1 The Impact of Mobile Advertising on Offline Purchase Intentions - The Influence of
Location-Specificity and Perceived Relevance
T. Landheer Master Marketing Student Number: S1915304
23-06-2014
Rijksuniversiteit Groningen - Faculty of Economics and Business
Master Thesis
2 Abstract
Marketing messages such as offers or promotions often reach consumers at inconvenient times or locations, which severely hinders the effectiveness of the advertising. Modern day smartphones are (commonly) equipped with GPS facilities, which allow for the customers being ‘traced’ to a certain location. Many scholars have hypothesized that this unique possibility of location-specificity of mobile phones could increase the effectiveness of advertising, especially in creating purchase intentions. Location-specificity of mobile advertising basically means targeting location-sensitive promotional offers to mobile device users when they are in the vicinity of the advertisers’ location. This thesis has empirically tested the impact of location-specificity of mobile advertising on offline purchase intentions, thereby testing both the impact of location-specificity as well as cross-channel effects. Moreover, the hypothesized moderating effect of perceived relevance was investigated. The results indicate that location-specificity of mobile advertising increases the offline purchase intentions, the moderating effect of perceived relevance over this relationship could not be established. The findings suggest that managers can significantly improve the effectiveness of mobile advertising by using location-specificity, a finding which might be of great importance as consumers become increasingly more reliant on their mobile phone for various activities. Keywords: Mobile Advertising – Mobile Marketing – Location-Specificity – Perceived Relevance Purchase Intention.
Word Count: 13.220
Word Count Main Body Text (i.e. excluding front page, abstract, table of content,
3 Table of Content
1. Introduction 4
2. Literature Review 7
- 2.1 Mobile Marketing, Mobile Advertising – Defining the concept 6 - 2.2 The Impact of Mobile Marketing on Customer Decision-Making 9 - 2.3 Location-Specificity/Geo-targeting of Mobile Advertising 12 - 2.4 The Personal Perceieved Relevance of Mobile Advertising 14
3. Research Design 15
- 3.1 Research Question 15
- 3.2 Theorized Conceptual Model 15
- 3.3 Research Method 16 - 3.3.1. Data Collection 16 - 3.3.2. Sample Frame 16 - 3.3.3. The Survey 16 - 3.3.4. Construct Development 17 - 3.3.5. Pre-Testing 19 4. Analysis 20 - 4.1 Two-Way Anova 20 - 4.1.1. Missing Data 20
- 4.1.2. Data Cleaning and Outliers 21
- 4.1.3. Descriptive Statistics of the Sample 23
- 4.2 Assumption Checks 26
- 4.3 Reliability of the scale 26
- 4.4 Manipulation Checks 27
5. Results 28
6. Discussion 31
7. Managerial Implications 32
8. Limitations and Reflection 33
9. Conclusion 34
10. References 35
Appendix A: Observation with Deviated Mean Times 40
Appendix B: Equal Variances and Normality Tests 41
4 1. Introducing New Media- The Rise of Mobile Marketing
The late 1990’s and beginning of the 21st century observed an enormous increase in the number of different media used by marketers to reach consumers (Winer, 2009). Whereas ‘traditionally’ the major marketing communications media options for marketers consisted of advertising (TV, print, outdoor, radio), promotions, direct marketing, personal selling and public relations and publicity, the internet had enlarged their opportunities in reaching consumers (Winer, 2009, Keller, 2010).
Although traditional media are still heavily used by marketers - US spending for 2011 was $111.5 billion - spending on these traditional media is expected to stagnate over the next years to an expected value of $112.7 for US advertising spending in 2016 (BIA/Kelsey via eMarketer.com).
Alternatively, spending on mobile advertising is taking off as more consumers are becoming increasingly dependent on their mobile device for their day-to-day needs (Shankar and Balasubramanian, 2009). Mobile revenues constituted 15% of total revenue in the first six months of 2013, compared to 7% in the first six months of 2012 and mobile revenue more than doubled from $1.2 billion in the first six months of 2012 to $3 billion this year, an increase of 145% (IAB, 2013). According to Rothenberg (2013): ‘’mobile advertising’s breakneck growth is evidence that marketers are recognizing the tremendous power of smaller screens’’ (Rothenberg, 2013).
The increased spending and revenues in the consumer environment, and expenditures in the firm environment have made ‘mobile advertising’ an attractive area for research (Varnali and Toker, 2010). Yet, the research is in its early stages and therefore highly inconsistent and thus the scope of mobile advertising is still vague (Varnali and Toker, 2010).
5 Marketing messages such as offers or promotions, especially online based advertising, often reach consumers at inconvenient times or locations, which severely hinders the effectiveness of the advertising (Banerjee and Dholakia, 2010). It is hypothesized that both this unique possibility of location-specification - allowing marketers to reach customers when they are in the vicinity of the advertisers’ location, and could therefore increase the effectiveness of advertising - as well as the progression in location-based services will lead to mobile advertising becoming the ’killer application’ (Kölmel, 2003; Bauer et al. , 2005; Banerjee and Dholakia, 2008; Shankar and Balausbramanian, 2009; Rothenberg, 2013).
Although several authors have proposed the impact of location-specificity (e.g. Bauer et al., 2005; Shankar and Balasubramanian, 2009), none of them have empirically tested their propositions.
This thesis attempts to test whether location-specificity using a mobile platform actually works in creating offline purchase intentions. Empirically testing this relationship result will not only bridge the gap between theory and empirical research surrounding location-specificity of mobile advertising, it will also provide managers with practical insights in discovering whether or not location-specific mobile advertising is an effective tool in creating offline purchase intentions. If this thesis would provide evidence in support of location-specificity, managers could increase offline purchase intentions of consumers by actively using location-specific mobile advertising.
6 More precisely, this thesis addresses the following research question: ’What is the impact of location-specificity of mobile advertising on offline purchase intention and is it strengthened by the perceived relevance of the mobile advertising?’
This research question caused two hypotheses which will be investigated to arise. The hypotheses are: H1: Mobile advertising has a stronger effect on offline purchase intention
when the advertising is location-specific and H2: Higher perceived relevance of the product
offering strengthens the relationship between mobile advertising and offline purchase intentions.
7 2. Literature Review
2.1 Mobile Marketing, Mobile Advertising – Defining the concept
A review conducted by Tahtinen (2006), illustrated the confusing usage by both practitioners and academics of the terminology surrounding the phenomenon of mobile-based marketing/advertising. Likewise, a review conducted by Leppäniemi et al. (2006) concerning mobile research discovered 21 different definitions. Both the reviews by Tahtinen (2006) and Leppäniemi et al. (2006) as well as criticism by Balasubramanian et al. (2002) discovered that the definitions used by academics and practitioners largely focused on the technology itself as opposed to trying to identify the aspects which differentiate mobile from ‘other’ online media and could be translated into marketing implications/opportunities. It follows that there is no consensus of an explicit definition of mobile marketing or mobile advertising which includes all facets of the phenomenon, resulting in the terms being used interchangeably. Therefore, a characterization of the terms used will be made, based on prevailing, yet not specific to mobile-context literature.
According to Baker (2008), a marketing strategy allows firms to focus their limited resources on the greatest opportunities to increase sales and attain a sustainable competitive advantage. Homburg et al. (2008) provide a more practical reasoning, ‘’the marketing strategy is implemented by systematically deploying the marketing mix’’ (Homburg et al., 2008). The marketing mix comprises out of four different components: product, price, communication and sales. It follows that, from this traditional perspective, a definition of mobile marketing should incorporate these components.
Advertising in a traditional sense is defined as any form of paid communication by an identified sponsor aimed to inform and/or persuade target audiences about an organization, product, service or idea (Belch & Belch, 2004; Tellis 2004; Yeshin, 2006; Fennis 2010). It is clear that the definition of advertising only involves one of the previously mentioned four components, namely the communication component. It follows that, from this traditional perspective, a definition of mobile advertising should solely incorporate the communication component.
8 i.e. solely utilizing the communication component using the mobile phone as a platform. Nonetheless, to better understand advantages offered by mobile advertising a classification of the distinguishing features of the mobile phone should be conducted.
An important benefit of a mobile phone is the size and weight. Mobile phones are small, they fit in one’s hand/pocket, and wireless (in both power supply and tethering) resulting in very easy transport, i.e. mobile phones offer easy portability (Balasubramanian et al. 2002; Shankar and Balasubramanian, 2009). Mobile phones allow consumers to reach (and be reached by) firms and other consumers (largely) regardless of time and place, i.e. mobile phones offer ubiquity (Hennig-Thurau et al., 2010). The portability distinguishes the mobile phone from desktops, which are usually wired. Moreover, this portability of a mobile phone allows for it to be a constant companion to consumers, who in turn become increasingly dependent on their mobile device for their day-to-day needs (Shankar and Balasubramanian, 2009). Nonetheless, the small screen size also has its disadvantages; the small sized screen constrains the possibility and effectiveness of information-intensive marketing communications (Shankar and Balasubramanian, 2009).
9 Lastly, due to technological advances, many mobile phones (particularly smartphones) have GPS capabilities which allow for physical location identification (Bauer et al, 2005; Shankar and Balasubramanian, 2009). This combined with the fact of the portability of the mobile phone, and the fact that consumers are becoming increasingly dependent on their mobile phone for day-to-day needs i.e. always carry their mobile phone with them, allow for eradication of geographical and informational barriers by reaching consumers when they are nearby advertisers’ locations (location-specificity/geo-targeting).
As illustrated in the sections above, research has indicated that the mobile phone has several inherent advantages; portability, uniqueness and personalization, interactivity and location-specificity, as compared other ‘new’ online media and traditional mass media. A definition of mobile advertising should therefore incorporate these aspects as well as the underlying component of communication identified earlier. Consequently two definitions are constructed based on previous definitions introduced by Tahtinen (2005), Leppäniemi (2008) and Shankar and Balasubramanian (2009), Belch & Belch (2004), Tellis (2004), Yeshin (2006) and Fennis (2010). The first definition regards mobile marketing: ‘Mobile marketing is the two-way or multi-way communication of an offer between a firm and its customer with the possibility for location-sensitive, personalized information using a portable, mobile phone’’.
The second definition regards mobile advertising: ’Mobile advertising is the paid communication of an offer from an identified sponsor with the possibility for location-sensitive, personalized information using a portable, mobile phone to inform and/or persuade target consumers’. These definitions will be the foundation on which this thesis will proceed.
10 2.2. The Impact of Mobile Marketing on Customer Decision-Making
The impact of mobile advertising on consumer behavior, customer preferences and decision-making has been a key issue in research over the last years. According to Shankar and Balusbramanian (2009) research compiled over the years has indicated that mobile marketing directly impacts all the five stages of customer decision-making. The five stages of customer decision-making are: need recognition, information search, alternative evaluation, purchase decision and post-purchase (Häubl and Trifts, 1999).
Rigby (2010) argued that mobile advertising can significantly increase awareness. Kim et al. (2007), Rasinger et al. (2007) discovered that mobile phones allow for a context-aware search and therefore mobile marketing directly impact the information search phase. Other studies have shown that both mobile advertising and mobile marketing impact the purchase decision (Al-Alak & Ibrahim, 2010; Peng, 2006, Banerjee & Dholakia, 2010). And lastly, several studies have discovered that mobile marketing impacts the post purchase phase (Pura, 2005; Pihlstrom, 2007; Chae et al., 2002).
During one of the five stages, the purchase decision stage, consumers are likely to form purchase intentions (Park et al., 2007). What is, for the purpose of this thesis, extremely interesting are findings indicated by Peng (2006); Li and Stoller, 2007; Lu & Su (2009); Al-Alak & Ibrahim (2010) who indicate that mobile marketing directly influence purchase intentions.
Numerous variables influencing the effect of mobile marketing on consumer purchase have been identified. The study by Al-Alak and Ibrahim (2010) discovered that there is a positive relationship between perceived usefulness-, perceived entertainment, and permission –and purchase intentions. Simultaneously, they discovered that there is a negative relationship between perceived extensiveness- and perceived personal usage –and purchase intentions. Peng (2006) discovered that both personalization and content-personalization of the mobile marketing communication positively affects, and privacy concerns negatively affect consumer purchase intentions (Peng, 2006). Lu and Su (2009) discovered that usefulness, enjoyment and compatibility of mobile SMS marketing have a positive influence on purchase intentions.
11 studies have attempted to discover the effect of mobile marketing on purchase intentions. Yet, most of these studies do not specify the channels for which the purchase intentions are formed. From a managerial perspective, estimating the cross-channel effects of advertising/marketing is crucial in constructing an effective marketing strategy (Dinner et al., 2011; Rigby, 2011).
12 2.3. Location-Specificity/Geo-targeting of Mobile Advertising
Marketing messages such as offers or promotions, especially online based advertising, often reach consumers at inconvenient times or locations, which severely hinders the effectiveness of the advertising (Banerjee and Dholakia, 2010). Technological advances are bridging the gap, GPS capabilities of smartphones allow consumers to be traced, which in turn could help marketers avoid reaching consumers at inconvenient locations.
The progression in locating technologies has led experts predicting that location-based service (LBS) for mobile marketing will become the ‘’killer application’’ (Kölmel, 2003). This is exactly why both researchers and practitioners have taken an interest in location-specificity.
However, as stated earlier, research concerning mobile advertising is still in its infant stages, and the terminology used is highly inconsistent (Varnali and Toker, 2010). This phenomenon also holds true for location-based services for mobile marketing. For the sake of uniformity, this thesis has adopted the definition which suits the research topic best as defined by Shankar and Balasubramanian (2009) as: ‘’targeting location-sensitive promotional offers to mobile device users when they are in the vicinity of the advertisers’ location’’ (Shankar and Balasubramanian, 2009).
Location-specificity is not a new concept, billboard messages along the roadside placed in the vicinity of the establishment for instance also allowed for location-based advertising (Shankar & Balasubramanian, 2009; Banerjee & Dholakia, 2010).
There is ample literature available which investigated the effects of location-based advertising. For instance a study by Kelly et al. (2008) discovered that (underage) adolescents who were (repeatedly) exposed to an outdoor advertising for alcohol in close proximity to the business selling the product had a higher chance of alcohol abuse.
13 Lastly, a study by Taylor et al. (2013) discovered that the location of a billboard is one of the most important aspects in creating tangible responses from the billboard. That is, the closer the billboard is to the business, the higher the likelihood of tangible responses. So, location-based advertising certainly has an effect.
Yet, due to the vast technological advances of and adoption by consumers of mobile phones, the term location has become ever more dynamic and traceable to one specific consumer (Banerjee & Dholakia, 2010). With the localization technologies, marketers can reach consumers when they are in the vicinity of the point-of-sales. By adjusting a mobile marketing communication to include both the location of the consumer and the location of the ‘offer’, the marketer conducts a pre-selection which consumers normally would need to do themselves (Bauer et al, 2005).
As stated earlier, marketing communications such as offers or promotions in general, but especially online based advertising, frequently reach consumers at inconvenient locations, which severely hinders the effectiveness of the advertising (Banerjee and Dholakia, 2010). Internet marketing is far more suitable for information-rich marketing communications, yet internet marketing lacks location-specificity (Shankar and Balasubramanian, 2009). ‘’In some sense location-specificity is the most important distinguishing feature of mobile marketing’’ (Shankar and Balasubramanian, 2009).
Academics have suggested that by utilizing localization technologies marketers can inform consumers about offers and promotions when they are in the vicinity of the point-of-sale, thereby persuading them to engage in impulse purchase (Bauer et al., 2005). It is hypothesized that when consumers are reached by mobile advertising when they are near the advertisers’ location, the intentions of consumers to purchase something at the advertisers’ location increases. Therefore, for mobile advertising to be truly effective there is a necessity for location personalization (Al-Alak and Ibrahim, 2010). Hence, the first hypothesis can be formulated:
H1: Mobile advertising has a stronger effect on offline purchase intention when the
14 2.4. The Personal Perceived Relevance of the Mobile Advertising
Advertising such as new offers or promotions are often irrelevant to customers (InsightsOne, 2013). A study conducted by InsightOne (2013) had discovered that 87 percent of the participants were annoyed with ‘online-junk’ advertising. Moreover, the study revealed 23 percent of the participants stated they would ignore a company after a single irrelevant advertisement, 43 percent would ignore a company after two irrelevant advertisements customers (InsightsOne, 2013).
The relevance-accessibility model offers a framework for reviewing marketing communication effectiveness (Baker and Lutz, 2000).What is established is that (in general), for marketing communication messages to most likely affect purchase intentions, the communication has to be highly relevant (Baker and Lutz, 2000). Moreover, it has been proven that when consumers perceive relevance, the relevance prompts encouraged attention and comprehension (Celsi and Olson, 1988).
It is clear that the (perceived) relevance of the marketing communication by consumers is of extreme importance in influencing the consumer in his/her decision making process, and thus in forming purchase intentions. This holds true for all types of ‘media’, and thus also for mobile advertising. Research has shown that the perceived relevance of the mobile communication is of extreme importance (Rettie et al., 2005; Doherty, 2007; Scharl et al., 2005; Doherty, 2007; Carroll et al., 2007, Pagnani, 2004; Nasco and Bruner, 2008).
Moreover, research has shown that consumers’ feelings of relevance also trigger more motivated attention and comprehension processes (Celsi and Olson, 1988). Furthermore, research by Rettie et al (2005) indicated that relevance was one of the main drivers of consumer acceptance of mobile advertising.
Therefore, it is reasonable to assume that when a mobile advertisement is more relevant to a consumer (by means of manipulation), the impact of mobile advertising on offline purchase intentions is amplified. Therefore, a second hypothesis can be formulated:
H2: Higher perceived relevance of the product offering strengthens the relationship between
15
3. Research Design
This chapter will explain the research questions, the research strategy, and the constructs used to collect data necessary for the analysis
3.1 Research Question
The goal of this thesis is to examine the impact of location specificity of mobile advertising on offline purchase intentions and how this relationship is affected by perceived relevance. As illustrated in the literature review, research has established that mobile advertising directly influences consumer purchase intentions. Nevertheless, this thesis could add to literature in the sense that this thesis is aimed at investigating the cross-channel effects of mobile marketing on offline purchase intentions. Moreover, this thesis will investigate the effect of, what is considered to be the ‘next big thing’, location-specificity and relevance on the previously mentioned impact. The research question for this thesis therefore is:
’What is the impact of location-specificity of mobile advertising on offline purchase intention and is it strengthened by the perceived relevance of the mobile advertising?’
In aiding the research to answer the research question, two hypotheses were identified can be identified for this research:
H1: Mobile advertising has a stronger effect on offline purchase intention when the
advertising is location-specific
H2: Higher perceived relevance of the product offering strengthens the relationship between
mobile advertising and offline purchase intentions.
3.2 Conceptual Model
Combining these two hypotheses, the following model can be constructed with offline purchase intention as the dependent variable, location-specificity of mobile advertising as the independent variable and perceived-relevance of the mobile advertising as the moderator. The conceptual model can be observed in figure 1.
H1 + H2 + Offline Purchase Intentions Location-Specificity Manipulated (perceived) Relevance
16 3.3. Research Method
This thesis has conducted an experiment based on a 2 x 2 between-subject, full factorial design in which the independent variable is location-specificity (i.e. location-independent or location-based) and the moderator is manipulated relevance (i.e. relevant or non-relevant). This thesis has also opted for a full factorial design, meaning both levels of the IV are crossed with both levels of the moderator, resulting in four experimental conditions. Moreover, this thesis has chosen for the between-participant design, meaning every participant only serves in only one condition, for unawareness purposes, which otherwise might distort the authenticity of the experiment. Participants were randomly assigned to one of the four conditions, to give every participant an equal chance of being assigned to one particular condition of an experiment, ultimately resulting in the experimental groups being similar on practically every dimension imaginable.
3.3.1. Data Collection
This thesis has opted for data collection by means of a web-based survey. The response rate of a web-based survey is commonly lower than that of other data collection methods (Millar & Dillman, 2011). Yet, the application of mixed methods, such as postal mail, telephone calls, social media, use of better presentation and style, follow-up emails, the length, and using graphic images can increase response rates (Yin, 2011; Calderwood, 2013; Millar & Dillman, 2011).This thesis tried to increase response rates by actively using social media and the usage of graphic images to evoke interest in the survey. The link to the survey was distributed via email and social media to the sample frame.
3.3.2. Sample Frame
The survey was mass distributed via social media to a total of 1886 potential respondents. Of these 1886 potential respondents a total of a response rate of 12.2% (230 respondents) was realized. The period of distributing the survey started from the 13th of May 2014 until the 16th of May 2014.
3.3.3. The Survey
17 structure. Accordingly, questions capturing information about the hypotheses were obtained first and other personal information later on. The entire experiment was set in an online context. The distinctive advantages of web-based surveys revolve around cost-saving and a low chance of missing data due to the enforcing of participants to answer a specific question. One problem which is likely to come up with an online survey is the aspect of anonymity (Blumberg, et al., 2011). Participants might express opinions deemed desirable, yet not honestly. To overcome this problem, the participants will be guaranteed full confidentiality and anonymity.
The survey will consist out of three sections. Firstly, the participants will be subjected to an introduction of the study. Secondly, a situation will be described to the participants in whom the participant is either at home (location-independent) or in the vicinity of an advertisers’ location (location-specificity/based) and receives an advertisement. Moreover, the relevance of participants will be manipulated, i.e. consumers will be primed in such a way that the advertisement is either relevant, or not relevant.
Participants will receive an offer, in the form of a special promotion endorsed via an interstitial, for running shoes on his or her mobile phone. However, the participants will be primed in such a way that for participants assigned to the high relevance condition, the offer is highly relevant, i.e. the participants will be primed to perceive themselves as runners, in need of new shoes. Simultaneously, priming for participants assigned to the low relevance condition will be conducted in such a way in which the offer is only marginally relevant. Thirdly, questions assessing the manipulation checks for location-specificity and relevance, and the dependent variable -purchase intentions- will be asked after each scenario. This allows for testing of H1: Location-specific mobile marketing increases the impact of mobile
marketing on offline purchase intentions, and for testing of H2: Manipulated relevance
increases the impact of both location-specific and location-independent mobile advertising on offline purchase intentions. To end, the respondents will be asked six demographical questions.
3.3.4. Construct Adoption
18 originally designed to measure time and effort perceived to be necessary to shop at a certain store on the basis of a 7-point Likert scale. The questions have been rephrased to make the scale more suitable to highlight the time and effort perceived to be necessary to travel the distance between the location of the participant and the location of the store. The items used were:
1. Getting to the store of Intersport would require a lot of effort.
2. I would have to sacrifice a great deal of time to get to the Intersport shop.
The construct shows good reliability with a Cronbach’s alpha of .762 by Baker et al. (2002)
The second scale used is also intended as a manipulation check, but for the relevance manipulation. Three semantic differentials are used to assess how useful and important something is to a consumer (Miyazaki, Grewal and Goodstein, 2005). The construct is operationalized by the three items taken from Miyazaki, Grewal and Goodstein (2005). The items used were:
1. Having read the scenario, the advertisement received on my mobile phone to me is:
Not very relevant/very relevant Not very useful/very useful
Not at all important/very important
The construct has reported to show good reliability (α = 0.90) by Miyazaki, Grewat and Goldstein (2005).
Scales to measure the first dependent variable, purchase intention, are typically categorized as 7-point Likert-like scales, often with between two and four items (Bruner, 2009). The scale was developed to measure cognitive, affective and conative components of one’s evaluation of an ad. The scale and anchors have been taken from Baker and Churchill (1977).
1. Would you, in this scenario, like to try out the shoes?
2. Would you, in this scenario, buy the shoes if you happened to see them in a store? 3. Would you actively seek out the shoes (in a store in order to purchase them)?
19 3.3.5. Pre-Testing
20
4. Analysis
This chapter will provide an overview of the analysis by addressing the missing data, cleaning processes and exploratory factor analysis for the two-way Anova.
4.1 Two-Way Anova
Given the setup of the experiment conducted within the scope of this thesis with one (assumed to be) metric dependent variable – purchase intention- and one independent variable – location-specificity of mobile advertising - and one moderator – perceived personal relevance of the mobile advertising -, in a full factorial design, a two-way Anova is most appropriate. Anova analysis allows the researcher to find out what the effect of one or more independent variable is on an interval-scaled dependent variable (Janssens et al., 2008). The primary goal of this two-way Anova is to examine whether or not there is an interaction between the independent variable and moderator, location-specificity and relevance respectively, on the dependent variable, offline purchase intention.
However, before a two-way Anova can be conducted exploratory several steps have to be undertaken. These steps include report the missing data, cleaning the data and handling outliers, reliability checks for the constructs and the assumptions (Janssens et al., 2008).
4.1.1. Missing Data
21 4.1.2. Data Cleaning and Outliers
Data cleaning was conducted on the basis of looking for unengaged responses. Unengaged responses are responses indicate very little variety in answering questions (Janssens et al., 2008). To check for unengaged responses the individual standard deviation per column was investigated, however, this investigation did not indicate any unengaged responses and as such did not lead to a deletion of responses.
To screen for outliers three activities were undertaken. The first activity undertaken was to check for the time needed to complete the survey. Based on three attempts in completing the survey by the researcher, it was determined that the mean time to complete the survey was 4 minutes and 23 seconds. However, the mean times of two respondents, also based on three attempts in completing the survey, were 6 minutes 18 seconds and 8 minutes and 6 seconds. So the boundaries have been set between 4 minutes and 23 seconds and 8 minutes and 6 seconds. A possible deviation of 20% was allowed in order to identify outliers. Based on these parameters, 17 cases were identified to be outside the boundaries (Appendix A). Of these cases 15 cases took longer than the boundaries permitted. These 15 cases filled out the survey in a normal manner (i.e. were correctly primed) and therefore were maintained in the dataset. A possible explanation for the longer completion time is that the respondents opened the survey, but did not immediately complete it as Qualtrics establishes the starting time of the survey as the moment in which the first page is opened. The remaining two cases showed shorter completion times than the means (Appendix A). These two cases also displayed signs that priming was successfully conducted and therefore were maintained in the database.
Secondly, observations with unrealistic values were identified. Unrealistic values are observations which have a z-score of greater than 3.29 (Janssens et al., 2008). This first activity did not identify any outliers. The first activity undertaken was to check for the time needed to complete the survey.
22 answered truthfully, their response being ‘I don’t know’. The remaining 5 responses answered incorrectly. 1 respondent answered that they were in a low relevance state whilst they were actually primed to be in a low relevance state. The remaining 4 respondents responded in a high relevance manner while they were primed to be in a low relevance state.
To check whether or not they should be treated as outliers, their (faulty) response was cross-examined with their mean scores on the items for the construct relevance. Based on this examination 4 out of the 12 ‘I don’t know’ cases were removed from the sample. The 1 respondent, who answered the instructional conduct in a low relevance state while actually being primed to be in a high relevance state, answered the items for the construct relevance in a high manner way and therefore was retained in the sample. Of the remaining 4 cases who answered the instructional conduct in a low relevance manner while they were primed to be in a high relevance manner 2 cases were deleted as their answers on the construct relevance revealed they actually were in a low relevance state. Accordingly, the sample after data cleaning is 191. An overview of the identified and removed outliers can be found in table 1.
LBS - Low Relevance LBS - High Relevance LIS - Low Relevance LIS – High Relevance
Total Total After exclusion
'No definitely not. Last night was the first and last time. I will never go running again’
42
1 (0 of 1 excluded)
42 0 85 85
'I definitely want to, but my running shoes are all worn out. They are completely destroyed. So, unless I can get some new running shoes before tonight, I will not be able to go running tonight' 1 (1 of 1 excluded) 50 3 (2 of 3 excluded) 46 100 98 I don't know 4 (0 of 4 excluded) 4 (4 of 4 excluded) 3 (0 of 3 excluded) 1 (0 of 1 excluded) 12 8 Total 47 55 48 47 197
Total after exclusion of outliers 47 51 46 47 191
23 4.1.3. Descriptive Statistics of the Sample
The description of the sample was based on gender, age, nationality, educational level, mobile phone Internet usage and likelihood of accessing the Internet on a mobile phone on certain places, see table 2. As indicated by the output, there was an almost perfectly even distribution of males to females. Moreover, the majority of respondents (80.1%) is between the ages of 18 and 25, of Dutch nationality (70.3%) and has a Bachelor degree (54.5%). The results are depicted in table 2 below.
Table 2. Description of the overall sample.
Having described the sample based on gender, age, educational level and nationality; two demographical aspects remain: the mobile phone Internet usage and likelihood of accessing the Internet on a mobile phone on certain places also see table 3. What becomes apparent is that the vast majority (80.6%) accesses the Internet on his/her mobile phone on a daily basis. Moreover, the highest likelihood of accessing the Internet of is when the participants were traveling. What is more interesting for this thesis is that the respondents indicate that the likelihood of them accessing the Internet on his/her mobile phone during a shopping trip is not that high (mean of likelihood 4.66). Consequently, interstitials might not be the most ideal form of advertising for location-specific mobile advertising.
Gender Frequency Percentage Nationality Frequency Percentage
Male 93 48.7% American 4 2.1%
Female 98 51.3% Australian 2 1.0%
Austrian 1 0.5%
Age Frequency Percentage British 1 0.5%
<13 0 0.0% Bulgarian 2 1.0% 13-17 1 0.5% Chinese 3 1.6% 18-25 153 80.1% Dutch 134 70.2% 26-34 22 11.0% French 1 0.5% 35-54 14 7.3% German 20 10.5% 55-64 2 1.0% Greek 1 0.5% >64 0 0.0% Hongkonger 1 0.5% Japanese 4 2.1%
Educational Level Frequency Percentage Kenyan 1 0.5%
None 0 0.0% Korean, South 2 1.0%
Lower Secondary 6 3.1% Latvian 1 0.5%
Upper secondary 37 19.4% Mongolian 2 1.0%
Associates/Vocational 21 11.0% Norwegian 6 3.1%
Bachelor Degree 104 54.5% Portuguese 1 0.5%
Master Degree 23 12.0% Saudi-Arabian 1 0.5%
Phd 0 0.0% Slovenian 1 0.5%
Swedish 1 0.5%
24
Table 3. Internet Access Overall Sample.
Having displayed the descriptive statistics of the entire sample, a description of the sample for the four experimental conditions will be made. It is important to have a similar distribution among all four experimental conditions. Similarly to the previous section, the gender, nationality, age and level of education of the respondents for the different experimental conditions will be discussed firstly, see table 4.
Table 4. Description of sample for the four conditions.
What becomes apparent from the dividing from one whole into the four experimental conditions is that there seem to be several (slight) deviations in percentages. Firstly, the male to female ratio - which is nearly 50-50 for the entire sample – is significantly different for the location-independent, low relevance experiment condition. In this condition there are is a considerably higher amount of females present. The male to female ratio of the three other experimental conditions is largely similar to the overall ratio, and is around 50-50. The nationality is also very similar to the overall ratio, which was 70-30, for three of the experimental conditions. However, the location-independent, high relevance condition has a
Internet access on mobile phone
Frequency Percentage Internet access on
following places
Mean (0-10)
Likelihood
Daily 154 80.6% At home 6.71 High Likelihood
2-3 times a week 20 10.5% At work 3.73 Low likelihood
Once a week 8 4.2% At school 5.87 Average likelihood
2-3 time a month 2 1.0% While traveling 8.22 Very high likelihood
Monthly 4 2.1% At shopping street 4.66 Average likelihood
Never 3 1.6%
Condition Gender Male Female Total Dutch Non-Dutch Total
LBS- Low Relevance 22- 46.8% 25- 53.2% 47 34-72.3% 13- 27.8% 47
LBS- High Relevance 29- 56.9% 22- 43.1% 51 38- 74.5% 13- 25.5% 51
LIS – Low Relevance 14- 30.4% 32- 69.6% 46 34- 73.9% 12- 26.1% 46
LIS – High Relevance 28- 59.6% 19- 40.4% 47 28- 59.6% 19- 40.4% 47
Overall 93- 48.7% 98 -51.3% 191 134- 70.3% 57- 29.7% 191
Condition Age 13-17 18-25 26-34 35-54 55-64 Total
LBS- Low Relevance 1- 2.1% 36- 76.6% 5- 10.6% 5- 10.6% 0 47
LBS- High Relevance 0 41- 80.4% 5- 9.8% 3- 5.9% 2- 3.9% 51
LIS – Low Relevance 0 40- 87.0% 3- 6.5% 3- 6.5% 0 46
LIS – High Relevance 0 36- 76.6% 8- 17.0% 3- 6.4% 0 47
Overall 1 – 0.5% 153- 80.1% 22- 11.0% 14- 7.3% 2 – 1.0% 191 Condition Education Lower Secondary Upper Secondary Associates Degree Bachelor Degree Master Degree Total LBS- Low Relevance 3- 6.4% 7- 14.9% 8- 17.0% 24- 51.1% 5- 10.6% 47 LBS- High Relevance 1- 2.0% 10- 19.6% 4- 7.8% 32- 62.7% 4- 7.8% 51
LIS – Low Relevance 2- 4.3% 9- 19.6% 5- 10.9% 25- 54.3% 5- 10.9% 46
LIS – High Relevance 0 11- 23.4% 4- 8.5% 23- 48.9% 9- 19.1% 47
25 slightly different ratio, 60-40. The distribution of the age and education level of the respondents among the four experimental conditions followed the same pattern as the overall distribution without any (significant) exceptions. Having described the respondents in the different experimental conditions based on gender, age, educational level and nationality; two demographical aspects remain: the mobile phone Internet usage and likelihood of accessing the Internet on a mobile phone on certain places also see table 5.
Table 5. Internet Access Overview for the four Conditions
What becomes apparent is that the vast majority of respondents (80.6%) which access the internet on their mobile phone on a daily basis are more or less similarly present when the respondents are assigned to the experimental conditions. The distribution is pretty similar to the overall distribution and therefore no (significant) underlying differences between the respondents in the experimental conditions could be identified based on internet access. However, when one looks at the likelihood of access the internet on their mobile phones while shopping of the respondents, - which is particularly important for this thesis- significant differences can be observed. What becomes apparent is that the means for the respondents assigned to the location-based/specific advertising (mean 5.22 & 5.06) are significantly higher than the means for respondents assigned to the location-independent conditions (mean 4.16 & 4.51). This might distort the results as one infer that respondents assigned to the location-independent conditions- who are less likely to access the internet on their mobile phone while shopping- are less open to interstitials while shopping.
Condition Internet Use Daily 2-3 times
a week
Once a week
2-3 times a month
Monthly Never Total
LBS- Low Relevance 39- 83.0% 4- 8.5% 2- 4.3% 1- 2.1% 1- 2.1% 0 47
LBS- High Relevance 43- 84.3% 6- 11.8% 1- 2.0% 0 1- 2.0% 0 51
LIS – Low Relevance 34- 73.9% 6- 14.0% 2- 4.3% 0 2- 4.3% 2- 4.3% 46
LIS – High Relevance 38- 80.9% 4- 8.5% 3- 6.4% 1- 2.1% 0 1- 2.1% 47
Overall 154- 80.6% 20- 10.5% 8- 4.2% 2- 1.0% 4- 2.1% 3- 1.6% 191
Condition Internet Use (mean 0-10) At home At work At school While traveling While shopping LBS- Low Relevance 6.78 4.11 6.16 8.14 5.22 LBS- High Relevance 6.58 3.73 6.16 8.49 5.06
LIS – Low Relevance 6.58 3.59 5.71 8.21 4.16
LIS – High Relevance 6.90 3.47 6.08 7.93 4.51
26
4.2. Assumption Checks
Several of the statistical procedures, including two-way Anova, are based upon the assumption that samples from different populations have the same variance (Janssens et al., 2008). To test for these equal variances one should perform Levene’s test of equal variance. Moreover, the normality of distribution should be checked by using the Kolmogorov-Smirnov test and Shapiro-Wilk test (Janssens et al., 2008).
The outcomes of the test can be found in Appendix B. Levene’s test indicated that there was homogeneity of variances (p = .533). Consequently, the null hypothesis of equal variances may not be rejected at a confidence level of 95%, and the first assumption is satisfied. The Kolmogorov-Smirnov test and Shapiro-Wilk test indicated that that purchase intention was not normally distributed for all group combinations of location-specificity and relevance (p =.033 and p = .005) respectively. Thus, both hypotheses of normality may be rejected at a confidence level of 95%. Subsequently, the second assumption of normality is not satisfied. Fortuitously, a two-way anova is not very sensitive to modest deviations of normality; the false positive rate is not disturbed very much by violation of the normality assumption (Kesselman et al. 1998). Based on the histogram, as shown in Appendix B, normality is substantially present. Therefore, it is allowed to continue working with the variables as they are (Janssens et al., 2008).
4.3. Reliability of the scale
Cronbach’s alpha is commonly used in determining how much a certain number of items measure the same underlying construct (Janssens et al., 2008). In this case the Cronbach’s alpha is used to determine whether or not the scales used to measure location-specificity, perceived relevance and purchase intention are reliable. The results are displayed in table 6 below.
The commonly accepted social-science cut-off point for the Cronbach’s alpha of the constructs is above the value of .75. If the alpha is above this value one can continue with using exploratory factor analysis, if the value is below the threshold one should conduct confirmatory factor analysis (Nunally, 1994). As can be seen from the results can in table 6,
Construct Cronbach’s Alpha
27 all constructs adhere to the threshold and therefore this research can continue by solely using exploratory factor analysis.
4.4. Manipulation Checks
This thesis has conducted a 2 (location-specificity) x 2 (perceived relevance) analysis of variance to test whether the priming (i.e. manipulations) was successfully conducted. Perceived relevance and location-specificity served as well as the interaction between both variables acted as the independent variables. The location-specificity- and the perceived relevance-manipulation checks served as the dependent variables.
What becomes apparent is that participants assigned to the location-independent experimental conditions rate the perceived effort to travel to the store significantly higher than the participants assigned to the location-based/specific experimental conditions (F(1,190) = 243.125, p = .000 MLI = .772 and MLBS = -.718). Moreover, the participants assigned to the
low-relevance experimental conditions rated their perceived relevance as significantly lower than participants assigned to the high-relevance conditions (F(1,190) = 171.436, p = .000 MLR
28
5. Results
This chapter will provide an overview of the results as given by the SPSS output for the two-way Anova.
When conducting a two-way Anova there are three hypotheses to test. Two hypotheses for the main effects of the two factors and one hypothesis for the interaction effect. For the first factor, location-specificity, the null hypothesis is H0: all group means are equal. For the
second factor, relevance, the null hypothesis is H0: all group means are equal. For the
interaction effect the null hypothesis is H0: the sum of all individual interactions equals 0, the
alternative hypothesis is that the sum of all individual interactions does not equal 0.
The main output of the two-way Anova is summarized in table 7. What becomes apparent is that there were statistically significant main effects for both location-specificity, F(1,187) = 2.114, p = 0.009 and relevance F(1,187) = 4.299, p = 0.000. However, the interaction effect is not/only marginally significant F(1, 187) = 1.638 p = 0.60 . Therefore, the null hypothesis may not be rejected, meaning that the sum of all individual interactions equals 0. To conclude, the second hypothesis made by this thesis (i.e. the perceived relevance of the product offering increases the impact of both location-specific and location-independent mobile advertising on offline purchase intentions) is not confirmed.
Table 7. Anova Output
There are two routes a researcher can take when there is no statistically significant interaction (Janssens et al., 2008). The first involves simply interpreting the main effects and conducting any post-hoc tests on the already produced Anova output, thereby including the interaction in the Anova model. The other approach is to remove the interaction from the Anova model, as the interaction was not significant.
This thesis has opted for the first option, i.e. interpreting the main effects while also including the interaction effect. This because using the second option is not universally recommended (Janssens et al., 2008).
Construct df Mean Square F Significance
29 Testing the main effects of location-specificity means testing for differences in purchase intention between location-based and location-independent mobile advertising regardless of relevance. The outcome indicated that there was a statistically significant difference in purchase intention score between the different levels of location-specificity, F(1,187) = 2.114, p = 0.009, partial η2= .482.
Testing the main effects of perceived relevance means testing for differences in purchase intention between low-relevance and high-relevance mobile advertising regardless of location-specificity. The outcome indicated that there was a statistically significant difference in purchase intention score between the different levels of perceived relevance F(1,187) =4.299, p=0.000, partial η2= .840.
The Anova indicated that (1) the interaction effect between location-specificity and perceived relevance was not significant and (2) that both main effects were statistically significant. However, the Anova did not explain where the differences in purchase intentions between location-specificity and perceived relevance lie. To discover where these differences lie- thus breaking down the main effects and analyzing the hypotheses - this thesis will conduct a contrast analysis.
A contrast analysis for a two-way Anova is, as opposed to a one-way Anova, restricted to a one of several standards. This thesis has, given the design, opted for a simple contrast using the first category as the reference category. However, for a contrast analysis to be appropriate, the Anova should produce significant results (Janssens et al., 2008). Given this assumption, this thesis will only perform a contrast analysis on the main effects and not on the interaction effect. The results of the contrast analysis are depicted in table 8.
Table 8. Contrast Analysis.
30 Table 8 indicates the contrasts in offline purchase intentions for location-independent and location-specific. The contrast analysis test whether the mean of location-independent (MLI
-.182) and the mean of location-specific (MLS.130), which is a difference of .312 as indicated
by the contrast estimate, is meaningful.
As indicated by the significance level (p = .003), the difference is statistically significant. Moreover, as indicated by the confidence interval (which does not crosses zero), this sample is assumed to be one of the 95 out of a 100 that produces a confidence interval containing the true value of the difference (Field, 2005). Hence, this thesis is able to conclude that the offline purchase intention is higher for location-specific mobile based advertising compared to location-independent mobile advertising.
Although the individual impact of perceived relevance was not hypothesized by this thesis – this thesis merely hypothesized the impact of perceived relevance in an interaction setting - for the sake of completeness, this thesis will also discuss the contrast results for perceived relevance. Table 8 indicates the contrasts in offline purchase intentions for low-relevance and high-relevance of the product offering. The contrast analysis test whether the mean of low-relevance (MLR -.712) and the mean of high-relevance (MHR.660), which is a difference of
-1.372 as indicated by the contrast estimate, is meaningful.
31
6. Discussion
The primary objective of this thesis was to study the impact of - what is considered to be the next-big thing in mobile advertising- location-specificity of mobile advertising on offline purchase intention and the moderating impact of perceived relevance over that relationship. These relationships have been conceptualized in the form of two research questions: H1:
Mobile advertising has a stronger effect on offline purchase intention when the advertising is location-specific and H2: Perceived relevance of the product offering increases the impact of
both location-specific and location-independent mobile advertising on offline purchase intentions.
The analysis provided support for the first hypothesis. The analysis indicated that offline purchase intention is higher when the mobile advertising is location-specific. Advertisements frequently reach consumers at inconvenient locations, which severely hinder the effectiveness of the advertising. As posited, when consumers are reached by mobile advertising when they are near the advertisers’ location, the intentions of consumers to purchase the offering at the advertisers’ location increases, thereby increasing the effectiveness of the advertising. This conceptualization was confirmed by the analysis.
32
7. Managerial Implications
Due to the increasing importance of the mobile (smart) phone in people’s lives, mobile advertising is also likely to become increasingly important. As a result, managers and marketers should make use of the unique capability for localization offered by mobile phones. Overall, the results found by this thesis have two managerial implications. Firstly, this thesis has indicated that for mobile advertising to increase its effectiveness in shaping purchase intentions, the mobile advertising should be location-specific. Given this result, managers should exploit the possibilities for location-specific advertising offered by mobile phones. For example, if managers want to increase the offline purchase intentions for a certain product, managers could offer location-based price promotions when they are near the stores.
Lastly, results by the analysis within this thesis have also indicated that for mobile advertising to increase its effectiveness in shaping purchase intentions, the mobile advertising should be highly relevant. To ensure this, managers should gather as much information as possible. Both managers and marketers should analyse search and purchase behaviour to provide the most relevant information at the most crucial times. For example if customers search for running shoes in two consecutive days, managers should target that customer with running shoes advertising on the third day. However, when the search for running shoes was two weeks ago and the customers has not searched for running shoes since, managers should not reach that customer with advertisements involving running shoes. Analysing the search behaviour could ensure that mobile advertisement have high relevance to consumers, which in turn – as indicated by this thesis – would result in higher amounts of offline purchase intentions.
33
8. Limitations and Future Research
Every piece of scientific work requires some critical reflection. The first limitation of this research is the fact that this thesis has used the offline purchase intentions by consumers based on the exposure to a certain mobile advertising. Generally speaking managers would consider an ad that succeeds in affecting intentions as very successful. Nevertheless, intentions only explain modest 28% variance in behavior (Sheeran, 2002). This provides a limitation in the applicability of the results towards ‘real-life’ situations as the location-specificity has been shown to impact the purchase intention, but this effect might be completely different for real purchase behavior. A second limitation of this research is the fact that the respondents assigned to the experimental conditions are not (completely) equally distributed. A third limitation is with regards to generalizability. The majority of the respondents were Dutch, which lowers the generalizability of this research. A fourth limitation lies in the fact that only one product category was analyzed. A fifth, and major, limitation of this paper is the fact that ‘only’ two variables were reviewed, when it is likely that there are numerous other variables who influence the effectiveness of mobile advertising on offline purchase intentions, yet due to restrictions on both time and word-quantity, these were not discussed.
34
9. Conclusion
In this thesis, the impact of location-specificity of mobile advertising on the offline purchase intentions of consumers was investigated. Moreover, the moderating impact of perceived relevance on this relationship was investigated. It has become apparent that both location-specificity and a high degree of relevance of mobile advertising enhance the offline purchase intentions of consumers. The moderating effect of perceived relevance however could not be established.
These findings bridge the gap that existed between theory and empirical research. This thesis has provided empirical evidence in support of the much theorized enhancing effect of
location-specific mobile advertising on purchase intentions.
In order to enhance generalizability, future research could test other variables and other product categories to see whether the effect holds over contexts. Moreover, in the future the mobile phone is likely to become even more important to consumers and the capabilities for marketers will have become even greater and (likely) to be more accepted. Given these developments, it would be interesting to see whether the same experiments in future years would report higher levels of purchase intentions. All in all, a further investigation of
35 10. References
AdSpeed (2014). ‘’What are interstitial ads?’’. (March 19).
Al-Alak, B. A., and Ibrahim, A. M. (2010). Mobile Marketing: Examining the Impact of Trust, Privacy Concern and Consumers' Attitudes on Intention to Purchase. International Journal of Business & Management.
Ammerman, S. D., & Nolden, M. (1995). Neighborhood-based tobacco advertising targeting adolescents. Western journal of medicine, pp. 514-.
Ayanwale, A. B., Alimi, T., and Ayanbimipe, M. A. (2005). The influence of advertising on consumer brand preference. Journal of Social Science, pp. 9-16.
Baker, M.J. and Churchill, G. (1977). The impact of Physically Attractive Models on Advertising Eveluations. Journal of Marketing Research. Pp.
Baker, W. E., and Lutz, R. J. (2000). An empirical test of an updated relevance-accessibility model of advertising effectiveness. Journal of Advertising, pp. 1-14.
Balasubramanian, S., Peterson, R.A. and Jarvenpaa, S.L. (2002). Exploring the Implications of M-Commerce for Markets and Marketing," Journal of Academy of Marketing Science, pp. 348–61.
Banerjee, S., and Dholakia, R. R. (2008). Mobile advertising: does location-based advertising work?. International Journal of Mobile Marketing.
Barnes, S. J., and Scornavacca, E. (2004). Mobile marketing: the role of permission and acceptance. International Journal of Mobile Communications, pp. 128-139.
Barwise, P. and Strong, C. (2002), Permission-Based Mobile Advertising, Journal of Interactive Marketing, pp. 14–24.
Baker, J. A., Parasuraman, D.G. and Voss, G.B. (2002), The Influence of Multiple Store Environment Cues on the Perceived Merchandise Value and Patronage Intentions, Journal of Marketing, pp. 120-141.
Bauer, H. H., Barnes, S. J., Reichardt, T., and Neumann, M. M. (2005). Driving consumer acceptance of mobile marketing: A theoretical framework and empirical study. Journal of electronic commerce research, pp. 181-192.
BIA/Kelsey via eMarketer.com, (2013), ‘Search Gets a Mobile Makeover’’, (April 18). Available: <http://www.emarketer.com/Article/Search-Gets-Mobile-Makeover/1009822> Last accessed: 09-03-2014.
Blumberg, B., Cooper, D. & Schindler, P., 2011. Business Research Methods. Third European Edition red. London: McGraw-Hill Higher Education
36 Calderwood, L., (2013). The Use of New Technologies on the British Birth Cohort Studies. International Journal of Market Research.
Carroll, A., Barnes, S. J., and Scornavacca, E. (2005). Consumers Perceptions and Attitudes towards SMS Mobile Marketing in New Zealand. In the Proceedings of the Fourth
International Conference on Mobile Business (ICMB 2005), pp. 434-440.
Chae, M., Kim, J., Kim, H., and Ryu, H. (2002). Information quality for mobile internet services: A theoretical model with empirical validation. Electronic Markets, pp. 38–46. Cooper, D. R. & Schindler, P. S., 2008. Business Research Methods. 10th red. New York: NcGraww-Hill/Irwin.
Dinner, I.M., Van Heerde, H.J. and Neslin, S.A. (2011), Driving Online and Offline Sales: The Cross-Channel Effects of Digital versus Traditional Advertising, Hanover, NH
Doherty, C. (2007). Mobile Marketing is Yet to Bloom. Revolution, November, pp. 51. Field, Andy. (2005). Discovering Statistics Using SPSS. London: SAGE Publications. Griffith, D.A. and Chen, O. (2004), The Intelligence of Virtual Direct Experience (VDE) on On-Line Ad Message Effectiveness, Journal of Advertising, pp. 55-68.
Häubl, G. and Trifts, V. (2000). Consumer decision making in online shopping environments: The effects of interactive decision aids. Marketing science, pp. 4-21
Hennig-Thurau, T., Malthouse, E. C., Friege, C., Gensler, S., Lobschat, L., Rangaswamy, A., and Skiera, B. (2010). The impact of new media on customer relationships. Journal of Service Research, pp. 311-33.
Homburg, C., Kuester, S, and Krohmer, H. (2008), Marketing Management: a Contemporary perspective. London: McGraw-Hill Higher Education.
InsightOne (2013). ‘’Irrelevant Internet Advertising’’. (May 8).
Interactive Advertising Bureau, (2013), ‘’Internet advertising revenue report - An industry survey conducted by PwC’’, (October 3).
Janssens, W., Wijnen, K., De Pelsmacker, P. and Van Kenhove, P. (2008), Marketing Research with SPSS. New Jersey: Prentice Hall.
Kasesniemi, E. L., and Rautiainen, P. (2002). Mobile culture of children and teenagers in Finland. Perpetual contact: Mobile communication, private talk, public performance. Keller, K. (2010), Brand Equity Management in a Multichannel, Multimedia Retail Environment,” Journal of Interactive Marketing, pp. 58-70.
37 Keselman, H. J., Huberty, C. J., Lix, L. M., Olejnik, S., Cribbie, R. A., Donahue, B.,& Levin, J. R. (1998). Statistical practices of educational researchers: An analysis of their ANOVA, MANOVA, and ANCOVA analyses. Review of Educational Research, pp. 350-386. Kim, H., Chan, H. C., and Gupta, S. (2007). Value-based adoption of mobile internet: An empirical investigation. Decision Support Systems, pp. 111–126
Kim, D. Y., Lehto, X. Y., and Morrison, A. M. (2007). Gender differences in online travel information search: Implications for marketing communications on the internet. Tourism management, pp. 423-433.
Kölmel, B., (2003). Location Based Services, Mobile Commerce, pp. 88-101.
Kilbourne, W.E. (1986), An Exploratory Study of Sex Role Stereotyping on Attitudes Toward Magazine Advertisements," JA1 pp. 43 46.
Kilbourne, W.E., Painton, S. and Ridley, D.Y. (1985), The Effect Of Sexual Embedding on Responses to Magazine Advertisements, Journal of Advertising, pp. 48 56.
Kireyev, P., Pauwels, K. and Gupta, S. (2013). Do Display Ads Influence Search? Attribution and Dynamics in Online Advertising, Harvard Business School, pp. 1-32.
Leppäniemi, Matti, “Mobile marketing communications in consumer markets”, Faculty of Economics and Business Administration, Department of Marketing, University of Oulu, 2008, pp. 21.
Leppaniemi, M., Sinisalo, J., and Karjaluoto, H. (2006). A review of mobile marketing research. International Journal of Mobile Marketing, pp. 2-12.
Lepkowska-White, E., Brashear, T.G., and Weinberger, M.G. (2003), ‘’A Test of Appeal Effectiveness in Poland and the United States: The Interplay of Appeal, Product and Culture " Journal of Advertising. pp. 57-67.
Li , H., and Stoller , B. (2007). Parameters of mobile advertising: A field experiment, International Journal of Mobile Marketing, pp. 4 – 11.
Millar, M. and Dillman, D., (2011). Improving Response to Web and Mixed-Mode Surveys. Public Opinion Quarterly, pp. 249 - 269.
Miyazaki, A.D., Grewal D. and Goodstein, C. (2005). The Effect of Multiple Cues on Quality Perceptions: A Matter of Consistency. Journal of Consumer Research, pp. 146-153
Mobile Marketer, (2013). ‘’Precise location, big data power charge marketers geotargeting strategies’’. (August 8).
