This study aims to investigate how Internet memes affect brand image.
The authors first used the Delphi method to refine Internet memes' constructs and dimensions and developed a scale for Internet memes. Second, the authors used a questionnaire to collect data from 348 valid Internet consumers.
The authors proposed four significant characteristics of Internet memes: humour, high positive emotional intensity, brand interactions and prestige, and high spreadability to prompt consumers to remake, share and spread memes. The study results indicate that Internet memes positively influence the brand image. However, not all meme characteristics were correlated with the brand image; only brand prestige, interaction and humour enhanced brand image.
For scholars in online marketing communication research, this study shifts the current paradigm of brand-generated and customer-passive to user-generated and customer-active. It also addresses the importance of pandemic effects of Internet memes on brand image. To be specific, this study presents the important symbolic values that Internet memes need to include to affect consumers' behaviour in response to perceived brand image by applying both mind infection and symbolic value perspectives. Finally, this study redefines the dimensions and measurements of Internet memes to address the lack of consensus and concrete scales regarding meme transmission characteristics and extending into online marketing communication strategies.
The peer review history for this article is available at: https://publons.com/publon/10.1108/OIR-11-2019-0364
Teng, H., Lo, C.-F. and Lee, H.-H. (2022), "How do internet memes affect brand image?", Online Information Review, Vol. 46 No. 2, pp. 304-318. https://doi.org/10.1108/OIR-05-2020-0192
Emerald Publishing Limited
Copyright © 2021, Emerald Publishing Limited
With the prevalence of Internet use, Internet memes have become a novel and indispensable online communications strategy (Wei et al., 2012; Dubey et al., 2018; Zannettou et al., 2018). Internet memes generally refer to a process by which text, images and videos are widely promoted, which allows online users to increase awareness of transmitted information through humorous and ironic texts (Miltner, 2018). They can also transmit information in an extremely short time (Wei et al., 2012; Dubey et al., 2018; Zannettou et al., 2018). Hence, academia has begun to pay more attention to their online propagation effect on information diffusion (Brubaker et al., 2018). Additionally, the marketing field now attaches importance to its application in marketing communication strategies (Chen, 2012). According to marketing communication literature, brand image is the first stage in which consumers become aware of new products and information via various channels (Rogers, 1995; Gökerik et al., 2018). Since the online environment has no geographical limitations and consumers can easily interact with each other, online information propagation can be faster and more extensive compared to traditional marketing channels. Scholars and marketers have started to focus on the effectiveness of online information diffusion and its application to marketing communication (Bakshy et al., 2012; Guille et al., 2013).
Although the literature proposes some effective interactive communication channels, such as two-way dialogues and online brand communities (Rowley, 2009), most of these strategies are brand-generated and customer-passive. However, the main characteristics of Internet memes, unlike traditional online marketing communication channels, are that they are user-generated and customer-active. The current literature has limited discussions on user-generated and customer-active communication channels, like memes and their impact on brand image. A review of most related literature is currently in Brubaker et al. (2018), which discusses relationships between memes and online public engagement. However, this study discusses how political issues spread at a macro-level and does not mention impacts on consumers at a micro-level. Furthermore, although existing literature has explored memes' propagation characteristics (Brubaker et al., 2018; Shifman, 2014), there is no consensus on the propagation characteristics and no explicit scale of measurement. Finally, most of the antecedents of brand images are related to consumers' internal psychological perceptions (Chinomona, 2016; Krishnamurthy and Kumar, 2018; Sääksjärvi and Samiee, 2011). However, these are not only subjective but also lack a solid understanding of the influence of external social stimuli such as Internet memes on brand image (Grönroos, 1997). How do Internet memes affect brand image? What is the vital insight that causes Internet memes to link to a brand image? The answers to these questions are still unknown and require further investigation. Therefore, the purpose of this study is to explore how Internet memes affect brand image.
According to some researchers, memes can be compared to a “social contagion” whereby consumers are unconsciously “infected” and “infect” others, and this eventually leads to a “pandemic effect” (Marsden, 1998; Robertson, 2017). Benaim (2018) has further proposed that the internet meme's nature has a symbolic value, and online users spread memes because of that. By combining these two perspectives, this study proposes that Internet memes must be built in terms of their symbolic value to result in consumers being infected in the mind. We believe that if the research gap is filled regarding the possible mind infection effect by Internet memes' symbolic value on brand image, advertisers using memes will be able to provide more comprehensive online communication marketing strategies that influence consumers' behaviour through brand image.
This study provides several essential contributions. For scholars in online marketing communication research, this study may shift the current paradigm of brand-generated and customer-passive to user-generated and customer-active. It also addresses the importance of the pandemic effect of memes on brand image and the drawbacks of brand image literature that primarily focuses on consumers' internal psychological perceptions without considering external social stimuli. Additionally, by applying both the mind infection and symbolic value perspectives, this study presents the important symbolic values that memes need to infect consumers' minds and thus change their behaviour based on a brand's image. Finally, this study redefines memes' dimensions and measurements to address a lack of standards regarding scale and propagation characteristics.
Theoretical background and hypothesis development
Internet meme and brand image
For a marketing communications strategy to be effective, consumers must fully understand the message conveyed by the brand and the message should be diffused widely. In other words, memes are designed to catch a consumer's attention (Pan, 2020; Li et al., 2020; Zhang and Lin, 2015). Hence, brand image awareness is the first stage of a consumer's behaviour when adopting innovative products and services (Gökerik et al., 2018). Brand image was initially defined as the perceived symbolism of a brand's commodity (Sommers, 1963), but is now mostly defined as a series of brand associations in a consumer's memory (Aaker, 1991). In this study, brand image is described as a synergistic and symbolic meaning decoded by consumers, based on Bivainienė and Šliburytė (2008) and Hauge (2015).
Factors that influence brand image include brand identity (Sääksjärvi and Samiee, 2011), word of mouth (Krishnamurthy and Kumar, 2018; Reza Jalilvand and Samiei, 2012) and brand communication (Chinomona, 2016). However, most of the factors affecting brand images are related to consumers' internal psychological perceptions. The present study proposes that consumers' perceptions are subjective. It is also critical to consider external social stimuli. Consequently, using only psychological perspectives to explain brand image is inadequate for understanding the external social environment's influence on brand image. External environmental factors that may influence brand image must also be identified (Gökerik et al., 2018).
As an external environmental factor, Internet memes allow messages to be widely propagated. Further, the transmitted messages are more user-generated and consumer-initiated due to the characteristics of intertextuality, remix and humour (Brubaker et al., 2018). This makes online marketing communication strategies more effective. Furthermore, academia has recently begun considering Internet memes as a novel marketing communication channel (Brubaker et al., 2018; Shifman, 2014). Dawkins (1976) first proposed the concept of the meme. A meme is defined as an idea, behaviour or style conveyed from person to person. Memes evolve by replicating (imitation), variation and selection (Dawkins, 1976). Since the online environment has no geographical limitations and consumers can easily interact with others, the online information propagation can be faster and more extensive compared to traditional marketing channels. The term “Internet meme” has been mentioned in digital media, massively publicised, broadcast on the internet and become well known. Internet memes are relayed quickly through blogs, message boards, Internet forums, search engines, email, social media or video-sharing websites (Shifman, 2014).
The current literature on Internet memes can be divided into two major types. The first type focuses on propagation methods and cascade (Dubey et al., 2018; Wei et al., 2012; Zannettou et al., 2018). How memes propagate information to different online users is typically explained by using model derivation. The other type of research discusses Internet memes' propagation characteristics (Brubaker et al., 2018; Díaz and Mauricio, 2013; Shifman, 2014). Model derivation is commonly used for determining the characteristics that memes must have to achieve extensive propagation. A review of the relevant literature indicates that regardless of the research type, memes' propagation methods or characteristics remain mostly focused on satire or controversial information regarding politics or morality. Users relate the most to issues they are interested in, contributing to user-generated content and active propagation (Brubaker et al., 2018; Zannettou et al., 2018). The specificity or meaning of memes may prompt online users to share (Sela et al., 2019) and even remake memes (Dubey et al., 2018), thereby achieving multiple-cascade propagation (Wei et al., 2012). Some studies have proposed that such propagation methods are mostly applicable to propagating political satire or messages on controversial moral issues. Such characteristics are also mainly focused on public opinion of macro-level issues (Boudana et al., 2017; Brubaker et al., 2018). However, in marketing communication, consumer awareness and perception of brands are generally at the micro-level; it is still necessary to explore whether the macro-level information can be extended to the micro-level.
From the perspective of mind infection, Marsden (1998) has proposed that memes can be compared to a “social contagion” whereby consumers are unconsciously infected and then infect others to contribute to a significant propagation (contagious) effect. Benaim (2018) proposed that memes have symbolic value and consumers propagate them due to this. According to this perspective, people view objects in the external environment as symbols and give them symbolic meanings and significance (Hauge, 2015). For example, when consumers receive an advertising message related to a brand, they view the brand as a symbol based on the advertisement's content. Further, they interpret the symbolic meanings and value and respond through cognition and behaviour (Entwistle and Rocamora, 2006; Hauge, 2015).
Certain studies focus on the propagation characteristics of Internet memes as related to symbolic value. For example, Berger and Milkman (2012) reported that online advertisements must induce strong positive (i.e. awe) or negative (i.e. anger or anxiety) emotions to be propagated extensively. Díaz and Mauricio (2013) proposed that Internet memes must have spreadability and reproducibility. Shifman (2014) stated that when Internet memes have positivity (and humour), have high emotional intensity, are easy to understand, have propagation source prestige, a propagation method and mutation level, users transmit them extensively. Brubaker et al. (2018) proposed that Internet memes must have quiddity, have humour, portray the brand and be interactive if they are to be propagated. By applying both the mind infection and symbolic value perspectives, this study further proposes that consumers become unconsciously infected by perceiving the symbolic meaning and then continuously infect others, thus exhibiting a significant propagation (contagious) effect. The consumers would then have a better interaction with the brand, thereby increasing consumers' positive brand associations, thus improving brand image.
The first characteristic of Internet memes is humour. Numerous consumers employ social media for social interaction and self-presentation; they like to spread content that makes other users feel happy. Therefore, they show their optimism and sense of fun (Shifman, 2014). When the propagation information of Internet memes has an entertainment characteristic, consumers consider them to have symbolic value and may reconstruct, share and propagate the meme to demonstrate they are optimistic and fun. When other users view, like or repost the meme because it is entertaining, they spread the meme further, resulting in more views. This also increases the brand's positive associations (Gelb and Zinkhan, 1986), and the brand image is improved.
The second characteristic of Internet memes is high emotional intensity (provoking high-arousal emotions). According to Shifman (2014), high positive emotional intensity is created when Internet consumers generate a feeling of elation in the face of something greater than oneself. Natural wonders, path-breaking scientific discoveries and people overcoming adversities are prominent examples of narratives that generate “Aw” or “Wow” responses. This behaviour further stimulates meme propagation, imitation and transformation (Berger and Milkman, 2012). When the disseminated information of Internet memes has high positive emotional intensity, consumers consider them “amazing” symbols that reflect their own positive or negative emotions. When other consumers view, like or repost a meme due to their emotions being aroused, the meme is further disseminated. This also means that the opportunities for interaction between the brand and consumers, and consumers' positive brand associations increase, resulting in improved brand image.
The third characteristic of an Internet meme is that it should be easy to propagate. Berger and Milkman (2012) reported that simple videos and jokes are more comfortable to share because users can easily understand them and assume others also easily understand them. When an Internet meme has the characteristic of being easier to understand, consumers consider it to have symbolic value and they may feel more comfortable propagating it. In this way, consumers propagate them more than through traditional marketing channels, thereby increasing consumers' positive associations with the brand, and thus improving brand image.
The fourth characteristic of Internet memes is brand prestige and interactions. Concerning online news stories, Shifman (2014) found that the more famous an author is, the more likely people are to spread the piece. Brubaker et al. (2018) further reported that interaction or transformation between consumers and brands increases consumers' sharing behaviours. Consequently, this study proposes that when the disseminated information of Internet memes has prestige and interactive characteristics, consumers consider them symbols of their prestige and interaction and then may reconstruct, share and propagate the memes to demonstrate that they are also prestigious and interactive. This results in more positive associations and improved brand image.
In summary, Internet memes must have humour, high positive emotional intensity, be easy to propagate and confer brand prestige and interactions. These significant characteristics mean that Internet memes have symbolic value, which leads to a kind of consumer mind contagion that spreads to others. Consumers interact with others by recreating, sharing and spreading the symbolic value of the memes. This spread results in more consumers viewing them, creating more positive associations and cognition for the brand; thus, brand image is improved. Therefore, this study proposes the following hypothesis:
Internet memes have a positive effect on brand image.
The purpose of this study was to explore how Internet memes affect brand image. In total, two studies were conducted: one for data collection and one for verification. In Study 1, scale development was designed by conducting expert interviews and using the Delphi method and content analysis. In Study 2, exploratory factor analysis was first used to select questions. Confirmatory factor analysis and structural equation modelling were then used to perform reliability and validity tests and test the proposed model.
Study 1– scale development
No consensus regarding Internet memes' constructs has yet been reached, and no reliable scale for measurement has yet been devised. We conducted semi-structured, in-depth, open-ended oral interviews with experts and scholars in industry and academia to understand preliminarily online consumers' opinions of Internet memes. The content was guided to be interviewee-oriented, and the constructs of Internet memes were identified. The Delphi method was then used for questionnaire content.
Based on the literature, three indicators were adopted to select expert group members: the willingness of experts and scholars to participate, professional ability or seniority in the field and diversification of experts' source fields (Skulmoski et al., 2007). The academic selection criteria were oriented towards e-commerce research, especially related to the online interactive communications field. Conversely, the industry selection criteria included having at least five years' practical experience in e-commerce and online interactive communications. We interviewed six experts and three scholars from industry and academia. This ratio was used because Internet memes are more extensively employed in industry than academia, and there are few scholars who were familiar with meme research. The background of these experts is in the Appendix. This study used semi-structured interviews to avoid the limitations of structured interviews, such as lack of flexibility and difficulty in having in-depth discussions of issues. Further, this avoids the time-consuming, laborious and difficulty of quantitative analysis of non-structured interviews.
The study lasted for nearly three months, from the invitation to the formal confirmation of the scale. We contacted the invited experts and scholars personally. We have received practical training on how to conduct the Delphi. Before the interviews, it was confirmed that the interviewees were familiar with memes. We provided the researchers' identities and explained the interview purpose and main interview content through interviews, phone calls and email. The interviewees were invited, the interviews were audio-recorded so that transcripts could be obtained and the interviewees were informed that the recorded data would be kept secure. The researchers recorded their necessary information and the interview content for safe preservation and to maintain the interviewees' privacy. The interviewees were represented using codes. After the interviews, each interviewee was given a department store coupon as a token of gratitude. The interview protocol was used (Skulmoski et al., 2007). The experts first confirmed their understanding of the definition of the research topic and then described their perspective on the characteristics of Internet memes.
Content and Delphi method analyses
The interview recordings were transcribed and then arranged and analysed according to the content mentioned above for discussion. The experts' opinions were analysed using the results of the content analysis. Additionally, they were asked to supplement the proposed scale and make corrections. The results of content analysis were then used for scale design by referring to similar items (Haryani and Motwani, 2015; Kumar and Paul, 2018; Lee et al., 2019; Phelps et al., 2004; Yang and Zhou, 2011) and the first version of the closed questionnaire for the Delphi method. The first version of the Delphi method questionnaire was divided into six parts, representing different characteristics. There were four sessions total, until all the experts and scholars agreed on the questionnaire's content. Items were selected preliminarily using the quartile method within the Delphi method (Holden and Wedman, 1993). Some of the 18 items had low consistency in the second round. The number of items was reduced to seven in the third round. Finally, in the fourth round, two items still had low consistency. They were “Reforming and sharing Internet memes gives me a sense of satisfaction” and “The brand referred to in the internet meme is very popular, so I wanted to share the meme with other people”. Consequently, these items were deleted, and finally, 16 items were retained.
This method was used instead of a focus-group interview because the Delphi method includes anonymous responses from experts. In this method, the experts participate in a survey and then exchange opinions about the questionnaire contents. They do not know who the other experts are. Thus, this method eliminates the concern of some participants blindly going along with the others in face-to-face interviews. Moreover, the modification of expert opinions is followed by an opinion review, enabling the experts to adjust their views where appropriate. Through communication and repeated feedback, different ideas can be collected to reach a consensus.
Furthermore, whether consistency and convergence have been reached among the experts' opinions can be determined through statistical analysis. However, the Delphi method requires the expert panel to repeat the same steps and involves a complicated process. Without adequate incentives or rewards, experts' cooperation is difficult to obtain. Consequently, in this study, only nine experts and scholars were recruited for the Delphi method survey. Despite the small number of experts, it was close to the standard of 10–15 people in the relevant literature (Delbeq et al., 1975). Additionally, the degree of cooperation among the experts was high in every round.
Study 2 – measurement confirmation and hypothesis testing
This study operationalised Internet memes as images, texts and videos, which have been heavily publicised and quickly spread (Shifman, 2014). We revealed the real Internet meme at the beginning of the questionnaire so that the participants understood it. This study conducted a pre-test to examine the reliability and validity of the scale items. A total of 62 questionnaires—22 Internet and 40 paper questionnaires—were distributed. The average variance extracted was between 0.51 and 0.63. Furthermore, most composite reliability values were between 0.64 and 0.83, which reached the threshold value of 0.7, thereby satisfying the conditions for convergent and discriminate validity (Fornell and Larcker, 1981). For the reliability test, all Cronbach's α values were between 0.81 and 0.91, which was higher than 0.7, the threshold value commonly accepted in previous studies (Nunnally, 1978). Therefore, the official questionnaire was further distributed.
Official questionnaire collection
The study subjects were consumers attracted by Internet memes who liked, shared, commented on and made purchases after viewing Internet memes on social websites. The cover of the questionnaire explained the purpose of the research and asked the respondents whether they had ever liked, shared or commented after viewing an Internet meme. This was asked to ensure that they had specific cognition and understanding of them. The responses in which the answer to this question was “Yes” were statistically analysed. This study referenced the Taiwan Electronic Commerce Yearbook (2013) and used sex and age as stratification variables for stratified sampling. Overall, 38% of the respondents were male, and 62% female; 8% were under 20 years old, 77.2% were 21–40 years old and 14.9% were 41–50 years old.
Additionally, this study also adopted Chang and Chen's (2008) sampling distribution as the second reference, which has quite a similar stratification ratio compared to the Taiwan Electronic Commerce Yearbook (2013). The ratio of the Internet to paper questionnaires returned was 6:4. Paper questionnaires were employed as well as Internet questionnaires because the latter can lead to duplicate responses. Thus, to ensure the questionnaires' validity, 40% of the questionnaires were in paper form. The researchers distributed the Internet questionnaires via different social media (e.g. Facebook, Instagram, LINE and Dcard). For paper questionnaires, the researchers conducted convenience sampling on streets in commercial areas. Because sample ratio control with stratification variables was exerted, convenience sampling was used for practical execution.
After confirming that the respondents had answered all the questions, the questionnaires were retrieved. A total of 516 questionnaires were distributed, and the first question was answered “Yes” in 405 of these. The final number of valid questionnaires was 348, with a valid response rate of 67.4%. The questionnaire responses were completed anonymously to maintain academic objectivity and neutrality. Before a respondent completed the questionnaire, the document informed them that an academic questionnaire was performed only for an educational purpose and that no gift would be provided. Regarding the internet questionnaires, Internet Protocol addresses were checked to prevent the problem of duplicate answering.
All researchers received research ethics training. The Delphi method, pre-test and questionnaire distribution were conducted per the guidelines of the institutional review board (IRB). The researchers clearly described the study and interview purpose to respondents. The respondents all provided informed consent to participate in the study. Anonymous questionnaires were adopted for both the pre-test and full-scale survey to protect the respondents' privacy. Respondent data remained private. Apart from the researchers, none of the respondents knew the other participating experts and scholars' names or related contact information for the Delphi method.
Scale measurement, reliability and validity analysis
This study provided a complete construct, dimensions and research structure after the content analysis and pre-testing illustrated in Figure 1.
In this study, only the Delphi method was used for designing an Internet meme scale. For the measurement of other constructs, the literature was referenced. Biel (1992) referred to the brand image scale and Park et al. (1986) to the concept of a brand image's symbolic meaning. The brand image scale was also referred to by Bİlgİn (2018). For the reliability test, all Cronbach's α values were between 0.81 and 0.92. Regarding validity, the factor loadings of different items in this study were all larger than 0.5 except that of item 24: “Other people consider me buying the brands marketed in memes to be immature”. This was a reverse question and was deleted because the factor loading did not conform to the standard (see Table 1).
Restructure components of Internet memes
This study analysed the relationships between Internet memes and brand image using linear structural equations (Figure 2). The results indicated that Internet memes and brand image were significantly positively correlated (β = 0.18, p < 0.01), and the hypothesis was supported. The factor loadings of each construct were then examined. High positive emotional intensity was discovered to be the most crucial construct (eigenvalue = 0.900), followed by high spreadability (eigenvalue = 0.698). Brand interactions was the third most important (eigenvalue = 0.900) and humour the least (eigenvalue = 0.619). The results indicated that Internet memes' high positive emotional intensity is the factor with the strongest influence on brand image.
According to the results mentioned above, the direct relationships of different dimensions of Internet memes with brand images were further analysed (Figure 3). Brand interaction and brand image were significantly positively correlated (β = 0.347, p < 0.01). Humour and brand image had a significant slight positive correlation (β = 0.148, p < 0.1). Surprisingly, high positive emotional intensity and high spreadability were not correlated with brand image despite being critical in meme propagation.
Effects of internet memes on brand image
This study conducted an exploratory factor analysis for the Internet meme scale. The results revealed four significant factors related to Internet memes. These factors' eigenvalues were all larger than 1, and the total accumulated variance explained was 70.4%. The first group of items was related to the brand prestige conferred by Internet memes, interaction with a brand meme post and the reposting of a brand's meme. The eigenvalue was 7.9, and the explanatory variance was 43.9%. Because the group's factors relate to the interaction between brands and consumers, the group was named “brand interactions”. The second group of items relates to the levels of emotion stimulated in consumers by Internet memes. The eigenvalue was 2.7, and the variance explained was 14.7%. This group was named “high positive emotional intensity”. The items in the third group are related to the humour in Internet memes. The eigenvalue was 1.1, and the variance explained was 6.1%. The group was named “humour”. Finally, the fourth group was related to memes' easy propagation. The eigenvalue was 1, and the variance explained was 5.7%. This group was named “high spreadability”. The factor loadings of all items were higher than 0.4, conforming to the factor-loading standard of exploratory factor analysis (Mata-Toledo and Gustafson, 1992) (see Table 2).
Based on Shifman's (2014) research, the present study considered six Internet meme constructs: humour, emotional intensity, easy to understand, propagation prestige, propagation method, mutation level and interactivity. The common Internet meme constructs in other articles were integrated, and a scale was developed using the Delphi method. Finally, four constructs were extracted: brand interactions, high positive emotional intensity, humour and high spreadability. Brand interactions were an essential construct regarding their explanatory power, followed by high positive emotional intensity. Humour was the third most important, whereas high spreadability was the least. This demonstrated that brand prestige and interactivity with consumers were vital to Internet meme propagation. Moreover, among the items related to brand interactivity, “wanting to transform and propagate”, “the brand is representative” and “the brand is my first choice” showed that prestige was an extremely critical factor in Internet meme propagation. Triggering the desire to transform a meme in consumers also boosts Internet meme propagation. Moreover, among the items for high positive emotional intensity, “exciting”, “interacting with others” and “reducing the sense of loneliness” revealed that if memes can increase consumers' sense of interactivity and excitement, propagation is more likely.
Discussions and implications
The research questions mainly address the following points: How do Internet memes affect brand image? What is a vital insight that causes Internet memes to be linked to the brand image by consumers? Hence, the purpose of this study was to explore how Internet memes affect brand image. This study assumed that Internet memes would positively affect brand image and consumers would be found to have their minds infected by an Internet meme's symbolic value. The results also indicated four essential propagation characteristics of Internet memes: humour, high positive emotional intensity, brand prestige and interaction and high spreadability. However, not all four meme characteristics were essential to brand image, only the incorporating symbolic values of brand prestige, interaction and humour.
The primary research contributions of this study are as follows. First, we responded to the lack of consensus in the current literature on the propagation characteristics of Internet memes (Brubaker et al., 2018; Shifman, 2014). This study organised the various meme propagation characteristics identified in previous studies into four major categories, namely brand interaction (Brubaker et al., 2018; Díaz and Mauricio, 2013), high positive emotional intensity (Berger and Milkman, 2012; Shifman, 2014), humour (Berger and Milkman, 2012; Shifman, 2014) and high spreadability (Díaz and Mauricio, 2013; Shifman, 2014). Based on symbolic value, this study further proposed that Internet memes must have four significant characteristics: humour, high positive emotional intensity, brand prestige and interaction and high spreadability. This stimulates a mind infection in consumers who then have the desire to spread, imitate and remake the memes rapidly on social media according to their symbolic values. Additionally, the Internet meme propagation characteristics identified in previous literature have been mainly applicable to political or public issues (Brubaker et al., 2018). No empirical research has focused on online marketing communication effectiveness. Therefore, this study enables the empirical measurement of Internet memes by using the Delphi method to develop a scale and apply it to online marketing communications effectiveness.
Second, by applying both mind infection and symbolic value perspectives, this study presents the important symbolic values that Internet memes need to include to affect consumers' behaviour in response to perceived brand image. The results indicate that when Internet memes exhibit better brand prestige and interactivity and the higher the symbolic value, the more possibility that consumers will be mind-infected and have the desire to propagate them. The memes enhance the interaction between consumers and the brand. Positive associations with the brand and brand image also increase (Krishnamurthy and Kumar, 2018; Jalilvand and Samiei, 2012). Furthermore, when an Internet meme's entertaining symbolic value is higher, consumers propagate the Internet meme to present themselves as optimistic and fun. Their positive associations with the brand increase (Gelb and Zinkhan, 1986) the brand's image. In sum, this study argues that brand prestige, interactivity and humour in Internet memes are critical factors for marketing communication strategies. However, not all dimensions of memes directly improve brand image. Although Internet memes' high spreadability and high positive emotional intensity can make consumers propagate them, consumers may not necessarily associate these symbolic values with the brand. Therefore, they do not improve brand awareness and hence not brand image. This may occur because these characteristics involve low symbolic values and weak connection with the brands. This study provides a complete explanation for Internet meme symbolism mechanisms that create a connection between the minds of consumers and brand image.
Third, for scholars in online marketing communication research, this study shifts the current paradigm of brand-generated and customer-passive to user-generated and customer-active. It also addresses the importance of pandemic effects of Internet memes on brand image. Moreover, Internet memes positively affect brand image and thus contribute to understanding the manipulable stimulus to the individual consumer rather than psychological factors of brand image. Finally, regarding managerial implications, this study highlights that online interactive communication strategies, like Internet memes, must be built on the basis of their symbolic value to affect consumers to the point where they are mind-infected. Specifically, this study suggests that meme advertisers should first design funny, witty and highly interactive brand memes. At the same time, the memes need to be widely disseminated to improve the brand reputation and highlight the memes' symbolic value. By doing this, the Internet memes' contagious effect on consumers would be magnified and consumers would quickly propagate the meme to others, and thus increase the brand image.
Limitations and further research
There are several limitations to this study. Brubaker et al. (2018) proposed the relationship between Internet memes and online public engagement. Although public engagement is not in the scope of this study, online consumers do generate engagement. This study suggests that researchers could further explore the possible relationship between Internet memes and customer engagement. Moreover, the research only focuses on Internet memes and brand image without also examining other possible mediating or moderating factors. We suggest further exploring whether there are other influencing factors to strengthen the mechanism of consumer cognition. Additionally, this study has not yet examined whether the Internet memes designed by different types of marketers will affect brand image. Further research can clarify the impact of marketers' characteristics on Internet memes and brand image. Fourth, this study only investigated a Taiwanese context. Therefore, future studies are recommended to see whether different study results are obtained in other countries. Finally, this study only examined a cross-sectional survey that might be a snapshot of information analysed within a specific period and might require a longitudinal study to have a complete understanding.
Reliability and validity tests
|Makes me have fun||0.87|
|Clearly understanding the content||0.81|
|Immediately understanding the content||0.84|
|Others can immediately understand||0.69|
|High emotional intensity||0.859||0.564||0.865|
|Cheers me up||0.76|
|Arouses my curiosity||0.80|
|Enables me to interact with others||0.67|
|Reduces my sense of loneliness||0.68|
|Brand Interaction and prestige||0.920||0.618||0.917|
|Makes others feel interesting||0.55|
|Makes me want to transform and propagate||0.90|
|Attract to transform and propagate||0.93|
|The meme brand is representative||0.70|
|The meme brand is my first choice||0.64|
|The brand is reliable||0.58|
|The brand is popular||0.74|
|The brand entertains me||0.89|
|The brand is with a lot of fun||0.88|
|The brand is familiar to me||0.83|
Note(s): N = 348. CR: composite reliabilities; α: Cronbach's α and AVE: average variance extracted. All CFA factor loadings are significant at p < 0.01
Exploratory factor analysis of Internet meme
|Factor||Brand interaction and prestige||High positive emotional intensity||Humour||High spreadability|
|Items||Make others feel interesting||Cheers me up||Interesting||Clearly understanding the content|
|Make me want to transform and propagate||Are exciting||Making me have fun||Immediately understanding the content|
|Attract to transform and propagate||Arouse my curiosity||Feel relaxing||Others can immediately understand|
|The meme brand is representative||Enable me to interact with others|
|The meme brand is my first choice||Reduce my sense of loneliness|
Background of the Dephi experts
|Identity||Code||Service units/University/Department||Job title||Average years of work experience and related academic research|
|Scholar||1||Education/Electronic commerce department||Assistant Professor||12|
|2||Education/Marketing department||Researcher in e-commerce||5|
|3||Education/Logistic and marketing||Professor||18|
|Industry personnel||4||Marketing/Electronic commerce department||Manager||7|
|5||International marketing/Electronic commerce department||Specialist||5|
|6||International business/Online marketing||Manager||10|
|7||International business/Online marketing||Assistant manager||5|
|8||Manufacturing/Online community promotion||Director||6|
Aaker, D.A. (1991), Managing Brand Equity, Free Press, New York.
Bakshy, E., Rosenn, I., Marlow, C. and Adamic, L. (2012), “The role of social networks in information diffusion”, in Proceedings of the 21st International Conference on World Wide Web, pp. 519-528.
Benaim, M. (2018), “From symbolic values to symbolic innovation: Internet-memes and innovation”, Research Policy, Vol. 47 No. 5, pp. 901-910.
Berger, J. and Milkman, K.L. (2012), “What makes online content viral?”, Journal of Marketing Research, Vol. 49 No. 2, pp. 192-205.
Biel, A.L. (1992), “How brand image drives brand equity”, Journal of Advertising Research, Vol. 32 No. 6, pp. 6-12.
Bİlgİn, Y. (2018), “The effect of social media marketing activities on brand awareness, brand image and brand loyalty”, Business and Management Studies: An International Journal, Vol. 6 No. 1, pp. 128-148.
Bivainienė, L. and Šliburytė, L. (2008), “The brand image as an element of brand equity”, Socialiniai tyrimai, Vol. 12 No. 2, pp. 22-31.
Boudana, S., Frosh, P. and Cohen, A.A. (2017), “Reviving icons to death: when historic photographs become digital memes”, Media, Culture and Society, Vol. 39 No. 8, pp. 1210-1230.
Brubaker, P.J., Church, S.H., Hansen, J., Pelham, S. and Ostler, A. (2018), “One does not simply meme about organizations: exploring the content creation strategies of user-generated memes on Imgur”, Public Relations Review, Vol. 44 No. 5, pp. 741-751.
Chen, C. (2012), “The creation and meaning of Internet memes in 4chan: popular Internet culture in the age of online digital reproduction”, Habitus, Vol. 3, pp. 6-19.
Chinomona, R. (2016), “Brand communication, brand image and brand trust as antecedents of brand loyalty in Gauteng Province of South Africa”, African Journal of Economic and Management Studies, Vol. 7 No. 1, pp. 124-139.
Dawkins, R. (1976), The Selfish Gene, Oxford University Press.
Delbeq, A., Van de Ven, A. and Gustafson, D.H. (1975), Group Techniques for Program Planning: A Guide to Nominal Group and Delphi Processes, Scott Foresman Company Glenview.
Díaz, C. and Mauricio, C. (2013), “Defining and characterizing the concept of Internet meme”, CES Psicología, Vol. 6 No. 2, pp. 82-104.
Dubey, A., Moro, E., Cebrian, M. and Rahwan, I. (2018), “MemeSequencer: sparse matching for embedding image macros”, in Proceedings of the 2018 World Wide Web Conference, pp. 1225-1235, doi: 10.1145/3178876.3186021.
Entwistle, J. and Rocamora, A. (2006), “The field of fashion materialized: a study of London fashion week”, Sociology, Vol. 40 No. 4, pp. 735-751.
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural equation models with unobservable variables and measurement error”, Journal of Marketing Research, Vol. 18 No. 1, pp. 39-50.
Gelb, B.D. and Zinkhan, G.M. (1986), “Humor and advertising effectiveness after repeated exposures to a radio commercial”, Journal of Advertising, Vol. 15 No. 2, pp. 15-34.
Gökerik, M., Gürbüz, A., Erkan, I., Mogaji, E. and Sap, S. (2018), “Surprise me with your ads! the impacts of guerrilla marketing in social media on brand image”, Asia Pacific Journal of Marketing and Logistics, Vol. 30 No. 5, pp. 1222-1238.
Grönroos, C. (1997), “Value driven relational marketing: from products to resources and competencies”, Journal of Marketing Management, Vol. 13 No. 5, pp. 407-419.
Guille, A., Hacid, H., Favre, C. and Zighed, D.A. (2013), “Information diffusion in online social networks: a survey”, ACM Sigmod Record, Vol. 42 No. 2, pp. 17-28.
Haryani, S. and Motwani, B. (2015), “Discriminant model for online viral marketing influencing consumers behavioural intention”, Pacific Science Review B: Humanities and Social Sciences, Vol. 1 No. 1, pp. 49-56.
Hauge, A. (2015), “Negotiating and Producing Symbolic Value”, Spatial Dynamics in the Experience Economy, Routledge, pp. 51-63.
Holden, M.C. and Wedman, J.F. (1993), “Future issues of computer-mediated communication: the results of a Delphi study”, Educational Technology Research and Development, Vol. 41 No. 4, pp. 5-24.
Hsin Chang, H. and Wen Chen, S. (2008), “The impact of online store environment cues on purchase intention”, Online Information Review, Vol. 32 No. 6, pp. 818-841.
Krishnamurthy, A. and Kumar, S.R. (2018), “Electronic word-of-mouth and the brand image: exploring the moderating role of involvement through a consumer expectations lens”, Journal of Retailing and Consumer Services, Vol. 43, pp. 149-156.
Kumar, A. and Paul, J. (2018), “Mass prestige value and competition between American versus Asian laptop brands in an emerging market—theory and evidence”, International Business Review, Vol. 27 No. 5, pp. 969-981.
Lee, H., Jang, Y., Kim, Y., Choi, H.M. and Ham, S. (2019), “Consumers' prestige-seeking behavior in premium food markets: application of the theory of the leisure class”, International Journal of Hospitality Management, Vol. 77, pp. 260-269.
Li, Y., Wu, R. and Li, D. (2020), “The influence of subjective characteristics of social network sites on consumers' word-of-mouth sharing”, Online Information Review, Vol. 44 No. 5, pp. 977-994.
Marsden, P.S. (1998), “Memetics: a new paradigm for understanding customer behaviour and influence”, Marketing Intelligence and Planning, Vol. 16 No. 6, pp. 363-368.
Mata-Toledo, R.A. and Gustafson, D.A. (1992), “A factor analysis of software complexity measures”, Journal of Systems and Software, Vol. 17 No. 3, pp. 267-273.
Miltner, K.M. (2018), Internet Memes, The SAGE Handbook of Social Media, pp. 412-428.
Nunnally, J.C. (1978), Psychometric Theory, McGraw-Hill, NY.
Pan, H. (2020), “Online community value co-creation”, Online Information Review, Vol. 44 No. 3, pp. 645-669.
Park, C.W., Jaworski, B.J. and MacInnis, D.J. (1986), “Strategic brand concept-image management”, Journal of Marketing, Vol. 50 No. 4, pp. 135-145.
Phelps, J.E., Lewis, R., Mobilio, L., Perry, D. and Raman, N. (2004), “Viral marketing or electronic word-of-mouth advertising: examining consumer responses and motivations to pass along email”, Journal of Advertising Research, Vol. 44 No. 4, pp. 333-348.
Reza Jalilvand, M. and Samiei, N. (2012), “The effect of electronic word of mouth on brand image and purchase intention: an empirical study in the automobile industry in Iran”, Marketing Intelligence and Planning, Vol. 30 No. 4, pp. 460-476.
Robertson, L.H. (2017), “The infected self: revisiting the metaphor of the mind virus”, Theory and Psychology, Vol. 27 No. 3, pp. 354-368.
Rogers, E. (1995), Diffusion of Innovations, 4th ed., Free Press, New York.
Rowley, J. (2009), “Online branding strategies of UK fashion retailers”, Internet Research, Vol. 19 No. 3, pp. 348-369.
Sääksjärvi, M. and Samiee, S. (2011), “Relationships among brand identity, brand image and brand preference: differences between cyber and extension retail brands over time”, Journal of Interactive Marketing, Vol. 25 No. 3, pp. 169-177.
Sela, A., Milo, O., Kagan, E. and Ben-Gal, I. (2019), “Improving information spread by spreading groups”, Online Information Review, Vol. 44 No. 1, pp. 24-42.
Shifman, L. (2014), Memes in Digital Culture, MIT Press.
Skulmoski, G.J., Hartman, F.T. and Krahn, J. (2007), “The Delphi method for graduate research”, Journal of Information Technology Education: Research, Vol. 6 No. 1, pp. 1-21.
Sommers, M.S. (1963), “Product symbolism and the perception of social strata”, in Greyser, S.A. (Ed.), Toward Scientific Marketing, American Marketing Association, Chicago, pp. 200-216.
Taiwan Electronic Commerce Yearbook (2013), The Applications of E-Commerce in Taiwan 2013, available at: http://ecommercetaiwanapp.blogspot.tw/2013/12/2013_4184.html (accessed 12 October 2019).
Wei, X., Valler, N., Prakash, B.A., Neamtiu, I., Faloutsos, M. and Faloutsos, C. (2012), “Competing memes propagation on networks: a case study of composite networks”, ACM Sigcomm Computer Communication Review, Vol. 42 No. 5, pp. 5-12.
Yang, H.C. and Zhou, L. (2011), “Extending TPB and TAM to mobile viral marketing: an exploratory study on American young consumers' mobile viral marketing attitude, intent and behavior”, Journal of Targeting, Measurement and Analysis for Marketing, Vol. 19 No. 2, pp. 85-98.
Zannettou, S., Caulfield, T., Blackburn, J., De Cristofaro, E., Sirivianos, M., Stringhini, G. and Suarez-Tangil, G. (2018), “On the origins of memes using fringe web communities”, in Proceedings of the Internet Measurement Conference 2018, pp. 188-202, doi: 10.1145/3278532.3278550.
Zhang, C.B. and Lin, Y.H. (2015), “Exploring interactive communication using social media”, The Service Industries Journal, Vol. 35 Nos 11-12, pp. 670-693.
Peres, R., Muller, E. and Mahajan, V. (2010), “Innovation diffusion and new product growth models: a critical review and research directions”, International Journal of Research in Marketing, Vol. 27 No. 2, pp. 91-106.