external. Even if we are experienced shoppers or not, we all have some degree of knowledge already stored in our memory regarding previous purchases or general information about products picked up from living in a consumer culture.

When first engaging in information search we may first use internal search to scan our own bank of memory to find anything useful to assemble information about a product. But in most cases we lack some information and need additional knowledge which leads us to engage in external search where we obtain information from advertisements, friends or watching what other people do. Many customers, especially veterans, enjoy browsing just for the joy of it and to be able to stay updated on what’s going on in the marketplace. These types of consumers could be described as being engaged in ongoing search. Two of the motives for consumers engaged in ongoing search is to experience fun and pleasure but also to build a bank of information. Two outcomes of being engaged in ongoing search could be increased impulse buying and increased satisfaction from search and other outcomes (Solomon, 2016).

Figure 1.1. Consumer decision-making process (Solomon, 2016).

(15)

2.6 Devices

2.6.1 Mobile

Mobile devices include smartphones and the most common and repeated form are Apple's iOS and Android from Google (Jiang et al., 2018).

The use of mobile devices increases and is mostly associated with all the daily activities that individuals are doing. Planning, socializing and communicating, browsing, among others (Jesdabodi & Maalej, 2015). Mobile devices are easy to bring, the possibility to alway have the device available at any place is an advantage (Islam et al., 2010). Add the ability to access the internet everywhere (Daurer et al., 2015). This means that consumers could be reached at any time and make purchases at any place (Islam et al., 2010).

Mobile devices are a central figure in consumers' lives today. With the availability mobile devices bring, it becomes more than just a tool to communicate with friends and family, it is being used as an extension of their own personality. While consumers see their device as an extended identity with private influence the marketers see it as a marketing channel (Persaud

& Azhar, 2012).

Mobile devices seem to be the most popular and most frequently used device among behavior of online shopping. Through the increase of mobile devices the behaviour has changed

according to a smaller screen, other functionalities etc. Consumers prefer to use their mobile device in the path of searching for products and prices online (Jiang et al., 2018). But, studies show that smaller screen, lowercase letters and the lower battery time entails increased search costs, on the contrary the ability to bring it anywhere could decrease search costs.

Mobile behavior consists of browsing the website and using the smartphone applications.

Through mobile applications the consumer could be reached precisely and individually relevant information through location-based systems through the application.

2.6.2 Desktop

Desktop could be distinguished and defined as a device where the user needs to have a constant Wi-Fi-connection in contrast to a mobile where the user can tap into a cellular network in order to stay connected to the internet (Kenney, n.d.).

Desktops are used in work related and personal conditions. Browsing the internet, watching video, communicating etc. are common things to do. With a bigger screen size and a larger

(16)

memory to store information, compared to mobile devices, it makes the desktop better functional in special occasions.

Consumers prefer to use a desktop when reading about various products and their details.

This is connected to the screen size, with a bigger screen size consumers feel to perceive more details on the site and about the product that their emotions is to make better and more rational decisions (Jiang et al., 2018).

2.7 Hypotheses

Hypothesis 1 (H1) ‘The more alternatives the consumers are exposed to in a digital context the less probability there is of them taking a decision.’

Hypothesis 2 (H2) ‘The consumers' attitudes towards the two products (eyewear and shoes) in the experiment will be different.’

(17)

3. Methodology

In this chapter methods and research approaches chosen for this study will be described and motivated. Methods will be explained and then subject for method critique.

3.1 Research approach

A hypothesis was formed to test in an empirical study in the form of an experiment in accordance with a deductive research approach (Söderbom & Ulvenblad, 2016).

Quantitative research methods could be simplified as research implying measurement, data collection, statistical processing methods and analytics methods with the intention to illustrate questions relating to quantity, frequency and then also the connection between variables, cause and effect (Patel & Davidson, 2019). Therefore a quantitative research method was believed to be most optimal for the study.

The authors are aware that deductive research connected with quantitative methods may involve conditions where theories are tested generally on those who are susceptible to the reach of quantification, with consequences that generalizability to theories may be unsure. To avoid this deductive theory being associated with positivism and bias from the authors, the hypotheses are tested to the resultant objective process (Bitektine, 2008).

3.2 The construction of the experiment

In order to create a control group and one experiment group A/B-testing (a.k.a. split-testing) was used. Two identical surveys were created, Survey A and Survey B, the only difference here was the amount of alternatives to choose from on each picture where Survey A (6 alternatives) had fewer alternatives than Survey B (24 alternatives). Survey A was supposed to simulate a mobile context with less alternatives and Survey B a desktop context with more alternatives representing respective screen sizes. The respondents were asked to rate how likely they were to purchase one of the products seen in the image. Two experiment questions (Q1 & Q2) and two images per survey were constructed, one image with shoes (Product A) and one image with eyewear (Product B). Two different kinds of products were used in order to study if there would be any difference of results depending on what type of product was displayed to the consumer.

(18)

All of the products in the images were unisex in order to make the product selection gender neutral. The images were supposed to resemble a digital shopping context scenario and the images therefore were supposed to look like a screenshot directly from an e-commerce site, in this case Nike and Chimi Eyewear. To be able to strengthen the validity of the experiment the additional alternatives added were not in terms of new products. Instead the alternatives in Survey A were duplicated four times in Survey B creating 6 respectively 24 samples. In other words the different surveys were designed as Survey A (6 samples, small screen size = mobile context) and Survey B (24 samples, large screen size = desktop context) and it should be noted that participants' use of the device did not influence whether they randomly received experiment A or B.

In order to randomize what survey each respondent would get the two links to each survey were merged into one link where the respondents were randomly redirected to one of the two surveys when clicking the link. The merging of the links was done using the website

www.allocate.monster.

The two experiment questions in the survey had answer alternatives designed with an uneven Likert Scale. The scale is structured with odd numbers to reach a neutral mainstay. A five point scale provides highest reliable answers and highest reliability to the study. The five point scale is to prefer with respect to (1) time for the participant to select an alternative and (2) Participants' satisfaction with their answer (Wakita et al., 2012). The choice of a five point scale also provides respondents ability to select a neutral alternative which could erase

inaccurate data. Potential of the scale is to measure opinions, attitudes and behaviours among respondents due to a particular theme. It is in the authors best interests to rely on respondents but also emphasize pitfalls correlated to inaccurate answers (Paulson & Ursing, 2020).

Likert Scale was in this experiment done by describing each value as followed:

1 = Stämmer inte alls 2 = Stämmer delvis

(19)

3.2.1 Web survey

The authors decided to use a quantitative method in the form of a web survey to conduct the experiment. A survey means that the research gets done on a larger defined group with the help of a questionnaire. This gives researchers the possibility to gather information about a larger number of variables and likewise it can give us a larger number of information about a limited number of variables (Patel & Davidson, 2003). Data is essential for a thesis to test theories and hypotheses to understand situations and processes. Situations and processes imply behaviour and changes that will emerge when the data becomes new available useful information (Stopher, 2012).

One of the cons of using surveys is that the respondents can see the whole survey and its answers and adapt their answers from that which leads to variables not getting studied independently from each other. Another con is that if the respondent doesn’t understand the question there is no one there to help which otherwise could have been done during an

interview (Patel & Davidson, 2019). Therefore the survey was created so only one question at a time could be answered with no possibility to view the rest. The respondents were asked questions as clearly as possible without difficult formulations or words opening up for personal interpretation.

Pitfalls tried to avoid when constructing the survey were long questions, leading questions, double-barrelled questions, prerequisites and ‘why’-questions (Patel & Davidson, 2019). The authors tried to avoid pitfalls like this because they wanted to gather as much reliable data as possible to make the study as purposive as possible.

The authors are aware of difficulties it could imply to get respondents to conduct the survey.

Respondents in general need to be encouraged by personal interests or gain benefits of conducting the survey. Also, importance in emphasising anonymity in the survey (Paulson &

Ursing, 2020). To solve this, it is important to encourage respondents by relating their personal interests to this studys’ purpose.

A decision was taken to make the experiment as a web survey because of the positive aspects a survey brings. First of all because it is easier to adapt the survey to respondents' needs in the form of for example time of completion. Secondly because of the administration of study that becomes easier and quicker to handle (Patel & Davidson, 2019).

Lugtig and Toepoel (2016) says that web surveys' outcome is affected by how respondents conduct the survey. Designers of the survey need to think of what devices the respondent

(20)

uses. Considerations on devices and different screen sizes imply new challenges for designers. Respondents' use of different devices could imply selection and measurement fauls.

3.4 Sample Selection

This survey was decided to be sent out via email to students at Halmstad University but also uploaded on a community site on Facebook with members allocated to Halmstad University.

The choice of using convenience sampling implies that the survey will reach a large number of participants. Where to distribute could be a critical issue, both concerning the number of participants and distribution to reach a representative survey response. Convenience sampling implies critical issues, especially connected to the fact that the survey participants may not be a representative population. This issue is critical for the study itself and reliability of the conclusions to be made (Bryman & Bell, 2011). Thus this issue, the authors considered convenience sampling as the most proper method and considered both options as representative although the whole sample was not investigated in detail.

Respondents were given the green light to spread the survey further in order to create a snowball selection (Patel & Davidson, 2019). As the survey was published in community groups and mail lists of students, authors believed respondents within the snowball selection would with high probability be people of the same age range.

There must be a link between the sample and the gap this research tends to investigate and if this sample is selected is representative of online consumers as a broader group. Since this report investigates online consumer effects due to COVID-19 and quarantine, students at higher level schools seem to be a proper group of individuals since all of the Universities are recommended to be closed and students to study from home (Regeringskansliet, 2020).

3.5 Research process

(21)

interesting to examine how a desktop context differs from mobile context through a consumer behaviour perspective. Authors found it relevant to focus on the size of the screen and how it could have an impact on what is displayed to the consumer. In this case they found it

interesting to research the area of alternatives connected to decision making and consumer behaviour. The authors began to read already published research in the area of alternatives and decision making and saw a gap where research in a digital context was not as much researched. The main theoretical framework found and used was the ‘consumer decision making process’ where the authors identified the second step ‘information search’ to be focused on. The authors decided to construct an experiment in order to collect empirical data for their research.

3.6 Data collection

Primary data

The primary data was received through an experiment with the help of a survey. The survey was sent out via email and uploaded on Facebook on a wednesday morning. It was a

conscious decision to invite participants to conduct the survey first on Wednesday morning.

Faugh et al. (2016) investigated how to reach a survey's best potential and claims timing to be a successful factor to enhance response rate. Results showed that a survey's best potential is significantly reached on Wednesday mornings which is the time where it obtains more responses than any other day, and in conclusion that timing matters. The authors decided to close the survey when it reached the amount of wanted participants.

Secondary data

A literature review was conducted to gain a greater overall understanding of the already existing research. The literature review was done using the database of Halmstad University and for a few occasions Google Scholar. Keywords used alone and combined being such as:

consumer decision making, alternatives, mobile & desktop devices, COVID-19, consumer behaviour, alternatives, options, selections, etc.

3.7 Quality of research

Reliability

(22)

The reliability of the study depends on indispensable quality measurement. Reliability through web surveys could imply failure data. Acceptance to incorrect data does not directly affect degree of reliability, but should be taken under consideration when estimating degree of coincidences (Patel & Davidson, 2019). To gain greater reliability of the study and experiment two different types of products acted as samples, in this case eyewear and shoes were tested to measure the product variable also. In order to get unbiased selection of

respondents the two respondents groups were randomised as explained in the construction of the experiment.

Validity

The validity of measurement for this thesis was behaviour in an online digital context. The measurement of online behaviour could imply critical issues considering participants' opinion towards certain products and may not therefore be in line with the participants actual online behaviour. Nevertheless, this issue was considered with alternatives which could imply avoidance from this issue. Internal validity was taken into consideration since this thesis aimed to measure decision-making in a digital context. The authors tried to prevent respondents from biases in their decision-making to reach the highest level of causality (Bryman & Bell, 2011). Therefore the experiments were A/B tested where respondents saw product A and product B in either a simulated mobile context (6 samples) or in a simulated desktop context (24 samples). The samples in both surveys were the same, only multiplied, in order to rule out the effect of the respondents seeing more unique alternatives that probably would increase the possibility of liking in the extensive assortment (desktop) survey.

Considering external validity, distribution between respondents' ages were located to a

younger audience. It was taken into consideration that the majority of younger individuals has had an experience as an online consumer once, often or on a regular basis. Therefore could the result from this thesis be applicable and generalized to online consumers as a group (Söderbom & Ulvenblad, 2016).

(23)

spent regardless of what their answer was, comparing it to Iyengar and Lepper (2000) study where a purchase decision actually was made or not.

The added samples could have been in the form of new samples. However the authors then thought the results would be too predictable because more samples meant more products that could catch liking and as stated above in a digital context and web survey where respondents do not lose money by “making a purchase decision” it would not study the desired behaviour and variables.

(24)

4. Results & Analysis

In this chapter results in the form of quantitative data from the experiment conducted will be presented in tables. The results will then be analysed using the main theoretical framework and previous research to gain a greater understanding of the results. M = mean SD = standard deviation.

4.1 Results from experiment questions

Below the results from the A/B-tested experiment questions will be presented in tables. Total number of respondents who answered the survey and were part of the experiment was 174 respondents.

4.2.2 Experiment 1 (Product A, eyewear)

Depicted from Figure 4.1 respondents’ answers from the first experimental question (Q1) are shown in tables. Results showed participants who saw 6 samples of the glasses (M = 1.09, SD

= 0.286) did differ in probability of making a decision to those who only saw 24 samples (M

= 1.29, SD = 0.454). Results showed a mean difference of 0.197 with a p-value of 0.000***.

Hypothesis 1 is therefore supported in Q1.

Survey A (6 samples) Survey B (24 samples)

Figure 4.1 Experiment responses from question “It is likely that I will buy a pair below”.

(25)

(M = 1.12, SD = 0.326). Results showed a mean difference of 0.148 with a p-value of 0.013*.

Hypothesis 1 is therefore not supported in Q2.

Survey A (6 samples) Survey B (24 samples)

Figure 4.2 Experiment responses from question “It is likely that I will buy a pair below”.

4.2.4 Comparison of attitude (Product A vs. Product B)

Depicted from Figures 4.1 and 4.2 respondents’ answers from the two experiments results showed participants' combined answers on Product A (M = 2.51, SD = 1.346) did differ compared to the combined answers on Product B (M = 3.41, SD = 1.386). Results showed a mean difference of 0.90 with a p-value of 0.000***. Hypothesis 2 is therefore supported.

4.3 Evaluating Statistical Significance

In order to conclude if there was any statistical significance an Independent Sample T-test was done to evaluate the statistical significance. Results presented in table below.

Hypothesis 1 will be using Likert Scale as followed:

(1), (2), (4), (5) = Consumer takes a decision Neutral mainstay (3) = Consumer takes no decision

(26)

Hypothesis 2 will be measured using Likert Scale in the common way comparing the differences of the mean of the scores in order to determine any differences between Product A and Product B.

A-B Q1 = Comparison of Product A 6 samples vs. 24 samples.

A-B Q2 = Comparison of Product B 6 samples vs. 24 samples.

A-B Q1-2 = Comparison of Likert Scale means between total scores on Product A vs total scores on Product B.

Compared Sig. (2-tailed) Mean difference

A-B Q1 0.000*** 0.197

A-B Q2 0.013* 0.148

A-B Q1-2 0.000*** 0.90

Figure 4.3 Retrieved data

4.4 Analysis of hypotheses

4.4.1 Hypothesis 1 (H1)

‘The more alternatives the consumers are exposed to in a digital context the less probability there is of them taking a decision.’

Experiment 1 (Product A, eyewear)

Results from the simulated online context experiment in A-B Q1 (6 image items vs. 24 image items) (mobile vs desktop) showed that participants in the environment with 6 samples image

(27)

to deciding to purchase and to not decide to purchase. It also goes in line with previous research which shows that a decreased number of alternatives are ultimate in an online context because it’s easier to process. This is because too many alternatives will make it too complex in the decision-making (Kahn, 2017, Jiang et al., 2018, Kuksov & Villas-Boas, 2010).

Experiment 2 (Product B, shoes)

Results in A-B Q2 (6 image items vs. 24 image items) showed a smaller number of decisions in 6 sample images (M = 1.27, SD = 0.445) compared to 24 sample images (M = 1.12, SD = 0.326). The difference was found to be significant p < .013 (*). Which would be considered to be the other way around from product A, the greater number of alternatives the higher the probability of making a decision, both positively and negatively. Somewhat similarly Baumol and Ide (1956) and Kahn and Lehman (1991) found that larger assortment size increases probability of consumer making a purchase decision because of the increase of probability to find something that they like.

In conclusion this means that the author's manipulation, with the 6 samples of Product B and Product A duplicated four times to 24 samples was successful in creating a difference.

Contrary, this thesis findings of Survey B (24 samples) Q2 (Product B) differ from other findings. A quote could explain a scenario which does not correlate to thesis findings in Q2:

“I was told never to show customers more than three pairs of shoes. If they saw more, they would not be able to decide on any of them.” (Wauxman, 2004).

There is a lack of findings from former research investigating decision-making in a digital context. Especially considering difficulties in decisions-making when exposed to extensive or limited items shown. In Q1 with Product A most decisions were made in the image with 6 samples and in Q2 with Product B most decisions were made in the image with 24 samples.

Similarly to this result, former research also indicates inequalities in this area. Results from Iyengar and Lepper (2000) who compared extensive assortment (30 samples) to limited assortment (6 samples) made the findings that an extensive assortment made it more difficult in the decision-making process compared to a limited assortment size. They also found that participants exposed to 30 alternatives saw it as “too many” alternatives compared to

participants exposed to 6 samples who saw it as the “just right” number of alternatives. Both

(28)

of these findings were significant (p < 0.0001). Contrary did Chernev (2003) and Mogilner et al. (2008) found in their results that a larger assortment could increase probability of

decision-making. This previous research combined with this thesis experimental findings shows how consumer behaviour connected to alternatives is diffuse and it is obvious as the authors expected that other factors play a role.

Söderlund et al. (2019) found that screen size has an influence on behaviour.Jiang et al.

(2018) said that consumers feel that they perceive more details on a larger screen. A larger screen size, which could be considered a desktop context, generates possibilities to offer a larger number of alternatives as the authors did in Survey B with 24 samples. Kim et al.

(2015) made findings that consumers behave differently on the various devices. They also made findings where smaller screens led to harder times to find the right information and also in processing the information which could be the case in Survey A and Q2 with Product B where consumers who saw 6 samples had a more difficult time to make a decision compared to Survey B and 24 samples of Product B.

According to Sweeney and Crestani (2006) a scenario caused by information search on a smaller screen leads to less information shown and an increased amount of scrolling to reach all the information available. Findings from Kim et al. (2015) also highlight that extended time is needed on smaller screens to obtain all the available information which could be an considered factor in inequalities between Survey A Q1 (Product A) to Survey B Q1 (Product B) where more decisions were made on an decreased number of alternatives. Contrary, results from Survey A Q2 to Survey B Q2 indicate the different conditions where more decisions were made on an increased number of alternatives. Jiang et al. (2018) also identified inequalities on the various devices. But on the contrary his results indicate that consumers who browse spend more time on desktop than on mobile. Findings also showed that desktops users also seem to be more rational thinkers in their decision making. Furthermore did mobile users seem to be less rational and more effective and base decisions on instinct and generally

(29)

4.4.2 Hypothesis 2 (H2)

‘The consumers' attitudes towards the two products (eyewear and shoes) in the experiment will be different.’

When comparing overall and combined attitudes from both Survey A and Survey B regarding the two different products a difference was found in the mean which turned out to be

significant p < .000 (***). Results showed that consumers have a more positive attitude towards product B (M = 3.41, SD = 1.386) compared to product A (M = 2.51, SD = 1.346).

This allows the authors to support hypothesis 2. The authors interpret the differences as respondents overall generally favoured product B over product A.

These results indicate that a consumers' attitude towards a product and in the end their preferences could affect their decision-making because the difference in attitude in H2 correlates to the contradictory results in H1.Findings in this study showed that Survey A Q1 had a higher frequency of making a decision compared to Survey B Q1 which showed lower frequency of making a decision. Other way around on Q2 where Survey A Q2 showed lower frequency of making a decision compared to Survey B Q2 where more samples showed higher frequency of making a decision. This indicates that the morepositive attitude towards a product the more decisions are being made in contrast to if the consumer has a less positive attitude it decreases the probability of them making a decision. Spassovaand Isen (2013) somewhat strengthen this with their results where neutral preferences make decision-making more complex on a larger assortment compared to positive preferences. In contrast neutral preferences on a smaller assortment increase the probability of making a decision compared to positive preferences. Positive or existing preferences then increase decision-making on a larger assortment. Conversely do neutral or non existing attitudes compex the

decision-making (Chernev, 2003,Mogilner et al., 2008).

As stated, results from Q2 showed an increased decision making when faced with a more extensive assortment and the products B in Q2 were overall more preferred than product A in Q1. Both Spassova and Isen (2013) and Kuksov and Villas-Boas (2010) found something similar where they found consumer preferences to be an influencing factor in

decision-making. This thesis research and findings could indicate that consumers who have

(30)

more positive preferences towards a product prefer a greater size of alternatives and a consumer with neutral preferences or less positive preference are less likely to make a decision in this case because of the complexity of choice, and therefore prefer a smaller size of alternatives. This could be connected to the consumer decision making process and information search and the phase of internal search where preferences are built in several ways as we build a bank of information in our memory and associate positive or negative preferences towards the product. The findings of this thesis indicate that a greater number of alternatives decrease consumers' decision intentions in Survey B (24 samples) Q1 (product A) which could then be because of a lack in bank of memory or a negative preference from memory. This could make decision-making more difficult and complex. And interestingly enough, Survey B (24 samples) Q2 (product B) indicates that a more positive preference towards the product seems to increase decision making.

Kuksov and Villas-Boas (2010) relates this to consumers in the searching phase, where one type of consumer maximizes utility from a larger number of alternatives and the other maximizes their utility with a decreased number of alternatives. In result consumers that are exposed to a large amount of assortment in their searching phase feel that there are too many alternatives and find it too complex and time demanding to find any alternative that matches their preferences. This leads in the end to consumers don’t making a decision at all and completely staying out of the market. The other way around where consumers feel that there are too few alternatives they may think that the alternatives do not maximize their utility and they can find other alternatives that match their preferences better.

Noteworthy both Iyengar and Lepper (2000) and Greifeneder et al. (2010) also report a disruption in the number of alternatives presented affected by consumer preference. It seems that depending on the product the number of alternatives plays a different role. It also seems that depending on the consumers the number of alternatives matters.

This thesis experimental findings could be further explained by information search and its category ongoing search. Outcomes from ongoing search is either impulse behaviour or

(31)

is an individual who is low at ongoing search, seen as a specific information searcher who always knows what to look for. The two contradictory findings from Survey A (6 samples) and Survey B (24 samples) could be analysed through the ongoing search perspective.

Participants in Survey B (24 samples), had a hard time making a decision in Q1 (product A), and could in this case be seen as a mainstream searcher but when comparing respondents answers in Q2 (product B) where they had an easier time making a decision it seems not so likely because then they are seen as a specific information searcher. This goes for

respondents in Survey A also. This is another indication that there is something else affecting the decisions made by the respondents.

Authors find another connection to previous research who say that consumers' preferences and liking towards the product influence decision making and has an impact on how many alternatives that are shown which could be correlated to devices and their various screen sizes (Spassova & Isen, 2013, Kuksov & Villas-Boas, 2010). This strengthens the importance of H2 because results in Q1 (product A) shows that participants seem to have, what Spassova and Isen (2013) calls a neutral preference for the products where a larger size of 24

alternatives (desktop) make decision-making more complex compared to smaller size of 6 alternatives (mobile). In contrast to Q2 (product B) where participants seem to have, what Spassova and Isen (2013) calls a positive preference towards the product, which leads to more decisions when exposed to 24 samples which indicates a desktop environment in contrast to 6 samples and a mobile environment.

(32)

5. Conclusion

In this chapter the authors will present their main conclusions together with possible implications. In the end the authors discuss future research and their desired direction of this.

The purpose of this study was to research consumer behaviour related to alternatives and then connect it to screen sizes of mobile and desktop devices. The reason for this was because of an obvious disruption in device usage and increase in desktop usage because of COVID-19.

Most of the previous research says that less alternatives shown is the way to go and an

increased number of alternatives are estimated as “too many” and makes the decision-making more difficult due to the TMC-effect (Iyengar & Lepper, 2000). The findings of this thesis provide evidence for and against this as it showed contradictory results from their two experimental questions, Q1 and Q2 in H1. The authors believe that just concluding that desktop because of its possibility of extensive assortment and more alternatives leads to less purchase decision is an oversimplification and this thesis proves that it is probably not correct. There is something else also affecting the consumer in the decision stage when faced with different amounts of alternatives.

Experimental Q1 (product A) indicates that participants made more decisions on the mobile simulated Survey A (6 samples) compared to experimental Q2 (product B) where more decisions were made on desktop simulated Survey B (24 samples).

The authors believe the results in H2 could be offered as an explanation to the contradictory results in H1 as there was a difference in attitude and then preference among the respondents when it came to the products. In other words the results in H1 could be explained by the respondents' preferences towards the product or brand (Nike or Chimi) where the preference seems to affect the effect of alternatives. Respondents were more likely to make a decision when faced with an extensive assortment (Survey B) the more positive attitude they had towards the product. Contrary, respondents were less likely to make a decision the less positive and more negative preference they had when faced with an extensive assortment

(33)

exactly clear what type of products are appropriate for displaying with more alternatives vs.

less alternatives to what customer and why. As stated the authors believe it has to do with preference perhaps influenced by the memories when conducting internal search. It could be the case in this study and experiment that the consumers already had a positive or negative perception of the brands (Nike and Chimi) which they fetched from their previous

interactions with the brand or that they just lacked interest in the product overall. In other words it could be that consumers have an easier time taking a decision when faced with large assortments and alternatives if they already have a positive perception about a brand or have a preference for a certain product in general. Vice versa consumers then have a harder time to make a decision when faced with large assortments and alternatives if they have a negative or neutral perception of the brand or lack of interest in the product overall.

When connecting this to usage of devices and their screen sizes two important insights are found. Access to a desktop extends the size of assortment faced by the consumer which could enrich decision-making when they have positive preferences toward a product. While

negative preferences or less positive preferences toward a product could decrease

decision-making. Conversely, access to a mobile limits the size of assortment faced by the consumer which decreases decision-making when they have positive preferences towards a product. While negative preferences or less positive preferences towards a product enrich decision-making.

5.1 Implications

It is worth stating that the authors of this thesis believe that for an e-commerce company a consumer making a decision is more favourable than a consumer not making a decision at all.

The findings of this thesis could indicate some implications valuable for e-commerce.

The most rational belief is that the more samples displayed increase the possibility that consumers find something that they like and in the end lead to more purchase decisions. But as this thesis showed, it is not that easy. E-commerce should be aware that depending on what product, what device and then in the end the individual consumer itself the amount of

alternatives to be displayed may have to be selected for each individual case. The authors propose different realistic recommendations for implications below.

Recommendation 1

(34)

While consumer behaviour shifted to desktop usage consumers were faced with an extensive number of alternatives. E-commerce retailers can easily collect data from consumers with high ongoing search and make an estimation about their preferences towards certain products.

They could see which products or brands consumers visit or buy regularly and make an estimation that most consumers have positive preferences towards these products or brands.

These products and brands should be offered with an extensive number of alternatives to increase decisions being made, and in the end enhance the probability of consumers making a purchase. Similarly through collecting data from consumers' ongoing search they could identify which of their products or brands consumers have a neutral preference for by seeing which products or brands consumers do not visit regularly. These products should be in a limited number of alternatives to increase decisions being made, and in the end enhance the probability of them making a purchase on these products.

Recommendation 2

E-commerce retailers should take consumer preferences into consideration. An estimation from the authors to solve consumer preferences to the amount of alternatives shown could be to help consumers orientate through categories. To reduce TMC-effect to consumers with negative or less positive preferences towards a product could be to offer consumers

possibilities to reduce the assortment to 6 samples through a category to increase decisions being made. Other ways around should retailers offer the possibility to increase the number of alternatives through a category to consumers with positive preferences to increase

decision-making. As a scenario the consumer might need to buy a pair of sunglasses as a gift to someone but know nothing about sunglasses and find them rather dull and boring. Then, if consumers are offered the possibility to filter to face the 6 most popular alternatives through a category it could increase decisions being made..

5.2 Future research

The ways of researching how alternatives affect consumer behaviour feels like a never ending

(35)

they affect the TMC-effect and alternatives would be highly interesting. Even though there are theories both from previous research and from the authors themselves a narrow and deeper dive into this area would be of interest.

(36)

6. References

Bahkir, FA. (2020). Impact of the COVID-19 lockdown on digital device-related ocular health. Indian J Ophthalmol. 68(11), 2378-2383.

https://www.ijo.in/article.asp?issn=0301-4738;year=2020;volume=68;issue=11;spage=2378;e page=2383;aulast=Bahkir

Barmpatsalou, K., Cruz, T., Monteiro, E., & Simoes, P. (2018). Current and Future Trends in Mobile Device Forensics: A Survey. ACM Journals, 51(3), 46.

https://dl-acm-org.ezproxy.bib.hh.se/doi/pdf/10.1145/3177847

Bauer, C., & Strauss, C. (2016). Location-based advertising on mobile devices: A literature review and analysis. Management Review Quarterly, 66(3), 159-194.

doi:http://dx.doi.org.ezproxy.bib.hh.se/10.1007/s11301-015-0118-z

Baumol, W. J., & Ide, E. A. (1956). Variety in Retail-ing. Management Science, 3(1), 93-101.

Bitektine, A., (2008). Prospective Case Study Design: qualitative methods for deductive theory testing. Organizational Research Methods, 11(1), 160-180.

https://journals-sagepub-com.ezproxy.bib.hh.se/doi/pdf/10.1177/1094428106292900

Boatwright, P., & Nunes, C. J. (2001). Reducing Assortment: An Attribute-Based Approach.

Journal of Marketing, 65(3), 50-63.

https://search-proquest-com.ezproxy.bib.hh.se/docview/227801566?pq-origsite=primo&acco untid=11261

Brasel, S. A., & Gips, J. (2014). Tablets, touchscreens, and touchpads: How varying touch interfaces trigger psychological ownership and endowment. Journal of Consumer

(37)

Canova, L., & Nicolini, M. (2019). Online Price search across desktop and mobile devices:

Evidence on cyberslacking and weather effects. Journal of Retail and Consumer Service. 47, 32-39.

https://www-sciencedirect-com.ezproxy.bib.hh.se/science/article/pii/S0969698918304570

Chernev, A. (2003). Product assortment and individual decision processes.Journal of Personality and Social Psychology, 85(1), 151-162.

Cramer-Flood, E. (2020). Digital Ad Spending by Industry. Emarketer. (website) Access at:

https://www.emarketer.com/content/digital-ad-spending-by-industry-2020[Accessed 20 October, 2020].

Daurer, S., Molitor, D., Spann, M., & Manchanda, P. (2015). Consumer search behavior on the mobile internet: An empirical analysis. Available at SSRN.

https://www.msi.org/wp-content/uploads/2020/06/MSI_Report_15-111.pdf

Dodson, I. (2016). The art of digital marketing : The definitive guide to creating strategic, targeted, and measurable online campaigns. ProQuest Ebook Central

https://ebookcentral-proquest-com.ezproxy.bib.hh.se

Donthu, N. & Gustavsson, A. (2020). Effect on COVID-19 on business and research.

Elsevier Public Health Emergency Collection, 117, 284-289.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7280091/

Faught, K., Whitten, D., & Green, K. (2016). Doing Survey Research on the Internet: Yes, Timing Does Matter. The Journal of Computer Information systems, 44(3), 26-34.

Feldman, A., Gasser, O., & Pujo, E. (2020). The Lockdown Effect: Implications of the COVID-19 Pandemic on Internet Traffic.

https://dl.acm.org/doi/pdf/10.1145/3419394.3423658

Greifeneder, R., Scheibehenne, B., & Kleber, N. (2010). Less may be more when choosing is difficult: Choice complexity and too much choice. Acta Pshycologica, 133(1), 45-50-

(38)

https://www-sciencedirect-com.ezproxy.bib.hh.se/science/article/pii/S000169180900119X

Haan, D E., Kannan, P.K.,, Verhoef, P C., & Wiesel, T. (2018). Device Switching in Online Purchasing: Examine the Strategic Contingencies. Journal of Marketing. 82, 1-19.

https://journals-sagepub-com.ezproxy.bib.hh.se/doi/pdf/10.1509/jm.17.0113

Islam, R., Islam, R., & Mazumder, A T. (2010). Mobile Application and Its Global Impact.

International Journal of Engineering & Technology, 10(6).

https://www.researchgate.net/profile/Dr_Md_Rashedul_Islam/publication/308022297_Mobil e_application_and_its_global_impact/links/5991fbafa6fdcc53b79b606d/Mobile-application-a nd-its-global-impact.pdf

Iyengar, S. S., & Lepper, R. M. (2000). When Choice is Demotivating: Can One Desire Too Much of a Good Thing? Journal of Personality and Social Psychology, 79(6), 995-1006.

https://faculty.washington.edu/jdb/345/345%20Articles/Iyengar%20%26%20Lepper%20(200 0).pdf

Jesdabodi, C., & Maalej, W. (2015, September). Understanding usage states on mobile devices. In Proceedings of the 2015 ACM international joint conference on pervasive and ubiquitous computing (pp. 1221-1225).

https://dl.acm.org/doi/abs/10.1145/2750858.2805837

Jessup, R. K., Veinott, E. S., Todd, P. M., & Busemeyer, J. R. (2009). Leaving the store empty‐handed: Testing explanations for the too‐much‐choice effect using decision field theory. Psychology & Marketing, 26(3), 299-320.

Jiang, T., Yang, J., Yu, C., & Sang, Y. (2018). A Clickstream Data Analysis of the Difference between Visiting Behavior of Desktop and Mobile Users. Data an Information Management.

(39)

Kahn, E. B. (2017). Using Visual Design to Improve Customer Perception of Online Assortment. Journal of Retailing, 93(1, 2017), 29-42.

https://search-proquest-com.ezproxy.bib.hh.se/docview/1881413655?pq-origsite=primo&acc ountid=11261

Kenney, S. (****). ‘Six Key Differences Between Mobile Vs. Desktop For Social Media’

Workspace.digital [online]. Available at:

https://workspace.digital/6-differences-using-mobile-for-social-media-vs-desktop/

[Accessed 9 December, 2020].

Kim, J., Thomas, P., Sankaranarayana, R., Gedeon, T., & Yoon, H-j. (2015). Eye-tracking analysis of user behaviour and performance in web search on large and small screens.

Research School of Computer Science and Technology, 66(3), 526-544.

https://asistdl-onlinelibrary-wiley-com.ezproxy.bib.hh.se/doi/full/10.1002/asi.23187

Klever, A. (2009). Behavioural Targeting: An Online Analysis for Efficient Media Planning?

In Behavioural Targeting: An Online Analysis for Efficient Media Planning? (1. Aufl.).

Diplomica Verlag GmbH.

Koeze, E. & Popper, N. (2020). ‘The Virus Changed the Way We Internet’ New York Times [online]. Available at:

https://www.nytimes.com/interactive/2020/04/07/technology/coronavirus-internet-use.html [Accessed 20 October, 2020].

Kolesova, S., & Singh, R. (2018). One Vs. Many: who wins? An empirical investigation of online product display. The International Review of Retail, Distribution and Consumer Research, 29(3), 285-305.

https://www.tandfonline.com/doi/pdf/10.1080/09593969.2019.1598465?needAccess=true Kotler, P., Armstrong, G., Harris, L. C., & Piercy, N. (2013). Principles of marketing European edition 7th. Essex: Pearson Education Limited.

(40)

Kuksov, D., & Villas-Boas, M. J. (2010). When More Alternatives Leads to Less Choice.

Marketing Science, 29(3), 507-524.

https://www-jstor-org.ezproxy.bib.hh.se/stable/40608162?seq=1#metadata_info_tab_contents

Lugtig, P., & Toepoel, V. (2016). The Use of PCs, Smartphones, and Tablets in a Probability-Based Panel Survey: Effects on Survey Measurement Error. Social Science Computer Review, 34(1), 78–94.

https://journals-sagepub-com.ezproxy.bib.hh.se/doi/full/10.1177/0894439315574248#articleC itationDownloadContainer

Mogilner, C., Rudnick, T., & Iyengar, S. S. (2008). The mere categorization effect: how the presence of categories increases choosers' perceptions of assortment variety and outcome satisfaction. Journal of Consumer Research, 35 (2) (2008), pp. 202-215.

https://academic-oup-com.ezproxy.bib.hh.se/jcr/article/35/2/202/1806103

Patel, R., & Davidson, B. (2003). Forskningsmetodikens grunder. Att planera, genomföra och rapportera en undersökning. Studentlitteratur.

Paulson, E., & Ursing, B. (2020). Sustainable Marketing in Fast Fashion: A quantitative research investigating sustainable marketing strategies’ impact on consumer behaviour. 1-45.

https://gupea.ub.gu.se/bitstream/2077/67689/1/gupea_2077_67689_1.pdf

Persaud, A., & Azhar, I. (2012). Innovative Mobile Marketing Via Smartphones. Are consumers ready?. Marketing Intelligence and & Planning. 30(4), 418-433.

https://www.researchgate.net/profile/Ajax_Persaud/publication/235307685_Innovative_mobil e_marketing_via_smartphones_Are_consumers_ready/links/54108d190cf2f2b29a410fdf.pdf Radova, K. (2016). Fruit or Candy? Ease of justification as a moderator of the effect of

A study of consumer behaviour in the digital world - The power of alternatives

Bachelor Thesis

Bachelor's Program in International Marketing, 180 credits