THE EFFECTIVENESS OF MOBILE ADS ON CONSUMER’S BRAND
AWARENESS AND PURCHASE INTENTIONS
Gonzalo Navarro Marin s3148963
Rijksuniversiteit Groningen (RUG)
1. INTRODUCTION
The use of smartphones has increased over the last years. The Statista report (2017) shows an increase of smartphone users since 2014. In that year, the number of users was greater than 1,5 billion, while in 2017 there were more than 2,3 billion. The same report illustrates the worldwide mobile commerce revenues from 2015 and the predictions until 2019. The data suggests an increase of more than 400 billion in the period 2017-2019. As a result of that, companies have increased the use of mobile ads since 2015. The report from Sales force (2017) shows a 142% growth on the use of SMS advertising as well as a 80% growth on the use of mobile display advertising. The CMO survey (2017) also reveals an increase of the 5,2% in mobile marketing budget expenditures and it predicts a 11,6% increase for the next three years.
companies to overspend marketing budget without choosing the correct touchpoint in which customers are mainly influenced.
The use of mobile devices as a communication channel is growing every year, bringing the opportunity to marketers to create more and more touchpoints to interact with customers at any time on an individual-oriented way (Bauer, Barnes, Reichardt and Neumann, 2005). When comparing with other media types, mobile marketing is probably the trendiest one to consider when studying the effectiveness of digital advertising, as it possesses several advantages compared to others: being a two-way communication channel (interaction), it allows to make exclusive offers thanks to the personalization and the individual targeting. In addition it is suitable for location-based technology that offers an even more personalized offer (Schrott, 2003).
As in other communication channel, mobile channel has different touchpoints that can affect consumers responses toward the brand advertised, therefore it is managerially valuable to consider different touchpoints and compare their effectiveness. This study will account for two touchpoints regarding mobile marketing, which are the two ways to deliver advertisements used by marketers: push-based method and pull-based method (Grewal, Bart, and Spann, 2016).
comparison between the two methods to deliver mobile ads on two different stages of the customer journey.
As stated in Grewal, Bart, and Spann (2016), marketers use different outcome variables to quantify the effectiveness of marketing campaigns at different stages of the customer journey. Therefore, this research addresses these issues by formulating the question of which of the two methods to deliver mobile ads is the most effective one during the customer journey stages.
In addition, continuing the line of reasoning behind the study of Bart, Andrew and Sarvary (2014) marketers need a deeper exploration of the factors that affect mobile marketing campaigns’ performance. The authors incorporate variables such as product type and product involvement as determinants for mobile display advertising effectiveness. This study adds value to the existing literature in two ways. First, it incorporates the same condition variables not only to the mobile display advertising field (pull-based method) but it also tests how these variables affect SMS mobile marketing campaigns (push-based method). Second, this study incorporates these variables together with another covariate variable, brand familiarity. Brand familiarity could be considered as another marketing effectiveness’ determinant because of its linkages with the outcome variables purchase intentions (Laroche, Kim and Zhou, 1996) and brand awareness (Axelrod 1968; Haley and Case 1979; Baker, Wesley, Danny, and Nedungad, 1986). The use of this variable as a condition brings managerial implications. Marketers of either a familiar or non-familiar brands could consider this variable when selecting which of the two methods to deliver mobile ads is the most effective at the different stages of the customer journey for one specific group of customers.
2. LITERATURE REVIEW AND THEORETICAL MODEL
SMS delivery service and apps delivery when consumers accept the permissions to receive messages from the app (Grewal, Bart, and Spann, 2016). The construction of the conceptualization for both methods helps to clearly differentiate them.
2.1 SMS ad conceptualization
In the category of push-based method, this study uses SMS (Short Message Service) as the way to deliver the ads. The SMS concept can be conceived from a marketing perspective as an advertising message in alphanumeric format up to 160 characters that can be stored in user’s handset, reviewed or forwarded to other users at later time (Weia, Xiaoming and Pan, 2010). The content of SMS is diverse, including special offers, teaser ads and product information requests (Barwise and Strong, 2002). This content can be personalized depending on the target consumer (Rohm and Sultan, 2005). Other characteristics from the SMS that differ from other mass communication media is the opportunity to facilitate viral marketing strategies (point-to-point interpersonal communication), includes ubiquity, location-based technology, interactivity, and results-oriented approach (e.g., e-coupons can be redeemed) (Weia et al., 2010).
To understand the effectiveness of SMS mobile ads, this study presents the framework in which SMS technology operates and how it works. As stated before, SMS ads correspond to push-based method category where consumer must give permissions to the advertiser in order to receive these ads. More concretely, SMS advertising assumes that customers will receive the message only when they accept the introduction of new products on their mobile phones coming from the advertisers (Kavassalis, Spyropoulou, Drossos, Mitrokostas, Gikas and Hatzistamatiou, 2003; Tsang, Ho, Liang, 2004).
2.2 Mobile Display Ads conceptualization
messages such as videos, images or text in a combination of them (Shankar and Balasubramanian 2009). Because of the importance of being visible for all customers and appeal to their acceptance without making the customer irritated about the presence of the ads, many MDA use simple and small images easy to render in almost all screens at even the lowest speed internet levels (Patel, Schneider, and Surana 2013). Hence, most of the MDA includes elements such as logos, symbols or slogans of brands (Bart, Stephen, and Sarvary, 2014).
2.3 Effectiveness of mobile ads
As stated in the introduction, previous studies aimed to measure mobile marketing campaign’s effectiveness focusing on the isolated effect of a concrete method to deliver the ad (pull or push) on a certain outcome variable (brand awareness, consumer’s attitudes, purchase intentions, engagement and conversion rates). The study from Grewal, Bart, and Spann (2016) proposes some metrics to measure the outcome variables used in previous literature; shares, likes, clicks, purchases, loyalty/NPS and digital WOM.
Prior investigations focused on the effectiveness of mobile ads on different scenarios by using different variables to measure it. Some of these studies covered the effectiveness of SMS mobile marketing campaigns addressing conversion rates and using moderators such as the social presence when consumers are exposed to the ad (Andrews, Xueming, Fang and Anindya, 2015), permission to receive the ads (Bacile, Ye, and Swilley, 2014), type of products advertised and timetable to receive the ads (Baker, Fang, and Luo, 2014), distance to store (Danaher, Peter J., Smith, Ranasinghe and S. Danaher, 2015) or location-based information (Fong, Fang, and Luo, 2015; Luo, Andrews, Fang and Phang, 2014a; Luo, Reinaker, Phang, and Fang, 2014b). Another research measured the SMS method’s effectiveness using variables such as brand awareness and engagement (Barwise and Strong, 2002).
and Sarvary, 2014), the interaction effect with other media (Ghose, Han, and Park, 2013) and geographical location, time and weather (Molitor, Reichhart, Spann and Ghose, 2015; Molitor, Reichhart, and Spann, 2014).
In all above-presented cases, the researchers have shown that mobile marketing campaigns affect significantly all outcome variables despite some low reliability exceptions. Although there is no doubt on the effectiveness of both methods, it could not be formalized that both perform equally good. Thus, the literature regarding factors and comprehensive framework in which both methods operate is required to measure the relative effectiveness between them.
Prior scholars have formalized conceptual models about the factors that influence mobile advertising effectiveness. Determinants such as the role of the mobile medium, the development of technology, personalisation of the message and regulation regarding privacy have been considered when studying consumers’ willingness to accept the mobile ad (Leppäniemi and Karjaluoto, 2005). Other studies included perceived risk and perceived utility as the key factors in mobile advertising effectiveness (Bauer, Barnes, Reichardt and Neumann, 2005). When consumers are exposed to the ad, variables such as personal characteristics, user motives, social norms, time, mode and location influence the way consumers process the information coming from the ad (Barnes, 2002).
Considering all these previous studies, authors as Park, Shenoy and Salvendy (2006) constructs the framework of effectiveness for mobile advertising as three broad categories of factors that may influence consumers’ behavioural responses: (1) advertisement factors, (2) environmental factors, and (3) audience factors. The interaction between all these factors is set in the context of the persuasive hierarchy model (Petty and Cacioppo, 1981, 1983; Vakratsas and Tim, 1999).
The second factor is the environment in which the advertisement is set up. The authors maintain that customers usually are not interested on the advertisement, and because of that, capturing their attention is necessary to achieve the intended goal of the ad. The context of the ad can be viewed as a source of distraction for customers. Recent evidence has shown that the more information included both in the context and the situation in which the ad is presented, the less mental resources are assigned to the ad (Shah, Mullainathan and Shafir, 2012).
The third factor is the audience. It is characterized for being subjective due to the individual differences among customers. This factor is analysed through the information theory (Proctor and Zandt, 1994), which shows that the ad works as a stimulus in the sensation stage (mainly influenced by the design of the ad), creating a perceived image for customers. Depending on whether the customer has been exposed to a familiar stimulus before or not, this stimulus, along with previously built attitudes and experiences, helps customers to retrieve memories from long-term memory and recognize the ad. During the process of sensation and perception, the environmental determinants play an important role determining the customer’s degree of involvement. The level of involvement plays an important role guiding advertisers to build improved messages to persuade customers (Petty and Cacioppo 1981, 1983). When consumers present low level of involvement, marketers should rely on the peripheral route to persuade them, which means using positive or negative cues to change their attitudes. Differently, when consumers are highly involved, marketers should use the central route of persuasion by employing strong arguments appealing to their reasoning and cognition.
as emotions, feelings and familiarities are used to evaluate the product (Aaker, Stayman, and Hagerty, 1986). Throughout the affective dimension, consumer’s behavioural responses such as purchase intentions, can be achieved. First, the affective processing generates attitudes towards the advertisement, which are simple evaluations that customers form towards the advertisement. These evaluations have an impact on consumers’ attitudes toward the brand (Edell and Burke, 1987), which are relatively enduring, one-dimensional summary evaluations of the brand that presumably energizes behaviour (Spears & Singh, 2004). Subsequently, the attitudes towards the brand influence consumer’s behaviour through the formulation of behavioural intentions (MacKenzie, Lutz, and Belch 1986; Batra and Ray 1986).
However, this framework only explains how advertising factors operate, hence further literature is needed to link the already explained factors and the two dependent variables tested in this study.
The two variables used as proxies to measure the mobile’s marketing effectiveness along two stages of the customer journey are the following: consumer’s brand awareness for the pre-purchase stage and consumer’s pre-purchase intentions for the pre-purchase stage. The concept of brand awareness is related to the strength of the brand node or trace in memory, as reflected by consumers' ability to identify the brand under different conditions (Rossiter and Percy, 1987). It is also related to purchase intentions because consumers tend to buy a familiar product. Greater levels of brand awareness lead consumers to retrieve brand names from memory easier, affecting subsequently consumers’ purchase intentions (Keller, 1993). This could be explained using the classic hierarchy effect model of advertising effectiveness. First awareness has to be built and then behaviour responses come after (Lavidge and Steiner, 1961; Palda, 1966).
Purchase intention is defined as a consumer decision-making process that studies whether to buy or not a particular brand (Shah, 2012). Previous scholars had examined determinants of purchase intentions, such as price or perceive quality and value to consumers (Zeithaml, 1988). This study uses the framework presented from Park, Shenoy and Salvendy (2006) to understand which are the drivers and factors that affect consumers’ purchase intentions. The reason behind this rests on the larger information that the model of these authors can provide when comparing the two methods to deliver mobile ads. In fact, next chapters explain how two of the three factors that affect mobile marketing effectiveness are crucial when hypothesizing which of the two methods to deliver mobile ads is the most effective in different circumstances.
2.4 Brand familiarity and Advertising effectiveness
Previous literature has defined brand familiarity as the extent of consumers’ direct (trial or using) and indirect experience (advertising or WOM) with a brand. Brand familiarity captures brand’s knowledge structures, which are the brand associations that consumers hold in their memory (Alba and Hutchinson 1987; Kent and Allen 1994; Campbell and Keller 2014). The experiences and brand knowledge structures that consumers may have could come from a variety of sources, such as previous exposures to marketing advertisements of the brand, previous usage or consumption, family and friends that use it, knowledge about the brand elements (logo, symbols, packaging) (Campbell and Keller 2014). Prior studies have argued the multidisciplinary construction of the brand’s familiarity concept. The dimensions in which this variable is created are: (1) interpersonal familiarity, (2) product familiarity and (3) brand communication’s familiarity (Krishnan, 1996; Korchia 2001). When considering brand familiarity, the omnipresent nature of the variable makes it suitable to be considered as a variable that characterize all brands in the market, and therefore it is an inescapable condition in which all brands can be classified. Consumers may have different brand familiarity levels with brands advertised, and this could have different consequences when considering which method to deliver mobile ads is the most suitable to achieve marketers’ objectives.
consumers attitudes towards the advertisement (stimulus) and subsequently the attitudes towards the advertised brand. Previous literature has linked brand’s familiarity and purchase intentions throughout the attitudes formed toward the brand (Laroche, Chankon and Lianxi, 1996). The model suggested by these authors incorporates the concept of brand familiarity as an antecedent of consumers’ attitudes towards the brand. The explanation of that rests on the positive relationship between repeated exposures to an object and the probability to have positive attitudes towards this object (Zajonc and Markus, 1982). In other words, when consumers are exposed to an ad that contains a familiar stimulus (eg. brand) the repetition of being bared again to that stimulus makes consumers have a positive attitude towards it.
In the case of brand awareness, when consumers are exposed to a familiar stimulus, the probabilities to recall and recognize that stimulus increases. In fact, studies have shown a high correlation between brand awareness and brand familiarity variables when testing brand awareness through brand recognition tests (Axelrod 1968; Haley and Case 1979; Baker, Wesley, Danny, and Nedungad, 1986). While brand familiarity remains on the formation of brand knowledge structures in consumer’s memory, brand awareness is based on the strength of these nodes when consumers encounter the brand. The correlation between them appears because the recognition task used to measure brand awareness can be biased by the already knowledge structures settled on consumers’ memory which come from other sources. To cope with that, more recent studies have used brand recall as the instrument to measure brand awareness (Barwise and Strong, 2002). By asking consumers to recall the brand name from a list, consumers’ familiarity with other brands play the same role as the one testing, consequently it eliminates familiarity biases and it allows to isolate the effect of a particular advertisement.
Figure 1: Broad conceptual model of the study
As stated before, when considering two stages of the customer journey, the outcome variables used to measure advertising’s effectiveness are different, thus to clarify the previous conceptual model, figures 2 and 3 show how the model varies depending on the stage of the customer journey used to compare both methods.
Figure 3: Conceptual model of study 2
Hypothesizing the presented conceptual model leads to use the already explained knowledge from previous literature. The research from Barwise and Strong (2002) shows that push-based method to deliver mobile ads affects positively consumer’s brand awareness. That is because the ads form or reinforce brand knowledge structures in consumers memory, making stronger the brand nodes held in consumer memory. This argument is used to cope with the lack of literature regarding the other method to deliver mobile ads. In general terms, both methods to deliver mobile ads could affect positively consumer’s brand awareness because either an SMS or mobile display advertisements deliver brand information to consumers that may create or reinforce consumer’s brand knowledge structures.
Previous literature results show that mobile ads delivered by pull-based method (mobile display advertising) affect positively consumer’s purchase intentions (Bart, Stephen, and Sarvary, 2014). In addition, parallel literature has found a positive influence of push-based method on consumer’s attitudes towards a brand (Chaniotakis, 2010). Then, using the model of Park, Shenoy and Salvendy (2006) both messages either an SMS or mobile banner could be used as the stimulus of information that is processed affecting attitudes toward the ad, and subsequently attitudes toward the brand and purchase intentions.
hypothesis. Regarding Brand awareness, a prior research about the comparison between verbal description messages and illustration messages shows that respondents perform better in memorability test when being exposure to images compared to verbal description (Rossiter and Percy 1980; Levin 1981; Levin, Anglin, and Carney, 1987; Park, Shenoy and Salvendy 2006). When considering both methods to deliver mobile ads, the characteristics of banner ads make more suitable to have a greater brand recall rate compared to SMS ads. In addition, the research from Adis and Kim (2013) shows that entertainment plays an important role when considering advertising effectiveness regarding brand awareness. Entertainment in the context of advertising is related to the ability to fulfil audience needs for escapism, diversion, aesthetic enjoyment or emotional release (McQuail, 1983; Ducoffe, 1996). Entertainment perceptions are influenced by the content media used on the advertising. Therefore due to the differences between the two methods to deliver mobile ads, it could be argued that banner ads, having richer content quality, have more probabilities to be perceived as more entertaining because compared to SMS ads. Thus, it is expected that banner ads have a greater effect on consumers’ brand awareness rates compared to SMS ads.
H1.a: Pull-based method to deliver mobile ads has a greater effect on consumers’ brand awareness than push-based method.
From the model of Park, Shenoy and Salvendy (2006), consumers’ purchase intentions are influenced by the attitudes towards the advertisement and the brands that customers form when processing the information captured from the stimulus (advertising). The results show that ad’s content and design are the most influential factors when considering consumers perception of the stimulus (sensation). When comparing the multimedia content of pull-based method and the text content of push-based method, the attention level to a stimulus is greater affected by images rather than text (Pieters and Wedel, 2004). Hence, it could be argued that because of the richer message’s design of banner ads compared to SMS ads, consumers can perceive easier the stimulus coming from pull-based methods. Considering this, the easier to perceive the ad, the larger probabilities to form both attitudes and purchase intentions towards the ad.
consumers’ attitudes toward the ad and the brand. Therefore, this finding reinforces the hypothesis that banner ads have a greater effect on purchase intentions compared to SMS ads.
H1.b: Pull-based method to deliver mobile ads has greater effect on consumer’s purchase intentions compared to push-based method.
Brand familiarity plays an important role considering the effectiveness of the ad. When consumers have already certain brand knowledge structures about a brand, the exposition to a marketing advertisement about this brand reinforces these structures making the brand nodes stronger. In fact, the research from Baker, Wesley, Danny, and Nedungad (1986) shows that brand familiarity enhances the probabilities to recognize and recall a brand from memory. Therefore, when consumers are exposed to mobile marketing ads, the effect of these ads on consumer’s brand awareness could depend on consumer’s brand familiarity levels with the brand advertised. Indeed, when consumers already have previous knowledge structures about a brand, either if the ad is delivered throughout pull or push based methods, the levels of brand awareness/recall about this brand should be greater than for non-familiar brands. Thus, it is expected that brand familiarity moderates positively the relationship between the two methods to deliver mobile ads and brand awareness/recall.
H2.a: Brand familiarity affects positively the relationship-effect of pull-based method to deliver mobile ads and brand awareness/recall.
H2.b: Brand familiarity affects positively the relationship-effect of push-based method to deliver mobile ads and brand awareness/recall.
Taking everything into account, it is expected a positive influence of brand familiarity on the relationship between the two methods to deliver mobile ads and purchase intentions.
H2.c: Brand familiarity affects positively the relationship-effect of pull-based method to deliver mobile ads and purchase intentions.
H2.d: Brand familiarity affects positively the relationship-effect of push-based method to deliver mobile ads and purchase intentions.
When considering brand familiarity as a factor that moderates the comparison between the two methods to deliver mobile ads, the lack of previous literature leads the study to apply the theory of competitive interference between brands. Competitive interference is defined as the extent to which the effect of advertising is diluted by the presence of competitor’s brands (Burke and Srull 1988; Keller 1991; Mandese 1991; Kent and Allen 2015). The study from Kent and Allen (2015) suggest that familiar brands are not negatively influenced in terms of brand recall neither by competitors’ familiar brand or competitors’ non-familiar brands presence. Nevertheless, non-familiar brands suffer competitive interferences regarding brand recall when being exposed together with both familiar and non-familiar brands from competitors. Considering these findings and comparing both methods to deliver mobile ads, this study assumes that because pull-based methods places advertisements in a context in which multiple brands can be presented (e.g., social media, web browsers and web pages), non-familiar brands should be delivered throughout pull-based methods (SMS). Therefore, it is expected that brand familiarity moderates positively the superiority effect of pull-based method when the aim is to increase consumers’ brand awareness.
H3.a: Brand familiarity moderates positively the larger effect of pull-based method on consumer’s brand awareness compared to push-based method.
with familiar brands, the level of attention to non-familiar brands is lower because consumers focus more on familiar advertising. Thus, it is expected that push-based method is the most effective method when advertising non-familiar brands.
H3.b: Brand familiarity moderates positively the larger effect of pull-based method on consumer’s purchase intentions compared to push-based method.
3. METHODOLOGY
The research is based on two separated studies with a sample of 135 students (62 males, 73 females) from Spain, aged 19-27 (M=25, SD=18,03). The participation rate was 100%. Demographic characteristics were accounted through a questionnaire provided before the actual study, containing questions regarding their gender, age and income level.
3.1 Pre-questionnaire
A previous research about mobile display advertising shows that utilitarian and high involvement products are more effective to be advertised through pull-based method compared to hedonic and low involvement products (Bart, Stephen and Sarvary, 2014). Therefore, a pre-evaluative test following the steps described on Bart, Stephen and Sarvary (2014) was conducted before the survey to consider which products were classified into these categories. With the results provided by this questionnaire, the analysis which tests the relationship between the IVs and DVs, was made more accurately.
from the toilet paper category, telecommunication contracts scored 2,5 points out of 12,5 as “very utilitarian”.
Regarding product involvement, the athletic shoes category obtained a larger score on product involvement than the other categories. Almost half of the sample (45%) indicated higher scores when answering the question regarding purchasing a pair of athletic shoes is an important decision for them. In addition, the toilet paper category was considered as the lowest product involvement among 65% of respondents.
Considering this, the empirical study which tests the effect of the conceptual model variables used athletic shoes as a very hedonic and high involvement product category, and toilet paper as very utilitarian and low involvement product category.
3.2 Study design
The research was composed by two studies, both testing the effectiveness of the two methods to deliver mobile ads but using different outcome variables. The first study tested the effectiveness in terms of brand awareness/recall (pre-purchase stage), while the second study tested the effectiveness in terms of purchase intentions (purchase stage). The participation rate for each study was 100%, meaning that all respondents were assigned to both studies.
The number of mobile ads for each sub-group is four; two well-established brands (one delivered by SMS and the other by banner) and two invented brands (one delivered by SMS and the other by banner).
Table 1: Study design: classification of the conditions by outcome variable, method and brands.
3.3 Procedure
The study started with a questionnaire to evaluate personal and demographic characteristics. Then, participants were asked about some distraction questions in order to entertain them until the measurement of the outcome variables took place. Finally, after respondents were randomly assigned to a certain condition, questions regarding the dependent variables and covariate variable (brand familiarity) were asked.
After the pre-survey questionnaire, participants were exposed to four mobile ads, where the variation regarding the ads remained on the type of brand (well-established vs invented) and method used to deliver the ads (pull vs push based methods). When consumers were exposed to the ads, the survey showed mobile screen shots to obtain higher external validity (Drossos, Giaglis, Lekakos, Kokkinaki and Stavraki, 2014). After that, questions measuring the impact of both methods to deliver the ads on both dependent variables were provided.
3.4 Variables
The independent variable is the method used to deliver mobile ads that is represented by mobile banner ads in the case of pull-based method, and SMS ads in the case of push-based method. Regarding covariates, the study manipulates brand familiarity by asking respondents about their familiarity level for both well-established and invented brands. The reason behind
CONDITIONS BRAND AWARENESS PURCHASE INTENTIONS
1.a: Athletic shoes category Banner ads: Adidas and Yax SMS ads: Nike and Voyons 1.b: Toilet paper category Banner ads: Scottex and
Suavemiel
SMS ads: Cottonelle and Aloen
2.a: Athletic shoes SMS ads: Adidas and Yax Banner ads: Nike and Voyons
2.b: Toilet paper category SMS ads: Scottex and Suavemiel
that rests on the broader spectrum of familiarity levels that consumers could have. The expected brand familiarity levels for well-established brands are larger than the invented brands, therefore, when considering both well-established brands and invented brands, the analysis of brand familiarity not only asses the differences between respondents with an specific brand type, but it also considers the differences on brand familiarity that consumers could have regarding a new released brand in the market as well as a well-established brand. The total number of brands is eight, four well-established brands and four invented brands.
The other covariate variable is product’s type and involvement. Previous studies have shown the impact of this variable on at least one of the dependent variables tested on this study. Customers could consume certain product because it provides utilitarian or hedonic value to them.
The dependent variables differ in both studies. The first study analyses the impact of the delivery method variable on consumer’s brand awareness together with the moderating effect of brand familiarity. The second study will test the effect of the delivery method variable on consumer’s purchase intentions, including the moderating effect of brand familiarity as well.
Because of the differences of both studies, the characteristics and content of the advertisements presented will vary depending on the ad goal set up by marketers when presenting the ad. Some ads will be oriented to create brand building, that is brand salience and visibility, while others will be designed to obtain direct response from respondents. Direct response ads aim to stimulate a direct order and traffic that results in sales of the product advertised (DMA, 2000; Barwise and Strong, 2002).
3.5 Measurement
scale to classified them (1=”mostly hedonic”, 9=”mostly utilitarian”). Regarding product involvement, two items on five-point Likert scales were used (1=”strongly disagree”, 5=”strongly agree”). In the empirical study, the variable is measured as a dichotomous variable (0=“utilitarian/low involvement”, 1=“hedonic/high involvement”) within subjects designed.
Following the experiment of Laroche, Kimand and Zhou (1996), two questions were formulated to indicate consumer’s familiarity with the brand advertised. The measurement was based on two nine-point scale (1=“no information”, 9=“great deal of information”; 1=“no previous experience”, 9=“a lot of previous experience”). Therefore, the brand familiarity variable was initially constructed as a continuous variable due to two reasons. First, it allows to examine the differences in terms of means between well-established and invented brands. Secondly, it allows to divide the data on two categories; low-middle and middle-high familiarity when assessing the interaction effect of the variable of each category. However, when making the comparison between these two categories, the consideration of the variable changes into a dichotomous variable (0=“non-familiar”, 1=“familiar”) between subjects designed. By including invented brands, the expected results of that could show how low familiarity levels moderates the effect of the different methods to deliver the ad on both dependent variables. On the other hand, well-established brands are expected to score higher familiarity levels. Therefore this kind of brands could help to explain the other part of the brand familiarity’s spectrum.
In addition, the operative system variable is going to be constructed as a dichotomous variable (1=“IOS”, 2=“Android”) and measured through a question with two options at the beginning of the survey.
semantic differential scale (interested”, 7=“very interested”; 1=“not-at-all-likely”, 7=“extremely likely”) (Putrevu and Lord, 1994; Xu, Oh and Teo, 2009).
In contrast, consumers brand awareness is measured following the study of Barwise and Strong (2002). The authors used consumers’ brand recall to test consumers’ memorability of the brand. Consistent with the definition of brand awareness coming from Rossiter and Percy (1987), the task to evaluate consumers’ brand recall consists on recalling brand names from a prompted list provided by the experimenter. Barwise and Strong (2002) study used 6 weeks between the exposure of the ad and the test of the variable. Because of practical reasons, this research used only extra time to make respondents focus on another task before being asked about the dependent variable. To do so, the study includes some questions regarding the design and style of the context in which the ads are set up, and four questions regarding the personal innovativeness level. After these distraction questions, respondents must choose which of the listed brands have been advertised during the survey. Therefore, this variable is constructed into a dichotomous variable (0=“not-recalled”, 1=“recalled”). During the data collection, the responses were initially grouped into two different variables: brand awareness of invented brands and brand awareness of well-established brands. These two variables were used when analysing brand recalled differences between low-middle brand familiarity levels (invented brands awareness) as well as brand awareness differences between middle-high familiarity levels (well-established brand awareness). In addition, when measuring the differences between these two groups, the data was structured into a longitudinal form, accounting for repeated measures of brand familiarity factor, and creating a total brand awareness variable. The first row of the variable corresponds to the measure of the invented brands while the second one corresponds to the measurement of the well-established brands.
3.6 Statistical Analysis
When contrasting the hypothesis, the nature characteristics of the two dependent variables together with the different approaches of brand familiarity variable make the analysis more complex. Table 5 illustrates the different statistical tools applied on each circumstance:
BRAND FAMILIARITY
APPROACH
DEPENDENT VARIABLE METHOD
Dichotomous
Brand awareness GEE: generalized estimated equations model of repeated measures with binary logistic linkages (Zeger, Liang and Albert, 1988). Purchase intentions Repeated measures ANOVA (Weinfurt,
2000).
Continuous
Brand awareness Two binary logistic regressions (Maroof, 2012).
Purchase intentions Two linear regressions (Montgomery, Peck and Vining, 2013).
Figure 4: statistical tools used to test the research hypothesis.
In addition, a mean comparison and cross-tabs tools were developed together with T-test and Chi-square tests to get deeper insights as well as for checking for robustness of the results. While continuous variables’ means were compared and validated throughout T-tests (Malhotra, 2010), dichotomous variables’ frequencies were evaluated throughout cross-tabs and validated by Chi-square tests (Malhotra, 2010).
4. DATA ANALYSIS AND FINDINGS
4.1 Data Reliability
familiarity for invented brands (see table 3 and figure 4 in the appendix). The results suggested that Adidas, Nike, Scottex and Cottonelle can be combined to form brand familiarity of well-established brands. In addition, in the case of Yax, Voyons, Aloen and Suavemiel brands, the results also suggested a reliable construct of brand familiarity for invented brands variable. Despite that, a T-test was developed to validate the statistical differences between the two groups of brands in terms of brand familiarity. The results from the test were statistically significant (see table 4 in the appendix) where invented brands scored 1,197 and well-established brands 4,091 in terms of means (see table 5 in the
appendix). Thus, as expected it could be argued that well-established brands are more familiar
than invented brands.
Concerning the purchase intentions variable, two items were asked to each participant. In this case, the reliability analysis was developed by analysing the reliability in the construction of purchase intention score for each brand (see table 6 in the appendix). Finally, after checking their reliability all brands were grouped together into two purchase intention variables, one for well-established and the other for invented brands.
The reliability analysis validates the construction of all variables. The analysis was developed with Cronbach Alpha tests following the validation study of Hair, Anderson, Tatham and Black (1998) where Cronbach Alpha’s scores higher than 0,7 means an accepted reliability of the construction.
4.2 Factors affecting Brand awareness
Figure 4 shows the classification of the different approaches by which the analysis of the two dependent variables can be undertaken. Then, two different analyses were performed to test the hypothesis and give the reader another perspective of the outcomes.
4.2.a First approach: Brand familiarity as a dichotomous variable
A general estimated equation model of repeated measures with binary logistic linkages was developed to analyse the main effect of the different methods of delivering mobile ads on the dependent variable. Because respondents were asked to respond to brand recall tests regarding both types of brands, the dependent variable was analysed in a longitudinal way. Each respondent was represented in two rows of data where brand recall for non-familiar brands (invented) corresponded to the first row while brand recall response for familiar brands (well-established) corresponded to the second row.
The estimated model was performed separately for each of the effects because of practical matters when interpreting the estimated coefficients and the odds ratio. In other words, the lack of hierarchical structure when introducing the variables on the GEE model make the analysis more suitable to be performed in various steps, introducing each of the factors that explain the dependent variableseparately.
The results from the analysis of the main effect of the ad delivery methods suggest that there are statistically significant differences (see table 7 in the appendix) on consumers’ brand recall probability between banner and SMS methods (χ²[1, N=135] = 12,246; p<0,01). In addition, the estimated parameters suggest that the probability to recall the brands is lower for banner (B=-0,864) compared to SMS method. The odds ratio is 0,421 (EXP(B)=0,421) meaning that the probability to recall the brands is 0,421 times lower for banner ads compared to SMS ads. These coefficients are statistically highly significant(p<0,01) (see table 8 in the
appendix).Hence, the hypothesis H1.a was rejected.
The results from the direct effect that brand familiarity variable has on consumers' brand awareness showthat there are statistically significant differences (see table 9 in the appendix) on brand recall probability between familiar and unfamiliar brands (χ²[1, N=135] = 29,238;
case of familiar brands than non-familiar brands (B=1,444). The odds ratio is 4,239 (Exp(B)=4,239) meaning that familiar brands are 4,239 times more probable to be recalled compared to non-familiar brands. These coefficients were statistically significant (p<0,01
)
(see table 10 in the appendix).
The moderating effect that brand familiarity has on the relationship between the ad delivery method variable and consumers' brand recall variable was statistically significant. There are significant differences on brand recall when consumers have been exposed to a banner versus an SMS of a familiar versus an unfamiliar brand (χ²[3, N=135] = 40,479; p<0,01) (see table
11 in the appendix).
The estimated coefficient for banner ads in the case of familiar brands were not statistically significant, therefore these coefficients as well as the odds ratios were not used to determine the magnitude of interaction effect (see table 12 in the appendix). However, a personalised table and Chi-square tests were performed to account for the significance of the differences between frequencies (see table 13 and 14 in the appendix). The results from the cross-tab show significant differences between familiar and non-familiar brands delivered through the banner method χ²[1, N=130] = 29,682; p<0,01). The frequencies show that familiar brands were more often recalled (50 recalls) compared to non-familiar brands (19 recalls). Then, familiar brands performed better (38% of respondents recalled the brands) than non-familiar brands (14% of respondents recalled the brands). Hence, the hypothesis H2.a was accepted. The estimated coefficients for SMS ads for familiar and non-familiar brands were statistically significant. Familiar brands delivered by the SMS method are 2,436 (Exp(B)=2,436) times more probable to be recalled compared non-familiar brands delivered by the same method (see table 12 in the appendix). A cross-tab together with Chi-square tests were performed to check for the robustness of the results. As the previous results suggest, the frequencies and the tests (χ²[1, N=140] = 5,201; p<0,05) validate the assumption of a moderating positive effect of the brand familiarity variable on the relationship between method and brand recall variables (see table 14 in the appendix). Familiar brands were recalled 57 times (40% of respondents recalled the brands) compared to 45 times in the case of non-familiar brands (32% of respondents recalled the brands) (see table 13 in the appendix). Hence, the hypothesis H2.b was accepted as well.
Because of the non-significance coefficient of the banner method for familiar brands (see
or non-familiar was performed using the cross-tab and the Chi-square tests. The non-significance of the test (χ²[1, N=135] = ,416; p>0,05) concludes the non-significance of the
differences on brand recall between both methods to deliver the ads when the brand is familiar (see table 14 in the appendix). Therefore, both methods perform equally well on brand recall tests when larger familiarity levels were observed (see table 13 in the appendix). Hence, hypothesis H3.a was rejected.
Although the research hypothesis did not cover the control variable of product type, part of the result section should be dedicated at least to contrast the findings that have already been uncovered in previous literature. The results coming from a crosstab together with the Chi-square tests suggest that there were no statistically significant differences on brand recall between an utilitarian/low involvement products and hedonic/high involvement products (χ²[1, N=270] = 2,562; p>0,05) (see table 16 in the appendix). The same results applied for the differences between each method, where there were no statistically significant differences on brand recall between both types of products delivered either by SMS method (χ²[1, N=140] = 0,075; p>0,05) nor by banner method (χ²[1, N=140] = 2,144; p>0,05). Furthermore, there were no significant differences on brand recall between SMS and banner methods when the brand advertised was hedonic/high involvement (χ²[1, N=140] = 2,972; p>0,05). However, this was not the case for utilitarian/low involvement (χ²[1, N=140] = 8,001; p<0,01) where the SMS method performed better than banner method on brand recall task (see table 17 and 18
in the appendix).
4.2.b Second approach: Brand familiarity as a continuous variable
The results from consumers' brand awareness for invented/non-familiar brands showed that only the full model test against a constant was statistically significant, indicating that the predictors as a reliable set deferred between respondents who recall and not recall the brands (chi square=23,436; p<0,05 with df=8). The Negelkerke’s R² of .213 indicates a low relationship between prediction and grouping. Prediction success overall was 63,7% (66,2% for non-recall and 60,9% for recall).
B Wald Df Sig. Exp(B)
Product type ,721 3,375 1 ,066 2,057
Method -1,542 15,874 1 ,000 ,214
Brand familiarity -1,514 ,003 1 ,954 1,065
Method*BF ,107 ,021 1 ,885 1,113
Constant -4,671 ,171 1 ,680 ,631
The Wald criterion demonstrated that only the method used to deliver the ad variable made a significant contribution to prediction (p=0,000). The method used to deliver the ad B coefficient indicates that the likelihood to recall the brand is higher for SMS ads compared to banner ads. Exp (B) value indicated that the odds ratio was 0,214 as large, and therefore the probability to recall the brands was 78,6% lower for banner ads compared to SMS ads. However, the rest of variables were not statistically significant predictors.
B Wald Df Sig. Exp(B) Product type ,177 ,155 1 ,694 1,194 Method -,127 ,081 1 ,776 ,881 Brand familiarity ,208 ,786 1 ,375 1,232 Method*BF -,319 ,950 1 ,330 ,727 Constant -3,352 1,297 1 ,255 ,035
The results coming from both regressions regarding the interaction effect and main effect of brand familiarity variable suggest that there are no significant differences on the probability to recall either familiar or non-familiar brands between the different levels of brand familiarity for each type of brand. In other words, there are no significant differences on the probability to recall a non-familiar brand between the different low-middle brand familiarity levels. The same occurred when only considering familiar brands, where no significant differences on the probability to recall a familiar brand between the different middle-high brand familiarity levels were found.
4.3 Factors affecting Purchase Intentions
4.3.a First approach: Brand familiarity as a dichotomous variable
A one way repeated measures ANOVA test was conducted to compare the effect of the method to deliver the ads variable on consumers' purchase intentions when the brand is familiar and unfamiliar. There was a significant effect of brand familiarity variable, Wilks’ Lambda=0,604 F(1,131)=85,758; p<0,01 (see table 19 in the appendix). A two-pair sample t-test was developed for the comparison between conditions. A first paired samples t-t-test (F(1,131)=85,758; p<0,01) indicated that there was a significant difference between purchase intentions of familiar (M=3,407) and non-familiar brands (M=2,144) (see tables 22 and 23 in
the appendix).
The results from the estimated marginal means of the interaction effect between the ad delivery method variable and the brand familiarity variable were statistically significant. There were significant differences on consumers' purchase intentions (see table 29 in the
appendix) when participants were exposed to a banner ad (F(1,131)=49,482; p<0,01; eta partial square=0,274) between familiar (M=3,356) and non-familiar (M=2,162) brands (see table 28 in the appendix). Therefore, hypothesis H2.c was accepted. Additionally, the same
occurred in the case of SMS method. The marginal estimated means revealed significant differences (see table 29 in the appendix) on consumers' purchase intentions when participants were exposed to an SMS (F(1,131)=36,995; p<0,01; eta partial squared=0,220) between familiar brands (M=3,356) and non-familiar brands (M=2,162) (see table 28 in the
appendix). Hence, the hypothesis H2.d was accepted.
When adjusting the Bonferroni pair-comparison of the interaction effect to the ad-delivery method variable (see table 30 in the appendix) the results illustrate the role of brand familiarity on the comparison between the two methods to deliver the ads. The results from the four-paired comparison show that there were no significant differences on consumers purchase intentions between respondents exposed to a banner or SMS when the brand was familiar (F(1,131)=0,030; p>0,05; eta partial square=0,000) (see table 30 in the appendix). In the case of non-familiar brands the results also showed that there were no significant differences on consumers' purchase intentions (F(1,131)=0,135; p>0,05; eta partial
square=0,001) when the respondents were exposed to an SMS or banner ads (see table 30 in the appendix). Hence, the hypothesis H3.b was rejected.
The estimated marginal means results, regarding the interaction effect of the product type variable and method variable, suggest that there are no statistically significant differences on consumers' purchase intentions means when the utilitarian/low involvement products (F(1,131)=0,021; p>0,05; eta partial square=0,000) are delivered by neither SMS or banner methods. The findings are similar for the hedonic/high involvement product category (F(1,131)=0,009; p>0,05; eta partial square=0,000) (see table 37 in the appendix). However, there were significant differences on consumers' purchase intentions means when analysing the product types against each other (F(1,131)=7,729; p<0,01; eta partial square=0,056) (see
table 35 in the appendix). The results from the univariate tests show that the hedonic/high
4.3.b Second approach: Brand familiarity as a continuous variable
Two different linear regressions were constructed to observe the different effect that could appear when comparing: (1) the different middle-high brand familiarity levels on purchase intentions for well-established brands, (2) the different low-middle brand familiarity levels for invented brands. Two hierarchical linear regressions were modelled to predict consumers' purchase intentions using age, income, gender, product type and mobile operative system in the first block. Method used to deliver the ad was included into the second block. Finally, brand familiarity and the interaction effect of brand familiarity with method were included in the third block of the regression.
The results of purchase intentions for well-established brands indicate a significant equation for the first, second and third blocks (F(8,126)=7,730; p=0,00), with an R² of 0,329. Respondents' predicted purchase intention is equal to 3,192 + 1,511 (ProdType), where product type is coded as 0=toilet paper, 1=athletic shoes. Participants’ purchase intentions were 1,511 points higher for athletic shoes than toilet paper. However, when examining the individual effect of the variables, neither the direct effect of the predictor method nor the interaction effect with the moderator was statistically significant.
Regarding purchase intentions for invented brands the results suggest a non-significant equation for all blocks (F(8,126)=1.997, p = 0,052), with an R² of 0,113. Besides, neither the direct effect of the independent variable nor the interaction effects with the moderator were statistically significant. Unstandardized coefficients Standardized coefficients t Sig. VIF Product type ,085 ,038 ,434 ,665 1,073 Method ,002 ,001 ,010 ,992 1,023 B. Familiarity -,177 -,080 -,301 ,764 9,982 BF*Method ,522 ,316 1,190 ,236 9,994 Constant 1,430 - 2,681 ,008 - 5. DISCUSSION
stages of the customer journey considered. Finally, after controlling for two variables that previous literature found to have important implications in the context of mobile marketing effectiveness, the research adds, once more, value to the existing literature and defines some managerial implications for the future of mobile marketing.
This chapter follows the previous structure of dividing in two separate blocks the conclusions for each of the two dependent variables analysed.
5.1.a Pre-purchase stage
The presented results negate the hypothesis that banner ads are more effective than SMS ads during brand recall tests. The percentages show a success rate of 73% in the case of SMS ads and 53% in the case of banner ads. Hence, independently from the type of brand considered, a push based method is the most effective method during the pre-purchase stage of the customer journey. These findings are partly consistent with previous literature. The study from of Barwise, Patrick and Strong (2002) measured consumers’ brand awareness through brand recall tests and found that SMS ads increased consumers' brand awareness up to 63%. The percentages’ differences between both studies could be due to the different time-frames employed for exposing the participants to the ads (six weeks vs few minutes) until the brand recall task took place. Besides, it could be assumed that one of the factors affecting this comparison between methods could be the lack of pre-attentive perception to the ad. The study from Dreze&Hussherr (2003) showed that certain elements have an influence on consumers' pre-attentive perception of the ad and consequently on consumers' brand awareness. One of the main elements that could have the greatest impact on consumers' brand recall during this study could be the small size and the bad location of the ad.
literature (Kent & Allen, 1994), where invented brands (lowest familiarity levels) are negatively affected by the competence interferences coming from the higher familiarity brands. In addition, both methods to deliver the ads perform equally well on brand recall tests when the brands where familiar. The same did not occur for low-middle familiarity brands, where a push-based method seems to be more effective than a pull-based method.
Besides, the type of product did not have an effect on brand recall variable. This is reflected in the lack of significance in probability of brand recall when comparing the two product types. Therefore, the findings contribute to existing literature by showing that both product categories performed uniformly when measuring the mobile marketing effectiveness during the pre-purchase stage. In addition, when observing the interaction effect with the method variable there were no statistical differences on the probability to recall the brand except for the case when comparing which method is the most effective to deliver utilitarian/low involvement products. The differences suggest that a push-based method is more effective than pull-based method for that product category.
5.1.b Purchase stage
The presented results suggest that independently from the brand familiarity, the behaviour of consumers exposed to ads either delivered through pull-based method or push-based method is not affected differently during the purchase stage of the customer journey. Analysing the data, it seems that both methods have performed equally bad with slightly non-significant differences.
display advertising campaigns do not present highly significant results as well as relatively bad performance on consumers' purchase intentions in terms of means.
Regarding the effects of the brand familiarity variable on consumers' purchase intentions, the results suggest that ads for middle-high familiar brands are more effective than low-middle familiar brands during the purchase stage of the customer journey. In addition, it has been probed that both ad delivery methods are more effective for middle-high familiar brands compared to low-middle familiar brands. These results are consistent with previous literature. Laroche, Kim and Zhou (1996) found that the more familiarity consumers have with a brand, the more confidence and positive attitudes towards the brand have. When comparing both methods in each scenario of familiarity, the results indicate a non-significant difference on effectiveness between both methods when the brands where middle-high familiar or low-middle familiar. However, a trend could be observed where the push-based method is more effective than the pull-based method for low-middle familiar brands. This trend contrasts with the trend observed for middle-high familiar brands, where the pull-based method seems to be more effective than the push-based method. Finally, the results from the two linear regressions show non-significant differences in the effectiveness of both methods to deliver the ads when comparing different participants from both low-middle and middle-high brand familiarity levels.
involvement product, and where toilet paper category was classified as utilitarian and low involvement product. In any case there is still a job to do regarding the comparison between both methods, because although the differences were not significant, there is a trend where pull-based method seems to be more effective than push-based method when advertising both product type categories.
5.2 Implications and managerial recommendations
When setting up a mobile marketing campaign marketers should account for previous information about the target group in the market. However, when the brand familiarity levels and product type considerations of the brand advertised are unknown, marketers should consider a push-based method as the most effective method during the pre-purchase stage when consumers are considering and evaluating brands before purchasing them. At the same time, marketers should not prefer any method to deliver mobile ads during the purchasing stage based on effectiveness. In fact, further research should account for other forms to deliver the ads for each method, because it has been proved that SMS ads and banner ads are not really effective for triggering purchase behaviour.
Brand familiarity is an important determinant for the success of a mobile marketing campaign. Marketers from newly launched brands should consider a push-based method instead of apull-based method in order to be more effective during the pre-purchase stage of the customer journey. At the same time, marketers from well-established brands that possess middle-high brand familiarity levels should not waste much time on deciding which method is the most effective to target consumers on the pre-purchase stage of the customer journey. However, when the consumer is about to purchase a product, marketers from both newly launched and well-established brands should not focus on distinguishing methods based on effectiveness, but rather try to combine both methods and/or search for more quality and rich media messages.
utilitarian/low involvement or hedonic/high involvement. On the other hand, marketers targeting consumers during purchase stage should consider a push-based method to be more effective than apull-based method when the product advertised is considered a utilitarian and low involvement product. When the product advertised is considered hedonic/high involvement, marketers should not make a distinction between methods because both perform uniformly when triggering purchase behaviour.
5.3 Limitations and further research
The comparison between the two methods is not broad enough to specify whether all mobile pull-based methods or push-based method are the most effectives for each of the circumstances explained above. Future researches should consider the inclusion of other forms of pull-based methods (video banners, native ads or interactive banners) as well as the possible combinations of the different ad elements that could have an impact on the dependent variables (size, colours, shapes, location).
When analysing consumers' brand awareness, brand recognition tasks where not considered during this study because of the problematic bias and high correlations that could be hadwith the brand familiarity variable. In addition, the purchase stage effectiveness was not measured throughout all outcome variables that encompass all actions that consumers take along the stage. Therefore, further researches should include more outcome variables to deliberate which method is the most effective on each of the stages of the customer journey.
Additionally, the scope of this research did not take account of an examination of the third stage of the customer journey because of time and complexity factors, as other outcome variables should be considered into the model. Thus, to complete the scope for the outcomes coming from this research, following researches should comprise the last stage of the customer journey.
consideration of a broader sample of participants ranging from different ages, professions and nationalities for future researches.
Finally, further analyses should be done accounting for a wider variety of product type categories to examine the different degrees by which this variable could have a different effect on the relationship between the researched variables.
ACKNOWLEDGMENTS
BIBLIOGRAPHY
Aaker, D. A. (1995). “Building Strong Brands”. Free press.
Aaker, D.A., Stayman, D.M., and Hagerty, M.R., (1986). “Warmth in advertising: measurement, impact, and sequence effects”. Journal of Consumer Research, 12 (4), 365–381.
Adis, Azaze-Azizi and Hyung Jun, Kim (2013). “Antecedents of Brand Recall and Brand Attitude towards Purchase Intention in advergames”, European Journal of Business and Management Vol.5, No.18.
Agarwal, R. and Prasad, J. (1998). ‘A conceptual and operational definition of personal innovativeness in the domain of information technology’, Information Systems Research, Vol. 9, pp.204–215.
Alba JW, Hutchinson J. Wesley (1987). “Dimensions of consumer expertise” J. Consum-Res 1987;13 411-54. Andrews, M., Luo, X., Fang, Z., & Ghose, A. (2016). “Mobile ad effectiveness: Hyper-Contextual targeting with crowdedness”. Marketing Science, 35(2), 218-233.
Axelrod, J. N. (1968). "Advertising Measures that Predict Purchase," Journal of Advertising Research, 8, 3-17. Bacile, Todd J., Christine Ye, and Esther Swilley (2014). “From Firmcontrolled to Consumer-contributed: Consumer Co-production of Personal Media Marketing Communication” Journal of Interactive Marketing, 28, 2, 117–33.
Bacile, Todd J., Christine Ye, and Esther Swilley (2014). “From Firm controlled to Consumer-contributed: Consumer Co-production of Personal Media Marketing Communication” Journal of Interactive Marketing, 28, 2, 117–33.
Baker william, Hutchinson J. Wesley, Moore Danny, and Nedungadi Prakash (1986). "Brand Familiarity and Advertising: Effects on the Evoked Set and Brand Preference", in NA - Advances in Consumer Research Volume 13, eds. Richard J. Lutz, Provo, UT : Association for Consumer Research, Pages: 637-642.
Bamba, Fatim and Barnes, Stuart, (2008). “Evaluating Consumer Permission in SMS Advertising". All Sprouts Content. 159.
Barnes, S. (2002).” Wireless digital advertising: Nature and implications”. International Journal of Advertising, 21(3), 399-420.
Bart, Yakov, Andrew T. Stephen, and Miklos Sarvary (2014). “Which Products are Best Suited to Mobile Advertising? A Field Study of Mobile Display Advertising Effects on Consumer Attitudes and Intentions,” Journal of Marketing Research, 51, 3, 270–85.
Barwise, Patrick and Colin Strong (2002). “Permission-based Mobile Advertising,” Journal of Interactive Marketing, 16, 1, 14–24.
Batra, R. and Ray, M.L., (1986). “Affective responses mediating acceptance of advertising”. Journal of Consumer Research, 13, 234–249.
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, 6 (3), 181– 192.
Chaniotakis, I., Lymperopoulos, C., & Soureli, M. (2010).” Consumers' intentions of buying own-label premium food products”. Journal of Product & Brand Management, 19(5), 327-334.
Churchill, G. (1978). “A paradigm for developing better measures of marketing constructs” (Wisconsin working paper, 6-78-16). Madison: Graduate School of Business, University of Wisconsin-Madison.
Culnan, M.J. (1995). ‘Consumer awareness of name removal procedures: implication for direct marketing’, Journal of Interactive Marketing, Vol. 9, pp.10–19.
Danaher, Peter J., Michael S. Smith, Kulan Ranasinghe, and Tracey S. Danaher (2015). “Where, When, and How Long: Factors that Influence the Redemption of Mobile Phone Coupons,” Journal of Marketing Research, 52, October, 710–25.
Dhar, Ravi and Klaus Wertenbroch (2000). “Consumer Choice Between Hedonic and Utilitarian Goods,” Journal of Marketing Research, 37 (February), 60–71.
Dimitris Drossos, Geroge M. Giaglis, George Lekakos, Flora Kokkinaki, and Maria G. Stavraki (2014). “Determinants of effective sms advertising: an experimental study”,
Journal of Interactive Advertising, Vol 7 No 2 (Spring 2007), pp. 16‐27.
DMA. (2000). “1999 Economic Impact: US Direct Marketing Today”. Direct Marketing Association.
Ducoffe, R.H. (1996). “Advertising value and advertising on the web”, Journal of Advertising Research, Vol. 36, No. 5, pp. 21-34.
Ducoffe, R.H., and E. Curlo (2000). “Advertising value and advertising processing”. Journal of Marketing Communications 6: 247–62.
Edell, J., & Staelin, R. (1983). “The Information Processing of Pictures in Print Advertisements”. Journal of Consumer Research,10(1), 45-61.
Ferris, M. (2007). “Insights on mobile advertising, promotion, and research”. Journal of Advertising Research, 47(1), 28–37.
Fong N.M., Fang Z. and Luo X.(2015). “Geo-conquesting: competitive locational targeting of mobile promotions” . Journal of Marketing Research, 52(5), 726-735.
Ghose, Anindya, Avi Goldfarb, and Sand P. Han (2013a). “How is the Mobile Internet Different? Search Costs and Local Activities,” Information Systems Research, 24, 3, 613–31.
Grewal, Dhruv, Yakov Bart, Martin Spann (2016).”Mobile Advertising: A Framework and Research Agenda” Journal of Interactive Marketing, 34, 3-14.
H. Xu, L.B. Oh, H.H. Teo, (2009). “Perceived effectiveness of text versus multimedia location-based advertising messaging”, International Journal of Mobile Communications 7 (2) 154–177.
Haley, R. 1. and P. B. Case (1979). "Testing Thirteen Attitude Scales for Agreement and Brand Discrimination” Journal or Marketing, 43, 20-32.
Hair, J.R; Anderson, R.E.; Tatham, R.L.; and Black, W.C (1998). “Multivariate Data Analysis”. Englewood Cliffs, NJ: Prentice Hall.
