Richard Dawkins coined the term ‘meme’ in 1976 on the analogy of ‘gene’, but with reference to memory. Memes were proposed to be the cultural equivalent of biological genes, and to explain differences with other species as a differentiating fact responsible for evolution (Dawkins, 2004) or, according to Torres (2015), as one of several elements.

It was proposed that memes organise themselves as cultural units with the ability to pass themselves on and survive in a given social ecosystem, and that, just like genes, they undergo mutations and combine with other memes. According to this theory, memes therefore are ideas, fashions, religions, idioms, sayings or ways of thinking that, like viruses, seem to multiply and spread from mind to mind in our cultural life.

Dawkins (1976, 2004) identifies three characteristics that underlie the concept of the meme: longevity (they have staying power), fecundity (they are capable of reproducing) and copy-fidelity (they stay faithful to the original, given that they undergo change during replication and mutate constantly).

Pérez, Aguilar and Guillermo (2014, p. 81), expounding Dawkins (1976), suggest that ‘memes are a sensory unit that can replicate both transversally (across geographically dispersed groups that are linked by communication and that exist more or less concurrently in time) and longitudinally (across various generations in time)’.

This replication happens for psychosocial reasons and takes place when different persons repeatedly use the meme. It then becomes part of the collective imagination of a given culture.

Two general tendencies can be identified in the study of the communicative aspect of memes: the instrumental-objective approach and the mentalist approach. The mentalist approach, taken by Dawkins himself (1976), takes an intangible perspective to memes; the instrumental-objective approach sees them as an observable, tangible phenomenon and underscores their tangible manifestations. Methodologically and following Pérez, Aguilar and Guillermo (2014, p. 84), Internet memes as cultural units can express themselves as objects of concrete observation that can exist in the minds of social users who share the memes.

This original concept has adapted to the new communicative and social context, having become an indispensable tool for users of the online world and of social media in particular. Davidson (2012, p. 122) describes the Internet meme as a cultural unit that typically takes the form of a joke or is comical in nature and that becomes influential through digital transmission. It could be an image, a song or a video. The description identifies the core qualities of the Internet meme: it is a cultural unit (which intends to communicate something), it almost always has a humorous tone and it is shared online. Its online distribution enhances its ability to go viral, as long as its content has a rhetorical relationship with the social reality it aims to represent. Lankshear and Knobel (2007, p. 224), referring to these new forms of literacy, point out that these are forms that are socially recognised and that generate, communicate and negotiate meaningful content using codified texts within participative contexts. Consequently, they are concrete ideological discourses that have a specific take on the world and that are spread by specific social or institutional groups. The reference points are expressed by way of shared meanings within these groups; as such, the meme acquires a specific purpose. Its success lies in finding the ideal references in connection with the intended context for the meme. ‘These allusions to certain subcultures, techniques and etiquettes create a balance between the norm or convention and possible deviations from it (Ruiz-Martínez, 2018, p. 1001). This essentially rhetorical contrast is addressed by El Groupe µ (1993), which describes the convention as a systematic type of change that is applied to the message in its entirety, requiring a predictable, repeated structure that enhances foreseeability, which in terms makes the meme recognisable and cohesive as it is being spread.

The deviation is described as a localised, surprising alteration of the norm, which reduces the meme’s foreseeability and gives it lasting impact, freshness and communicative usefulness for the Internet users. Shifman (2014, p. 100) talks of different levels of literacy: some can be understood (and created) by almost anyone, while others require detailed knowledge of a meme subculture. Appropriating and spreading memes anonymously, then, involves not only the imitation or appropriation and the mixing of content, but also knowing how to share the content. The phenomenon of the meme cannot be studied without looking at communities of users. It is this community that has the ability to convert content into a meme and that can give us information about which content generates engagement. Just having the image macro format, for example, does not make content liable to be shared.

According to Davidson (2012), for a meme to be successful it must be communicable as well as easy to configure. The first characteristic refers to its reach: many more people share it; the second implies a direct proportionality with time: the more people take it over, the more uses and meanings it can be given so that the meme persists and does not go out of fashion. This constant mutation has led authors such as Shifman (2014) to study memes not as isolated units but as groups that have a number of shared characteristics. They are not just isolated or closed objects, but also open elements that change in small ways over the course of being repeated.

Memes can also be characterised as strongly intertextual elements that connect with each other by way of their own replication. Because of this characteristic the meme acquires new contexts and meanings, and its information becomes more diverse. In this way memes lodge themselves in the collective imagination, weaving a web of references by which they can be identified. These references are often related to mass culture as distributed by Internet users. Because memes are intertextual elements, they not only reproduce through being cited or referenced, but their content also relates an idea or story with an element of popular culture, fitting in with the collective imagination a user must be part of to be able to understand it.

It is not easy to be initiated into understanding memes, given that they are encoded as complicated jargon in the eyes of those who venture for the first time into the process of sharing and creating this type of elements. One of the semiotic structures most often repeated by Internet memes is the one that has become known as the image macro, a static image accompanied by a text. However, a given item will not inevitably be shared just because it has a specific format. Also, to cite Ruiz-Martínez (2018, p. 1011): ‘this text is sometimes divided in two: a protasis or clause at the top of the image and an apodosis at the bottom that serves as a conclusion, refutation or paradox’.

This structure reproduces a common semiotic framework that is essential to comprehension: the images need a stereotyped emotion that is considered funny, they need a humorous tone or they need to show a snapshot of an action being carried out by a person, animal or thing in an amusing context.

These images become containers of emotions and connotations with a very powerful persuasive and visual charge.

The main objective, to evaluate whether memes lead to engagement that causes their content to go viral, was subdivided into the following objectives:

1. O-1. Determining the level of engagement of memes in relation to other resources Netflix uses for its Twitter account (@NetflixES).

2. O-2. Determining the formats and contents of the memes that are most widely used and that generate the most engagement on the Twitter account of Netflix (@NetflixES).

3. O-3. Identifying the communication objectives in the use of memes.

The following hypotheses were formulated based on user experience on Twitter:

1. H-1. Tweets that contain a meme create greater engagement than those that do not.

2. H-2. The image macro is a more frequently used format than the video clip.

3. H-3. Promotion of the Netflix brand is the main communication objective in the use of memes.

Content analysis was carried out to review the memes included in tweets published by the Netflix Spain account (@NetflixES). This produces quantitative and/or qualitative data from which a number of inferences can be drawn based on selected variables (Krippendorff, 1990, p. 11).

The analysis focused on the tweets Netflix published in the fourth quarter of 2019. All tweets originally published by the company were analysed, as well as the retweets, likes and replies they generated. The sample consisted of 307 tweets from the account @NetflixES, in which 112 memes were identified through content analysis.

This sample made it possible to infer the audience response to the memes published in that year; that is, their level of engagement in relation to other communication tools used on Twitter, as well as the characteristics of the memes in terms of their format and content. At the time of the study Netflix had one million followers and had published approximately 16,400 tweets since creating its account in April 2015.

The analysis took place in three phases:

Phase 1. Determining the level of engagement memes generate in relation to the other resources Netflix Spain uses on its Twitter account (objective 1). To this end, the resources used in the tweet and the number of interactions these generated are identified, using the classification by Fernández-Gómez and Martín-Quevedo (2018 a, p. 1,293) as a reference. Their classification is as follows:

1. Social media links to other posts by the platform, or tweets that refer to another person’s posts so as to embed these in the content.

2. Link to own website that redirects users to the content.

3. External link intended to show content other than that of Netflix on Twitter or on its website.

4. Image: including, as per Linares (2015), images of news in the media, advertising posters, premieres or previews, all aimed at promoting own content.

5. Video: including trailers or teasers, ‘making of’ or ‘behind the scenes’ videos, videos of news in the media and videos of premieres or previews.

6. Hashtag: represented by the hash symbol (#), which according to Martínez-Rolán and Piñeiro-Otero (2015) functions as a hyperlink that ties the social conversation generated on Twitter to a specific theme.

7. Emoticons: predesigned images, mostly of facial expressions, inserted into electronic messages.

This stage also involves an analysis of the level of engagement of tweets that contain memes compared to tweets that do not, defined by De Aguilera, Baños and Ramírez (2015) as the intrinsic force that leads the user to make decisions related to the brand and that guides the user’s behaviour. According to the Advertising Research Foundation (2007), cited by González-Bernal (2016, p. 786): ‘the foremost requirement for generating consumer engagement is the emotional opening. According to this logic, engagement is based on the emotions felt by users, which in turn reinforce their thoughts and ultimately lead to a change in behaviour or to certain decisions’.

With reference to the social network that is Twitter, the level of engagement can be compared to the ‘degree to which the user interacts with the brand’ (Del Pino-Silva, 2017, p. 47) as expressed by the indicators of retweets, likes and replies. In addition, the interaction coefficient was calculated which, according to Martínez-Rolán and Piñeiro-Otero (2017, p. 65), ‘is the result of dividing the total number of interactions (retweets, likes and replies) by the number of followers of the output-generating user. This constitutes a value index for analysing the virality of content’. That is to say, the level of virality produced synthetically by this indicator was used to measure the level of engagement with the published contents among users of this network.

The @NetflixES account had 1,050,312 followers at the time of recording the data. Afterwards, the number of times each tweet contained the aforementioned resources was counted, to which were added the interaction rates generated by each resource (total interactions for each resource). These two values (total interactions for each resource and number of times each resource was repeated) were divided among each other to obtain the average interaction rate per tweet for each resource.

Phase 2. Determining which meme formats and contents generate the most use and engagement with the @NetflixES Twitter account (objective 2). To this end, a content analysis of the following variables was carried out (Martínez-Rolán and Piñeiro-Otero, 2015, p. 150):

1. 1. The manifestation: ‘is the materialisation of the meme in itself, that is, its external phenomenon’. The following formats used for the memes were analysed:

   • Visual text: A type of image that can be assigned more than one meaning, as a result of which it can be configured in different ways. The key to this is in the text, which is the element that defines the image. Each person interprets an image in accordance with the mental framework he or she uses to decode all of its components.

   • Image macro: in this format, the function of the image is not to illustrate and reduce the text that accompanies it; instead, the image ‘amplifies it and unlocks its meaning by way of a paradox’ (Ruiz-Martínez, 2018, p. 1,014). In other words, the image macro provides the text with cultural frameworks and an entire collective imagination. Generally speaking, such images are accompanied by a superimposed text or a text in a black box.

   • Photo collage: Photos are inserted that make reference to content on the platform; the photo and the content form a single message.

   • Graphic. A concrete message is put across schematically through shapes, symbols or emoticons.

   • Table. A framework is produced that provides information from the various contents.

   • Screenshot. Specific shots from the series, film or documentary are used.

   • To round out the study, in addition to all of these formats the analysis included another category in which different memes could be grouped:

   • Video clips showing fragments of the contents published by the brand.

2. 2. Behaviour: The various actions users perform to keep the meme in circulation; i.e., sharing it ‘as is’ or altering the image in various ways. Two types of variables were formulated.

   • Image alteration: whether its formal aspects were encoded with editing software.

   • Integration of text with the image: this increases the possibilities for reconfiguring the original content of the image in question. The following subvariables were identified for its encoding:

   • Original text

   • Paraphrases the meme

   • No text

   • Identical to the original

3. 3. Aim, which is the concept underlying the meme (Ruiz-Martínez, 2018, p. 1001). Two variables were formulated for this aspect:

   • Communication objective: followed by the Netflix account

   • Topic: to this end the memes were tagged based on their formal aspects; elements such as hashtags and other textual elements were also taken into account to configure their meaning.

Phase 3. Identifying which communication objectives are pursued by use of the meme (objective 3). To this end, the data collected under the rubric of ‘aim’ in the previous phase were analysed according to the model proposed by Fernández-Gómez and Martín-Quevedo (2018 b), which catalogues the tweets published by the Netflix account according to the objectives they pursue. These objectives are described by Del Pino-Silva (2017), who condenses them into six points in order to fit them into the communication strategies used for the promotion of audiovisual products:

1. Promotion of the Netflix brand, which includes tweets with the following objectives:

2. Providing information about news or services offered by the platform, such as the dates of upcoming premieres or the announcement of events.

3. Greeting or appealing to followers with informal messages and a humorous tone so as to reach them better.

4. Launching messages that directly promote the Netflix brand, containing for example the hashtag #SoloEnNetflix or referring to series marathons.

5. Promoting the audiovisual content of the catalogue: messages providing information or aimed to promote Netflix series, films and documentaries.

6. Starting a conversation with users: inviting them to participate in discussions, to respond to a question or to reply to tweets posted by followers themselves.

7. Sharing alternative content: messages that are unrelated to the brand and deal with an external topical issue.

8. Suggesting competitions, games or quizzes to followers: the brand invites them to participate, aiming to get more followers and to increase its visibility.

9. Inviting followers to take part in surveys in order to know their opinion about a given theme using Twitter’s survey tool.

A review was carried out of the 307 original tweets published by the account @NetflixES, which shed light on the platform’s use of memes and its interaction with the audience. Following this, the data yielded by the resource analysis were examined.

Table 1:

Degree of use and engagement for each resource used by Netflix Spain on Twitter. (Authors, 2020). Source: Prepared by the authors.

Following a review of the 307 tweets, we found – as shown in table 1 – that the most used resource is the image, used in 33% of the cases and repeated 154 times, followed by memes (24%, repeated 112 times) and hashtags (16%, repeated 74 times).

With regard to the level of engagement generated by the tweets among followers of Netflix, however, we see that tweets containing an emoticon come first, with an interaction coefficient of 0.0040. This means that for each publication that included an emoticon we counted an average of 4,223 interactions. These tweets are followed by the link on social media, with a coefficient of 0.0038 and an average of 4,036 interactions per posting and, thirdly, by the meme, with an interaction coefficient of 0.0022 and an average of 2,322 interactions.

Graphic 2:

Interaction coefficient for each resource. Source: prepared by the authors.

A total of 112 memes were found, making them the second-most used resource by the streaming platform. For the purpose of a detailed analysis of the level of use and engagement, the study looked at the three aforementioned differential aspects: manifestation, behaviour and aim of the meme. Each of these aspects in turn included previously encoded subvariables. The collected data are summarised in table 2, which seeks to analyse the manifestation of the 112 analysed memes.

Table 2:

Analysis of the degree of use and engagement of memes from Netflix Spain on Twitter (Díaz-Molina and Arroyo-Almaraz, I. (2020). Source: Prepared by the authors.

Regarding the variable of formats, we see that the most used format is the visual text, used in 42% of the cases (47 publications). Their preponderance suggests that the platform makes use of a number of highly stereotyped arrangements of image and text. The text here is ultimately that which gives meaning to the image and provides a reference framework for it. The second-most used format is the image macro, present in 31% of the memes (35 publications), while the third-most used format is the video clip, used in 11% of the cases (12 publications). All of the formats were characterised by including images or video fragments from the platform’s series, films and documentaries. They were published, moreover, with messages that emulate the communicative dynamics of the audience on this platform.

The format that reached the highest level of engagement, however, was the graphic, with an interaction coefficient of 0.0079 (an average of 8,837.75 interactions per post), followed by the photo collage, with a coefficient of 0.0023 (an average of 2,410.5 interactions), and the video clip, with a coefficient of 0.0022 (an average of 2,315 interactions). The symbolism used by the graphics to express a concrete idea and connect emotionally with the audience was characterised by its originality and led to greater interaction with the content among followers.

Graphic 3:

Interaction coefficient per format. Source: prepared by the authors.

With regard to the level of use of alteration techniques, we find that for the great majority of publications alterations made with editing programs were not effective. This applied to 87% (97 publications), whereas in 13% of the cases (15 publications) compositional aspects were in some way altered with programs such as Photoshop. This confirms the deliberate way in which Netflix goes about showing original content, selecting images strategically and choosing easily identifiable images.

With regard to the level of engagement, images that had been altered in some way had an interaction coefficient of 0.0008 (an average of 873 interactions), whereas images that had not been altered had an interaction coefficient of 0.0024 (an average of 2,568 interactions). We see that users interacted more with images that were not altered.

The two variables of the meme’s communication objective were analysed: the objective of Netflix and the meme’s topic.

The objective of Netflix

Promotion of the Netflix brand was the foremost objective of 57% of the memes (64 publications). These memes were characterised by a message with a humorous or informal tone that made reference to the brand, or by aspects that identified the followers in an emotional way. The insights of the audience and the idiom used became the central element with which the attention of the audience was drawn.

This was followed by the promotion of the catalogue’s audiovisual content, at 35% (39 publications), for which various resources were used: hashtags, the text or the image itself.

Lastly, there is the objective of conversations with the user, at 8% (9 publications). We note the absence of memes intended to suggest competitions or games, or memes intended to promote alternative content and to invite the audience to take part in questionnaires.

Graphic 4:

Degree of use of the communication objectives Source: prepared by the authors.

With regard to the level of engagement, values were inverted compared to the level of use, given that the least used communication objectives obtained the highest interaction coefficients. Starting a conversation with the user generated an interaction coefficient of 0.0024 (an average of 2,517 interactions per tweet). This concerned posts that initiate a discussion or directly invite users to share their opinion. The effort made by the platform to mobilise the audience resulted in a greater number of interactions.

Topics

In the first place, in 77% of the cases, memes were used to define an emotion, behaviour or situation (86 publications). The memes were designed principally to appeal directly to the emotions or behaviours of the public, as well as to emotions or behaviours with which they identify. In the words of Ruiz-Martínez (2018, p. 2,012): ‘a basic semiotic characteristic is created in order for the meme to be understood: a highly expressive face (human or animal), clearly (and stereotypically) expressing an emotion (confusion, disgust, hilarity, horror), almost always ridiculous or laughable’.

In the second place, at 14% (16 publications), there were those memes that refer to elements from series. In these cases, rather than referring to the viewer, the meme took as its central aspect elements related to Netflix content and framed these in a discursive structure that imitated a meme.

Thirdly, at 9% (10 publications), there were those memes referencing a behaviour that was contrary to what would be expected. This topic is expressed strongly through the figure of speech of the paradox, expressed as a highly symbolic image. The image encapsulates a complete iconic universe that rounds out its meaning and has a humorous effect.

As shown in Table 2, in terms of the level of engagement we found that the highest interaction coefficient (at 0.0023, an average of 2,450 interactions) was obtained by memes that expressed an emotion, behaviour or situation. These were followed closely (at 0.0022, an average of 2,402 interactions) by memes reflecting a behaviour that was contrary to what was expected. Lastly, there were those memes that referred to elements from the series, with a coefficient of 0.0016 (an average of 1,720 interactions). It follows that memes that appeal directly to emotions or behaviours are likely to have a greater interaction index than memes that simply refer to elements of the content created by Netflix.

Graphic 5:

Interaction coefficient of the memes’ topics. Source: prepared by the authors.