Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. Is digital advertising effective under conditions of low attention? The impact of low attention processing on consumer brand consideration and choice A thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Marketing at Massey University, Palmerston North, New Zealand Irene Irawaty Santoso 2020 ii Abstract A crucial task for digital advertising is to influence choice despite consumers' lack of attention. Although lack of attention can reduce advertising effectiveness, recent research suggests that incidental exposure to ads while accessing digital content can lead to some outcome for the exposed ads. This evidence prompts four critical questions: (1) is digital advertising effective if processed at low attention; (2) can low attention processing increase brand consideration and choice; (3) what specific brand/product characteristics embedded in the ads are likely to influence the effect; and (4) what measures are appropriate to capture the low attention effects. To address the questions, three experimental studies (n = 1,423) were conducted in laboratory and online settings. The research manipulates two conditions for low attention processing, namely divided attention and incidental attention. The results show that, at least in the Twitter environment, advertising is effective even under conditions of low attention. Although focused attention still drives the greatest impact, low attention significantly increases the likelihood of target brands being included in the brand consideration and selected as preferred brand choice more than ‘no exposure’. The low attention effects were obtained without subsequent correct respondent recognition. This shows that brand consideration and choice measures were capable of capturing the low attention effects that the recognition measure failed to do. However, the results for source factors – factors that can moderate the effect of stimuli on the outcome – are more nuanced. Brand familiarity, utilitarian/hedonic products, rational/emotional appeals, and (mis)matching between appeals and brands affect the results in some unexpected ways when they interact with low attention. The thesis makes substantive contributions to the application of attention theory in advertising research, testing methodology for ads that are not actively processed, and design of advertising that can work at low attention. The findings are particularly relevant to address current phenomena such as multitasking, multiscreening, and ad avoidance behaviour. Unless advertisers understand how to make advertising work at low attention, the practice of bombarding consumers with attention-grabbing ads will continue to rise, and ad avoidance will accelerate, which in turn, will put advertising at greater risk of being wasted. iii Acknowledgements I wish to thank all the people whose assistance was a milestone in the completion of my PhD project. First and foremost, I would like to express my deepest gratitude to my research supervisors, Professor Malcolm Wright, Dr Giang Trinh, and Dr Mark Avis. I am highly indebted to Professor Wright for his understanding, generosity, patience, enthusiasm, and continuous encouragement, and for pushing me further than I could imagine. It really is a great honour to work under his supervision. I also wish to express my immense gratitude and thanks to Dr Trinh who helped me get over the hurdle of the analysis of data and the interpretation of results. Dr Trinh was a dedicated mentor whose commitment and patience were second to none. Also, I particularly thank Dr Avis for his invaluable support and assistance in writing this thesis and for expanding my perspectives through insightful discussions. My sincere thanks also go to Dr Daniela Rosenstreich for her support and guidance in the earlier stage of my study. I would like to acknowledge with gratitude all the support that I received from the School of Communication, Journalism, and Marketing at Massey University for the research and conference grants and for the teaching and tutoring opportunities. This thesis would not materialise without the financial support from the Directorate General of Resources for Science, Technology, and Higher Education, Ministry of Research, Technology, and Higher Education of the Republic of Indonesia. My thanks and appreciation also extend to my fellow PhD students who made my learning journey more fun and enjoyable. Finally, I would like to dedicate this thesis to my late Mom who stays in a very special part of my heart. I would like to acknowledge with gratitude the support and love of my family – my Dad, my husband, and my beloved children Nicholas and Patricia. They all kept me going and their love made the hardship of writing the thesis worthwhile. Lastly, I wish to apologise to them for being absent from their lives in the past four years. iv List of Publications from the Thesis Santoso, I., Konopka, R., Rosenstreich, D., Wright, M., & Avis, M. (2017). Conscious and Non-conscious Influences on Consumer Choice. Australia and New Zealand Marketing Academy (ANZMAC) Conference Proceedings, pp. 141 – 144, Melbourne, Australia. Santoso, I., Rosenstreich, D., & Wright, M. (2017). Are we paying too much attention on attention? Australia and New Zealand Marketing Academy (ANZMAC) Conference Proceedings, pp. 481 - 485, Melbourne, Australia. Santoso, I., Wright, M., Trinh, G., & Avis, M. (2019). Are branded tweets effective under the conditions of low attention? Exploring the effects of brand familiarity and product category. Paper presented at the 52nd Academy of Marketing, London, United Kingdom. (Best paper in eMarketing & Digital Marketing track). Santoso, I., Wright, M., Trinh, G., & Avis, M. (2020). Is digital advertising effective under conditions of low attention? Journal of Marketing Management, 1-24. v Table of Contents Abstract …..……………………………………………………………………………….... ii Acknowledgements ……...…………………………………………………………………..iii List of Publications from the Thesis..….……………………………………………………iv Table of Contents…………………………………………………………………...………..v List of Abbreviations ……………………………………………………………………...xiii List of Tables …………………………………………………………………..…………..xiv List of Figures …………………………………………………………….………………..xvi CHAPTER 1: INTRODUCTION ............................................................................. 1 1.1. BACKGROUND ................................................................................................................ 1 1.2. STATEMENT OF THE PROBLEM ....................................................................................... 5 1.3. RESEARCH OBJECTIVE .................................................................................................... 6 1.4. RESEARCH QUESTIONS ................................................................................................... 7 1.5. RESEARCH DESIGN ......................................................................................................... 9 1.6. KEY CONSTRUCTS ........................................................................................................ 10 1.6.1. Low attention ........................................................................................................ 11 1.6.2. Brand consideration set ......................................................................................... 12 1.6.3. Brand choice ......................................................................................................... 12 1.7. SIGNIFICANCE OF RESEARCH ........................................................................................ 13 1.7.1. Theoretical implications ........................................................................................ 13 1.7.2. Methodological implications................................................................................. 15 1.7.3. Practical implications ............................................................................................ 15 1.8. STRUCTURE OF THESIS ................................................................................................. 16 CHAPTER 2: UNDERSTANDING ADVERTISING EFFECTIVENESS ........ 19 2.1. INTRODUCTION ............................................................................................................. 19 2.2. A BRIEF HISTORY OF ADVERTISING .............................................................................. 20 2.2.1. The development of advertising formats and media ............................................. 20 vi 2.2.2. The development of advertising content ............................................................... 22 2.2.3. The development of advertising concepts and definitions .................................... 23 2.3. HOW ADVERTISING WORKS: CURRENT STATUS OF THEORY ......................................... 24 2.3.1. The controversy of advertising theories ................................................................ 25 2.3.1.1. The Strong versus Weak theories ................................................................... 25 2.3.1.2. The Elaboration Likelihood Model of Persuasion .......................................... 27 2.3.1.3. The Integrative Framework of Persuasion Theories ....................................... 28 2.3.1.4. The Low Attention Processing Model ............................................................ 29 2.3.2. The controversy of cognitive versus affective processing in advertising .............. 29 2.3.2.1. Cognitive reactions to advertising challenged ................................................ 30 2.3.2.2. A renewed interest in affective reactions to advertising ................................. 31 2.3.2.3. Emotion and advertising responses ................................................................ 32 2.3.3. A rethinking of attention ....................................................................................... 36 2.3.4. Attention, emotion, and memory ........................................................................... 36 2.3.5. Theoretical advancement based on memory theory .............................................. 39 2.3.5.1. Brand cloud (brand as a neural network of memory) ..................................... 40 2.3.5.2. Mild emotional attachments ........................................................................... 41 2.3.5.3. Brand salience ................................................................................................ 42 2.4. EVALUATING ADVERTISING .......................................................................................... 43 2.4.1. Problems in evaluating advertising effectiveness .................................................. 44 2.4.2. Conceptualising advertising effectiveness ............................................................. 45 2.4.3. Measures of advertising effectiveness ................................................................... 47 2.4.3.1. Traditional methods ........................................................................................ 48 2.4.3.2. Contemporary advertising testing methodology ............................................. 52 2.4.4. Advertising testing methodology used in the current research .............................. 58 2.4.5. Summary of methods and measures of advertising effectiveness ......................... 62 2.5. CHAPTER SUMMARY ...................................................................................................... 63 vii CHAPTER 3: LOW ATTENTION ADVERTISING PROCESSING ................ 65 3.1. INTRODUCTION ............................................................................................................. 65 3.2. ATTENTION ................................................................................................................... 65 3.2.1. Attention as an information filter .......................................................................... 66 3.2.2. Attention as processing capacity ........................................................................... 69 3.2.3. How does attention work? .................................................................................... 70 3.2.3.1. The Spotlight theory and The Gradient Model .............................................. 70 3.2.3.2. The role of attention in advertising processing .............................................. 72 3.2.3.3. Selective visual attention to advertising ......................................................... 77 3.2.3.4. Lessons from eye-tracking research ............................................................... 79 3.2.3.5. Lessons from neuroscience research .............................................................. 82 3.3. LOW ATTENTION ADVERTISING PROCESSING ............................................................... 84 3.3.1. Background ........................................................................................................... 84 3.3.2. Conceptualisation .................................................................................................. 84 3.3.3. Low attention conditions ....................................................................................... 85 3.3.4. How ads create effects under conditions of low attention .................................... 86 3.4. KEY THEORIES OF LOW ATTENTION PROCESSING ......................................................... 88 3.4.1. Dual-process theory .............................................................................................. 88 3.4.1.1. Intuition and reasoning................................................................................... 90 3.4.1.2. Intuitive choices under low attention conditions ........................................... 92 3.4.1.3. Going beyond the dual-process theory ........................................................... 93 3.4.2. Mere exposure effect ............................................................................................. 94 3.4.3. Processing fluency ................................................................................................ 96 3.4.4. Hedonic marking of processing fluency ............................................................... 97 3.5. CHAPTER SUMMARY .................................................................................................... 99 CHAPTER 4: CONTEXTUAL FACTORS AND SPECIFIC HYPOTHESES 103 4.1. INTRODUCTION ........................................................................................................... 103 viii 4.2. DIGITAL ADVERTISING ................................................................................................ 103 4.2.1. Advertising on social media ................................................................................ 107 4.2.2. Advertising on Twitter ........................................................................................ 110 4.3. CONDITIONS AFFECTING BRAND CONSIDERATION AND BRAND CHOICE .................... 112 4.3.1. Brand familiarity ................................................................................................. 113 4.3.2. Utilitarian and hedonic product type ................................................................... 114 4.3.3. Rational versus emotional appeal in advertising ................................................. 115 4.3.4. Utilitarian and hedonic brand type ...................................................................... 117 4.3.5. Matching and mismatching between appeals and brand type ............................. 118 4.3.6. Rational versus intuitive judgements ................................................................... 120 4.4. CHAPTER SUMMARY ................................................................................................... 121 CHAPTER 5: METHODOLOGICAL APPROACH ......................................... 123 5.1. RESEARCH OBJECTIVES ............................................................................................... 123 5.2. METHODOLOGICAL CHALLENGES AND APPROACHES IN THIS RESEARCH ................... 123 5.2.1. Theoretical and methodological perspectives ...................................................... 124 5.2.2. The ecological context of research ...................................................................... 125 5.2.3. Experimentation .................................................................................................. 127 5.2.4. Laboratory versus online experiment .................................................................. 128 5.3. RESEARCH QUESTIONS AND HYPOTHESES .................................................................. 129 5.4. METHODOLOGY .......................................................................................................... 132 5.4.1. Overview of studies ............................................................................................. 132 5.4.2. Variables .............................................................................................................. 135 5.4.3. Treatments ........................................................................................................... 136 5.4.4. Reliability ............................................................................................................ 137 5.4.5. Validity ................................................................................................................ 138 5.4.6. Confounding factors ............................................................................................ 140 5.4.7. Controlling for confounding factors .................................................................... 140 ix 5.4.8. Post hoc examinations ......................................................................................... 141 5.5. CHAPTER SUMMARY ................................................................................................... 142 CHAPTER 6: STUDY 1 – METHOD, RESULTS, DISCUSSION ................... 145 6.1. INTRODUCTION ........................................................................................................... 145 6.2. METHOD ..................................................................................................................... 146 6.2.1. Sample ................................................................................................................ 146 6.2.2. Pre-tests ............................................................................................................... 147 6.2.2.1. Product selection .......................................................................................... 147 6.2.2.2. Brand selection ............................................................................................. 148 6.2.2.3. Stimuli selection ........................................................................................... 148 6.2.3. Main experiment ................................................................................................. 149 6.2.3.1. Experimental procedure ............................................................................... 149 6.2.3.2. Dependent measures .................................................................................... 152 6.2.4. Data analysis ....................................................................................................... 154 6.3. RESULTS ..................................................................................................................... 154 6.3.1. Manipulation check ............................................................................................. 154 6.3.2. Main effects ........................................................................................................ 156 6.3.2.1. Low attention – Main effects ....................................................................... 157 6.3.2.2. Brand familiarity – Main effect and interactions ......................................... 158 6.3.2.3. Product type – Main effect and interactions ................................................ 159 6.3.2.4. Three-way interaction – low attention, brand familiarity, and product type 159 6.4. DISCUSSION ................................................................................................................ 160 6.5. CHAPTER SUMMARY ................................................................................................... 163 CHAPTER 7: STUDY 2 – METHOD, RESULTS, DISCUSSION ................... 165 7.1. INTRODUCTION ........................................................................................................... 165 7.2. METHOD ..................................................................................................................... 166 7.2.1. Sample ................................................................................................................ 166 x 7.2.2. Pre-tests ............................................................................................................... 168 7.2.2.1. Brand selection ............................................................................................. 169 7.2.2.2. Stimuli selection ........................................................................................... 170 7.2.3. Main experiment .................................................................................................. 174 7.2.3.1. Experimental procedure ................................................................................ 174 7.2.3.2. Dependent measures ..................................................................................... 176 7.2.3.3. Data analysis ................................................................................................. 178 7.2.4. Reliability and validity ........................................................................................ 178 7.2.4.1. Reliability testing .......................................................................................... 178 7.2.4.2. Validity testing ............................................................................................. 178 7.2.4.2.1. Exploratory and Confirmatory Factor Analyses ................................... 178 7.2.4.2.2. Face validity of the instrument .............................................................. 181 7.2.4.2.3. Demand effects ..................................................................................... 183 7.3. RESULTS ...................................................................................................................... 183 7.3.1. Manipulation check ............................................................................................. 183 7.3.2. Main effects ......................................................................................................... 185 7.3.2.1. Attention – Main effect ................................................................................ 187 7.3.2.2. Appeals – Main effect and Interactions ........................................................ 188 7.3.2.3. Products – Main effect and Interactions ....................................................... 189 7.3.3. Post hoc examination ........................................................................................... 190 7.3.3.1. Rational and intuitive judgements ................................................................ 190 7.3.3.2. Response latency (reaction time) .................................................................. 192 7.4. DISCUSSION ................................................................................................................ 193 7.5. CHAPTER SUMMARY ................................................................................................... 197 CHAPTER 8: STUDY 3 – METHOD, RESULTS, DISCUSSION ................... 199 8.1. INTRODUCTION ........................................................................................................... 199 8.2. METHOD ...................................................................................................................... 201 xi 8.2.1. Sample ................................................................................................................ 202 8.2.2. Pre-tests ............................................................................................................... 203 8.2.2.1. Brand selection ............................................................................................. 204 8.2.2.2. Stimuli selection ........................................................................................... 206 8.2.3. Main experiment ................................................................................................. 209 8.2.3.1. Dependent measures .................................................................................... 210 8.2.3.2. Data analysis ................................................................................................ 212 8.3. RESULTS ..................................................................................................................... 212 8.3.1. Manipulation check ............................................................................................. 212 8.3.2. Main effects ........................................................................................................ 215 8.3.2.1. Attention – Main effect ................................................................................ 217 8.3.2.2. Matching – Main effect and Interactions ..................................................... 218 8.3.2.3. Brand type and shopping situation - Main effects ....................................... 220 8.3.3. Post hoc examination .......................................................................................... 221 8.3.3.1. Rational and intuitive judgements ................................................................ 222 8.3.3.2. Response latency (reaction time) ................................................................. 225 8.4. DISCUSSION ................................................................................................................ 227 8.5. CHAPTER SUMMARY ................................................................................................... 230 CHAPTER 9: DISCUSSION AND CONCLUSION ........................................... 233 9.1. GENERAL DISCUSSION ................................................................................................ 233 9.1.1. Summary of key findings and hypotheses testing ............................................... 237 9.1.2. Answering the research questions ....................................................................... 240 9.1.2.1. RQ1: Do digital advertisements increase the likelihood of the advertised brand being included in the consideration set and being selected as brand choice under conditions of low attention?............................................................................ 240 9.1.2.2. RQ2: Do brand familiarity and product type enhance the effectiveness of digital advertisements under conditions of low attention? ....................................... 241 xii 9.1.2.3. RQ3: Do emotional (rational) appeals enhance the effectiveness of digital advertisements under conditions of low attention? ................................................... 243 9.1.2.4. RQ4: Does the matching (mismatching) between appeal and brand type enhance the effectiveness of digital advertisements under conditions of low attention? ................................................................................................................................... 246 9.1.2.5. RQ5: Does attention in the processing of digital advertisements affect the likelihood of people making rational versus intuitive judgements? .......................... 247 9.2. CONTRIBUTIONS OF THE THESIS ................................................................................. 248 9.2.1. Theoretical contributions ..................................................................................... 248 9.2.2. Methodological contributions .............................................................................. 251 9.2.3. Practical and managerial contributions ................................................................ 253 9.3. LIMITATIONS OF RESEARCH ........................................................................................ 255 9.4. RECOMMENDATIONS FOR FUTURE RESEARCH ............................................................ 257 9.5. CONCLUSIONS ............................................................................................................. 258 REFERENCES ........................................................................................................ 263 xiii List of Abbreviations AIDA Attention, Interest, Desire and Action ANOVA Analysis of Variance ATR Awareness – Trial – Reinforcement ARF Advertising Research Foundation AVOC Audible and Visible on Complete CFA Confirmatory Factor Analysis CPCV Cost Per Completed View DA Divided Attention EFA Exploratory Factor Analysis EEG Electroencephalography ELM Elaboration Likelihood Model FA Focused Attention fMRI Functional Magnetic Resonance Imaging HED/UT Hedonic/Utilitarian HSD Honest Significant Difference IA Incidental Attention IFPT The Integrative Framework of Persuasion Theories KMO Kaiser-Meyer-Olkin LAP Low Attention Processing MEE Mere Exposure Effect SSTS Situation-Specific Thinking Style UK United Kingdom US United States USP Unique Selling Proposition xiv List of Tables Table 2.1 Methods, measures, and key components of advertising processing ................................... 62 Table 4.1 Summary of key findings of digital adevertising research from 2008 to 2018 .................... 106 Table 5.1 Summary of research questions and hypotheses ................................................................. 131 Table 5.2 Experimental details with between and within subjects factors .......................................... 134 Table 5.3 Summary of variables used in the current research ............................................................. 135 Table 5.4 Summary of experimental treatments ................................................................................. 137 Table 6.1. STUDY 1 - Research questions and hypotheses ................................................................... 146 Table 6.2 STUDY 1 - Recognition ......................................................................................................... 155 Table 6.3 STUDY 1 - One-way ANOVA on recognition ........................................................................ 156 Table 6.4 STUDY 1 - Recognition across groups .................................................................................. 156 Table 6.5 STUDY 1 - Binary logistic regression on brand consideration and brand choice .................. 157 Table 7.1 STUDY 2 - Research questions and hypotheses .................................................................... 166 Table 7.2 STUDY 2 - Sample ................................................................................................................. 167 Table 7.3 Exploratory factor analysis - Rational and intuitive judgements ......................................... 179 Table 7.4 STUDY 2 - Demand effects .................................................................................................. 183 Table 7.5. STUDY 2 - Hits, Misses, False alarms, Correct rejections, and Recognition rates ............... 184 Table 7.6 STUDY 2 - One-way ANOVA: Recognition across attention ................................................. 185 Table 7.7 STUDY 2 - Multiple comparisons - Recognition across attention ......................................... 185 Table 7.8 STUDY 2 - Descriptive - Brand consideration and brand choice ........................................... 185 Table 7.9 STUDY 2 - Logistic regressions on brand consideration and brand choice ........................... 186 Table 7.10 STUDY 2 - Descriptive- Rational and Intuitive judgements ................................................ 190 Table 7.11 STUDY 2 - Tukey HSD: Rational and Intuitive judgements ................................................. 191 Table 7.12 STUDY 2 - T -test paired samples: Rational and Intuitive judgements .............................. 191 Table 7.13 STUDY 2 - Descriptive - Response latency .......................................................................... 192 Table 7.14 STUDY 2 - ANOVA - Response latency ............................................................................... 193 Table 8.1 Study 3 - Research questions and hypotheses ..................................................................... 201 Table 8.2 STUDY 3 - Research design ................................................................................................... 202 Table 8.3 STUDY 3 - Sample ................................................................................................................. 203 Table 8.4 STUDY 3 - Hits, Misses, False alarms, Correct rejections, and Recognition rates ................ 213 Table 8.5 STUDY 3 - ANOVA - Recognition rates across groups .......................................................... 214 Table 8.6 STUDY 3 - Multiple comparisons - Recognition across attention ......................................... 214 Table 8.7 STUDY 3 - Demand effect ..................................................................................................... 215 Table 8.8 STUDY 3 - Descriptive - Brand consideration and brand choice ........................................... 216 Table 8.9 STUDY 3 - Logistic regressions on brand consideration and brand choice ........................... 217 Table 8.10 STUDY 3 - Logistic regression: matching versus mismatching ........................................... 219 Table 8.11 STUDY 3 - Descriptive - Rational and Intuitive judgements ............................................... 222 Table 8.12 STUDY 3 - Multiple comparisons - Rational and Intuitive judgements across groups ....... 224 xv Table 8.13 STUDY 3 - T-test paired sample results: rational versus intuitive judgements .................. 224 Table 8.14 STUDY 3 - Response latency across attention .................................................................... 225 Table 8.15 STUDY 3 - ANOVA - Response latency ................................................................................ 226 Table 9.1 Summary of hypotheses testing ........................................................................................... 239 xvi List of Figures Figure 3.1 A comparison between theories of attention ...................................................................... 67 Figure 6.1 STUDY 1 - Experimental Procedures ................................................................................... 149 Figure 6.2 STUDY 1 - Brand consideration set – cars .......................................................................... 153 Figure 6.3 STUDY 1 - Brand consideration set - shoes ........................................................................ 153 Figure 7.1 STUDY 2 - Experimental procedure .................................................................................... 175 Figure 7.2 STUDY 2 - Brand consideration set and choice – shampoo .............................................. 177 Figure 7.3 STUDY 2 - Brand consideration and choice – toothpaste .................................................. 177 Figure 7.4 Component plot of rational and intuitive judgements ....................................................... 180 Figure 7.5 Rational and Intuitive Judgements Scale ............................................................................ 181 Figure 8.1 STUDY 3 -Brand consideration set ...................................................................................... 211 Figure 8.2 STUDY 3 - Brand choice ...................................................................................................... 212 1 CHAPTER 1: INTRODUCTION This chapter provides contextual information on the present gap in understanding advertising effectiveness, the circumstances where low attention processing of advertising potentially occurs, and the challenges low attention processing causes for the effectiveness of advertising. In particular, the chapter discusses the impact of low attention processing on consumer brand consideration and brand choice. It then briefly introduces the research questions, key constructs, and the implications of research for the development of marketing theory and practice. Finally, an overview of the thesis structure is presented. 1.1. Background Advertising effectiveness has been a primary concern among advertising scholars and practitioners for decades. For advertising to work, one of the major challenges is to gain consumers’ attention. Attention to advertising is defined as “the ability to focus on advertising and also suppress attention to other things in the environment” (Bellman, Nenycz-Thiel, Kennedy, Hartnett, & Varan, 2019, p. 295). Given the dynamics of advertising landscape and consumer behaviour (Dahlén & Rosengren, 2016), an increasing effort has been put into the development of advertising testing methodology. One such effort is a rethinking about consumer attention to advertising and how it relates to effectiveness. Recent findings have revealed that advertising on digital media relies on a different type of attention than traditional media (Rosengren, 2016). However, the literature on newer concepts of advertising attention, such as the Low Attention Processing (LAP) model (Heath, 2007), has predominantly focused on traditional media (television). Moreover, much focus has been directed towards 2 overcoming the problem of inattention rather than accepting it, despite the evidence that positive low-attention advertising effects do occur (Heath, Brandt, & Nairn, 2006). Low attention to advertising may, therefore, not be a problem it is assumed to be. Thus, the focus should not be on overcoming the problem of inattention, but rather on understanding, measuring, and managing it. The review of the literature has identified several gaps. Research has offered some new insights into the ways that we acquire and store knowledge of the world around us, and the impact of these on our decision-making process. For example, ninety-five percent of our thinking process is unconscious (Zaltman, 2003), and most of our judgements and decisions are intuitive rather than rational (Kahneman, 2003). These insights have dramatically improved our understanding of how advertising is processed, yet they are not fully assimilated into the methods for evaluating advertising effectiveness. There is an acceptance that attention forms the base of how advertising ‘works’ (Davenport & Beck, 2000; Nelson-Field, 2020; Stipp, 2016; Venkatraman et al., 2015). Consistent with this, Wright (2016) considers attention as a key element in his model of advertising processing. Here, attention can have conscious (e.g. the brand is remembered) as well as unconscious resultant effects (e.g. mild emotional attachments to the brand). Likewise, most advertising research has focused its investigations on a set of core constructs, namely attention, affect, memory, and desirability (Venkatraman et al., 2015). These constructs can affect advertising effectiveness independently or in combination. Although attention has been recognised as an important factor for advertising effectiveness, there are competing theories to explain how attention to advertising works: the Strong theory, the Weak theory, and affective processing. 3 The first theory regards advertising as a strong force with aggressive intentions (Jones, 1990). Corresponding with this approach is the AIDA model, denoting that advertising persuades people to go through the stages of Attention, Interest, Desire and Action (Strong, 1925). Gaining attention is, therefore, the first and most critical step to advertising effectiveness, and what people remember about the ad or the brand is the test of advertising effectiveness. Following this reasoning, advertising would be effective if it can successfully hold attention. Attention is important because brand learning occurs more readily when people pay attention to the advertisement. The second theory regards advertising as a weak force. Advertising nudges through repetitions to create brand awareness in a sequence of ATR- Awareness, Trial, and Reinforcement (Ehrenberg, 2000). Most advertising is processed at lower attention, and advertising works without having to be recalled (Heath, 2007; Krugman, 1986). Thus, it is legitimate to enquire if what people remember consciously or unconsciously is the test for advertising effectiveness. More recent research and theory has taken a different avenue by focusing on affective reactions as opposed to purely cognitive reactions (Hasford, Hardesty, & Kidwell, 2015) and on the role of emotion in advertising (Poels & Dewitte, 2019). For example, Heath (2012) argues that advertising does not work through a cognitive process, but through an automatic affective process called subconscious seduction. Similarly, advertising works by changing the way consumers feel about the brand (Cramphorn, 2004) or by developing mild emotional attachments to the brand (Wright, 2016). These emotional attachments help make the brand become salient and increase the probability that the brand will be considered for purchase in a buying situation (Romaniuk & Sharp, 4 2004). The overall changes in brand feeling and salience will establish a brand-person relationship effect (Cramphorn, 2004) that indicates the effectiveness of advertising. However, there are some questions marks concerning such an approach. First, it is not clear whether advertising needs to gain active attention in order to be effective. In prior research, low attention to advertising is associated with low sales response (Bellman et al., 2017) or low recall and recognition (Angell, Gorton, Sauer, Bottomley, & White, 2016). The ‘limited capacity framework’ frequently used to examine low attention (e.g. Bellman, Rossiter, Schweda, & Varan, 2012; Jeong & Hwang, 2012; Lang, 2000) proposes that humans have limited attentional capacity to process information and low attention inhibits the processing and storage of that information. However, other theorists provide a different account, with Krugman (1986) considering that learning takes place “even if by quick looks, short attention, and unrecalled exposure or perception” (p. 79). Thus, there is some possibility that advertising can be effective despite low attention. Second, as the purpose of advertising is to establish a brand-person relationship through emotional reactions, there is still a debate about whether emotive advertising works more effectively at lower or higher levels of attention. Heath et al. (2006) argue that emotive advertising will be more effectively processed at lower levels of attention. In contrast, Du Plessis (2005) holds that the main role of emotional appeal in advertising is to attract attention, therefore “it is unlikely that the more emotional an advertisement is, the more it will become low-attention processed” (p. 141). Third, Wright (2016) proposes that the overall testing of advertising effectiveness should cover the whole process: opportunity to see (exposure), attention, activation, 5 and retrieval. However, advertising processing consists of a set of mental processes, some operating at an unconscious level, so the difficulty concerns methodology for evaluating those processes and the effects that occur. Tellis and Ambler (2007) hold that advertising effects are generally weak, difficult to observe, and easily lost and biased by the use of inappropriate measurement. The problem of weak effects will be exacerbated in the context of low attention and raises an interesting question as to whether the behavioural diagnostic measures of brand consideration and choice (Nedungadi, 1990; Trinh, 2015) are capable of capturing the subtle effects from low attention advertising on consumer behaviour. 1.2. Statement of the problem Theories and models currently used to evaluate advertising effectiveness have not provided a plausible explanation for how advertising might work effectively. At one end, advertising is believed to work by grabbing our attention and persuading us to remember why a brand is the best. At another end, advertising is thought to establish emotional connections with a brand that can influence our attitudes, decisions, and behaviours in ways we cannot consciously detect. In this view, high attention to advertising itself does not support this process, and may actually inhibit it. The competing theoretical approaches above yield some managerial concerns. First, advertisers are unsure whether to design ‘attention-grabbing’ or ‘low-attention’ advertising. Second, while much of advertising may be processed at lower attention, particularly in the context of digital media, the academic literature provides little guidance on how to improve advertising effectiveness under such conditions. Specifically, whether there are factors associated with brand/product characteristics, such as brand familiarity and utilitarian/hedonic choices, that should be considered. 6 Third, various measurements for advertising effectiveness have been established but no validated measurement is available for the effects under conditions of low attention. Brand consideration and choice have been proposed as alternative measures for advertising effectiveness (Wright, 2016) but have not been tested for low attention effects. 1.3. Research objective The objective of the current research is to investigate (i) whether digital advertising is effective if it is processed at low attention, (ii) the effect of this type of processing on consumer decisions, (iii) what variables are likely to influence the effect, and importantly, (iv) how to measure the effects in a meaningful way. Therefore, the overarching research question is: Is digital advertising effective under the conditions of low attention? Although a general understanding of low attention processing is useful, the expected benefits of low attention in enhancing advertising effectiveness depend on specific conditions; that is, the contextual factors associated with particular brand/product characteristics, advertising properties, and the measurement used to determine the effect. Advertisers need to know these conditions to ensure the success of their advertising. 7 1.4. Research questions The digital landscape is vast and digital advertising is broadly defined as “advertising utilising digital media” (Lee & Cho, 2019, p. 3). To limit the scope of research, the current research investigates the low attention processing of branded tweets. The reasons why Twitter provides an appropriate context for this research are detailed in Chapter 4. While many refer to branded messages on Twitter as social media advertising, several authors highlight the significant differences among the platforms and advocate the term ‘digital advertising’ instead of ‘social media advertising (Rodgers & Thorson, 2017; Voorveld, Noort, Muntinga, & Bronner, 2018). Rodgers and Thorson (2018) describe ‘digital advertising’ as a variety of forms of branded content, including any branded content that appears in social media contexts, messages from companies in the form of blogs, tweets, Facebook posts, and comments that consumers make about brands in digital contexts. The overarching question is operationalised as follows: RQ1: Do digital advertisements increase the likelihood of the advertised brand being included in the consideration set and being selected as brand choice under conditions of low attention? Advertising can subconsciously trigger emotions such as some sort of familiarity with the brand, which can increase preferences for the advertised brand (Heath, 2007). However, there is a more fundamental reason to investigate brand familiarity. Digital media provide unprecedented opportunities for new/small/unfamiliar brands with limited budget and resources to level out the ‘playing field’ and to advertise side-by- side with some of the biggest brands in the world. Thus, it is interesting to know 8 whether advertising for familiar or unfamiliar brands will gain more advantages at low attention. In addition, consumers may undertake different processing strategies when evaluating advertising regarding hedonic versus utilitarian products. For utilitarian products, the inherent product features are important (Mittal, 1989), therefore, information processing is likely to be highly elaborative (Dhar & Wertenbroch, 2000). Alternatively, for hedonic products, consumers are more likely to rely on psychological interpretations, including ego gratification, social acceptance, and sensory stimulation (Mittal, 1989). As such, it can be assumed that ads for hedonic products may be more effective under low attention conditions than those for utilitarian products that require deliberation to process factual information. Thus, RQ2 is posed: RQ2: Do brand familiarity and product type enhance the effectiveness of digital advertisements under conditions of low attention? The literature highlights a controversy surrounding emotive advertising (Du Plessis, 2005; Heath, 2007), in particular, whether emotional appeal can be effectively processed at low attention. To help resolve this issue, RQ3 is posed: RQ3: Does emotional (rational) appeal enhance the effectiveness of digital advertisements under conditions of low attention? A commonly held conception is that marketing communication, including advertising, should strive for consistency in their messages (Bhat & Reddy, 1998 ; Park, MacInnis, Priester, Eisingerich, & Iacobucci, 2010), that is, using rational appeal for a utilitarian brand, and emotional appeal for a hedonic brand (Johar & Sirgy, 1991; Shavitt, 1992). 9 However, there is also evidence that incongruity can lead to better outcomes (Klein & Melnyk, 2016; Lange & Dahlén, 2003). The in-salience hypothesis (Guido, 2001) posits that incongruent stimuli attract more attention and lead to greater awareness of the information. Attempting to resolve the controversy, RQ4 is posed. RQ4: Does the matching (mismatching) between appeal and brand type enhance the effectiveness of digital advertisements under conditions of low attention? It is tempting to think that low attention processing may increase the probability of the brand being considered or chosen when the rational, deliberative choice process is rejected in favour of an intuitive, automatic process. However, the empirical support for this argument is lacking. The theoretical core corresponds to the dual-process theories of information processing described as System (type) 1 and System (type) 2 processing (Kahneman, 2003). Therefore, RQ5 is posed to address the issue. RQ5: Does attention in the processing of digital advertisements affect the likelihood of people making rational versus intuitive judgements? 1.5. Research design Experimental research on attention has often been criticised as being unrealistic and having little relevance to address real-life issues (Morales, Amir, & Lee, 2017; Romaniuk & Nguyen, 2017). The current research adopts a naturalistic, quantitative approach with a classic experimental design, which allows for manipulation of attention and measurement for the corresponding effects. The research design allows the natural variability of attention and a realistic state of distraction from the environment to be integrated into the experiments, with the aim to increase (1) the 10 ecological validity of the research (Bryman & Bell, 2015), that is, how much the processes appearing in the experiment reflect the real-world situations, and (2) the relevance of the findings to address the managerial problem concerning declining advertising effectiveness due to inattention to advertising. Study 1 investigates the main and interaction effects of brand familiarity (familiar versus unfamiliar brands) and product categories (utilitarian versus hedonic products) on consideration set and brand choice under low attention conditions (divided and incidental attention) and compares that to control (RQ1 and RQ2) Study 2 examines the main and interaction effects of advertising appeals (emotional versus rational) on consideration set and brand choice under low attention conditions (divided and incidental attention) and compares that to high attention condition (focused attention) and control (RQ1, RQ3, and RQ5). Study 3 explores the main and interaction effects of matching versus mismatching between advertising appeals (emotional versus rational) and brand types (rational versus hedonic) on consideration set and brand choice under low attention conditions (divided and incidental attention) and compares that to high attention condition (focused attention) and control (RQ1, RQ4, and RQ5). 1.6. Key constructs The key constructs in the current research are low attention, brand consideration set, and brand choice. A general description of each of these constructs is presented here, while more detailed description will be presented in subsequent chapters. 11 1.6.1. Low attention Attention is described as a brain mechanism that selects a small subset of inputs available in our environment to be further processed while ignoring others (Davenport & Beck, 2013). Attention mediates the processing of, the response to, and the effects of advertising on consumer choice (Chandon, Hutchinson, Bradlow, & Young, 2009; Janiszewski, Kuo, & Tavassoli, 2013; Pieters & Wedel, 2004). Therefore, attention is a vital consideration in any consumer research (Romaniuk & Nguyen, 2017), in particular, advertising research. Low attention describes a condition where attention is not fully directed at the stimulus and thereby less deliberate processing takes place. Low attention conditions may occur as a result of the subjects' lack of motivation to process the information, dual-task performance (doing multiple tasks simultaneously), or perceptual blindness (failure to perceive a stimulus because it is irrelevant to the current task). In the current research, low attention is operationalised as divided attention (Spataro, Cestari, & Rossi-Arnaud, 2011), where participants are exposed to different information and have to process the information simultaneously; and as incidental attention (Shapiro, MacInnis, & Heckler, 1997), in which the target information is not the focus of attention. The delineation of divided and incidental attention provides greater conceptual clarity than the umbrella term of ‘low attention. While prior research has looked at the low attention phenomenon as a ‘strategy to avoid advertising’, the current research diverges from such an approach and considers low attention as a unique ‘strategy to approach advertising’ in digital media. 12 1.6.2. Brand consideration set A brand consideration set refers to a set of all available brands which the consumer thinks of on a particular choice occasion (Roberts & Nedungadi, 1995). Brand consideration may be seen as a process of narrowing down alternatives in a specific consumption situation, with the set typically ranging from one to seven brands (Trinh, 2015). Wright (2016) proposes that brand consideration and choice measures may be sensitive to the conscious as well as subconscious effects of attention and, therefore, are better alternatives for measuring advertising effectiveness than sales, given sales data can be confounded by variables other than advertising. Brand consideration set has been considered as a measure for a combination of consumers’ cognitive, affective and behavioural responses to brand messages (Franzen, 1999). Consequently, methods of preference measurement such as attitude towards the brand and purchase intention may produce misleading results if brand consideration set is not taken into account (Bremer, Heitmann, & Schreiner, 2017, citing Gilbride and Allenby, 2004). 1.6.3. Brand choice Given limited cognitive resources, incorporating consideration sets into the choice process better explains the decisions made by the consumers (Kardes, Kalyanaram, Chandrashekaran, & Dornoff, 1993; Shocker, Ben-Akiva, Boccara, & Nedungadi, 1991). A rational view to brand choice submits that brand evaluation is a function of brand utility, and people make a choice based on the attributes that are considered important (Nedungadi, 1990), and the choice decisions are often made based on combined inputs: inputs available in the environment as well as information retrieved from memory (A. Y. Lee, 2002). These inputs serve as retrieval cues for the brand by 13 increasing the strength of activation of the brand node in memory (Anderson & Bower, 1974). In this view, the consumer deliberates about the brands in the consideration set to arrive at a final choice (Shocker et al., 1991). An alternative account known as the affect-as-information theory posits that people tend to rely on their feelings as a source of judgement-relevant information when cognitive resources are limited (Schwarz, Bless, & Bohner, 1991). In addition, Kahneman (2003) argues that most of our decisions are made intuitively based on simple decision rules (heuristics). Thus, in the context of low attention processing in which encoding of brand information is restricted and brand cognitions are lacking, the consumers will more likely rely on their affect or heuristics rather than brand cognitions to arrive at a final choice. 1.7. Significance of research The current research has substantive implications in terms of theory, methodology, and practice, and should be of great importance to advertisers and researchers alike. 1.7.1. Theoretical implications First, the current research contributes to advertising research on attention. It advances an understanding of advertising effectiveness that reduced attention to advertising is not always detrimental. A significant contribution is to show how advertising effectiveness changes with the amount of attention paid to process the advertisements. While prior research has shown that processing multiple ads or ignoring ads are detrimental to advertising effects, the current research finds positive effects of advertising in such circumstances. Although FA still drives the greatest increase in brand consideration and choice, the findings show that actively ignoring advertising has a greater effect than processing multiple ads simultaneously at a given time. 14 Additionally, exposure to advertising at low attention, even with few repetitions and different ads, significantly has a higher effect compared to the ‘no attention’ condition where there is no exposure. Importantly, the effects occur despite a lack of recognition of the exposed ads, supportive of the notion that advertising can work through affective processing without the need to activate working memory (Cramphorn, 2004; Wright, 2016). The findings particularly challenge the role of memory theory in advertising effectiveness measurement. Furthermore, specific brand/product characteristics, such as brand familiarity and hedonic versus utilitarian product/brand type, and ad properties, such as rational versus emotional appeals, and the (mis) matching between appeals and brand types, moderate the effect of low attention on brand consideration and choice in unexpected ways than what has been studied previously. Second, the current research contributes to understanding the attention theory in psychology and neuroscience. The Spotlight theory and the Gradient Model explaining how visual attention works are compared to determine which theory provides a better prediction for the low attention effects. The findings suggest that the Gradient model of attention offers a better explanation for the low attention effects, with a greater effect found in the focused attention where the stimuli are the centre of attention, and a weaker effect in the divided and incidental attention. Third, the current research makes an important contribution to the application of attention theory in advertising. The findings suggest that attention is allocated to the object and not necessarily to the location. It advances an understanding of ‘banner blindness’ (Hervet, Guérard, Tremblay, & Chtourou, 2011) and ‘looking without seeing’ (Mack, 2003) that we rarely see what we are looking at unless attention is directed to it. As such, the current research concerns the fate of advertising that is seen, 15 yet not fully attended, and postulates that a certain level of processing occurs for these ads although many details of this processing fail to register into consciousness. Overall, the research contributes to the advertising effectiveness theories as well as to the understanding of the role of visual attention in advertising processing and decision- making. 1.7.2. Methodological implications This study offers a novel testing methodology capable of capturing the effects of advertising that is processed at a low level of attention. Evaluating the effectiveness of advertising is particularly challenging because the low-attention effects are subtle, easily lost or biased by the use of the wrong analysis or measurement method, and the effects can occur immediately or at a later time (Tellis, 2004). Importantly, the proposed methodology properly predicts the effects of advertising under natural conditions. It incorporates the natural variability of consumer attention and realistic environmental distractions (Romaniuk & Nguyen, 2017), thereby increasing the relevance of research to address current issues surrounding consumer inattention and resistance to advertising. The simplicity of the method allows for easy replication by both researchers and practitioners. 1.7.3. Practical implications The findings provide insights into capitalising the low attention processing and leveraging advertising effectiveness. First, the study offers a practical tool for assessing the effectiveness of advertising which can be used for pretesting or testing the on-going advertising. Brand consideration and brand choice measures can be alternative measures for advertising effectiveness as they are sensitive to the subtle low attention 16 effects. Second, it offers insights into the issue of declining advertising effectiveness due to consumer inattention and resistance to advertising. Unless advertisers understand that specific ads can work at low attention, the practice of bombarding consumers with ‘attention-grabbing advertisements’ will continue to rise, and so advertising avoidance will accelerate. Third, several source factors that have been investigated provide specific guidelines for designing an ad and media planning. For example, the results suggest that ads for unfamiliar (new) brands might be more effective if they are published in high rather than low cluttered media; whereas ads for well-established utilitarian brands might be more effective if targeted to high multitaskers. 1.8. Structure of thesis The thesis has nine chapters which are organised as follows: Chapter 1 introduces the research topic and describes the impetus and rationale behind it. Chapters 2, 3, and 4 present the literature review pertinent to the research. The review is organised under three titles (i) Understanding advertising effectiveness, (ii) Low attention advertising processing, and (iii) Contextual factors and specific hypotheses in this research. Chapter 5 outlines the methodological approach. The choice for a realistic methodological approach (Romaniuk & Nguyen, 2017) is briefly explained. Other decisions made in relation to the research method, experimental procedure, data collection and analysis are justified. 17 Chapters 6, 7, and 8 detail Study 1, Study 2, and Study 3 respectively, including the method, analysis, results, discussions, and limitations. Chapter 9 discusses key findings, draws general conclusions of the findings, and presents the theoretical, methodological, and practical implications, limitations of the current research and recommendations for future research. 18 19 CHAPTER 2: UNDERSTANDING ADVERTISING EFFECTIVENESS 2.1. Introduction Advertising is a huge industry with spending growing rapidly, surpassing 560 billion U.S. dollars in 2019 (Enberg, 2019), but many criticisms of advertising are about its effectiveness. Lack of visual attention has been cited as the main reason why advertising is less effective than expected (Liu-Thompkins, 2019). The problem of attention was seriously discussed at the 2016 Cannes Advertising Festival and the Wall Street Journal wrote, “One of the biggest topics at Cannes this week has been how to win consumers’ attention amid a swiftly changing technology landscape and backlash against a perceived overload of advertising” (Perlberg, 2016). While advertisers have always been concerned with how to capture and keep attention focused on advertising, given the level of distraction on digital media, most advertising is, arguably, processed at lower attention as consumers largely ignore and avoid advertising (Cho & Cheon, 2004; Duff & Lutchyn, 2017; Heath, 2007). Currently, there is no clear evidence whether advertising is effective under low attention conditions. This chapter reviews theories, models, and measurements pertinent to evaluating advertising effectiveness, with a specific focus on effectiveness under low attention conditions. To better understand the advertising concepts, it is necessary to first examine a historical perspective on advertising and how the formats, media, contents, and definitions have evolved over the years. 20 2.2. A brief history of advertising 2.2.1. The development of advertising formats and media McDonald and Scott (2007) describe the development of traditional advertising in four periods: the earliest commerce, the industrial revolution and mass marketing, the mid- 19th century, and post-World War II. Then, Rodgers and Thorson (2017) describe the development of modern advertising on digital media. The earliest form of commerce. The earliest form of advertising known as 'outdoor advertising' was used as tradesman's signs and tavern signs in the ancient civilisations of Egypt, Mesopotamia, Greece, and Rome, as well as political advertising-thought graffiti found on stone walls in Rome as early as 4000 BC. The industrial revolution and mass marketing. Between 1760 and 1830, the industrial revolution prompted the development in printing, allowing for mass distribution of advertising through handbills, posters, and later, newspapers. One important contribution of advertising in this era was the creation of brand, defined as a label that designates an individual product and differentiates it from its competitors. Advertising generated selective consumers that would specifically ask for the brands they knew, leading retailers to stock products of the brands more than those of non-brands. The mid-19th century. This period was characterised by the introduction of the concept of Unique Selling Proposition (USP), a summary statement used to meaningfully differentiate the brand from the competition. Print media became more established, and the advertising's role for branding became more prominent, due to the range of products competing for consumers. Each product was branded with a unique name and 21 consumers asked for a particular product by mentioning the brand. The brand, the packaging, and the unique features (USP) were communicated to the consumers through advertising. The function of advertising evolved from selling to persuading, creating demands for the brand. In addition to printed ads, radio advertising started to emerge in the U.S in the 1920s, marking the era of broadcast advertising. The post-World War II. After World War II, advertising was in high demand for developing new brands and introducing new products. In the 1940s, the introduction of television made it possible for advertising to reach a larger number and different kinds of audiences. The use of radio and television as advertising media has dramatically changed the nature and focus of advertising by blurring the lines between entertainment and advertising, a strategy used to increase the receptivity of the message by the audience (Ducoffe & Curlo, 2000). Modern advertising. Since 1994, the Internet has become an attractive medium for advertising (Rodgers & Thorson, 2017). Internet advertising later evolved into digital advertising by utilising mobile devices and digital signage as advertising media. Digital ads take interactive formats such as advergames, search, banners, pop-ups, and social media ads (Dahlén & Rosengren, 2016). Among a variety of digital channels, advertisers have particularly recognised social media for its capacity to disseminate branded-content fast to a large audience, to increase brand awareness and brand recall, and to build brand loyalty (Goncalves et al., 2016). The competition for consumers' attention is more severe in digital media than in the traditional media as consumers have more control over their attention and exposure to advertising and they can easily move away at the click of a mouse or the touch of a finger on mobile devices (Rodgers & Thorson, 2000). Thus, the greatest challenge of today's advertising is to cope with 22 the consumer's limited attention and to design content that can favourably influence consumers despite limited attention. 2.2.2. The development of advertising content Advertising content can broadly be categorised into rational, informational ads and emotional, transformational ads. Rational, informational ads. From the 1800s to the early 20th century, advertising served as an information provider in which ads were mostly text-based, classified, and informational, and directly told the consumers what was for sale, how much it cost, and where it was sold (McDonald & Scott, 2007). Later, advertising included some persuasive information (reason-why), known as the 'salesman in print' approach. This approach was the key principle in the AIDA model - Attention, Interest, Desire, and Action (Strong, 1925) and other 'hierarchy of effect’ models . According to the models, ads must first introduce the brand name, then arouse consumer's interest toward the brand by describing the physical characteristics, and then convince the consumer that the brand is worth purchasing (Weilbacher, 2001). These models are not without criticism. The concept of hierarchical (temporal sequence) effects of advertising upon which the model is based cannot be empirically supported (Vakratsas & Ambler, 1999), and the models are regarded as an intuitive, non-validated explanation of how advertising works (Weilbacher, 2001). Emotional, transformational ads. In the mid-1900s, researchers started to believe that emotion played an important role in decision-making and more artistic ads with enhanced originality and creativity known as 'soft-shell' advertising (Armstrong, Lukeman, & Patnaik, 2010) were proposed to appeal to emotion. In this respect, 23 advertisers believe that the audience would respond more positively to emotional appeals than logical arguments. For this reason, ads were primarily designed for entertainment rather than explicitly selling (McDonald & Scott, 2007). Moreover, Heath et al. (2006) propose that emotional content of advertising is more effective if processed at lower attention; conversely, Du Plessis (2005) argues that emotional content serves to attract attention, so it will be unlikely that it is processed at lower attention. While content can be an important moderator for advertising effectiveness (Gavilanes, Flatten, & Brettel, 2018), the content itself does not determine the advertising outcomes. Rather, the subsequent cognitive, affective, or behavioural reactions resulting from the interaction between ad content and the processing variables, namely attention, activation, and retrieval (Wright, 2016), modify effectiveness. 2.2.3. The development of advertising concepts and definitions Traditionally, advertising was defined as “a paid, mediated form of communication from an identifiable source, designed to persuade the receiver to take some action, now or in the future” (Dahlén & Rosengren, 2016, p. 334). The primary purpose of advertising was to help communicate the brand to target consumers, with specific objectives to inform, persuade or remind the consumers about the brand. Recently, to address the rapid shifting of advertising resources from traditional to digital channels, Dahlén and Rosengren (2016) provide an updated definition of advertising as “brand- initiated communication intent on impacting people” (p. 334). This definition has broadened the objectives of advertising: in addition to informing and persuading, advertising aims to create consumer-brand connection, involving all the cognitive and 24 affective associations and behavioural reactions to the brand (Cramphorn, 2004). Thus, successful advertising modifies the way people relate to the brand so that people will think more favourably about the brand after viewing the ads. This new definition is particularly relevant to address the nature of digital advertising. In particular, it eliminates the word 'paid' from the previous definition, suggesting that advertising can be organic, non-paid, or paid (Fulgoni, 2015). Organic ads are delivered via the company's own website or the brand's fan page in social media, and include not only brand-sponsored messages, but also messages posted by the consumers, such as comments about particular brands (Rodgers & Thorson, 2018). Consumers can interact with the brands, exhibit their enthusiasm for the brands, and convince others that the brands are worth purchasing (Kwon, Kim, Sung, & Yoo, 2014). However, due to the algorithmic settings of digital media that favour paid over organic content, advertisers still rely on paid advertising for higher reach and better targeting (Fulgoni, 2015). Overall, the definitional differences above imply that advertising can now appear in different media channels and in many different formats. Advertising can be published by brands, but can also be created, co-created, and distributed by and among the consumers, which is sometimes beyond the control of the advertisers. As the current research is situated in the context of advertising effectiveness, it is necessary to examine relevant theories and models in advertising. 2.3. How advertising works: Current status of theory Despite decades of research in advertising, the Advertising Research Foundation (ARF) has recently commissioned a major research project entitled ‘How advertising works’ (Stipp, 2016), suggesting that it remains an important subject of inquiry to date. 25 2.3.1. The controversy of advertising theories There are competing theoretical considerations about how advertising might work: (1) advertising is a strong force working through persuasion and conversion as suggested by 'the Strong theory', and (2) advertising is a weak force working through nudging as suggested by 'the Weak theory or through subconscious seduction as proposed by the LAP model. 2.3.1.1. The Strong versus Weak theories The Strong theory The Strong theory assumes that advertising is a strong force with aggressive intentions – to boost sales, to attack competitors and increase market share, or to drive loyalty (Jones, 1990). Successful advertising needs to attract attention and present compelling brand-related arguments which will be encoded and stored in memory for later retrieval in a purchasing context. This process assumes strong engagement and interest with the ads, namely, attentiveness. Consistent with this theory is the sequential models of advertising, the hierarchy of effects models. The most popular, the AIDA model, denotes that advertising persuades people to go through the stages of Attention, Interest, Desire and Action (Strong, 1925). Gaining attention is, therefore, the first and most critical step to advertising effectiveness. Similar assumptions can be found in more recent assessments of advertising effects. For example, in their seminal book, Persuasive Advertising: Evidence-based Principles, Armstrong et al (2010) propose that advertising should, “ask customers to remember the brand name and arguments, especially when customers’ actions are likely to occur much later” (p. 137). The assumption is that if the consumer remembers the brand, the brand will be more likely considered for purchase in a buying situation. Therefore, at the heart of the Strong 26 theory is the view that advertising is a process of active learning (Heath, 2007), which is essentially effortful, and advertising that does not receive much attention and does not change attitudes will, therefore, be less effective. Endorsing the Strong theory, advertisers incorporate novelty and creativity in ads to capture and hold attention (Pieters, Warlop, & Wedel, 2002; Underwood & Everatt, 1996). If the Strong theory were right, repetitive advertising would not be effective as it lacks novelty and would fail to attract attention. In contrast, Ehrenberg (2000) asserts that advertising effects are generally weak and are built through repetitions. According to this stance, even if advertising seems ineffective initially, time and repetitions will ensure its ultimate success. The Weak theory provides an explanation for this process. The Weak theory According to the Weak theory, advertising works in subtle ways through a process called nudging, or passive learning. Specifically, advertising nudges through repetitions over time to reinforce brand awareness, in a sequence of ATR – Awareness- Trial-Reinforcement (Ehrenberg, 2000). The theory assumes that the consumer's mind is not a blank sheet, but already contains some memories of the brand/product purchasing, usage, and advertisements. The main role of advertising is, therefore, to reinforce associations with these memories so that the brand becomes salient in a purchase situation (Romaniuk & Sharp, 2004). This process of reinforcing memories is called 'activation' (Wright, 2016). In this view, consumers absorb a great deal of advertising information without actively searching for it or consciously evaluating it. As such, advertising works through a weak mechanism that does not require much attentional resource (Barnard & Ehrenberg, 1997). After purchase and use, attitude may or may not be reinforced, but it changes as a result of brand experience rather than as a 27 direct result of the advertising effect. Briefly, advertising works to create brand awareness, then the consumer trials the product; if they like it, they may continue to purchase and advertising can reinforce the attitude. Although the notion of repetitive advertising seems plausible, the amount of repetition is another issue of debate. Ehrenberg (2000) argues that the small, weak effects of advertising persist and will be visible following many repetitions; whereas others suggest that very few (two or three) exposures are more effective (Jones, 1998; Krugman, 1984; Tellis, 1997). More exposures can produce wear-out effects so that advertising can lose its effectiveness and produce a negative impact (Schmidt & Eisend, 2015). In addition, repetitive advertising can also activate counter-argument, scepticism, and annoyance (Grunert, 1996), leading to a negative attitude toward the brand. One criticism about the Weak theory is that this theory focuses on the cognitive process to build brand-related memory and largely ignores the importance of emotions such that repetitions can generate boredom. 2.3.1.2. The Elaboration Likelihood Model of Persuasion The differences between the strong and weak theories are reconciled in the Elaboration Likelihood Model (ELM) (Petty, Cacioppo, & Schumann, 1983), which is assumed to be the process of advertising persuasion and is a highly influential framework used by advertising researchers (Kitchen, Kerr, Schultz, McColl, & Pals, 2014). The basic tenet of the ELM model is the presence of two persuasion routes: the central and peripheral routes. The central route applies the principle of the Strong theory; while the peripheral route applies the notion of the Weak theory that allows superficial attitude change under conditions of low engagement (Petty & Cacioppo, 1986). The route that an individual takes called the elaboration likelihood depends on the nature and amount of critical 28 thinking that the individual gives to the persuasive message (i.e. advertisement). The more thinking is given, the better the individual can be influenced and persuaded by the message, and thus the more effective the message is. Attention is the mechanism that facilitates thinking, and therefore enhances elaboration, which in turn, improves the persuasiveness of advertising (Berger, Wagner, & Schwand, 2012). However, arguably, the ELM cannot really be called a reconciliation model between the Strong and Weak theories, as the effects of the peripheral route are implicitly considered inferior to the central route (e.g. the attitude change is not enduring). 2.3.1.3. The Integrative Framework of Persuasion Theories Meyers-Levy and Malaviya (1999) introduced the Integrative Framework of Persuasion Theories (IFPT) as an extension to the ELM. Their model considers three levels of processing. The first two, systematic and heuristic, share similar properties to the central and peripheral routes in the ELM. The third level is experiential processing, in which the amount of cognitive resource (attention) that people devote to this processing is so meagre that only scant information processing occurs. Although they admit that this type of processing plays very little role in the current information processing theory, the IFPT provides initial support to the notion that there is another type of processing besides effortful (high attention) processing. In summary, although less constrained than the Strong theory, the ELM and IFPT both advocate principles of the Strong theory for the most effective advertising. A key limitation of the Strong and Weak theories is likely to stem from the assumption that most of our thinking process (learning about brands) takes place in our conscious minds (cognition). The Strong theory suggests high attention, while the Weak theory proposes low but repetitive attention. Likewise, the ELM implies that the attitude change through 29 central versus peripheral routes results from active thinking, which demands a high level of cognitive resources. However, these theories are silent about the role of affect (emotion) in advertising. 2.3.1.4. The Low Attention Processing Model Heath et al. (2006) questioned the assumption of the Strong models, in particular, the focus on cognitive informational learning and the assumption that high attention drives advertising effectiveness. In particular, for the most part, our thinking process takes place in our unconscious minds – all those memories, emotions, thoughts, and other processes that we are not aware of or that we cannot articulate (Zaltman, 2003), and most of our decisions are intuitive rather than rational (Kahneman, 2003). The LAP model posits that advertising works through passive or implicit learning without demanding active attention to process (Heath, 2007). The main distinction between LAP and previous models is that LAP emphasises processing at lower attention levels, while the previous models endorse high-attention processing for advertising effectiveness. A detailed description of the model is presented in Chapter 3. 2.3.2. The controversy of cognitive versus affective processing in advertising In addition to the controversy whether advertising is a strong or weak force, another question that has bothered researchers for years is does advertising have more impact on consumers’ thoughts or feelings? The attempts to answer this question have led to an ongoing debate whether cognition or affect is more predictive of consumer behaviour, and to the dissociation between cognitive and affective processing (Morris, Woo, Geason, & Kim, 2002). The debate has a parallel in cognitive psychology concerning whether cognition dominates affect, and thus mediates the impact of affect 30 on decision-making (Lazarus, 1982), or whether affect precedes and is independent of cognition, and therefore, influences decision-making differently from cognition (Zajonc & Markus, 1982). 2.3.2.1. Cognitive reactions to advertising challenged For many years, the analysis of advertising effects has been dominated by the cognitive models of information processing (Bettman, 1979; Grunert, 1996), favouring conscious, effortful processes and recall metrics to assess the processes. In this view, affective reactions to advertising are always mediated by cognition (Greenwald & Leavitt, 1984; Morris et al., 2002). Cognitive models are commonly used to explain ad processing and consumer decisions. Cognition is particularly important due to the potential interval between ad exposure and purchase behavior, so advertising is required to change the long-term memory (Ambler & Burne, 1999). One way of changing the memory content is through repetitive exposure. Repetitive advertising will increase brand awareness (Ehrenberg, 2000) by rehearsing neural network related to the advertised brand in memory many times, and so strengthening the network and making the brand more memorable and more easily retrieved. As exposure to advertising increases, the cumulative temporary effects that occur at each exposure eventually build into a substantial effect, namely attitude change. Therefore, cognitive reactions to advertising are often reflected in the constructs such as attitude towards the ad (Biehal, Stephens, & Curio, 1992; Brown & Stayman, 1992) or attitude towards the brand (Spears & Singh, 2004). In this context, attitude is considered a function of cognitive beliefs which are a predictive intention of behaviour (Fishbein & Ajzen, 1977). Cognitive models have been applied not only to 31 ad processing but also to consumer decision-making. In these models, the various information gathered from ads is integrated to help a consumer to make a choice, and the choice process is usually described as the process of comparing attributes of the alternative brands and then trading off those attributes according to their importance or relevance to the individual's goals (Bettman, 1979). From quite a different perspective, Zajonc (1980) argued that affect could directly influence behaviour without involving cognition. Similarly, Schwarz, Bless, and Bohner (1991) introduced the affect-as-information framework suggesting that feelings may serve as affective feedback that guides judgement, decision-making, and information processing. In this respect, people perceive feelings as containing valuable information to help them make a good judgement. Furthermore, Pham (1998) has suggested that many choice decisions are based on a 'How-do-I-feel-about-it? heuristic' which involves examining feelings to a stimulus representation in the mind, so even if the stimulus is not physically present, people can still perform their evaluations. These developments signal the importance of affect in information processing, including ad processing, and foster the emergence of affective models in advertising. 2.3.2.2. A renewed interest in affective reactions to advertising The influential work of Damasio, Everitt, and Bishop (1996) reveals that decision- making process is primarily associated with the brain region that deals with emotion rather than cognition, leading to the concept that emotion is an essential element for rational thinking and behaviour. People's first reaction to an object would be an instinctively positive or negative emotion known as a 'somatic marker' (Damasio et al., 1996). Somatic marker is some sort of feeling that is formed by an experience that gets 32 connected to specific events. If the marker has a negative emotional tone, it acts as an alarm that a negative outcome will occur, leading to avoidance behaviour. Conversely, if the marker has a positive tone, it acts as a motivator for pursuing an action (an approach behaviour). Thus, the somatic markers are implicit emotional triggers which are retrieved from memory fast and involuntarily and lead people to react to an event as 'good' or 'bad' without conscious thought. As such, it is tempting to think that affect rather than cognition is key to advertising effectiveness. In line with the ‘somatic marker’ hypothesis, Zajonc (1980) argues that affective processing can bypass cognitive processing and affect may separate from content but still remain. Thus, the feelings that the consumer has about the brand when viewing the ads are often readily accessible although the consumer may have forgotten the ad or the exposure. This is because the consumer has formed an attitude linked to the brand that is congruent with the positive feelings (Hasford et al., 2015). At the point-of- purchase, these feelings can serve as cues that facilitate the salience of the brand. The important consideration is that ad-induced emotion will still play an important role in influencing consumer choice, even when the consumer may have forgotten the ad or even when the ad is not physically present. However, the feelings may become negative if people pay more attention, because deliberative processing can trigger counter- arguments (Heath et al., 2006; Segijn, Voorveld, & Smit, 2016), which can decrease brand attitude. Obviously, emotion plays an important role in advertising. 2.3.2.3. Emotion and advertising responses Damasio et al. (1996) distinguished emotions (body state) from feelings (mental state); whereas Cohen, Pham, and Andrade (2008) described affect as genuine subjective 33 feelings or moods (e.g. I'm sad), rather than evaluative thoughts about an object or event (e.g. I like this ad). To avoid confusion, the thesis uses the terms emotion, feeling, and affect interchangeably, but in contrast with cognition. There are numerous definitions of emotion; however, the most comprehensive one is summarised by Kleinginna and Kleinginna (1981): Emotion is a complex set of interactions among subjective and objective factors, mediated by neural/hormonal systems, which can (a) give rise to affective feelings of arousal, pleasure/displeasure; (b) generate cognitive processes such as emotionally relevant perceptual effects, appraisals, labeling processes; (c) activate widespread physiological adjustments to the arousing conditions; and (d) lead to behavior that is often, but not always, expressive, goal-directed, and adaptive. (p. 355). Based on the definition, emotion has three key functions: (1) the discrete, sub- conscious, unevaluated response to a stimulus; (2) an automatic activator or driver of behaviour; and (3) as an appraisal (feeling-based evaluation) that highlights the significance of an event or stimulus and connects cognitive and behavioural tendencies. In general, the term emotion can refer to two things: (1) the emotional quality of the stimulus, or (2) the emotional state of the individual. Applied to advertising, the first denotes the emotive content; while the second represents the effect of advertising on the audiences, namely any feeling, association, or memory about the brand. Types of emotions. Advertising can elicit different types of emotion: integral, incidental, lower-order, and higher-order. Integral emotion refers to a specific emotional appeal used in an ad that is deliberately selected by advertisers to induce specific feelings to the audiences, such as fear, guilt, or anger; while incidental emotion refers to non-specific emotions (e.g. good or bad) 34 that are evoked by unrelated sources or events but have the potential to influence decisions (Poels & Dewitte, 2019). Advancing cognitive accounts to emotion, Shiv and Fedorikhin (1999) distinguish lower-order emotions that involve automatic, spontaneous, largely unconscious and involuntary reactions to a stimulus; and higher-order emotions which require some deliberation and cognitive processing and reinforce action tendencies. Looking back to the key functions of emotion, lower-order emotions induce subconscious, unevaluated responses to advertising which can be a driver for ‘intuitive’ behaviour (e.g. implusive buying); whereas higher-order emotions reflect the cognitive evaluation or interpretation of an advertising stimulus. In marketing, lower-order emotions occur when affective cues (e.g. background music in a store) influence shopper behaviour. For example, French music led to French wines outselling German ones; whereas, German music led to more sales for German wines (North, Hargreaves, & McKendrick, 1999). In contrast, higher-order emotions can be defined by the perception of future events in relation to the cognitive appraisal of the current situation (Lerner & Keltner, 2000). Chang and Pham (2013) found that people rely on their feelings when outcomes are proximate, but discount them for more distant outcomes. The types of emotions that are most relevant to the low-attention processing in the current research are the incidental and lower-order emotions, so the following discussion will focus on these types. The role of emotions in advertising. Emotive advertising can induce emotions that will have subsequent consequences on advertising outcomes (Vakratsas & Ambler, 1999), and emotional reactions to advertising are a strong predictor of purchase intention and brand attitude (Morris et al., 2002). Poels and Dewitte (2006) reported an interesting 35 relationship between ad exposure and ad or brand familiarity such that previously seen ads lead to more positive emotional reactions (pleasantness and less boredom) compared to unfamiliar, first-seen ads. Moreover, the effect of ad-evoked emotions was found stronger for unfamiliar than familiar brands. They explain that attitude toward a familiar brand is already established in the consumer’s mind so it is not strongly affected by ad exposure. In other words, emotion-based advertising would be more effective for unfamiliar than familiar brands. How does emotive advertising work? Ohme (2009) uses the term ‘unconscious affect’ to suggest that advertising can tap into the unconscious mind to create a desire for the brand/product that spurs consumers into purchase-related behaviour. In other words, advertising can trigger a behavioural drive to approach the brand/product and eventually grab it from the shelf. To do so, advertising can utilise subtle cues, such as sound, music, or an image that is largely ignored by the conscious mind/attention during exposure, but can subconciously influence consumers to want the brand/ product, even if they cannot explain rationally why they want it. Ohme (2009) provides an example in which Sony used 'the frog scene' displaying the moment when a frog jumped out from a rain pipe in their ads for Sony Bavaria television. When this scene was discarded from the ad, the preferences for the brand/product significantly decreased, showing that 'the frog' was instrumental in evoking favourability. This supports the notion of low-attention, affect-based processing because 'the frog' that has no logical/rational connection with the brand/product has a significant impact on brand favourability. Also, more recent advertising research found a higher brand attitude under low rather than under high attention (Goodrich, 2014). The discussion points to an important consideration that ad-evoked emotions will more likely occur if low rather than high attention is given to the ads. 36 2.3.3. A rethinking of attention The notion that attention is closely linked to emotion is not new as Davenport and Beck (2000) note, “One of the most important factors for gaining and sustaining attention is engaging people’s emotions” (p. 123). Moreover, advertising equity (Rosengren & Dahlén, 2015) holds that prior experience with advertising influences willingness to approach future advertising. If the consumer has a positive experience with the brand’s ads initially, he/she would b