Graduate School of Communication
Master’s programme: Corporate Communication
Master Thesis
Startups’ cross-cultural communication on social media
Analysing the effects of cultural congruence between the content and the audience on endorsement.
Valentina Neri Serneri 11181389
Supervisor: Theo Araujo 30-06-2017
Abstract
This research addresses startups’ cross-cultural communication on social media, specifically focusing on the extent to which an audience endorses a piece of content when it is culturally congruent with them. Several studies about online cross-cultural communication have been carried out analysing websites in different languages while in this study we examine social media messages. In addition to that, literature on the topic only considered multinational, established companies whereas here we analysed startups’ communication. To stay in the market, it is fundamental for startups to find the best strategy to receive positive endorsement, increase the brand visibility and build a strong brand. A quantitative content analysis of 300 Facebook posts published by 30 startups based in three countries was carried out. We selected countries that score differently in Hofstede’s dimension of
individualism/collectivism, USA, Italy and Chile. To analyse the posts we used Hofstede’s dimension of individualism/collectivism, Hall’s low/high context communication styles and Markus and Kitayama’s independent and interdependent self-construal and values. The findings showed that cultural congruence particularly increases the comments and the shares ratio while it does not affect the likes ratio and the valence of the comments posted by users. Our findings extend academic knowledge about cultural antecedents of endorsement and contribute to the scarce literature about startups. This paper concludes by providing some practical implications as well as limitations and suggestions for future research.
Keywords: cross-cultural communication, cultural antecedents, startup, social media, endorsement
Startups’ cross-cultural communication on social media
Analysing the effects of cultural congruence between the content and the audience on endorsement
Eighty-eight percent of the marketers agree that social media brought benefit to their businesses thanks to this tool’s power to easily connect companies with clients (Byberg, 2016). However, many practitioners still struggle with successfully engaging their followers (Sklar, 2013) especially when considering that consumers communication preferences vary from country to country (Tsai & Men, 2017). Then, it is fundamental for companies to understand why consumers would engage with their social media content. Literature defines consumer engagement as the act of establishing a collaborative dialogue between a company and its consumers (Dhanesh, 2016).
As culture plays an important role in shaping the communication between a company and its stakeholders (Men & Tsai, 2012; Tsai & Men, 2017), we will explore the cultural antecedents of engagement on social media. We analyse startups’ messages on social media to understand whether consumers are more likely to endorse them when they express the cultural characteristics of the country where they are published. We define endorsement as an act a consumer does to signal his/her affiliation with a brand (Bernritter, Verlegh, & Smit, 2016).
Scholars argue that adapting an organization’s online content to the cultural values of the target audience is rewarding in terms of positive attitude toward the content itself and toward the organization consequently (Mooij, 2015; Sinkovics, Yamin, & Hossinger, 2007). Several studies about online cross-cultural communication have focused on web-sites (Baack & Singh, 2007; Okazaki & Rivas, 2002; Okazaki & Taylor, 2013) whereas in this study social media messages will be analysed to fill the literature gap. Furthermore, scholars have carried out cross-cultural studies on social media to understand antecedents of use, identity
construction and more (Barry & Bouvier, 2012; Chu & Choi, 2011; Park, Jun, & Lee, 2015) while little is known about the influence of the national culture on engagement (Tsai & Men, 2017). In addition to that, to our knowledge, no previous studies on cross-cultural
communication have analysed endorsement on social media and related it to the cultural congruency between the user and the content itself. Moreover, studies on online cross-cultural communication have concentrated on multinational, big companies (Gevorgyan & Manucharova, 2009; Nacar & Burnaz, 2011; Okazaki & Rivas, 2002; Singh & Matsuo, 2004; Vyncke & Brengman, 2010) while, as of today, no research on the topic has been carried out regarding startups on which existing literature is scarce (Bresciani & Eppler, 2010).
Startups represent an interesting case because they have no developed identity and they lack of internal structures and financial resources (Bresciani & Eppler, 2010; Rode & Vallaster, 2005). Furthermore, it is fundamental for startups to find the best communication strategy to build a brand, a reputation, to acquire new customers and investors (Bresciani & Eppler, 2010). Even though it was found that startups do not prioritize having a defined communication strategy over other financial and administrative issues (Bresciani & Eppler, 2010), entrepreneurs acknowledge the importance for new ventures to have a good
communication strategy to growth and stay in the market (Rode & Vallaster, 2005; Timmons, 1999). In this paper social media messages will be analysed because of the advantages social media offer especially to startups (Harris & Rae, 2009) thanks to their low-costs and the opportunity to reach a wide audience.
This research will bring innovative insights into the extent to which the cultural context influences how consumers react to social media messages. This study also has some
important managerial implications as it will provide practical recommendations to startups’ communication and marketing practitioners to understand how to better create content to increase consumer engagement on their social media pages.
In this study we focus on endorsement which will be analysed both as the interactions users have with a post and in terms of valence of the comments written on posts. We will treat the cultural characteristics of the organization and its audience as antecedents (Walsh, Shiu, & Hassan, 2014) to understand how culture influences online communication and endorsement. We will use Hofstede’s dimension of individualism/collectivism, Hall’s low/high context communication styles and Markus and Kitayama’s independent and interdependent self-construal and values (Gudykunst et al., 1996) to examine messages. Individualism/collectivism indicates how tight interpersonal relations are in each society (Hofstede, 2001); the communication style concerns how explicit or implicit communication between individuals is (Hall, 1976), while self-construal and values drive the behaviours of members of a group (Markus and Kitayama, 1991). Literature found that
individualism/collectivism and Hall’s communication styles are the most prominent dimensions to explain a culture’s foundation (Pineda, Hernández-Santaolalla, & Rubio-Hernánde, 2015).
Based on the previous discussion, the research question leading this study is formulated as:
RQ: To what extent do startups’ social media messages framed according to the national cultural dimension of individualism/collectivism, self-construal and values and communication style affect the audience in terms of endorsement and positive sentiment?
Theoretical Framework
This study aims at investigating startups’ communication on social media with a cross-cultural perspective and specifically focuses on consumer endorsement. There is little agreement among scholars and practitioners on the definition of startup company as every
academic study “developed an own ad hoc definition” (Luger & Koo, 2014, p. 17) . Luger and Koo (2014) were the firsts to propose a definition of startup that could be widely accepted. They define a startup company as a new, active and independent venture. “New” refers to the date in which the company was founded. “Active” refers to the fact that the company should be operating and trading goods or services. “Independent” means that the new venture should not be controlled by another company.
As aforementioned, this study will analyse startups’ communication on social media. Scholars have pointed out the importance for startups to deploy a consistent communication strategy since the early phases (Rode & Vallaster, 2005) to create a brand image and to promote and develop their business (Harris & Rae, 2009). Social media are a strategic tool for startups as they have no entry barriers, they are low cost, they have a great potential reach (Bresciani & Eppler, 2010) and allow to establish a dialogue with consumers (Cvijikj,
Spiegler, & Michahelles, 2011). Moreover, social media marketing also leads to financial benefits when effective (Goh, Heng, & Lin, 2013; Pletikosa Cvijikj & Michahelles, 2013). By effective, we mean when consumers positively endorse the company (Pletikosa Cvijikj & Michahelles, 2013).
An overview of the literature on endorsement, cross-cultural communication and the cultural elements analysed in this study will be presented in the next section. However, academic literature refers to these topics in general and not specifically for startups as studies on startups’ communication are scarce (Bresciani & Eppler, 2010).
Endorsement
Social media have changed the way companies interact with their audiences (Kaplan & Haenlein, 2010). Literature strongly suggests startups to leverage on social media power to reach customers and build a relation with them (Bresciani & Eppler, 2010), foster awareness
and enhance engagement (Byberg, 2016). Receiving consumer endorsement is fundamental for organizations to be perceived as trustworthy and authentic (Men & Tsai, 2014).
Endorsement is defined as a positive and interactive act an individual does to signal his or her affiliation with a brand on social media (Bernritter et al., 2016; Brodie et al, 2011). Endorsing a brand is a public behavior that can perceived by others, in contrast to other forms of private engagement such as subscribing to a newsletter that, being private, are not considered
endorsement (Bernritter et al., 2016).
Literature found that online content adapted to the cultural values of the target audience produces positive outcomes in terms of consumers’ attitude toward web-site’s design features and its overall effectiveness for example (Gevorgyan & Manucharova, 2009; Sinkovics et al., 2007; Vyncke & Brengman, 2010). This is also true in regard to social media where culturally tailored content successfully engage users as it meets their
self-authentication needs, namely the need to express their true self (Arnauld & Price, 2000). From a cultural perspective on the motives of consumption on social media (Beverland, Dobele, & Farrelly, 2015; Men & Tsai, 2012), we assume that consumers will endorse messages that allow them to express something about their personal identity (Sunny Tsai & Rita Men, 2013) and their own cultural values. Literature shows indeed how people use social media to express their individuality (Arnould & Thompson, 2005; Dunne, Lawlor, & Rowley, 2010) and personal identity (Hollenbeck & Kaikati, 2012; Muntinga, Moorman, & Smit, 2017). Consumers endorse brands, among other motives, to express their self (Bernritter et al., 2016) and their “values proposition” (Beverland et al., 2015, p. 660).
Users can engage with organizational-related content on Facebook by liking, sharing o commenting it (Kim & Yang, 2016). These behaviors are also three forms of endorsement. These actions have different levels of brand-related activeness which are consuming,
behavior that involves little effort by the user (Kim & Yang, 2016) who can simply apprecciate the content of a post. Sharing and commenting are contributing behaviors. By sharing a post an user transmits the content to his or her social group (Cho, Schweickart, & Haase, 2014). Commenting involves some effort in writing an actual contribution (Cho et al., 2014). These behaviors are very much identity-related as they allow an individual to self-express, self-present and self-assure (Muntinga et al., 2017).
In addition to the number of likes, shares and comments, the valence of the comments users leave under a post will also be investigated in this study. It is important to monitor how users talk about a brand on social media to understand their opinion and attitude toward the brand itself (Neri, Aliprandi, Capeci, Cuadros, & By, 2012). Users can write positive, negative or neutral comments. It has been found that positive comments lead to a more positive attitude toward the brand for people who read those comments (Barnes & Jacobsen, 2014; De Vries, Gensler, & Leeflang, 2012) and it should therefore be a priority for
companies to foster positive comments under their posts (Hoffmann, Röttger, Ingenhoff, & Hamidati, 2015). Monitoring sentiment on social media is a relatively new concept (Barnes & Jacobsen, 2014). However, while post characteristics that drive comments have been studied (De Vries et al., 2012; Lam & Lee, 2009), scholar have not specifically considered the relation between the cultural dimensions expressed in the post and the valence of the comments user write on it.
How culture influences communication
Cross-cultural communication. Many definitions of culture have been proposed in literature. Hofstede (2011) defines it as “the collective programming of the mind that distinguishes the members of one group or category from people of others" (p. 3). From this definition it emerges that culture is a collective phenomenon shared by multiple individuals
(Mooij, 2015). Hofstede’s definition also highlights both the existence of similarities shared by individuals belonging to a certain group and the fact that those characteristics make groups different.
One of the prominent elements of a culture is communication. Communication is a mechanism individuals use to express and interpret a message (Craig & Douglas, 2006). Individuals communicate differently according to their cultural values, perceptions and behaviours (Gudykunst, 1997; Gudykunst et al., 1996; Triandis, 2004) and depending on the norms they are socialized (Cayla & Arnould, 2008; Gudykunst, 1997) which varies across different cultures.
It has been proved that national culture also has an influence on consumer behaviour (Craig & Douglas, 2006; De Mooji, 2010) by stimulating the psychological process related to the framework individuals use to make sense of events (Möller, Eisend, & Ller, 2010). Hence, several scholars (Cayla & Arnould, 2008; Özsomer & Simonin, 2004; Walsh et al., 2014) have underlined the importance for companies to adapt their communication to the cultural context of their audience and elaborate messages that are culturally congruent with the culture of their consumers. Adapting the communication means placing certain cultural values in messages (Holland & Gentry, 1999). It represents the process of integrating the core cultural norms of a social group into the structure or format of online messages directed at that group (Gevorgyan & Manucharova, 2009). Scholars and practitioners have been discussing for over forty years whether adapting communication messages to the cultural dimensions of the target audience could be rewarding or not in terms of return on investments of the budget spent on adapting the content to the different cultures (Baack & Singh, 2007; Singh & Matsuo, 2004; Sinkovics et al., 2007) and now it is generally accepted that an adaptation strategy successfully impacts marketing efforts (Mooij, 2015; Sinkovics et al., 2007).
In this research the adaptation of the text and the images of social media messages will be investigated. Regarding the text, we will pay particular attention to the use of the language. It has been argued that language is a primary factor of communication as it provides a schema to encode and decode messages (Craig & Douglas, 2006). A language shared among the members of a group facilitates interactions and consequently acts as an unifying factor of those members and of a society in general (Craig & Douglas, 2006). The linguistic element is particularly relevant in this study as language plays a fundamental role in stimulating psychological processes related to consumer behaviour such as perception,
judgement and choice (Craig & Douglas, 2006). However, language is only one of the two dimensions of communication, being the other one images (Craig & Douglas, 2006; Hall, 1976). Studies indicate that visual messages are framed differently across countries and cultures as well (Jin, 2010; Okazaki & Rivas, 2002) demonstrating that cultural cues can be transmitted in ways that go beyond words.
Elements of a culture
Individualism/collectivism As aforementioned, Hofstede’s framework is the most effective framework to analyse cultural variance in business and marketing researches (Baack & Singh, 2007; Mooij, 2015; Steenkamp, 2001). By collecting empirical data, Hofstede distinguished cultures upon six different dimensions which are power distance, uncertainty avoidance, masculinity/femininity, individualism/collectivism, long/short term orientation and indulgence/restraint.
As individualism and collectivism lie at the foundation of any culture and shape an individual psychological process, this study will focus on this dimension. Individualism and collectivism are two opposite sides of a continuum and mainly refer to the degree to which people in a society are integrated into groups. Tendencies toward individualism or
collectivism exist in every culture even though one predominates (Gudykunst, 1997).
Collectivistic societies stress the importance of in/out groups dynamics and therefore group’s priorities are more important than the individual ones. Individuals in collectivistic cultures normally belongs to a limited number of groups such as family or the religious circle. These few groups have a great influence on the single individual behaviour being the ties between in-group members very strong. Conversely, individualistic societies value individual
priorities and people belong to multiple and more specific groups which have little influence on the single’s behaviour (Gudykunst, 1997; Hofstede, 1980; Hofstede, 2011). As Hofstede’s dimension of individualism-collectivism also explains consumers’ cultural differences in online communication preferences (Goodrich & de Mooij, 2013), then we hypothesize that:
H1: The more culturally congruent in individualism the posts are in respect to the country where they are published, the higher the number of (a)likes, (b) shares, (c) comments and (d) more positive the valence of the comments.
H2: The more culturally congruent in collectivism the posts are in respect to the country where they are published, the higher the number of (a)likes, (b) shares, (c) comments and (d) more positive the valence of the comments.
Communication style The extent to which a culture is individualistic/collectivistic affects its communication practices (Gudykunst, 1997). Aspects of a culture specifically related to communication can be analysed by taking into account Hall’s communication styles (Soares, Farhangmehr, & Shoham, 2007). Hall (1976) made the distinction between high-context and low-context communication. In high-context societies members share the same broad communication code. Communication tends to be implicit, the voice tone and
other contextual factors have a great influence on the way messages are encoded and decoded. Conversely, in low-context societies communication tends to be more explicit, the content of a message is more important than the context in which it has been communicated. We know that consumers prefer a communication style that is consistent with their cultural background (Tsai & Men, 2017). As communication preferences also drive consumption behaviours, then we hypothesize that:
H3: The more culturally congruent in the communication style the posts are in respect to the country where they are published, the higher the number of (a)likes, (b) shares, (c) comments and (d) more positive the valence of the comments.
Self-construal and values A culture’s individualism/collectivism has both a direct and an indirect effect on the communication style of members of that culture (Gudykunst et al., 1996). The indirect effect of individualism/collectivism on communication style is mediated by an individual self-construal and values. Culture and the self-concept are “dynamically inter-related” (Hansen et al., 2012, p. 223). Values differ from one culture to another. As a culture lies in a society beliefs, institutions and practices, then it will shape a society individuals’ ideas and values (Gudykunst, 1997; Hansen et al., 2012) which work as drivers of people behaviours (Luna & Gupta, 2001). Markus and Kitayama (1991) introduced the concepts of independent and interdependent self-construal and values. The former
emphasizes the oneness and singularity of each individual and it is found more in
individualistic societies while the latter emphasizes the tight relations between members of a group and it is found more in collectivistic societies (Okazaki & Rivas, 2002; Singelis, 1994). As consumers endorse brands on social media to express their own values proposition
H4: The more culturally congruent the independent self-construal and the values expressed in the posts are in respect to the country where they are published, the higher the number of (a)likes, (b) shares, (c) comments and (d) more positive the valence of the
comments.
H5: The more culturally congruent the interdependent self-construal and the values expressed in the posts are in respect to the country where they are published, the higher the number of (a)likes, (b) shares, (c) comments and (d) more positive the valence of the
comments.
Figure 1. Representation of the model tested in this study.
Method
Research design To conduct this research about startups’ cross-cultural
2012) of Facebook posts. This method allows the researcher to identify the characteristics of a message (Holsti, 1969) and it has already been extensively used in studies about online cross-cultural communication (Jin, 2010; Okazaky & Rivas, 2002; Singh & Matsuo, 2004).
Sample To have a nuanced overview of the phenomenon of cross-culture communication, we selected startups operating in countries that score differently on
Hofstede’s dimension of individualism/collectivism, USA, Italy and Chile. While the USA is a very individualistic country, Italy is less individualistic and Chile is considered a
collectivistic country (Hofstede, 2001). On Hofstede’s scale of individualism versus
collectivism, from 0 being very collectivistic and 100 being very individualistic, Chile scores 26, Italy 76 and US 961.
To draw the sample, data were collected from 30 startups’ Facebook profiles, 10 startups’ profiles per country. We retrieved startups by Angellist.com, a website that connects new ventures, investors and job seekers. To be included in the sample a startup had to fulfil with five criteria. First, they had to operate in one of the three selected countries. Second, as Luger and Koo (2014) define a startup as a new company, the Italian norm2 was applied, where a company to be considered as startup must be established for less than 5 years. Third, the startups needed to operate business to consumer. Swani, Milne, Brown, Assaf and Donhtu (2016) found that social media users are more likely to endorse B2C content than business to business one. Fourth, the startup had to have a Facebook profile. We chose Facebook among other social media networks because of the highest number of active users and for its major importance for business (Pereira, de Fátima Salgueiro, & Mateus, 2014). Lastly, startups had to communicate in the local language hence Spanish for the Chileans, Italian for Italians and English for the US ones. Selected startups are presented in Appendix A.
1
Scores retrieved from: https://geert-hofstede.com/countries.html
2
http://www.sviluppoeconomico.gov.it/images/stories/documenti/Executive_Summary_Italy_ Startup_Act_02_05_2016.pdf
Procedure By using the application Netvizz (Rieder, 2013), we downloaded files containing data about the last 10 posts of each startup. Then we uploaded each file on a Jupyter notebook to be read and we wrote Python commands to extract the needed data. Data were prepared with Excel and IBM SPSS and later analysed by using IBM SPSS.
Data for the comments’ sentiment analysis were collected by using the same procedure: the text of the comments was extracted by using a Jupyter notebook and then automatically analysed with SentiStrength (Thelwall, Buckley, Paltoglou, & Cai, 2010).
Inter-coder reliability Inter-coder reliability was also carried out together with a fellow communication science student with knowledge of content analysis and fluent in the three languages used here, English, Spanish and Italian. Based on literature on the topic we developed a first version of the codebook which was used to code 20% (N = 60) of the sample on Qualtrics. Being Krippendorff’s Alpha lower than 0.60 for about half of the variables, the codebook has been redefined (see Appendix B) and inter-coder reliability carried out a second time. The results of the second test showed a strong reliability for all the variables (see Appendix C).
Operationalization First, we coded basic information about the startup as the unitary code assigned to each company, the sector in which it operates (Finance, n = 1; Biotech, n = 1; Education, n = 4; Retail, n = 5; Food and beverage, n = 8; Travel, n = 6; Other, n = 5), the number of likes the Facebook page has (N = 100, M = 44669.45, SD = 78185.02) and the number of interactions each post had (M = 47.67, SD = 194.492).
Then, we coded variables concerning single posts (n = 300) and used as dependent variables: the number of likes per post (M = 37, SD = 166.80); the number of shares per post (M = 5.95, SD = 30.3), the number of comments per post (M = 2.67, SD = 8.96) and the valence of comments (n = 540, M = 0.67, SD = 1.44). Biases could arise when using likes, shares and comments as integer variables due to the different number of fans pages have and
therefore the endorsement some content could reach. For this reason, we decided that using ratio variables could be a better solution than just controlling for the number of likes the Facebook page has and therefore we calculated likes ratio (M = 0.45, SD = 2.24), share ratio (M = 0.04, SD = 0.16) and comments ratio (M = 0.02, SD = 0.08) by dividing the number of likes, shares and comments a post has by the number of followers of the page. All the variables were non-normally distributed: likes ratio had a skewness of 8.27 (SE = 0.14) and kurtosis of 72.9 (SE= 0.28); share ratio had a skewness of 7.23 (SE = 0.14) and kurtosis of 67.52 (SE= 0.28) and comments ratio had a skewness of 6.31 (SE = 0.14) and kurtosis of 48.16 (SE= 0.28). As all the variables were positively skewed, they have been log-transformed to be used in the analysis.
Next, we coded variables used as independent starting with the score of a country on Hofstede (2001) scale of individualism/collectivism. To test hypothesis 1 and 3 we used as a predictor the level of a country individualism (IDV) on Hofstede’s scale. To test hypothesis 2, 4 and 5, we computed a new variable, level of a country collectivism (COL), by subtracting the IDV value of each country to 100.
The first independent variable we analysed was individualism which indicates how individualistic the post is. For each post, the text and the image were coded separately and then computed together. We calculated individualism in text by counting the frequency of certain words that were extracted from scales by Singelis, Triandis, Bhawuk and Gelfand (1995) and Triandis and Gelfland (1998). The list of the words is available in the codebook (Appendix B). To calculate individualism in image we used items from Callahan (2005), Jin (2010) and Würtz (2005) were used. Individualism in text and individualism in image were computed together to have the variable individualism in post (from 0, not individualistic to 2, very individualistic). 58.7% of the posts were not individualistic, 37.3% were weakly
individualistic, 3.7% were moderately individualistic and 0.3% were strongly individualistic (M = 0.31, SD = 0.43).
Collectivism indicates how collectivistic the post is. For collectivism in text another list of words inspired by Triandis and Gelfland (1998) and Singelis, Triandis, Bhawuk and Gelfand (1995) was used. This list is also available in the codebook. To calculate collectivism in image we used items from Callahan (2005), Jin (2010), Singh and Matsuo (2004) and Würtz (2005). The two variables were computed together in collectivism in post (from 0, not collectivistic to 2, very collectivistic). 45.5% of the posts were not collectivistic, 47.6% were weakly collectivistic, 6% were moderately collectivistic and 0.7% were strongly collectivistic (M = 0.43, SD = 0.48).
Communication style measures how low or high the communication style used in the post is. Communication style in text and communication style in image were adapted by Okazaki and Rivas (2002), Singh and Matsuo (2004), Würtz (2006). They were computed together into communication style (from 1= low to 3 = high). 39.5% of the posts had a low communication style, 26.7% were neutral and 33.8% had a high communication style (M = 1.68, SD = 0.71).
Independent construal and values measures how individualistic are the self-construal and the values expressed in the post. A scale proposed by Konsky, Eguchi, Blue & Kapo (2000) containing 5 independent values (power, achievement, hedonism) was used for independent self-construal and values in text and independent self-construal and values in image. These two variables were then computed together to have independent self-construal and values (from 0, not expressing any individualistic self-construal and values to 5,
expressing 5 individualistic values). 53.9% of the posts did not express any independent self-construal and value, 31.4% of the posts expressed one independent self-self-construal and value,
13.7% expressed two independent self-construal and values and 1% expressed three independent self-construal and values, (M = 0.70, SD = 0.60).
Interdependent construal and values measures how collectivistic are the self-construal and the values expressed in the post. For both interdependent self-self-construal and values in text and interdependent self-construal and values in image items from Okazaky and Rivas (2002) were adopted (importance of nature, tradition, loyalty). The two variables were then computed together to have interdependent self-construal and values (from 0, not
expressing any interdependent self-construal and value to 4, expressing 4 interdependent values). 60% of the posts did not express any interdependent self-construal and value, 28.6% expressed one interdependent value, 9% expressed two interdependent values and 2% of the expressed 3 interdependent values, (M = 0.67, SD = 0.80).
An overview of the correlations between the variables studied here is presented below in Table 1.
Table1
Pearson’s Product Moment Correlations for the Main Effects
Main effects Likes Shares Comments
Valence of the comments IDV COL Individualism .08 .09 .07 .04 -.03 Collectivism .04 -.05 .22 .04 .03 Independent self-construal and values .15† .16** .14* .02 -.2*** Interdependent self-construal 0† -.03 .04 -0.2 .5**
and values Communication style -.03 -0.07 -.06 .04 -.18* IDV .05 .04 -.03 .00 COL -.06 -.05 .22 -.01 Note. † = p < .10, * = p < .05, ** = p < .01, *** = p < .001. N = 300.
Analysis To test the hypothesis and answer the research question, we carried out a series of multiple regressions. To avoid multicollinearity issues occurring when many interaction effects are used together as predictors, each independent variable, the IDV/COL and the interaction effect between them have been analysed in separate multiple regressions. Results of the analysis are presented in the next section.
Results
We carried out four multiple regressions with individualism, IDV and the interaction effect as independent variables to test hypothesis 1. The multiple regression with likes ratio as dependent variable is non-significant F(3, 295) = 0.42, p = .736, as well as the one with share ratio as dependent variable F(3, 295) = 1.96, p = .119, the one with comments ratio as dependent variable F(3, 295) = 0.63, p = .595, and the one with valence of the comments F(2, 537) = 1.10, p = .352. Hypothesis 1 is therefore not supported. Results of the multiple
regressions are shown in Table 2.
Table 2
Regression Table with Individualism, IDV and the Interaction Effect as Predictors
Likes ratio Shares ratio Comments ratio
Valence of the comments
Variable b SE B b* b SE B b* b SE B b* b SE B b* Constant .08 .03 .02 .01 .01 .01 .51 .08 IDV .00 .00 -.01 .00 .00 -.04 .00 .00 .01 .00 .00 .04 Individualism -.02 .05 -.04 -.01 .01 -.11 .00 .01 -.05 .51 .29 .18† IDV* Individualism .00 .00 .10 .00 .00 .24† .00 .00 .07 -.01 .00 -.17 R2 .004 .01 .006 .01 F .42 1.96 .63 1.10
Note. † = p < .10, * = p < .05, ** = p < .01, *** = p < .001. N = 300 for likes ratio, shares ratio, comments ratio; N = 540 for valence of the comments.
To test Hypothesis 2, we performed four multiple regressions with collectivism, COL and the interaction effect as independent variables.
The model with likes ratio as dependent variable is non-significant F(3, 295) = 1.32, p = .269 as well as the one with share ratio as dependent variable, F(3, 295) = 0.36, p = .781.
Hypothesis 2.a and 2.b are not supported. The multiple regression with collectivism, COL and the interaction effect as independent variables and comments ratio as dependent variable, shows that the model is significant, F (3, 295) = 3.53, p = .015. Collectivism, b = 0.01, b* = -0.21, t = -2.01, p = .037, COL, b = -0.00013, b* = -0.12, t = -1.67, p = .096, and the
interaction effect of collectivism and COL, b = 0.00036, b* = .034, t = 3.14, p = .002, predict 3.5 per cent of the variation in the comments ratio (R2 = .035). Hypothesis 2.c is supported: the result indicates that the higher the level of collectivism in the post, the comments ratio increases the more collectivistic the country is. The multiple regression with valence of the
comments is non-significant, F (3, 536) = 0.32, p = .809. Hypothesis 2.d is also not supported. Results of the multiple regressions are shown in Table 3.
Table 3
Regression Table with Collectivism, COL and Interaction Effect as Predictors
Likes ratio Shares ratio Comments ratio
Valence of the comments Variable b SE B b* b SE B b* b SE B b* b SE B b* Constant .09 .02 .02 .00 .01 .00 .71 .15 COL -.00 .00 -.10 .00 .00 -.06 -.00 .00 -.12† -.00 .00 -.03 Collectivism -.03 .04 -.08 -.00 .01 -.00 -.01 .00 -.21* -.08 .24 -.01 COL* Collectivism .00 00 .19† .00 .00 .06 .00 .00 .34** .00 .00 .07 R2 .01 .004 .03 .00 F 1.32 .36 3.53 .32
Note. † = p < .10, * = p < .05, ** = p < .01, *** = p < .001. N = 300 for likes ratio, shares ratio, comments ratio; N = 540 for valence of the comments.
Next, we performed other four multiple regressions to test Hypothesis 3. The multiple regression with communication style, COL and the interaction effect as independent variables and likes ratio as dependent variable is non-significant F(3, 295) = 1.44, p = .233 as well as the one with share ratio, F(3, 295) = 1.19, p = .313. Hypothesis 3.a and 3.b are not supported.
The multiple regression with communication style, COL and the interaction effect as independent variables and comments ratio as dependent variable, shows that the model is significant, F (3, 295) = 2.65, p = .049. COL, b = -0,00029, b* = -0.32, t = -2,221, p = .027, communication style, b = -0.1, b* = 0.26, t = -2.718, p = .007, and the interaction effect, b = 0,00019, b* = 0.38, t = 2.526, p = .012, predict 2.6 per cent of the variation in the comments ratio (R2 = .026). Hypothesis 3.c is supported: the comments ratio increases when the post uses a high communication style when the communication style of a country is higher. The multiple regression with valence of the comments is non-significant, F (3, 536) = 0.32, p = .561. Hypothesis 3.d is not supported. Results of the multiple regressions are shown in Table 4.
Table 4
Regression Table with Communication Style, COL and Interaction Effect as Predictors
Likes ratio Shares ratio Comments ratio
Valence of the comments Variable b SE B b* b SE B b* b SE B b* b SE B b* Constant .17 .05 .04 .12 .02 .00 .70 .23 COL -.00 .00 -.26† .00 .00 -2.19 -.00 .00 -.32* -.00 .00 -.01 Communication style -.05 .03 -.20* -0.01 .06 -.17† -.1 .00 .26* -.01 .13 -.01 COL* Communication style .00 .00 .27† -.00 .00 .19 .00 .00 .38* .00 .00 .01
R2 .14 .01 .02
F 1.44 1.19 2.65
Note. † = p < .10, * = p < .05, ** = p < .01, *** = p < .001. N = 300 for likes ratio, shares ratio, comments ratio; N = 540 for valence of the comments.
Then, we performed four multiple regressions with independent self-construal and values, IDV and the interaction effect as independent variables. The multiple regression with likes ratio as dependent variable is non-significant F (3, 295) = 1.81, p = .145. Thus,
hypothesis 4.a is not supported.
The multiple regression with independent self-construal and values, IDV and the interaction effect as independent variables and shares ratio as dependent variable, shows that the model is significant, F (3, 295) = 4.38, p = .005. The interaction effect of independent self-construal and values and IDV is significant, b = 0,00031, b* = 0.27, t = 2.101, p = .037. The model predicts 4.3 per cent of the variation in the comments ratio (R2 = .043). Hypothesis 3.b is supported. The shares ratio increases for posts expressing more independent
self-construal and values when the level of individualism of the country increases.
Also, the multiple regression with independent self-construal and values, IDV and the interaction effect as independent variables and comments ratio as dependent variable, shows that the model is significant, F (3, 295) = 2.91, p = .035. The interaction effect of independent self-construal and values and IDV is marginally significant, b = 0,00014, b* = 0.21, t = 1.651, p = .100. The model predicts 3 per cent of the variation in the comments ratio (R2 = .03). Hypothesis 4.c is supported as the comments ratio increases for posts that express more independent self-construal and values when the level of individualism of the country increases.
Next, the multiple regression with valence of the comments as dependent variable was tested. The model is non-significant F (3, 536) = 0.97, p = .977. Thus, hypothesis 4.d is not supported. Results of the multiple regressions are shown in Table 5.
Table 5
Regression Table with Independent Self-construal and Values, IDV and the Interaction Effect as Predictors
Likes ratio Shares ratio Comments ratio
Valence of the comments Variable b SE B b* b SE B b* b SE B b* b SE B b* Constant .04 .04 .02 .01 .01 .06 .56 .25 IDV .00 .00 .01 .00 .00 -.08 .00 .00 -.12 .00 .00 .04 Independent self-construal & values .03 .04 .09 -.00 .01 -.06 -.00 -.01 -.04 .07 .19 .04 IDV* Independent .00 .00 .05 .00 .00 .27* .00 .00 .21† -00 00 -.43 R2 .008 0.4 .03 .00 F 1.81 4.38 2.91 .97
Note. † = p < .10, * = p < .05, ** = p < .01, *** = p < .001. N = 300 for likes ratio, shares ratio, comments ratio; N = 540 for valence of the comments.
Lastly, to test Hypothesis 5.a we run a multiple regression with interdependent self-construal and values, COL and the interaction effect as independent variables and likes ratio as dependent variable which resulted to be non-significant F(3, 295) = 1.04, p = .376, as well as the one with share ratio as dependent variable, F(3, 295) = 1.63, p = .183. Both
hypotheses 5.a and 5.b are not supported.
The multiple regression with interdependent self-construal and value, COL and the interaction effect as independent variables and comments ratio as dependent variable, shows that the model is significant, F (3, 295) = 3.74, p = .012. Interdependent self-construal and values, b = -0.01, b* = -0.25, t = -2.354, p = .019, COL, b = -0,00011, b* = -0.12, t = -1.70, p = .090 and the interaction effect of collectivism and COL, b = 0,00022, b* = 0.39, t = 3.25, p = .001, predict 3.7 per cent of the variation in the comments ratio (R2 = .037). Hypothesis 5.c is supported: the comments ratio increases for posts that express more interdependent self-construal and values when the level of collectivism of a country increases.
Finally, we run a multiple regression with valence of the comments as dependent variable. The model is non-significant, F (3, 536) = 0.08, p = .097, hence hypothesis 5.d is not supported. Results of the multiple regressions are shown in Table 6.
Table 6
Regression Table with Interdependent Self-construal and Values, COL and the Interaction Effect as Predictors
Likes ratio Shares ratio Comments ratio
Valence of the comments Variable b SE B b* b SE B b* b SE B b* b SE B b*
Constant .1 .02 .02 .00 .012 .03 .71 .14 COL .00 .00 .20 .00 .00 -.13† -.00 .00 -.12† -.00 .00 -.01 Interdependent self-construal & values -.02 .02 -.12 -.01 .00 -.21† -.01 .00 -0.25* -.05 .11 -.04 COL* Interdependent .00 .00 .19 .00 .00 .25* .00 .00 .39*** .00 .00 .02 R2 .01 .01 .37 .00 F 1.04 1.63 3.74 .08
Note. † = p < .10, * = p < .05, ** = p < .01, *** = p < .001. N = 300 for likes ratio, shares ratio, comments ratio; N = 540 for valence of the comments.
In the next section, some conclusions will be drawn. We will also discuss the practical implications of the results, the limitations of this study and we will provide suggestions for future research.
Conclusion and discussion
This research aimed at understanding the extent to which the cultural context influences how consumers react to social media messages. We examined the texts and the images of 300 posts published on Facebook by startups operating in three countries that score differently in Hofstede (2001) dimension of individualism and collectivism, US, Italy and Chile, and analysed the endorsement these posts received.
Even though not all the hypotheses we proposed were confirmed, overall the results validated our expectations: culture congruence between the audience and social media
content published by a startups influences the extent to which users endorse that content. The results of our study will be discussed below.
The first key finding refers to the first cultural element analysed in this study,
individualism/collectivism. Contrary to our expectation, the more individualistic the content is when the level of individualism of a country increases, it receives no higher endorsement. Conversely, it is confirmed that the endorsement does increase when the level of collectivism of the content increases in more collectivistic countries. To understand why
individualism/collectivism differently influence endorsement, it could be argued that as one of the main functions of social media is to foster relationships (Kietzmann, Hermkens, McCarthy, & Silvestre, 2011), then users would prefer to endorse content that values interpersonal relations, even though they belong to an individualistic culture.
The second key finding refers to the comments ratio. Validating our expectations, the results indicate that the comments ratio increases when the content is congruent with the level of collectivism, the self-construal and values and the communication style of the country where the content was published. A correlation logic (Nasim, Ilyas, Rextin, & Nasim, 2013) could explain our findings: a user could be more willing to write a comment when the post’s content is culturally close to his/her own culture. Furthermore, to write a comment, users need to make a cognitive effort, thus to be creative and active (Kim & Yang, 2016). This result confirms Kim and Yang (2016) findings, according to whom users are more likely to make cognitive efforts when the content is congruent with their cultural background.
Another key finding indicates that, unlikely hypothesized, the likes ratio does not increase when the content of a Facebook message is congruent with the cultural background of the audience. This finding extends what was previously known about this behaviour. In accordance to van Dijck (2013) and Kim and Yang (2016), liking is an intuitive and immediate behaviour that requires little efforts from users. For this reason, users can easily
like content even though it does not reflect their self, whereas endorsing a content by commenting or sharing is more difficult when it is not congruent with the own identity. Literature suggests that sharing and commenting are similar behaviours when are cognitive reactions to a message (Kim & Yang, 2016). We could argue again that users are more motivated to make cognitive efforts when a message is congruent with their cultural identity. Additionally, we know that individuals use social media to self-express and to promote themselves to others (van Dijck, 2013). When users comment a content on Facebook, this will visible on their network’s news feed while when they share a content, this will also be visible on their personal Timeline. As individuals tend to carefully evaluate how to publicly present to others on social media (Rui & Stefanone, 2013; van Dijck, 2013), then we can assume that they will be more likely to present a self that is in line with the cultural background of their group by sharing or commenting culturally congruent content.
Interestingly, we also saw how the valence of the comments could not be predicted by any of the elements of a culture considered here, with no distinction between individualistic and collectivistic countries. This can be justified considering that Zhang, Weare, Koh and Chen (2016) found that when commenting on social media, audience can leave aside their cultural values. Future research needs to obtain additional insights on this topic.
In general, the results show that not only the different cultural dimensions influence differently the actions users take to endorse a content, but even the two sides of the same cultural element, for instance low and high context communication, can influence different behaviours. This can be better understood when considering that even though members of a society share the same values or communication style, these are only predominant in their culture: individualistic or collectivist cues, independent or interdependent self-construal and values, a high or low communication style are present in any society to different extents
(Gudykunst, 1997; Gudykunst et al., 1996). Therefore, we could argue that it is legit to find some inherent inconsistencies in the way culture influences individuals’ communication.
Lastly, we saw that the models we proposed could predict only a small percentage of the variance in the dependent variables studied here. That is because in this research we specifically focused on cultural aspects while many other factors influence how content is framed and the extent to which users endorse it such as the creative strategy applied (Ashley and Tuten, 2015) or the content type (Pletikosa Cvijikj & Michahelles, 2013). Nevertheless, cultural aspects did result to be significant in driving endorsement as previously found in literature (Tsai & Men, 2017) and should therefore be considered. In fact, our findings extend academic knowledge about cultural antecedents of endorsement and contribute to the scarce literature about startups.
Managerial implications
The results of this study provide some practical recommendations to startups’ communication and marketing practitioners, startup founders or members in general. We found that an adaptation strategy is particularly relevant to increase the comments ratio on a brand Facebook page. Specifically, our findings suggest that startups practitioners of
individualistic countries should prioritize the expression of independent self-construal and values in their communication. This would increase both the shares and the comments ratio. Practitioners of collectivistic countries should emphasize every cultural aspect reflecting their own culture: we saw how expressing more collectivism, interdependent self-construal and values and adopting a high-communication style, all increase the comments ratio. Facebook values differently the actions users can take: sharing is valued more than commenting, commenting is valued more than liking. Consequently, thanks to the way the Facebook
algorithm works, by increasing the comments ratio practitioners can boost the reach of their messages, increase the brand visibility and support the startup’s growth.
Limitation and suggestions for future research
This study has some limitations we need to acknowledge. First, due to the disagreement in literature about the definition of startup, the findings could not be fully generalizable. We invite scholars to agree on a definition of startup by developing Luger and Koo (2014) research. This would bring to more accurate studies on the phenomenon of startups which would better help practitioners.
Second, the results of this research are limited to Facebook posts. It would be interesting to make a comparison between different social networks and test whether the same results we obtained apply to different platforms.
Third, it must be pointed out that analysing a culture by analysing national states is a simplification needed to perform the research as already previously done (Jin, 2010; Okazaki & Rivas, 2002; Steenkamp, 2001; Würtz, 2005). We are aware that cultures do not have national boundaries nor that citizens of a country are culturally homogenous. To implement this study, we used three countries that score differently on Hofstede’s dimension of
individualism/collectivism to assure different cultures were included. Future research could consider a higher number of countries including both western and eastern cultures. In
addition to that, our findings can be developed by studying the other five Hofstede’s cultural dimensions.
Lastly, this study might be limited as some scholars criticize Hofstede’s framework (Baack & Singh, 2007) and Hall’s work (Cardon, 2008), the former to be data-based and not theory based, the latter to focus on groups rather than on cultures. Future research could also
consider Schwartz’s cultural values as it has been found that they could better predict differences between countries (Imm Ng, Anne Lee, & Soutar, 2007).
To conclude, it is important for scholars to deepen into the understanding of how cultures drives engagement to support practitioners in implementing an effective
communication strategy. The number of startups funded is growing every day and only those which are successful in receiving positive endorsement from their consumers will be able to develop a strong brand and stay in the market.
References
Arnould, E. J., & Thompson, C. J. (2005). Consumer culture theory (CCT). Twenty years of research. Journal of Consumer Research, 31(4), 868–882.
https://doi.org/10.1086/426626
Arnould, E.J. & Price, L.L. (2000). Authenticating acts and authoritative performances: questing for self and community. In Ratneshwar, S., Mick, D.G. and Huffman, C. (Eds), The why of consumption: contemporary perspectives on consumer motives, goals, and desires (pp. 140-163). London, UK: Routledge
Ashley, C., & Tuten, T. (2015). Creative strategies in social media marketing: An exploratory study of branded social content and consumer engagement. Psychology & Marketing, 32(1), 15-27. https://doi.org/10.1002/mar.20761
Baack, D. W., & Singh, N. (2007). Culture and web communications. Journal of Business Research, 60(3), 181–188. https://doi.org/10.1016/j.jbusres.2006.11.002
Barnes, N. G., & Jacobsen, S. L. (2014). Missed eWOM opportunities: A cross-sector analysis of online monitoring behavior. Journal of Marketing Communications. Taylor & Francis. https://doi.org/10.1080/13527266.2013.797788
Barry, W. a., & Bouvier, G. (2012). Cross-cultural Communication: Arab and Welsh
students’ use of Facebook. Journal of Arab & Muslim Media Research, 4(2), 165–184. https://doi.org/10.1386/jammr.4.2-3.165
Bernritter, S. F., Verlegh, P. W. J., & Smit, E. G. (2016). Consumers’ online brand endorsements. Advertising in New Formats and Media, 189–209.
https://doi.org/10.1108/978-1-78560-313-620151009
Beverland, M., Dobele, A., & Farrelly, F. (2015). The viral marketing metaphor explored through Vegemite. Marketing Intelligence & Planning, 33(5), 656–674.
Bresciani, S., & Eppler, M. J. (2010). Brand new ventures? Insights on start-ups’ branding practices. Journal of Product & Brand Management, 19(5), 356–366.
https://doi.org/10.1108/10610421011068595
Bryman, A. (2012). Social research methods. Oxford, UK: Oxford University Press. Brodie, J.R., Hollebeek, L., Juric, B. & Ilic, A. (2011a). Consumer engagement: conceptual
domain, fundamental propositions and implications for research. Journal of Service Research, 14(3), 252-271. DOI: 10.1177/1094670511411703
Cardon, P. W. (2008). A critique of Hall’s contexting model: A meta-analysis of literature on intercultural business and technical communication. Journal of Business and Technical Communication, 22(4), 399–428. https://doi.org/10.1177/1050651908320361
Cayla, J., & Arnould, E. J. (2008). A cultural approach to branding in the global marketplace. Journal of International Marketing, 16(4), 88–114. https://doi.org/10.1509/jimk.16.4.86 Cho, M., Schweickart, T., & Haase, A. (2014). Public engagement with nonprofit
organizations on Facebook. Public Relations Review, 40(3), 565–567. https://doi.org/10.1016/j.pubrev.2014.01.008
Chu, S.-C., & Choi, S. M. (2011). Electronic word-of-mouth in social networking sites: A cross-cultural study of the United States and China. Journal of Global Marketing, 24(3), 263–281. https://doi.org/10.1080/08911762.2011.592461
Craig, C. S., & Douglas, S. P. (2006). Beyond national culture: implications of cultural dynamics for consumer research. International Marketing Review, 23, 322–342. https://doi.org/10.1108/02651330610670479
Cvijikj, I. P., Spiegler, E. D., & Michahelles, F. (2011). The effect of post type, category and posting day on user interaction level on Facebook. Proceedings - 2011 IEEE
International Conference on Privacy, Security, Risk and Trust and IEEE International Conference on Social Computing, PASSAT/SocialCom 2011, 810–813.
https://doi.org/10.1109/PASSAT/SocialCom.2011.135
De Vries, L., Gensler, S., & Leeflang, P. S. H. (2012). Popularity of brand posts on brand fan pages: An investigation of the effects of social media marketing. Journal of Interactive Marketing, 26(2), 83–91. https://doi.org/10.1016/j.intmar.2012.01.003
Dhanesh, G. S. (2016). Putting engagement in its PRoper place: State of the field, definition and model of engagement in public relations. Public Relations Review, (September 2016), 1–9. https://doi.org/10.1016/j.pubrev.2017.04.001
Dunne, Á., Lawlor, M.-A., & Rowley, J. (2010). Young people’s use of online social networking sites – a uses and gratifications perspective. Journal of Research in Interactive Marketing, 4(1), 46–58. https://doi.org/10.1108/17505931011033551 Executive summary of the new Italian legislation on innovative startups (2016, May 2).
Retrieved from
http://www.sviluppoeconomico.gov.it/images/stories/documenti/Executive_Summary_It aly_Startup_Act_02_05_2016.pdf
Gevorgyan, G., & Manucharova, N. (2009). Does culturally adapted online communication work? A study of american and chinese internet users’ attitudes and preferences toward culturally customized web design elements. Journal of Computer-Mediated
Communication, 14(2), 393–413. https://doi.org/10.1111/j.1083-6101.2009.01446.x Goh, K., Heng, C.-S., & Lin, Z. (2013). Social media brand community and consumer
behavior: quantifying the relative impact of user- and marketer- generated content. Information Systems Research, (24 August 2014), 88–107.
https://doi.org/10.1287/isre.1120.0469
Goodrich, K., & de Mooij, M. (2013). How “social” are social media? A cross-cultural comparison of online and offline purchase decision influences. Journal of Marketing Communications. Taylor & Francis. https://doi.org/10.1080/13527266.2013.797773
Gudykunst, W. (1997). Cultural variability in communication.Communication Research, 24, 327–348. https://doi.org/10.1177/009365097024004001
Gudykunst, W., Matsumoto, Y., TingToomey, S., Nishida, T., Kim, K., & Heyman, S. (1996). The influence of cultural individualism-collectivism, self construals, and individual values on communication styles across cultures. Human Communication Research, 22(4), 510–543. https://doi.org/10.1111/j.1468-2958.1996.tb00377.x Hall, E. T. (1976). Beyond Culture. New York, NY: Anchor Press.
Hansen, N., Postmes, T., Van Der Vinne, N., & Van Thiel, W. (2012). Information and communication technology and cultural change: How ICT changes self-construal and values. Social Psychology, 43(4), 222–231. https://doi.org/10.1027/1864-9335/a000123 Harris, L., & Rae, A. (2009). Social networks: the future of marketing for small business.
Journal of Business Strategy, 30(5), 24–31. https://doi.org/10.1108/02756660910987581 Hoffmann, J., Röttger, U., Ingenhoff, D., & Hamidati, A. (2015). The rehabilitation of the
“nation variable”. Corporate Communications: An International Journal, 20(4), 483– 499. https://doi.org/10.1108/CCIJ-10-2014-0071
Hofstede, G. (1980). Culture and organizations. International Studies of Management & Organization, 10(4), 15-41. Beverly Hills, CA: Sage.
Hofstede, G. H. (2015b). Country Comparison. Retrieved from https://geert-hofstede.com/countries.html
Hollenbeck, C. R., & Kaikati, A. M. (2012). Consumers’ use of brands to reflect their actual and ideal selves on Facebook. International Journal of Research in Marketing, 29(4), 395–405. https://doi.org/10.1016/j.ijresmar.2012.06.002
Imm Ng, S., Anne Lee, J., & Soutar, G. N. (2007). Are Hofstede’s and Schwartz’s value frameworks congruent? International Marketing Review, 24(2), 164–180.
Jin, C.-H. (2010). An empirical comparison of online advertising in four countries: Cultural characteristics and creative strategies. Journal of Targeting, Measurement and Analysis for Marketing, 18(3–4), 253–261. https://doi.org/10.1057/jt.2010.18
Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of Social Media. Business Horizons, 53(1), 59–68.
https://doi.org/10.1016/j.bushor.2009.09.003
Kietzmann, J. H., Hermkens, K., McCarthy, I. P., & Silvestre, B. S. (2011). Social media? Get serious! Understanding the functional building blocks of social media. Business Horizons, 54(3), 241–251. https://doi.org/10.1016/j.bushor.2011.01.005
Kim, C., & Yang, S. U. (2016). Like, comment, and share on Facebook: How each behavior differs from the other. Public Relations Review, 43(2), 441–449.
https://doi.org/10.1016/j.pubrev.2017.02.006
Lam, D., & Lee, A. (2009). The effects of cultural values in word-of-mouth communication. Journal of International Marketing, 17(3), 55-70. https://doi.org/10.1509/jimk.17.3.55 Luger, M., & Koo, J. (2014). Defining and Tracking Business Start-UpS. Small Business
Economics, 24(1), 17–28.
Luna, D., & Gupta, S. F. (2001). An integrative framework for cross-cultural consumer behavior. International Marketing Review (Vol. 18).
https://doi.org/10.1108/02651330110381998
Men, L. R., & Tsai, W.-H. S. (2014). Perceptual, attitudinal, and behavioral outcomes of organization–public engagement on corporate social networking sites. Journal of Public Relations Research, 26(5), 417–435. https://doi.org/10.1080/1062726X.2014.951047 Men, L. R., & Tsai, W. H. S. (2012). How companies cultivate relationships with publics on
social network sites: Evidence from China and the United States. Public Relations Review, 38(5), 723–730. https://doi.org/10.1016/j.pubrev.2011.10.006
Möller, J., Eisend, M., & Ller, J. M. (2010). A Global investigation into the cultural and individual antecedents of banner advertising. Source Journal of International Marketing, 18(2), 80–98.
Mooji, M. de. (2010). Consumer behaviour and culture: consequences for global marketing and advertising. Thousand Oaks, CA: Sage.
Mooij, M. de. (2015). Cross-cultural research in international marketing: clearing up some of the confusion. International Marketing Review, 32(6), 646–662.
https://doi.org/10.1108/EL-01-2014-0022
Muntinga, D. G., Moorman, M., & Smit, E. G. (2017). Introducing COBRAs. International Journal of Advertising, 30(1), 13–46. https://doi.org/10.2501/IJA-30-1-013-046
Nacar, R., & Burnaz, S. (2011). A cultural content analysis of multinational companies’ web sites. Qualitative Market Research: An International Journal, 14(3), 274–288.
https://doi.org/10.1108/13522751111137505
Nasim, M., Ilyas, M. U., Rextin, A., & Nasim, N. (2013). On commenting behavior of Facebook users. Proceedings of the 24th ACM Conference on Hypertext and Social Media, 179–183. https://doi.org/10.1145/2481492.2481513
Neri, F., Aliprandi, C., Capeci, F., Cuadros, M., & By, T. (2012). Sentiment analysis on social media. Proceedings of the 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2012, 919–926. https://doi.org/10.1109/ASONAM.2012.164
Okazaki, S., & Rivas, J. A. (2002). A content analysis of multinationals’ web communication strategies: cross-cultural research framework and pre-testing. Internet Research, 12(5), 380–390. https://doi.org/10.1108/10662240210447137
Okazaki, S., & Taylor, C. R. (2013). Social media and international advertising: theoretical challenges and future directions. International Marketing Review, 30(1), 56–71.
Özsomer, A., & Simonin, B. L. (2004). Marketing program standardization: A cross-country exploration. International Journal of Research in Marketing, 21(4), 397–419.
https://doi.org/10.1016/j.ijresmar.2004.06.003
Park, C., Jun, J., & Lee, T. (2015). Consumer characteristics and the use of social networking sites. International Marketing Review, 32(3/4), 414–437. https://doi.org/10.1108/IMR-09-2013-0213
Pereira, H. G., de Fátima Salgueiro, M., & Mateus, I. (2014). Say yes to Facebook and get your customers involved! Relationships in a world of social networks. Business Horizons, 57(6), 695–702. https://doi.org/10.1016/j.bushor.2014.07.001
Pineda, A., Hernández-Santaolalla, V., & Rubio-Hernánde, M. (2015). Individualism in Western advertising : A comparative study of Spanish and US newspaper
advertisements. European Journal of Communication, 1–17. https://doi.org/10.1177/0267323115586722
Pletikosa Cvijikj, I., & Michahelles, F. (2013). Online engagement factors on Facebook brand pages. Social Network Analysis and Mining, 3(4), 843–861.
https://doi.org/10.1007/s13278-013-0098-8
Rieder, B. (2013). Studying Facebook via data extraction. Proceedings of the 5th Annual ACM Web Science Conference on - WebSci ’13, 346–355.
https://doi.org/10.1145/2464464.2464475
Rode, V., & Vallaster, C. (2005). Corporate branding for start-ups: The crucial role of entrepreneurs. Corporate Reputation Review, 8(2), 121–135.
https://doi.org/10.1057/palgrave.crr.1540244
Rui, J. R., & Stefanone, M. A. (2013). Strategic image management online. Information, Communication & Society, 16(8), 1286–1305.
Singelis, T. M. (1994). The Measurement of independent and interdependent self-construals. Personality and Social Psychology Bulletin. https://doi.org/10.1177/0146167294205014 Singh, N., & Matsuo, H. (2004). Measuring cultural adaptation on the Web: A content
analytic study of U.S. and Japanese Web sites. Journal of Business Research, 57(8), 864–872. https://doi.org/10.1016/S0148-2963(02)00482-4
Sinkovics, R., Yamin, M., & Hossinger, M. (2007). Cultural adaptation in cross border e-commerce: a study of German companies. Journal of Electronic Commerce Research, 8(4), 221–235.
Soares, A. M., Farhangmehr, M., & Shoham, A. (2007). Hofstede’s dimensions of culture in international marketing studies. Journal of Business Research, 60(3), 277–284.
https://doi.org/10.1016/j.jbusres.2006.10.018
Steenkamp, J.-B. E. M. (2001). The role of national culture in international marketing research. International Marketing Review, 18(1), 30–44.
https://doi.org/10.1108/02651330110381970
Sunny Tsai, W.-H., & Rita Men, L. (2013). Motivations and antecedents of consumer engagement with brand pages on social networking sites. Journal of Interactive Advertising, 13(2), 76–87. https://doi.org/10.1080/15252019.2013.826549
Thelwall, M., Buckley, K., Paltoglou, G., & Cai, D. (2010). Measuring use and creation of open educational resources in higher education. Journal of the American Society for Information Science and Technology, 61(12), 2544–2558. https://doi.org/10.1002/asi Timmons, J.A. (1999). New Ventures Creation. Boston, MA: McGraw Hill.
Triandis, H. C. (2004). The many dimensions of culture. Academy of Management Executive, 18(1), 88–93. https://doi.org/10.5465/AME.2004.12689599
Tsai, W.-H. S., & Men, L. R. (2017). Consumer engagement with brands on social network sites: A cross-cultural comparison of China and the USA. Journal of Marketing
Communications, 23(1), 2–21. https://doi.org/10.1080/13527266.2014.942678
van Dijck, J. (2013). You have one identity: performing the self on Facebook and LinkedIn. Media, Culture & Society, 35(2), 199–215. https://doi.org/10.1177/0163443712468605 Vyncke, F., & Brengman, M. (2010). Are culturally congruent websites more effective? An
overview of a decade of empirical evidence. Journal of Electronic Commerce Research, 11(1), 14–29.
Walsh, G., Shiu, E., & Hassan, L. M. (2014). Cross-national advertising and behavioral intentions: a multilevel analysis, 22(1), 77–97.
Würtz, E. (2005). Intercultural communication on web sites: a cross-cultural analysis of web sites from high-context cultures and low-context cultures. Journal of
Computer-Mediated Communication, 11(1), 274–299. https://doi.org/10.1111/j.1083-6101.2006.00013.x
Zhang, Y., Weare, A.M., Koh, H., & Chen, L. (2016) Cultural trends of audience online interaction with vocal talent shows: a comparative study between China and the US. Journal of Intercultural Communication Research, 45(3), 196-213.
Appendix A
List of the startups included in the sample.
Country Name Sector
Chile VoyHoy Travel/logistic Te Queja Suma Other Edoome Education
Eggless Food and beverage
First Job Other
Duko Other Wisboo Education Babytuto Retail Linguician Education Vendiste.com Retail Italy
Cortilia Food and beverage
Le Cicogne Other
Gnammo Food and beverage
Bauzaar.it Retail
Teeser Retail
Pinktrotters Travel/logistic
Faberest Food and beverage
Satispay Finance
Sgnam Food and beverage
Wanderio Travel/logistic
USA
uBiome Biotech
Shyp Travel/logistic
Forkable Food and beverage
DoorDash Food and beverage
Glose App Education
Vayable Travel/logistic
Bento Food and beverage
Classpass Other
Scout Alarm Retail
Appendix B Codebook
Q1
Startup code = read the Excel file. Example: CH1 Q2 Country Chile Italy USA Q3
Sector = read the Excel file Q4
Post code = Insert the startup code + the post number as in the Jupyter notebook. Example: CH1_00 (for the first post of the CH1 startup, CH1_01 for the second one) Q5
Post ID as in the Jupyter notebook. Q6
Number of likes the Facebook page has = see the third row on the right “sum” in the followers df in the Jupyter notebook
Q7
Number of likes the post has = see “like” value in the Jupyter notebook. Q8
Number of shares the post has = see “share” value in the Jupyter notebook. Q9
