JID: AMJ

ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

Australasian Marketing Journal xxx (xxxx) xxx

Contents lists available at ScienceDirect

Australasian Marketing Journal

journal homepage: www.elsevier.com/locate/ausmj

What do people “like” on Facebook? Content marketing strategies

used by retail bank brands in Australia and Singapore
Pengji Wang a,1,∗, Breda McCarthy b,1
a
College of Law, Business and Governance, James Cook University, Singapore Campus, Singapore
b
Department of Economics and Marketing, James Cook University (JCU), Townsville, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Previous research has yet to examine the effectiveness of social media content strategies in the finan-
Received 26 July 2019 cial services sector. This study advances knowledge of content strategies used by retail bank brands
Revised 23 April 2020
by examining how the type of message, the post format and the source of the content influence cus-
Accepted 28 April 2020
tomer engagement independently and interactively. A total of 541 posts from the Facebook pages of ten
Available online xxx
banks in Australia and Singapore are analysed using multiple regressions. Results suggest that in the
Keywords: Singaporean sample, persuasive-only content and mixed content lead to more likes and emoji responses,
Digital content marketing whereas informative-only content is effective in generating comments and questions. Furthermore, videos
Facebook and third-party sources strengthen the relationship between informative-only content and customer en-
Customer engagement gagement. In the Australian sample, informative-only content and mixed content are more effective than
Retail bank brands persuasive content in engaging customers. The findings support an integrated view of content design
and highlight the importance of adjusting the content strategy to suit the utilitarian nature of banking
services and engage high-involvement customers.
© 2020 Australian and New Zealand Marketing Academy. Published by Elsevier Ltd. All rights reserved.

1. Introduction suasive and informative content positively influence passive en-

gagement (likes), but neither of them influences active engagement
The term ‘content marketing’ has been used by practitioners for (comments). The visual format of content is reported to have either
many years (Pulizzi and Barrett, 2010). In fact, the phrase ‘content a positive effect on customer engagement (de Vries et al., 2012) or
is king’ is one of the most popular slogans in the digital economy no relationship with it (Lee et al., 2018). These mixed findings may
(Rancati et al., 2016). Despite general agreement that customer en- be attributed to the fact that one single dimension, when inter-
gagement is an increasingly important concept in modern market- acting with another, may lead to different results. However, very
ing (Abdul-Ghani et al., 2019), there is a lack of understanding of little research, if any, has taken an integrated view to examine the
the effect of content on the engagement behaviours of social me- interactive effect of multiple dimensions on customer engagement.
dia users (Ashley and Tuten, 2015). In addition, most existing studies on digital content market-
Informed by persuasive communication theories, numerous ing have investigated content marketing across hedonic sectors,
studies have examined how the various dimensions, such as type, such as wine (Dolan et al., 2019) and fashion (Gamboa and
format and source of content that marketers use to craft and de- Goncalves, 2014), as well as the content strategies adopted by lead-
liver online content separately influence the effectiveness of digital ing brands in the food, car manufacturing, retailing and hospitality
content marketing, yet scholars report mixed findings about the ef- sectors (Villarroel Ordenes et al., 2018). Customers purchase hedo-
fect of each dimension. For example, categorizing the type of con- nic brands largely for the affective gratification and hence are more
tent as either informative/rational (directly shifting beliefs about likely to engage with a company’s social media content if it offers
a product’s existence or prices) or persuasive/emotional (shifting entertainment and social bonding benefits. Unlike hedonic brands,
preferences in general), Lee et al. (2018) show that persuasive con- utilitarian brands in sectors such as banking and petrol are devoid
tent, rather than informative content, is the key to effective en- of meaning and emotion (Hollebeek, 2013). Prior studies posit that
gagement, while Dolan et al. (2019) give evidence that both per- the inherent characteristics of products or services – for example,
whether they fulfil a utilitarian or hedonic need – have implica-
∗
tions for the effectiveness of social media marketing (Hollebeek,
Corresponding author.
2013; Schulze et al., 2014). Applying the findings obtained from
E-mail address: [email protected] (P. Wang).
1
Both authors contributed equally to this work. predominantly hedonic sectors to utilitarian sectors without ad-

https://doi.org/10.1016/j.ausmj.2020.04.008
1441-3582/© 2020 Australian and New Zealand Marketing Academy. Published by Elsevier Ltd. All rights reserved.

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

2 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

justments and/or empirical validation is not constructive. However, marketing is to engage customers online; (2) content marketing is
few studies have investigated digital content marketing in utilitar- about developing and distributing information; and (3) the infor-
ian sectors. mation must be of relevance and deliver value to customers.
To address the theoretical lacunae in the content marketing lit- Scholars and practitioners are starting to take note of the ben-
erature, this study analyses how message type (informative, per- efits that content marketing can bring to both customers and
suasive or a mix of both), format (photo, video and text) and companies. Given the wide reach of digital channels at rela-
source (company-derived or from a third party) independently and tively low cost, content marketing represents a significant oppor-
interactively influence online customer engagement in a utilitarian tunity for brands to offer added value at reduced marketing costs
sector, namely, retail banking services. The study draws on the lit- (Malthouse et al., 2013). Engaging content assists in search engine
erature on persuasive message strategy as well as on theories re- optimisation, thus helping marketers grow their earned audience
lating to persuasive communications. (Hollebeek and Macky, 2019). However, in practice, studies have
Banking services, often perceived as complex and utilitarian noted the inability of large corporate brands to build large fan
rather than hedonic, tend to be ‘high-involvement’ services and bases (Hodis et al., 2015; Sitta et al., 2018) and a low level of online
require more cognitive effort from consumers (Aldlaigan and But- customer engagement (Rissanen and Luoma-Aho, 2016). Hollebeek
tle, 20 01; Berger and Messerschmidt, 20 09). Furthermore, due to and Macky (2019) state that significant practitioner interest in the
the erosion of the public’s trust in financial services since the concept of content marketing exists, yet “academic digital content
global financial crisis of 20 07–20 08 (Mogaji and Danbury, 2017) marketing research is lagging, resulting in an important knowledge
and the increase in fraudulent online activities, such as hoaxes gap” (p. 27).
and scams (Australian Competition & Consumer Commission, n.d), Existing literature on content marketing comes from both
banks are increasingly using social media to build trust and deepen practitioners’ literature and scholarly work. Practitioner-orientated
relationships with their customers (American Bankers Association guidelines include adopting a more informal style of communi-
[ABA], 2017). For example, many banks create content to establish cation (Charello, 2017), providing valuable information to address
‘thought leadership’ (e.g. releasing commentary on trends in finan- customers’ needs (Peppers and Rogers, 2011; Pulizzi and Barrett,
cial markets) or engage with the community (e.g. running an on- 2010), inviting comments, developing polls for interaction (Pulizzi
line financial literacy programme for high school students). How- and Barrett, 2010), using tangible incentives (Ashey and Tuton,
ever, empirical evidence is less encouraging regarding banks’ mar- 2014; Heller, Baird, & Parasnis, 2011) and optimising the allocation
keting efforts on social media (Rinaldi, 2015; Vejačka, 2017). More- of customer-orientated and company-orientated content (Charello,
over, scholarly research on banks’ social media marketing is still 2017). These guidelines, while insightful, tend to be fragmented
limited and tends to focus on the motivators of, and barriers to, and lack solid empirical evidence to demonstrate their effective-
engaging in social media marketing (Mitic and Kapoulas, 2012), the ness.
effect of social media marketing on customer loyalty (Laksamana, Scholarly work falls into three strands (for more details of the
2018) and effective segmentation strategies for social media mar- literature, see Appendix 1). First, some conceptual studies highlight
keting (Durkin et al., 2015). Little guidance on the optimal design the unique characteristics of digital content marketing. Unlike tra-
of digital content marketing strategies in the financial services sec- ditional advertising, digital content marketing relies on the princi-
tor exists in the scholarly literature. Therefore, it will be interesting ple of consumer consent or opt-in, or is permission-based (Godin,
to investigate the extent to which the most commonly used digital 1999). To ensure that consumers engage voluntarily, the content
marketing content tactics are effective in such a context. must be diverse in terms of information and format (Chaffey &
Empirically, this study examines the Facebook content market- Chadwick, 2019; Harris, 2017), meaningful (Deighton and Kornfeld,
ing practices of ten banks in Australia and Singapore. The scope of 2009), interactive (Calder et al., 2009) and framed in a way that
the study is limited to Facebook because it has the greatest share facilitates online conversations (de Ruyer et al., 2018).
of all social media networks in the two countries (StatCounter, The second strand of scholarship explores the mechanism
2019a; 2019b). Also, Facebook is the most frequently examined so- through which content marketing influences the customer’s cogni-
cial media platform in the existing literature (Alalwan et al., 2017). tive and emotional response and engagement, by drawing on well-
The two countries selected for study are similar in terms of high established persuasive models such as the elaboration likelihood
social media usage (Hashmeta, 2018; Sensis, 2018) while different model (ELM) (Chang et al., 2015), uses and gratification theory
in culture, thus providing insights on international differences in (UGT) (Bu et al., 2020; Calder et al., 2009) and social impact theory
content marketing strategies and customer reactions. (Bu et al., 2020). These studies mainly used survey data to mea-
In Section 2 we provide a critical review of the literature on sure the theoretical constructs. For example, drawing on the ELM,
digital content marketing, while the theoretical foundations and Chang et al. (2015) asked respondents to evaluate the constructs
research hypotheses are outlined in Section 3. Section 4 introduces such as argument quality, post popularity and post attractiveness
the empirical context; Section 5 describes the sample, data, cod- of a cooking company’s Facebook posts, and linked the constructs
ing and analytical approach; Section 6 reports the results; and with the number of likes and shares through respondents’ percep-
Section 7 presents a discussion of the results. Section 8 reviews tion of usefulness and preference.
the theoretical and practical implications of the research findings. The third strand of research focuses on persuasive message
Section 9 closes the paper by identifying the study’s limitations strategies. Studies within this set typically conduct content analysis
and suggesting directions for future research. based on companies’ digital activities on some platforms, such as
Facebook (e.g., Dolan et al., 2019; Villarroel Ordenes et al., 2018),
2. Literature review and research questions Twitter (Veale et al., 2015) and Instagram (Klassen et al., 2018).
Online activities or messages are coded in different ways to high-
2.1. Content marketing in practice and in the scholarly arena light the different nature or dimensions of the messages, drawing
on theories such as UGT and dual processing theory (Dolan et al.,
Content marketing, also termed ‘digital content marketing’ 2019), self-presentation theory (Brech et al., 2017) and advertising
(KoisoKanttila, 2004; Rowley, 2010; Hollebeek and Macky, 2019) or message theory (Lee et al., 2018). For example, Dolan et al. (2019),
‘content engineering’ (Lee et al., 2018), has been defined in differ- based on UGT, categorize the content of Facebook posts into ratio-
ent ways by scholars (see Appendix 1 for definitions), but all def- nal message appeals (informational content and remunerative con-
initions highlight three points: (1) the purpose of digital content tent) and emotional message appeals (including entertaining con-

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 3

tent and relational content), and test their relationship with Face- bate on the extent to which social media marketing practices are
book metrics such as likes and comments. Lee et al. (2018) use a universal.
similar categorization based on the advertising content literature.
Our study falls into, and brings new insights to, the third strand 3. Hypotheses development
of research in at least three ways. First, besides testing the ef-
fectiveness of different types of content following existing studies This study draws on well-established theories relating to per-
(Lee et al., 2018; Dolan et al., 2019), we include additional dimen- suasive message strategy to explain the effectiveness of digital con-
sions highlighted by persuasive communication theory that could tent in generating customer engagement. A persuasive message
affect the persuasiveness of a message, and investigate whether strategy involves the use of verbal and non-verbal messages to
and how these other dimensions (format of the post and source of influence attitudes and behaviour and, as with interpretive tech-
the information) and content type interactively affect online cus- niques, the message may be designed to influence both affective
tomer engagement. and cognitive domains (Dainton, 2005). Rooted in social psychol-
Second, most of the existing studies focus on hedonic or high- ogy, theories on persuasion can be categorized into two groups.
identity sectors, such as the wine industry in Australia (Dolan The first group explains how the audience assesses and processes
et al., 2019), consumer brands, sports and entertainment (Lee et al., persuasive messages; it includes social judgement theory (Sherif
2018). Brands perceived as predominantly hedonic in nature are et al., 1965), ELM (Petty and Cacioppo, 1986), cognitive dissonance
largely evaluated on the basis of their specific sensory, experiential theory (Festinger, 1957) and narrative paradigm (Fisher, 1984).
and emotive aspects (Hirschmann and Holbrook, 1982; Dhar and The second group encompasses persuasive communication theo-
Wertenbroch, 20 0 0; Jones et al., 20 06). As such, hedonic brands ries, which examine how message senders craft and deliver mes-
may be used and purchased mainly for the affective gratifica- sages to achieve effective persuasion through communication. Our
tion consumers receive from the brand’s sensory attributes, includ- study is in line with the latter group. Persuasive communication
ing entertainment, escapism or relaxation (Overby and Lee, 2006). theories suggest that an intricate web of factors play a role in de-
Seeking connection, self-reinforcement and entertainment (Dahl, termining the effectiveness of a persuasive message (Ajzen, 1991).
2018; Hodis et al., 2015), customers of hedonic brands are more A few contextual factors that have been identified are the content
likely to engage with the social media content developed by a type, format and source credibility of the message.
company.
In contrast to existing studies, our study focuses on a utilitar- 3.1. Type of content
ian sector. Brands that are viewed as mainly utilitarian are as-
sessed primarily based on their perceived functional, instrumen- Persuasive communication theories suggest that the impact of
tal benefits, thus reflecting relatively non-emotive, ‘means-to-an- a message depends on the type of content it contains. Drawing
end’ decision-making processes (Babin et al., 1994; Overby and Lee, on UGT, scholars have identified two types of information sought
2006; Jones et al., 2006). As Hollebeek’s (2013) qualitative study by consumers: consumers acting on their functional motives seek
points out, utilitarian brands in the retail banking and petrol sec- informative content, while consumers driven by hedonic motives
tors are commonly viewed as devoid of meaning and emotion and seek entertainment content (Dahl, 2018; Hollebeek and Macky,
hence tend to generate less consumer engagement and consumer 2019). In a similar vein, the ELM (Petty and Cacioppo, 2012) pre-
value levels, compared to hedonic brands. Applying the findings dicts that consumers, when purchasing high-involvement prod-
obtained from predominantly hedonic sectors to the context of ucts, are motivated to seek a wide range of information related
retail banking without adjustments and/or empirical validation is to the product/service and follow the central route of decision-
risky. Therefore, our research aims to enrich the understanding of making (such as the strength of the arguments in the message). In
digital content, specifically, whether and how the commonly iden- low-involvement situations, consumers are not motivated or lack
tified features of persuasive messages shape the utilitarian sector’s the ability to process information and hence rely on peripheral
content marketing effectiveness. cues (such as subjective impressions) and non-factual or emotional
Third, our study investigates two countries, Australia and Sin- messages (i.e. those that rely on music, humour or sex appeal to
gapore. Both countries are digitally savvy, yet they differ in all of draw attention to the advertised products) to make purchase deci-
the six cultural dimensions when referring to the classic Hofstede’s sions.
model, i.e., power distance, individualism, masculinity, uncertainty In line with these studies, Anderson and Renault (2006), Lee
avoidance, long term orientation and indulgence (Hofstede et al., et al. (2018) and Dolan et al. (2019) identify two categories of
2010). Many researchers suggest that people act differently as per content that are present in both advertising and digital marketing
their respective cultural values, which have a profound impact on messages: (1) Informative content aiming to shift beliefs about a
marketing (de Mooij, 2017; House et al., 2004). For instance, na- product’s existence or prices directly. This type of content includes
tional cultural values influence service evaluations (Tsiotsou, 2019), information explicitly about the product/service/company. It facili-
co-creation outcomes (Grott et al., 2019) and online WOM (Chiu tates consumers’ understanding of the company and its offerings in
et al., 2019; Park and Jeon, 2018). As Park and Jeon (2018) con- order to encourage decision-making through the central route. (2)
cluded, consumers in East and West show different reactions to Persuasive content aiming to shift preferences in general. The in-
online reviews due to their different thinking styles. On the other formation typically provided in this type of content does not pro-
hand, globalization combined with the digitalization revolution has mote the company and its offerings directly, but seeks to create
resulted in the emergence of global consumers (see Merz et al., positive feelings and a more enduring relationship with customers,
2008, for a review and analysis), which minimizes the impact of thus influencing consumers’ decision-making by a peripheral
national cultures and facilitates cross-country segmentation and route.
targeting. One might expect customer engagement patterns across Following this categorization, a few studies examine the re-
Singapore and Australia to converge, due to the developed na- lationship between types of content and customer engagement,
ture of their economies and the trend towards the internation- but report inconsistent empirical evidence. For example, Lee et al.
alisation of the banking sector. However, inadequate attention is (2018) find that persuasive content has a positive impact on en-
given to social media tactics in a cross-cultural setting (Cummins gagement, while informative content has a negative effect on en-
et al., 2016; Nguyen et al., 2014). Our comparative study aims to gagement, unless it is combined with attributes of persuasive con-
provide some preliminary empirical evidence to inform the de- tent. Dolan et al. (2019) find that rational appeals in Facebook

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

4 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

posts facilitate likes and shares from the audience, whereas emo- senses (sight and hearing), and hence increase the vividness and
tional appeals facilitate likes only. De Vries et al. (2012) find a neg- attractiveness of the message, cutting through the noise from the
ative relationship between entertainment content and the number environment (de Vries et al., 2012). Messages leveraging multiple
of likes (probably because the content was unrelated to the brand), senses could also facilitate the audience’s learning and cognitive
but no relationship between the provision of information and the process. We thus hypothesize as follows:
number of likes.
Studies of the impact of different types of content in the re- Hypothesis 2a: Posts with photos and videos will lead to more
tail banking sector have yielded similarly contrasting findings. On customer engagement than posts with text only.
the one hand, there is evidence that banks use social media to Hypothesis 2b: All types of content (informative-only,
highlight their socially responsible side to gain legitimacy (Ozdora- persuasive-only and mixed content) presented with video
Aksak and Atakan-Duman, 2015). On the other hand, industry re- or photo will have a stronger positive relationship with
ports suggest that providing informational content on financial customer engagement than those presented with text only.
products and services can drive engagement (ABA, 2017). In sum-
3.3. Source of content
mary, it is unclear whether persuasive or informative approaches,
or a hybrid approach, can increase customer engagement.
Persuasive communication theories also assume that beliefs
The degree of customer engagement falls on a continuum rang-
and attitudes can be influenced by the perceived credibility of
ing from passive engagement (e.g. liking a page on Facebook) to
the message’s source (Hovland and Weiss, 1951; Petty et al.,
more active engagement characterised by customer participation in
1981). Perceived source credibility is defined as judgments made
co-creation activities (e.g. writing reviews) (Malthouse et al., 2013;
by a perceiver concerning the believability of a communicator
Muntinga et al., 2011). We argue that content with an informa-
(O’Keefe, 1990) or, in this case, a brand. Kelman (1961) and
tive component (informative-only as well as mixed content con-
McGuire (1969) suggest that the more credible the message source,
taining both informative and persuasive components) aims to trig-
the more likely the audience will be persuaded to change their at-
ger and facilitate customers’ cognitive processing and central route
titudes. Three factors can influence the perceived credibility of the
of decision-making, and hence will generate customer engagement
source of a message: trustworthiness (the perception that a person
in a more active form, such as asking a question or commenting on
will tell the truth if they know it), expertise/competence (the per-
a post, since it demands a greater cognitive effort from the user.
ception that a person knows the truth) and goodwill/caring (the
Posts with only persuasive content induce consumers’ decision-
perception that a person cares about the perceiver) (McCroskey
making through a peripheral route; without much cognitive effort
and Teven, 1999). Following this line of thinking, despite the banks’
being committed, customers are more likely to passively engage
high expertise in structuring and explaining their offerings, cus-
with a brand in an easier and quicker way, for example, by clicking
tomers may suspect that banks will present only favourable infor-
the ‘like’ icon or adding an emoji. We hence propose the following
mation about their offerings because they care more about prof-
as our first hypothesis:
itability than about customers. Compared with bank-produced con-
Hypothesis 1a: Persuasive-only content leads to more passive tent, third-party content may be perceived as more independent,
engagement (i.e. more likes and emoij) than informative- objective and credible because third parties have fewer conflicts
only content or mixed content. of interest. A growing number of researchers in the United States
Hypothesis 1b: Informative-only content and mixed content (Johnson and Kaye, 2004; Cheong and Morrison, 2008; Chu and
lead to more active engagement (i.e. more shares, com- Kamal, 2008) argue that information on the Internet created by
ments, questions) than persuasive-only content. third-party sources (such as user-generated content, third-party
websites, etc.) is more credible than that produced by companies.
As a result, customers will not be responsive to the bank-generated
3.2. Format of content
content than that generated by third party.
We argue that this credibility concern is more pertinent for in-
In regard to persuasion and attitudinal change, scholars pro-
formative content than persuasive content. In other words, cus-
pose that for a persuasive appeal to work, the message recipient
tomers rely on product-related information when choosing bank-
must pay attention to, comprehend, learn, accept and remember
ing products and look for more reliable and objective information
the message. However, a highly complex message might be too
to aid their evaluations. This leads to the third set of hypotheses:
difficult to comprehend and therefore impossible to learn, accept
or retain; or there might be too much distraction from the envi- Hypothesis 3a: There is a negative relationship between bank-
ronment in the form of ‘noise’, preventing the audience from pro- generated content and customer engagement.
cessing and accepting the message (Ajzen, 1992). Therefore, per- Hypothesis 3b: The negative relationship between bank gen-
suasive communication is most effective when the format of the erated content and customer engagement is stronger for
message is tailored to the recipient’s information-processing pro- informative-only content than for persuasive-only and
clivities (Bakker, 1999). mixed content.
The existing literature suggests that customers prefer posts that
contain visual aids over those without. For example, Klassen et al. 4. Empirical context
(2018) find that the posts with a video or images featuring an at-
tractive human body are positively associated with Facebook inter- Our research focuses on two countries, Australia and Singapore,
actions. De Vries et al. (2012) find that vividness, such as the use of for a variety of reasons. First, social networks are absorbing an
video, is positively related to the number of likes. We argue that in ever-greater share of consumers’ leisure time in both countries. For
banking sector, customers have the same preference to posts with instance, 8 out of 10 people (79%) in Australia now use social me-
visual aids due to two reasons. First, social media users might be dia (Sensis, 2018), and 7 out of 10 Singaporeans are active social
distracted since they tend to face a high level of ‘noise’, i.e., the media users (Hashmeta, 2018). Second, in both countries banks are
large volume of information coming from multiple sources. Sec- increasing their use of digital channels and data insights to im-
ond, in the banking sector, products and services tend to be more prove the customer experience (Deloitte, 2018). With over 200 fi-
complicated and difficult to understand and evaluate. Posts with nancial institutions, Singapore is one of the most established cap-
visual aids, including photos and videos, may stimulate multiple ital markets in the Asia-Pacific region. Singapore’s ‘Smart Nation’

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 5

strategy is designed to harness digital technologies to drive eco- 5.3. Coding for content
nomic growth (Smart Nation, 2018). Likewise, the Australian gov-
ernment recognises that digital technologies will reshape finan- The content of each post was analysed to extract four variables,
cial services (Investment Commission 2016). Third, a cross-cultural as follows.
sample allows insights to be gained on international similarities Type of content. We followed the work of Resnik and
and differences in content marketing strategies and customer reac- Stern (1977) and Lee et al. (2018) to code the banks’ Facebook
tions. posts into two main categories: informative or persuasive. Content
We focus on Facebook because it is the dominant social net- was categorised as informative if it included information about
working platform in both countries: as of June 2019, it had the products, promotions, availability, price and product-related as-
largest share of the social media market in both Australia (62.46%) pects that could be used to optimise a purchase decision.
and Singapore (63.94%) (StatCounter, 2019a, 2019b). Analysis of the Persuasive content includes a wider range of purposes, such
sample banks’ social media efforts reveals that all ten banks have as humour, emotional appeal, casual banter or information on the
significantly more followers on Facebook than on other platforms brand’s philanthropic outreach. These messages are not specifically
(see Appendix 2), suggesting that Facebook is the retail banking about the company’s offerings but are shared to modify customers’
industry’s preferred platform in these two countries. overall beliefs, values and attitudes and to create liking, trust and
relationships with the company/brand. Lee et al. (2018) further
classify persuasive content as broadly seeking to influence con-
5. Methodology sumers by appealing to ethos, pathos and logos. Ethos is an appeal
of credibility or character through the use of a celebrity to endorse
5.1. Sample a product or gain goodwill (e.g. via small talk or banter). Pathos is
an appeal to emotion through philanthropic content. Logos appeals
The study employed content analysis to gain information on the to reason by supplying general facts that capture the consumers’
Facebook content marketing strategy of leading banks in Australia attention and influence them to adopt a product.
and Singapore. This method has proven useful for understanding Sixty-six of the posts contained elements from both the infor-
creative social media strategies (Ashley and Tuton, 2015). The fo- mative and persuasive categories.2 A typical example is from DBS:
cus was on the top five banks operating in each of the two coun- “You’ve probably heard many things about Investment-Linked Poli-
tries, with the assumption that the more valuable brands have cies (ILPs). But what’s the real story? Let’s dig through the evidence
more resources to invest in social media marketing (Sitta et al., to uncover the facts about ILPs. The verdict might surprise you. –
2018). Appendix 2 shows the profiles of the ten banks examined Lawrence Leu, DBS Treasures Wealth Planning Specialist.” In this post,
in this study. Three months’ worth of data (from 1st January to a video with educational information on ILPs was shown highlight-
31st March 2019) were collected from the official Facebook pages ing DBS Treasures as the service provider for such ILP products.
of these ten banks, including all posts and the customer engage- Another example is from National Australia Bank (NAB): “Dona-
ment for each post. A total of 541 posts were analysed. tion for floods in Townsville: Our customers who’ve been impacted by
floods in Townsville have suffered significantly. We’re offering $500 m
in concessional loans to those affected…” In this post, NAB’s cor-
5.2. Engagement metrics porate social responsibility (CSR) initiative and information about
concessional loans were conveyed at the same time. We therefore
Traditional engagement metrics developed by Facebook – the created a third category of content: mixed content.
number of likes, emoji, comments and shares – were used to In the end, we created three dummy variables indicating three
quantify users’ engagement with posts, similar to previous studies categories of content: persuasive-only posts, mixed content and
(Lee et al., 2014; Klassen et al., 2018). A like occurs when a user informative-only posts.3 In the regression, informative-only post is
clicks the ‘like’ button, indicating that the post has resonated with the base category. Moreover, under each main content category, we
the user in some way (Dahl, 2018). ‘Emoji’ represent a reaction identified six level-2 sub-categories. For example, informative mes-
feature introduced by Facebook in February 2016; five emoji are sages were further classified as product, incentives and banks’ in-
available, corresponding to five predefined emotions: ‘love’, ‘ha-ha’ terest rates/system information. Persuasive messages were further
(laughter), ‘wow’ (amazement), ‘sad’ and ‘angry’. They permit users classified as philanthropy (i.e. CSR campaigns), general facts and
to articulate their feelings (either positive or negative) wordlessly entertainment/seasonal greetings. The level-2 subcategories were
(Smieško, 2016). A comment is an action whereby a user leaves further classified into 13 level-3 categories. The detailed classifi-
a comment or replies to a post. In our study, each comment was cation scheme is shown in Appendix 3.
further classified as positive, negative, neutral or question reply. Post format. Post format was broken down into three types of
We acknowledge that the designation of a comment as positive or posts: posts ‘with text only’, ‘with a photo’, and ‘with a video’. Two
negative may be subjective and, furthermore, that some comments dummy variables were generated for posts with a photo or a video,
may be both positive and negative at the same time. However, all while posts with text only were treated as the base category.
the comments were very short, simple and easy to categorise, so
a composite code, such as ‘both positive and negative’, was not
needed. A question reply is defined as a fan post with a question
2
mark. A share refers to an action by a user who shares or reposts All of the sample banks in our study followed the principle of using eye-
the post. Among these metrics, likes and emoji represent passive catching and concise posts, usually containing 4–8 lines of either persuasive or in-
formative content. This approach is consistent with studies about the ideal length
customer engagement while comments and raising questions rep-
of Facebook posts, which suggest around 40–80 characters (Track Social, n.d.; Bullas,
resent active customer engagement. An overall engagement mea- n.d.). Few posts that are so short can combine both informative and persuasive con-
sure was developed by adding all five of the engagement metrics tent.
together (likes + emoji + comments (positive, negative and neu- 3
Another way to code the content is to have only two types of content, i.e.
tral) + question replies + shares). To ensure the normal distribu- persuasive content and informative content, in which case 66 posts contained
both variables equal to 1. However, putting these two highly correlated variables
tion of the main dependent variables, log transformation was ap- (corr = 0.8) into the regression at the same time would lead to multicollinearity,
plied to measures such as overall engagement, number of likes and which would distort the estimation. Therefore, we opted to create three categories
number of emoji responses. of content.

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

6 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

Fig. 1. Number of posts and customer engagement per bank. Note: The size of the bubbles indicates the number of followers at the beginning of the observation period.

Source. Each post was categorised according to source. The and Singapore. As shown in Panel 1a, banks in Singapore publish
dummy variable Source was coded as 1 if the post content was cre- significantly more bank-generated Facebook posts, in different for-
ated by banks themselves and 0 if banks shared posts created by mats and with more diversified types of content, than those in
other users. Australia. Panel 1b shows that, overall, Singaporean banks’ posts
Links. We also noticed that banks often use links to provide fur- receive more engagement from customers, particularly more likes,
ther information, such as availability of products, promotions, lo- but fewer comments, than Australian banks. However, in terms
cations or a full article from a third-party website. We therefore of engagement per post, Panel 1c shows that the posts generated
developed the variable Link, coded as 1 when a link is given in the by Australian banks are more engaging than those by Singaporean
post and 0 otherwise. banks in that they receive significantly more emoji responses, com-
In regard to inter-code reliability, the coding was done manu- ments and shares per post than the latter.
ally by four post-graduate students. We noted that the designation
of a post as informational or persuasive may be subjective and
6.2. Relationship between content type, post format and source
that posts may contain both elements. To reduce the subjectivity
credibility, and customers’ engagement levels
of coding, we asked two graduate students to do the coding for
each post consecutively, and the author completed the third round
We then conducted ordinary least squares (OLS) and negative
of checking and made the final decision if any discrepancies were
binomial regressions to test whether the content type, format and
identified.
source of the posts affect customers’ engagement levels. Table 2
5.4. Statistical analysis shows that the independent variables have low or no correlations;
thus, multicollinearity is not an issue for the regression.
Descriptive statistics provided characteristics of the content Regression results are reported in Tables 3 and 4. Hypothesis
for the banks across the two countries. Regression models were 1a predicts that persuasive-only content will lead to more pas-
constructed to explore associations between the post content sive forms of engagement than informative-only or mixed content,
and the Facebook engagement metrics with STATA. The Breusch- while Hypothesis 1b predicts that informative-only and mixed con-
Pagan/Cook-Weisberg test revealed heteroscedasticity. As a result, tent will lead to more active forms of engagement than persuasive-
regression with robust standard errors was employed to address only content. In Singapore, persuasive-only content is positively
the concern of heteroscedasticity. related with engagement measures in models 1, 3 and 5 in
Table 3 (model 1: β = 0.59; p < 0.001; model 3: β = 0.75; p
6. Results < 0.001; model 5: β = 0.64; p < 0.001), while negatively sig-
nificant in models 13 and 15 in Table 4 (model 13: β = −1.38;
6.1. Banks’ content marketing efforts and outcomes in Australia and p < 0.001; model 15: β = −0.54; p < 0.05). These results sug-
Singapore gest that persuasive-only content leads to more likes and emoji re-
sponses, but fewer comments and questions, than informative-only
Fig. 1 shows that all of the ten banks in our study were ac- content. Mixed content is positively related to engagement mea-
tive on Facebook during the observation period, but they experi- sures in models 1 and 3 in Table 3 (model 1: β = 0.47; p < 0.1;
enced different outcomes. In general, banks in Singapore (yellow model 3: β = 0.44; p < 0.1). It is also positively significant in mod-
bubbles) put more effort into their Facebook content marketing els 9 and 11(model 9: β = 1.51; p < 0.01; model 11: β = 3.34; p
and achieved better levels of customer engagement than their Aus- < 0.001), but negatively significant in model 13 (β = −1.46; p <
tralian counterparts (green bubbles). Further Pearson’s correlation 0.01) in Table 4. These results suggest that mixed content leads
tests show that the number of posts per week is moderately cor- to more likes, emoji responses and positive comments, but fewer
related with all engagement measures (correlation coefficients be- negative comments, than informative-only content.
tween 0.36 and 0.52, p < 0.001), except number of comments (cor- Overall, the results suggest that in Singapore, embedding per-
relation coefficient = 0.15, p > 0.1). These results suggest that more suasive messages in posts (persuasive-only or mixed content) can
frequent posting leads to better overall engagement. effectively engage consumers in the passive form of inducing likes
Table 1 compares the banks’ Facebook content marketing ac- and emoji; having informative messages in posts (mixed content or
tivities and their effectiveness in engaging customers in Australia informative-only content) can effectively induce consumers to ac-

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 7

Table 1
Features of weekly posts in Australia and Singapore.

Mean of banks in Australia Mean of banks in Singapore MeanAU – MeanSG

Panel 1a. Facebook marketing activities

Number of posts 2.63 6.34 p < 0.001
Number of informative-only posts 0.61 2.97 p < 0.001
Number of persuasive-only posts 1.53 2.72 p < 0.01
Number of posts with mixed content 0.49 0.65 p > 0.1
Number of sub-content categories 1.88 2.94 p < 0.001
Number of bank-generated posts 2.45 6.12 p < 0.001
Number of cited posts 0.18 0.22 p > 0.1
Number of posts with text (no photo or video) 0.27 0.31 p > 0.1
Number of posts with photo 0.98 2.09 p < 0.001
Number of posts with video 0.65 1.75 p < 0.001
Number of posts with link 1.02 3.69 p < 0.001
Panel 1b. Overall engagement

Number of engagement actions 1404.80 2287.54 p <0.1

Number of emoji responses 140.78 75.6 p <0.05
Number of likes 599.65 2010.6 p < 0.01
Number of comments 545.39 90.22 p < 0.01
Number of positive comments 273.94 32.73 p < 0.1
Number of negative comments 59.27 4.37 p < 0.01
Number of questions 18.82 5.38 p < 0.1
Number of comment replies 38.82 17.28 p < 0.1
Number of shares 118.98 111.12 p > 0.1
Panel 1c. Per-post engagement

Engagement per post 512.90 409.55 p > 0.1

Likes per post 234.59 370.49 p > 0.1
Emoji per post 50.03 10.28 p < 0.001
Comments per post 185.82 13.61 p < 0.001
Positive comments per post 72.09 3.63 p < 0.05
Negative comments per post 28.22 0.66 p < 0.01
Questions per post 8.04 1.06 p < 0.05
Comments reply per post 17.25 3.15 p < 0.05
Shares per post 42.45 15.16 p < 0.01

Table 2
Descriptive and correlation matrix.

1 2 3 4 5 6 7 8 9 10 11 12 13 14

1 Overall engagement 1
2 Likes 0.93∗ 1
3 Emoji 0.79∗ 0.75∗ 1
4 Shares 0.44∗ 0.25∗ 0.24∗ 1
5 Comments 0.28∗ 0.18∗ 0.29∗ 0.07 1
6 Positive comments 0.20∗ 0.10∗ 0.20∗ 0.04 0.93∗ 1
7 Negative comments 0.23∗ 0.18∗ 0.28∗ 0.10∗ 0.46∗ 0.13∗ 1
8 Questions 0.17∗ 0.15∗ 0.20∗ 0.07 0.45∗ 0.13∗ 0.85∗ 1
9 Persuasive-only content 0.14∗ 0.18∗ 0.19∗ −0.01 0.02 0.05 −0.09∗ −0.05 1
10 Informative-only content −0.19∗ −0.21∗ −0.19∗ −0.03 −0.01 −0.05 0.09∗ 0.07 −0.78∗ 1
11 Mixed content 0.06 0.04 0.00 0.07 −0.01 0.00 −0.01 −0.02 −0.35∗ −0.31∗ 1
12 Sources −0.13∗ −0.10∗ −0.11∗ −0.08 −0.02 −0.01 0.00 −0.02 0.01 −0.06 0.09∗ 1
13 Photos 0.19∗ 0.18∗ 0.17∗ 0.01 0.05 0.08 −0.07 −0.05 0.18∗ −0.19∗ 0.01 −0.07 1
14 Videos 0.02 −0.02 0.02 0.15∗ 0.01 −0.02 0.06 0.10∗ −0.01 0.04 −0.05 −0.02 −0.44∗ 1
15 Links −0.14∗ −0.14∗ −0.21∗ 0.01 −0.09∗ −0.07 −0.06 −0.05 −0.28∗ 0.25∗ 0.05 0.03 −0.29∗ −0.06 1
Mean 4.80 4.13 1.80 34.23 60.24 28.74 5.89 2.35 0.47 0.41 0.12 1.04 0.34 0.27 0.55
Std. Dev. 1.45 1.59 1.47 27.54 429.48 351.59 43.57 20.38 0.50 0.49 0.33 0.20 0.47 0.44 0.50
Min 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0
Max 9.80 9.79 7.01 84 8134 7688 851 458 1 1 1 2 1 1 1
∗
Note: p < 0.05.

tively post comments or raise questions. The three types of content tralian consumers, particularly in the form of encouraging com-
do not show significant differences in leading customers to share ments and questions. In other words, Australian consumers pre-
posts. These results generally support our prediction in Hypotheses fer posts with informative content. Mixed content is more effective
1a and 1b. than both persuasive-only and informative-only content in encour-
In the Australian sample, persuasive-only content is negatively aging Australian consumers to share posts with their communities.
significant in models 2, 10, 14 and 16 (model 2: β = −0.5; p These results partially support our Hypothesis 1b, which predicts
< 0.05; model 10: β = −1.09; p < 0.01; model 14: β = −1.9; that informative-only and mixed content lead to more active cus-
p < 0.001; model 16: β = −0.82; p < 0.1) but not significant tomer engagement than persuasive-only content.
in other models. Mixed content is positively significant only in Hypothesis 2a predicts that posts with photos and videos will
model 8 (β = 11.89; p < 0.1) but not significant in other mod- be more effective in engaging customers than posts with text only.
els. These results suggest that persuasive-only content is less ef- The results in Table 3 show that, for the Singaporean sample,
fective than informative-only and mixed content in engaging Aus- posts with photos lead to more overall engagement, likes, emoji re-

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

8 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

Table 3
Content type, post format, source of posts and customer engagement level, as measured by likes, emoji responses and shares.

Overall Engagement Likes Emoji 4 Shares

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8

Singapore Australia Singapore Australia Singapore Australia Singapore Australia

Persuasive-only 0.59∗ ∗ ∗ (0.15) −0.5∗ (0.23) 0.75∗ ∗ ∗ (0.18) −0.35 (0.25) 0.64∗ ∗ ∗ (0.14) −0.44 (0.36) 1.94 (2.84) −4.4 (5.2)
content
Mixed content 0.47+ (0.25) 0.13 (0.25) 0.44+ (0.26) 0.21 (0.29) 0.14 (0.21) 0.13 (0.38) 1.88 (4.34) 11.89+ (6.08)
Source −1.49∗ ∗ ∗ (0.31) 0.02 (0.43) −1.46∗ ∗ ∗ (0.31) 0.23 (0.47) −1.01∗ ∗ ∗ (0.14) −0.3 (0.73) −15.93∗ ∗ (5.7) −1.84 (8.92)
Photo 0.62∗ ∗ (0.2) 0.58∗ (0.23) 0.4+ (0.23) 0.91∗ ∗ ∗ (0.24) 0.37∗ (0.16) 0.75∗ (0.33) 8.37∗ (3.43) 0.54 (6.54)
Video 0.27+ (0.15) 0.61∗ (0.25) −0.02 (0.17) 0.76∗ ∗ (0.28) 0.09 (0.13) 0.72+ (0.37) 12.28∗ ∗ ∗ (3.23) 12.62∗ (5.79)
Link 0.13 (0.17) −0.49∗ (0.2) −0.02 (0.19) −0.35 (0.22) −0.02 (0.14) −0.94∗ ∗ (0.29) 6.15∗ (2.94) −3.49 (5.58)
Constant 5.49∗ ∗ ∗ (0.39) 5.51∗ ∗ ∗ (0.49) 5∗ ∗ ∗ (0.41) 4.17∗ ∗ ∗ (0.52) 2.16∗ ∗ ∗ (0.23) 3.16∗ ∗ ∗ (0.79) 37.58∗ ∗ ∗ (7.58) 42.38∗ ∗ ∗ (11.36)
R-squared 0.12 0.14 0.10 0.15 0.11 0.16 0.05 0.09
N 412 129 412 129 412 129 412 129

Notes:
1. ∗ ∗ ∗ p < 0.001; ∗ ∗ p < 0.01; ∗ p < 0.05; + p < 0.1.
2. OLS regression with robust standard errors was employed to address the concern of heteroscedasticity.
3. Informative-only content is the base category.
4. We followed existing studies to group emoji, including ‘love’, ‘ha-ha’ or ‘wow’ reactions on Facebook, into positive emotions, as they indicate enthusiasm and invite
people to consider new ideas more favourably (Fredickson, 2001), while ‘sad’ and ‘angry’ were grouped into negative emoji. We then conducted the analysis by regressing
the number of positive and negative emoji received for banks’ Facebook posts on our concerned independent variables. The results are similar to those for likes and overall
emoji in Table 4 in the manuscript, in that persuasive-only content is more effective in triggering consumers to give both positive and negative emoji than informative-only
content in Singapore, while the effect of mixed content is not significantly different from informative-only content. However, in Australia, the three types of content do not
show significant differences in terms of their impact on customers’ emoji reactions.

Table 4
Content type, post format, source of the posts and customer engagement level, as measured by comments and questions.

Comments Positive comments Negative comments Questions

Model 9 Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 16

Singapore Australia Singapore Australia Singapore Australia Singapore Australia

Persuasive-only content −0.29 (0.22) −1.09∗ ∗ (0.36) 0.59 (0.33) 0.05 (0.65) −1.38∗ ∗ ∗ (0.35) −1.9∗ ∗ ∗ (0.37) −0.54∗ (0.26) −0.82+ (0.45)
Mixed content 1.51∗ ∗ (0.48) −0.19 (0.49) 3.34∗ ∗ ∗ (0.73) −0.47 (0.8) −1.46∗ ∗ (0.49) −0.37 (0.55) −0.48 (0.4) −0.38 (0.58)
Source −1.2∗ ∗ (0.41) 0.89 (0.75) −3.88∗ ∗ ∗ (1) 2.21∗ (1.11) −1.56 (1.18) 2.09∗ (0.84) 0.13 (0.7) −1.87+ (1.11)
Photo 0.27 (0.29) 1.38∗ (0.61) 0.35 (0.4) 2.05∗ (0.84) −0.56 (0.48) −0.6 (0.4) 0.25 (0.28) −0.57 (0.41)
Video 0.49+ (0.28) 0.78+ (0.41) 0.97∗ (0.44) 1.06+ (0.57) 0.25 (0.46) 0.42 (0.47) 0.58∗ (0.28) 1.5∗ (0.66)
Link 0.73∗ ∗ ∗ (0.2) −1.31∗ ∗ (0.42) 0.71∗ (0.31) −2.3∗ ∗ ∗ (0.64) 0.13 (0.34) −1.18∗ ∗ (0.41) 0.68∗ (0.27) 0.06 (0.37)
Constant 2.96∗ ∗ ∗ (0.54) 4.57∗ ∗ ∗ (0.77) 3.15∗ ∗ (1.1) 1.19 (1.1) 1.64 (1.33) 2.1∗ (0.85) −0.77 (0.8) 3.65∗ ∗ (1.17)
Wald chi2 42.65∗ ∗ ∗ 35.30∗ ∗ ∗ 46.51∗ ∗ ∗ 32.63∗ ∗ ∗ 26.05∗ ∗ ∗ 43.94∗ ∗ ∗ 23.26∗ ∗ ∗ 24.31∗ ∗ ∗
N 412 129 412 129 412 129 412 129

Notes:
1. ∗ ∗ ∗ p < 0.001; ∗ ∗ p < 0.01; ∗ p < 0.05; + p < 0.1.
2. Negative binomial regression was employed as the dependant variable, i.e. number of comments are counted variables with an over-dispersion.

sponses and shares than those with text only (model 1: β = 0.62; Hypothesis 3a predicts a negative relationship between bank-
p < 0.01; model 3: β = 0.4; p < 0.5; model 5: β = 0.37; p < generated content and customer engagement. Table 4 shows that
0.05; model 7: β = 8.37; p < 0.05). Posts with videos lead to more bank-generated posts are negatively related to customer engage-
shares in model 7 (β = 12.28; p < 0.001). In Table 4, posts with ment in the Singaporean sample (in all models except 13 and 15),
videos lead to a higher number of positive comments and ques- lending support to Hypothesis 3a. In Australia, on the other hand,
tions in models 9, 11 and 15 (model 9: β = 0.49; p < 0.1; model the source of the posts does not show a significant result, except
11: β = 0.97; p < 0.05; model 15: β = 0.58; p < 0.05), while posts in models 12 and 14 (model 12: β = 2.21; p < 0.05; model 14:
with photos do not show any significant result. In summary, Hy- β = 2.09; p < 0.05). These results suggest that Australian con-
pothesis 2a is generally supported in Singapore: posts with videos sumers show more interest in posting both positive and negative
can lead to a more active engagement (such as sharing posts and comments in response to bank-generated content than third-party
offering comments), while posts with photos are helpful in gener- content, thus going against the prediction in Hypothesis 3a.
ating more passive engagement (such as likes and emoji).
In the Australian sample, in Table 3, posts with photos lead 6.3. Interactive effect of content type, post format, source of posts
to more likes and emoji (model 2: β = 0.58; p < 0.05; model and customer engagement level
4: β = 0.91; p < 0.001; model 6: β = 0.75; p < 0.05) but not
shares (model 8: β = 0.54; p > 0.1). Posts with videos lead to We then investigated whether the source and format
more likes, emoji and shares (model 2: β = 0.27; p < 0.5; model of posts strengthen or weaken the effect of persuasive-
4: β = 0.76; p < 0.01; model 6: β = 0.72; p<0.1; β = 12.62; p < only/mixed/informative-only content in engaging online customers
0.05). In Table 4, the results show that posts with photos lead to by adding the interactions between type of content and source,
more positive comments in model 12 (β = 2.05; p < 0.05). Posts photo and video. Table 5 shows the results.
with videos lead to more positive comments and questions (model Hypothesis 2b predicts that all types of content (informative-
10: β = 0.78; p < 0.1; model 12: β = 1.06; p < 0.01; model 16: only, persuasive-only and mixed content) presented with video or
β = 1.5; p < 0.05). In summary, in Australia, posts with photos photo will have a stronger positive relationship with customer en-
and videos seem to be helpful in generating a greater level of en- gagement than those presented with text only. In the Singaporean
gagement in different forms than those with text only, supporting sample, the interaction between mixed content and video leads
Hypothesis 2a. to more comments from customers in model 25 (β = 2.4; p <

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 9

0.05), while the main effect of video is positively significant in

models 17 and 23. These results suggest that content involving in-

−2.36∗ ∗ (0.74)

138.59∗ ∗ ∗ (12)
7.26∗ ∗ ∗ (1.27)
−1.4+ (0.79)
1.84+ (0.97)
−2.97 (2.29)
−1.92 (1.26)
−1.89 (1.54)

−0.54 (1.81)

−0.83 (1.81)
formation, including informative-only content and mixed content,

−0.2 (1.09)

−0.07 (1.2)
1.73 (1.45)

2.9∗ (1.38)
Model 26
Australia
engages customers more effectively when presented with videos,
supporting Hypothesis 2b. These results highlight the importance

129
of using vivid presentation in functional messages to reduce the
audience’s cognitive burden.

−3.77∗ ∗ ∗ (0.61)
−3.16∗ ∗ ∗ (0.82)
−3.9∗ ∗ ∗ (0.34)
3.31∗ ∗ ∗ (0.26)

3.29∗ ∗ ∗ (0.32)
In the Australian sample, in models 18, 20 and 22, the interac-

64.24∗ ∗ ∗ (12)
0.72∗ ∗ (0.26)
5.8∗ ∗ ∗ (0.69)
1.35 (1.13)
0.17 (0.49)
0.33 (0.44)

0.13 (0.41)
0.15 (0.34)
2.4∗ (1.18)
Comments

tion of persuasive-only content and photo is positively significant

Singapore
Model 25

(β = 1.62; p < 0.05; β = 1.65; p < 0.05; β = 3.04; p < 0.01).

412
In models 18, 20 and 22, the interaction of mixed content and
photo is positively significant (β = 1.17; p < 0.1; β = 1.37; p <

3. Models 25 and 26 are negative binomial regression models and Wald chi2 is reported as the dependant variable, i.e. number of comments are counted variables with an over-dispersion.
0.1; β = 3.2; p < 0.01). In models 18, 20 and 26, the interaction of
−11.63 (16.18)

50.18∗ (15.35)
−17.36 (4.87)
12.06∗ (3.75)

mixed content and video is negatively significant (β = −1.35; p <

−6.31 (5.55)
−3.14 (7.39)

−8.11 (12.8)
15.89 (9.37)

17.83 (9.59)
5.75 (10.78)

2.49 (10.73)
3.8 (9.28)
0.1 (7.08)
Model 24

0.1; β = −1.8; p < 0.05; β = −2.36; p < 0.01). The main effect of
Australia

video is positively significant in models 18, 20 and 26 (β = 1.26; p

0.12
129
< 0.05; β = 1.58; p < 0.01; β = 1.84; p < 0.1). These results sug-
gest that, in Australia, persuasive-only content is more effective in
−26.73∗ ∗ ∗ (1.15)

engaging customers if it is presented with photos. Mixed content

16.55∗ ∗ ∗ (1.45)
−12.88 (18.69)
−11.73 (21.79)
−2.48 (18.25)

47.63∗ ∗ ∗ (3.7)
6.39+ (2.74)
−7.14 (5.59)

−9.91 (9.72)
13.6 (16.03)

12.09 (9.63)

is more effective in engaging customers if it is presented with pho-

5.58 (5.87)

6.52 (5.45)
Singapore
Model 23

tos but not videos. Informative-only content is more engaging if it

Shares

is presented with videos. This can be attributed to users’ appreci-

0.07
412

ation of video content and its ability to communicate complicated

financial information in an engaging way that is easily understood.
−1.57+ (0.91)
−4.24+ (2.27)

−0.77+ (0.39)

The results generally support Hypothesis 2b.

3.04∗ ∗ (1.05)

3.43+ (2.05)

−1.83 (1.96)
−0.06 (1.12)

−0.89 (1.07)
−1.73 (2.16)
3.2∗ ∗ (1.12)

1.15 (0.76)
0.37 (2.03)

5.1∗ (2.04)

Hypothesis 3b predicts that the negative relationship between

Model 22
Australia

bank generated content and customer engagement will be stronger

0.31
129

for informative-only content than for persuasive-only and mixed

content. In the Singaporean sample, the interaction between per-
−0.85∗ (0.21)

suasive content and source of the post is significantly positive in

2.07∗ ∗ (0.44)
−0.24 (0.29)

−0.04 (0.42)

−0.05 (0.67)

−0.01 (0.15)
−0.07 (0.2)
0.26 (0.32)
0.35 (0.34)

0.51 (0.62)
0.46 (0.32)

model 25 (β = 3.31; p < 0.001). Interaction between mixed con-

0.24 (0.3)
0.7 (0.49)
Singapore
Model 21

tent and source shows positive significance in models 17 and 25

Emoji

0.11

(β = 0.98; p < 0.05; β = 3.29; p < 0.001). On the other hand,

412

the main effect of source is negatively significant in all models

in the Singaporean sample. These results suggest that persuasive-
5.54∗ ∗ ∗ (1.45)
1.58∗ ∗ (0.57)
1.37+ (0.74)

−1.37 (1.55)
−0.74 (0.82)

−2.01 (1.68)
−1.02 (1.41)
−0.37 (0.56)
−1.8∗ (0.85)
1.65∗ (0.67)

−0.4 (0.29)
0.71 (1.47)

only content and mixed content created by the banks them-

2.16 (1.5)
Model 20
Australia

selves will encourage customers to post more comments, while

0.29
129

customers react less to informative-only content created by the

banks themselves. The results support Hypothesis 3b, suggesting
Interaction of content type, post format, source of the posts and customer engagement level.

that when customers seek functional content (information about a

−1.42∗ ∗ (0.18)

4.94∗ ∗ ∗ (0.44)
−0.22 (0.41)

bank’s offerings) on which to base their decisions, they may sus-

−0.21 (0.6)
0.17 (0.55)
0.08 (0.64)

0.86 (0.46)

0.22 (0.31)
0.07 (0.26)
0.72 (0.86)
0.41 (0.77)

0.07 (0.22)
0.01 (0.18)
Singapore
Model 19

pect that the bank-generated information is biased, hence their

Likes

response is limited. However, when they seek entertainment (ei-

0.11
412

ther entertaining, emotion-evoking posts or posts with general

facts), they appreciate banks’ efforts in creating fresh and original
−1.35+ (0.69)

6.95∗ ∗ ∗ (1.15)

2. Models 17–24 are OLS regression models and R-squared is reported.

−0.51∗ (0.25)

content.
1.17+ (0.69)

−1.66 (1.22)
−2.21 (1.35)
−1.24 (1.08)
−0.54 (0.71)

−0.62 (0.57)
1.62∗ (0.65)

2.09+ (1.2)
0.72 (1.13)

1.26∗ (0.5)

In Australia, the interaction between persuasive-only content

Model 18
Australia

and source, as well as the main effect of source, is always insignif-

0.27
129

icant. In models 18, 22, 24 and 26, the interaction between mixed
Overall engagement

content and source of post is significantly positive (β = 2.09; p

< 0.1; β = 3.43; p < 0.1; β = 12.06; p < 0.05; β = 2.9; p <
−1.89∗ ∗ ∗ (0.1)
−0.71+ (0.29)

5.85∗ ∗ ∗ (0.27)

1. ∗ ∗ ∗ p < 0.001; ∗ ∗ p < 0.01; ∗ p < 0.05; + p < 0.1.

−0.36 (0.27)
0.98∗ (0.25)

0.52∗ (0.16)
0.46 (0.69)
0.08 (0.33)

0.54 (0.45)
0.11 (0.38)
0.23 (0.88)

0.15 (0.17)
0.43∗ (0.1)

0.05), suggesting mixed content will be more effective in engag-

Singapore
Model 17

ing customers if it is created by the banks themselves, while for

0.12
412

persuasive-only and informative-only content, Australians seem to

be indifferent regarding the source of the content. Hypothesis 3b
is not supported in the Australian sample.
Persuasive-only content ∗ photo
Persuasive-only content ∗ source

Persuasive-only content ∗ video

6.4. Subcategories of content

Persuasive-only content
Mixed content ∗ source
Mixed content ∗ photo
Mixed content ∗ video

R-squared /Wald chi2

The link between engagement and different subcategories of

content was examined. Fig. 2 shows the number of posts (lines)
Mixed content

and customer engagement level (bars) for each level-3 subcategory

Constant

of content for Australia (in orange) and Singapore (in blue), respec-
Source

Video
Photo
Table 5

tively. With regard to the five Singaporean banks, the top three
Link

Notes:

types of content that engage customers most are puzzles/games,

N

inspirational stories and campaigns about community. The chart

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

10 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

Fig. 2. Number of posts (lines) and customer engagement (bars) for each content subcategory – Australia (in orange) and Singapore (in blue).

shows that the three categories of content most frequently used by highly intangible and complex service-based offerings (Colgate and
the banks are product-related (i.e. offering information on finan- Stewart, 1998). The purchase of financial products/services tends to
cial services/products/promotion/availability), community-focused be of high risk and has a long-term impact. As a result, perceived
campaigns and general facts. Hence, a mismatch can be seen be- value is multi-dimensional, and the functional, emotional and so-
tween what works well for the bank’s customers and what the cial aspects of financial services are important to consumers (Roig
banks are doing in relation to content creation. et al., 2006). In such a high-involvement situation, people are mo-
With regard to the Australian banks, the top three types of con- tivated to seek, process and clarify comprehensive information and
tent that are most engaging to customers are puzzles/games, in- respond better to facts and logical arguments directly related to
spirational quotes and products. While the Australian banks gener- their concerned product/services rather than peripheral cues (emo-
ated a higher volume of posts that are community-orientated and tional, entertaining or general educational messages). These find-
product-related, they generated fewer posts with puzzles/games ings are similar to those of Chang et al. (2015), who find that the
and inspirational quotes. Therefore, banks in both Singapore and usefulness of post information increased customer engagement on
Australia can enhance customer engagement by closely examining the Facebook page of a cooking website, which is predominantly
the category of content used on their Facebook pages and adjust- a utilitarian service. Our study suggests that customer engagement
ing their posts accordingly. can be deepened by using informative types of content to meet the
needs of high-involvement customers, highlighting the wisdom of
matching cognitive messages to high-involvement consumers and
7. Discussion emotional messages to low-involvement consumers (Areni, 2003;
McMillan et al., 2003).
Noticing the sparse research on the optimal content marketing
tactics for utilitarian and service-orientated brands, such as those
in financial services, this study draws on persuasive communica- 7.2. Format of content
tion theories to investigate how the three commonly identified
features of content marketing, i.e. content type, format of content When it comes to the format of the posts, existing empirical
and content source, interactively influence customer engagement literature gives conflicting evidence about the impact of photos
in Australia and Singapore. Our findings enrich the existing studies and videos in posts. On the one hand, studies show that pho-
on content marketing in several ways. tos and videos are found to increase the vividness of persuasive
digital content (de Vries et al., 2012), thus engaging customers
7.1. Type of content more effectively. On the other hand, Lee et al. (2016) find that
posts with photos are more likely to be commented upon, while
With regard to the type of information presented in a post, posts with videos get less likes as compared to a simple status
prior studies based on hedonic sectors yield mixed evidence. Lee update.
et al. (2018) show that persuasive content, rather than informa- Our study tackles the incongruent evidence by showing that
tional content, is the key to effective engagement, while Dolan videos and photos strengthen posts with informative content or
et al. (2019) provide evidence that both persuasive and infor- persuasive content in different ways. In both Australia and Sin-
mative content positively influence passive engagement (likes), gapore, videos help the informative-only content to induce cus-
but not active engagement (comments). Our findings are differ- tomers to share posts and offer comments, which is a more cog-
ent from prior studies, showing that although persuasive-only nitively demanding type of engagement. This result suggests that
content leads to more passive engagement, such as more shares videos help communicate complex, bank-related information in an
amongst Australian customers and more likes/emoji amongst Sin- engaging way that is easy to understand, reducing customers’ cog-
gaporean customers, informative content, be it presented together nitive effort and facilitating the central route of decision-making.
with persuasive information or alone, is important in driving cus- This finding is consistent with studies in neuroscience that high-
tomer engagement in a more active and cognitively demanding light the role of visualisations in promoting deep learning (Bokek
way (such as giving comments and raising questions) in both and Tversky, 2016). In Australia, photos, not videos, help persuasive
countries. posts gain more likes and emojis. We draw on the work of de Vries
Our findings can be explained by the ELM and the special fea- et al. (2012) to explain this finding: persuasive content presented
tures of the financial services sector. This services sector provides with a photo is more vivid and attractive than a persuasive post

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 11

without a photo, as it ‘cuts through’ the noise from the environ- and certified, and to understand their fee structures prior to hir-
ment, constituting a more powerful peripheral cue. Due to time ing their services. Thus, they prefer more objective information
constraints, customers who seek entertainment instead of prod- released by third-party sources. In contrast, Australia scores 21
uct information may not want to spend time watching an entire on long-term orientation and therefore Australians exhibit a rel-
video. atively low propensity to save for the future and a greater focus
on achieving immediate results. Data show that Australians tend
7.3. Source of content to save less compared to Singapore (World Bank, n.d.). A typical
consumer may not be capable of doing due diligence and eval-
Existing research on the credibility of advertising information uating complicated financial information and hence depends to
suggests that customers are sceptical of, and less inclined to trust, a greater extent on financial advisors’ and financial institutions’
messages that come from marketers. Scepticism may come from a expertise.
general stereotype (that all advertising exaggerates product ben-
efits since firms prioritize profit over consumers) or a context- 8. Theoretical and managerial implications
specific stereotype (that Internet advertising should not be trusted
because the Internet contains a lot of fake information) (Friestad 8.1. Theoretical implications
and Wright, 1999). There are inconsistent research findings in re-
lation to the perceived credibility of the message sender on social Persuasion communication theories posit that an intricate web
media. Kumar et al. (2016) offer evidence that firm-generated con- of factors play a role in determining the effectiveness of a per-
tent in social media in the United States has a positive and sig- suasive message (Ajzen, 1991). Informed by these theories, numer-
nificant effect on customers’ behaviour. Using an Australian sam- ous studies examine how various dimensions used by marketers
ple, Sparks et al. (2013) show that, in the tourism industry, firm- to craft and deliver online content separately influence the effec-
generated content is less effective in changing consumers’ attitudes tiveness of digital content marketing. These studies, taken together,
than customer-generated content. report mixed findings, which may be attributed to the fact that a
Our results show that Australian customers generally welcome single dimension, when interacting with another, may lead to dif-
bank-generated content, particularly if it contains both persua- ferent results. In addition, firms offering hedonic products/services
sive and informative content at the same time. Singaporean cus- commonly use social media’s entertainment functions to engage
tomers react favourably to persuasive-only and mixed content cre- customers, but little is known about which tactics are most ap-
ated by banks themselves; however, when it comes to informative- propriate for utilitarian brands such as banks. This study focuses
only content, they prefer it to come from third parties. These re- on how several important dimensions in digital content market-
sults suggest that customers may seek entertainment from per- ing (type, format and source of content) interactively influence
suasive content created by marketers and, hence, generally ap- customer engagement in an under-explored sector, namely, retail
preciate banks’ efforts in producing original and creative con- banking.
tent. However, when it comes to informative content that solely Our findings first throw light on the debate on whether infor-
aims to promote a bank’s offerings, source credibility of the prod- mative or persuasive approaches to digital content creation gen-
uct information dominates customers’ concerns, particularly in erate a stronger response from customers – a particularly impor-
Singapore. tant topic given the scarce and inconclusive literature in this area
(Lee et al., 2018; Dolan et al., 2019; Klassen et al., 2018; de Vries
7.4. Cross-country comparison et al., 2012). Our findings show that, in contrast to other hedonic
or high-identity products for which customers favour persuasive
When comparing the findings from Australia and Singapore, we content, in the highly utilitarian banking sector, persuasive content
find more similarities than differences in terms of customers’ pref- can only engage customers at a superficial level (e.g. when a cus-
erences for types of information and format of the posts. For ex- tomer clicks the ‘like’ icon). Informative content seems to be more
ample, in both countries informative content leads to more active effective in engaging customers at a more cognitively demanding
customer engagement, while persuasive content helps engage cus- and advanced level (e.g. when a customer asks a question or leaves
tomers in a passive way. Customers in both countries value the role a comment).
of videos and photos in helping them process information in an The differentiation between informative and persuasive con-
exciting, sensory and user-friendly way. tent extends our understanding of how the format of a mes-
The only difference between the two countries is customers’ sage helps online persuasive efforts. Although previous studies
perception of the credibility of the information source. While Aus- show the importance of using visuals as a catalyst for engagement
tralians generally react to bank-generated content favourably, Sin- (Villarroel Ordenes et al., 2018), the question of when and where
gaporeans prefer third-party content, particularly for informative to use photos and videos is not clearly answered. Our results offer
content. This finding can be attributed to cultural differences. Re- insights on the use of photos and videos to assist different types of
ferring to Hofstede’s national cultural dimensions (Hofstede in- messages. While videos allow informative-only content to be bet-
sights, n.d.), Singapore and Australia differ substantially in long- ter processed by customers, for posts with persuasive content, pho-
term orientation. Singapore scores 72 for long-term orientation, re- tos enhance customer engagement more than videos. These ten-
flecting cultural qualities that support long-term investment, such dencies are shared by both Australian and Singaporean banking
as perseverance and thrift. This mindset encourages a pragmatic consumers.
and sceptical approach to financial investment. Singaporeans are Empirical findings on the issue of message source and whether
the world’s biggest savers, topping the saving-to-GDP ratio rank- online customers appreciate firm-generated content are inconsis-
ings (World Bank, n.d.). Driven to be thrifty, Singaporeans tend tent (Kumar et al., 2016; Sparks et al., 2013). Our study shows that
to spend and invest wisely, so they are cautious when evaluat- customers favour persuasive content generated by banks, but be-
ing financial advisors and financial institutions. For example, in cause they are concerned about credibility, they prefer informative
a study released in 2018 by the CFA Institute of Singapore, only content generated by third parties. Hence, the differentiation be-
about 10% of respondents agreed that financial advisers put their tween informative and persuasive content in this study helps ex-
clients’ interests first (Heong, 2018). Accordingly, they typically do plain the conflicting findings on the impact of the source of mes-
their due diligence to check if financial planners are independent sage on engagement.

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

12 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

Our cross-country research context allows us to reveal the sim- to provide entertainment value to their customers. However, in
ilarities and differences of customers’ reactions to social media light of Singaporean customers’ sensitivity to the source of mes-
marketing practices. Despite a recent surge in research on elec- sages, we recommend that banks in Singapore, when posting con-
tronic word-of-mouth and digital marketing, there is a surpris- tent about their offerings, consider using more information gen-
ing scarcity of cross-cultural insights in the marketing literature erated by third parties to reduce customers’ doubts and increase
(Banerjee and Chai, 2019). By investigating the leading banks’ Face- credibility.
book marketing in Australia and Singapore, our study shows that
online customers in both countries show similarities in their reac-
tions to banks’ online content in several key aspects. For example,
in both countries, informative content leads to an advanced level of
engagement, while persuasive content leads to superficial engage- 9. Limitations and future research
ment; videos enhance informative content, while photos make per-
suasive content more engaging; and customers prefer persuasive We acknowledge several limitations of this study. It confined
content generated by the banks themselves. The only difference itself to the study of Facebook posts from ten banks and collated
between the two countries is customers’ perception of the credi- only three months of data. Due to the short sampling window, we
bility of the information source: due to their concerns about cred- identified only 66 posts with a combination of persuasive and in-
ibility, Singaporeans prefer informative content that comes from a formative content; by extending the research time span, we could
third party, while Australians do not appear to have such concerns. have located more mixed content. Therefore, we should be cau-
These findings support a universal approach to social media mar- tious when generalising the results. Future studies may expand
keting, with some localization. the dataset by examining more companies’ social media market-
Lastly, our study offers a coding framework, including 13 sub- ing activities over a longer period of time. In addition, we only fo-
categories of content, which can be used as a guide by banks that cused on the retail banking industry in the hope of reflecting the
are planning social media campaigns. uniqueness of social media marketing in high-involvement sectors.
Future research could compare high- and low-involvement sec-
tors and confirm whether a significant difference exists between
8.2. Managerial implications them.
Next, different techniques for deciphering and analysing con-
In practical terms, our findings shed light on best practices tent can be used. For example, content posted by brands in their
for social media marketers in banks operating in Australia and messages could be analysed in terms of framing, such as positive
Singapore. With regard to what to post in utilitarian or high- messaging/negative messaging (Klassen et al., 2018); appeals used
involvement industries like financial services, it is important to in developing the content, including experiential or exclusivity ap-
create posts that educate customers about products and services, peals (Ashey and Tuton, 2014); components of post presentation,
and such complex information should be presented with the assis- such as the quality of the content; or the use of live video and
tance of videos. Such posts will encourage customers to take the emoji. Future research could test broader characteristics and the
central route to process the message, which will help them make presentation components of content, and analyse their impact on
an informed decision. We also recommend that banks in Singapore customer engagement.
create informative and ‘sharable’ posts, such as posts with contests, In addition, persuasive communication theories suggest that
as our sub-content analysis shows that in Singapore, customers are other than on the content type, format and source of the mes-
responsive to these types of posts. In Australia, our recommenda- sage, the effectiveness of persuasion also depends on individual re-
tion is to create more informative posts highlighting banks’ offer- ceivers’ characteristics, such as personality, sex, social status, intel-
ings because customers there seem to appreciate, and to some ex- ligence, involvement and so forth. Any attribute of the audience, or
tent depend on, such information to manage their financial invest- combination of attributes, may provide a context that contributes
ments. to the effectiveness of the message. The explanation for this fac-
At the same time, banks should not ignore the benefits of is- tor is that people make evaluations (judgments) about the con-
suing posts with persuasive content. Consumers do not purchase tent of messages based on their anchors, or stance, on particular
banking services as frequently as they buy fast-moving consumer topic messages (Sherif and Hovland, 1961; Sherif et al., 1965). Fu-
goods, so information related to products and services is not rel- ture research could incorporate online customers’ characteristics to
evant to them all the time. Although persuasive content does not develop a more comprehensive understanding of how firms’ con-
facilitate the central route of decision-making, it may help a bank tent marketing strategies affect different customers.
to provide the necessary level of online presence and maintain Lastly, our choice of two countries is based on the fact that
the interest of customers. In addition, the findings suggest that these are two countries with very different cultural values, follow-
banks are under-utilising certain types of persuasive content and ing classic Hofstede’s cultural dimensions (Hofstede et al., 2010).
should strive to use approaches that engage, entertain and serve But at the same time, we are aware of the convergence in con-
the needs of users. Banks in both Singapore and Australia can boost sumer behaviour due to globalisation. This study hence aims to
the popularity of their posts by fine-tuning emotional categories, provide some preliminary empirical evidence on the extent to
such as storytelling, a type of content that connects with peo- which social media marketing practices are universal or influenced
ple on a human level. They can also generate more entertaining by national cultural values. However, we believe more could be
quizzes, polls and games to drive engagement (Pulizzi and Bar- done to deepen our understanding of how the cultural differences
rett, 2010), particularly for user segments such as ‘entertainment influence content marketing.
chasers’ (Hodis et al., 2015). Given the positive responses to banks’
posts in community and employee-oriented campaigns, we concur
with other scholars who recommend that brands publicise their
CSR programmes on Facebook to enhance their brand image (Sitta
et al., 2018). Appendix 1
Our findings tally with Chaffey and Smith (2013), who recom-
mend that banks should generate their own persuasive content Literature on Content Marketing

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 13

Area Findings Author (s) Research method Theoretical concepts,

models or frameworks

Definitions of content Content marketing is “a strategic brand storytelling du Plessis Conceptual and Academic definition of
marketing technique aimed at changing consumers’ inactive (2015). empirical. content marketing.
behaviour through unobtrusive, engaging brand Qualitative analysis.
conversations in earned media”. Content analysis of
Content has six key elements: medium; strategic; literature.
formation; intrinsic, communication, corollary.
Content marketing “requires a cultural change from Holliman & Conceptual and Empirically grounded
“selling” to “helping”, which in turn requires Rowley (2014, empirical. definition of digital
different marketing objectives, tactics, metrics and p.269). Qualitative research. content marketing
skills to those associated with more traditional Semi-structured
marketing approaches”. interviews.
Focus on B2B content.
“B2B digital content marketing involves creating,
distributing and sharing relevant, compelling and
timely content to engage customers at the
appropriate point in their buying consideration
process, such that it encourages them to convert to
a business building outcome,”
Three types of content: ‘not paid for’, ‘paid for’ and
social digital content.
Digital content marketing (DCM) denotes the Hollebeek & Conceptual. Conceptual framework of
creation and dissemination of relevant, valuable Macky (2019) Literature review. DCM.
brand-related content to current or prospective Uses and Gratification
customers on digital platforms to develop their (U&G). Functional and
favourable brand engagement, trust, and Hedonic Motives.
relationships. Brand-related
DCM has cognitive, emotional and behavioural sense-making; brand
elements. identification, citizenship
behaviours, brand trust
and brand attitude.
Unique characteristics of Content marketing relies on the principle of Godin (1999). Conceptual Permission marketing or
digital content marketing consumer consent, permission or opt-in. customer-initiated
and implications for Permission is a way to enhance targeting and targeting.
marketing practice and improve the customer database.
theory. Incentives help secure permission.
Digital content is defined as: ‘bit-based objects Rowley (2008) Conceptual. Digital content supports
distributed through electronic channels’. It has Literature review. value creation and the
nine (9) unique characteristics, value is contextual; “value-in-use” concept.
reproducibility and multiplicability; interactivity;
repackageability; delivery and technology;
perishability; homogeneity and tangibility.
Mechanism Consumer engagement with a website is a Calder et al. Empirical. Customer Engagement,
collection of experiences with the site. A total of (2009). Quantitative Uses and Gratification
12 experiences are identified: stimulation; social research. Survey. (U&G) theory.
facilitation; temporal; self-esteem and civic
mindedness; intrinsic enjoyment; utilitarian;
participation and socializing; community; personal
engagement’ social-interactive engagement. Two
main types of online engagement were identified,
personal and social-interactive, which increases
advertising effectiveness.
Interactivity is seen as the hallmark of digital Deighton & Conceptual. Interactive practices
media which, in turn, is central to giving meaning Kornfeld (2009).
to a brand and facilitating people’s identity
projects.
Five roles for the interactive consumer: thought
tracing; activity tracing; property exchanges, social,
and cultural exchanges.
Significant gaps exist between what business think Heller Baird & Empirical. Practical orientated
consumers want from social media and what Parasnis (2011). Quantitative
consumers actually want from their social media research. Online
interactions with companies. Motives for customer surveys.
engagement include purchasing and receiving
incentives.
Four distinct segments of Facebook users are Hodis et al. Empirical. Segmentation.
identified: attention seekers; devotees; connection (2015). Qualitative research Customer engagement.
seekers and entertainment chasers. Content Focus groups and a
marketing entails building Facebook brand survey.
communities.
Digital content marketing can generate social Bu et al. (2020) Conceptual and Uses and Gratification
impact, which in turn, leads to e-WOM. Positive empirical. theory
associations are found between content Quantitative Social impact theory.
entertainment and informational social impact, and research: survey of
social impact mediates the relationship between 707 Chinese tourists.
digital content marketing and e-WOM.

(continued)

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

14 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

Area Findings Author (s) Research method Theoretical concepts,

models or frameworks
Features of digital Analysis of the factors that drive brand post popularity. De Vries et al. Empirical. Persuasive message
content that generate Key findings are that vividness, position of a post and (2012). Quantitative research. strategy
customer engagement interactivity enhance the number of likes. Entertainment Content analysis of 355
behaviours. content has a marginally negative relationship with brand posts from 11
number of likes marginally, while information content has international brands
no significant relationship. over nine months.
Sector: six product categories, such as cosmetics, alcoholic
beverages, mobile phones, leisure, accessories and food.
Tweets with interactivity features (such as hashtags with Burton et al. Empirical. Persuasive message
popular topics), along with positive stimuli (i.e. parties, (2013) Mixed methods. A total strategy
music and lifestyle content) are more likely to generate of 200 tweets analysed.
engagement than those without those features.
Sector: alcohol brands and advocates of safe drinking
Increasing post frequency is correlated with more page Houk & Empirical. Persuasive message
likes and user engagement. Multi-media posts draw the Thornhill (2013) Quantitative research. strategy.
most interest from fans. Content analysis of
Sector: University library. Facebook posts over 17
months.
Analysis of creative strategies, including appeals, Ashey & Tuton Empirical. Persuasive message
user-generated content and sales promotion, used by top (2015). Content analysis of the strategies.
brands on social media. top 100 brands, with
Brands use multiple appeals and experiential, image and one week of content
exclusivity appeals work well. posted to Facebook,
Incentives, frequent updates and fresh content are Twitter, MySpace,
important for participation. forums and blogs;
video and photo
content, resulting in
996 pages/
accounts/channels.
Persuasive messages (i.e., argument quality, post popularity Chang et al. Empirical. Persuasive message
and post attractiveness) leads to customer engagement (2015). Quantitative research. strategy.
behaviours and post popularity works through the central Survey of 392 fans from Elaboration Likelihood
and peripheral route in the ELM model. Facebook. Model.
Sector: cooking websites
Post frequency increases customer engagement, along with Veale et al. Empirical. Persuasive message
individualised interaction, encouraging conversation, (2015). Mixed methods. strategy
celebrity involvement and multi-media content. Analysis of 100
Sector: health promotion Facebook and Twitter
profiles.
Frequency of updates (4 posts a week) is crucial to Brech et al. Empirical Self-presentation theory.
individual fan engagement and longer posts receive more (2017) Content analysis of 159
interactivity. Facebook brand pages.
Sector: University.
Analysis of post features, such as timing, popularity (likes), Pinto & Yagnik Empirical. Persuasive message
engagement (comments, shares, and hastags), format of (2017). Quantitative research. strategy
post (i.e. text, video), elements included (i.e. models used), Content analysis of 421
advertising appeal (i.e. rational or emotional), marketing Facebook posts over
promotions and selling points. Primary message appeal four months.
was emotional and there was no significant difference
shown for likes or comments across message appeals.
Sector: fitness industry (fitness tracker brands).
Post content, post format and message strategy were Klassen et al. Empirical. Persuasive message
analysed and positive emotion-inducing strategies lead to (2018). Mixed methods, strategy.
engagement. Different strategies were most effective for exploratory design.
Facebook and Instagram. Content analysis of the
Sector: health promotion, food industry and lifestyle top 10 food and
brands lifestyle brands, and a
total of 227 Facebook
and Instagram posts
over a 30 day period
were analysed.
Persuasive content, such as emotional and philanthropic Lee et al. Empirical. Persuasive message
content, increases engagement with a message. (2018). Quantitative research. strategy.
Informative content, in isolation reduces engagement, but Content coding of
increases engagement when combined with persuasive 106,316 Facebook
attributes. Results are moderated by industry type (i.e. messages.
celebrity pages or type of company)
Sector: broad industry categories
Analysis of message content and impacts of assertive, Villarroel Or- Empirical. Speech act theory
expressive and directive messages on consumer sharing. denes et al. Content analysis of
Visuals, and rhetorical styles that facilitate social (2018) 12,374 Facebook posts
interactions, are important for message sharing. and 29,413 tweets by
Sector: food, retailing, manufacturing and hospitality. eight brands.
Rational appeals in social media have a superior effect in Dolan et al. Quantitative research. Persuasive message
generating active and passive engagement, whereas (2019). Content analysis of strategy.
emotional appeals facilitate passive rather than active 2236 Facebook posts. Uses and gratification
engagement behaviour. theory.
Sector: wine industry. Dual processing theory.

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 15

Appendix 2

Profiles of the banks used in this study

Residential Headquarter Established Total assetsa Facebook Instagram Twitter YouTube

country followersb followersb followersb subscribersb

SG bank 1 Singapore Singapore 1968 403.05 billion SGD 904,017 11,968 24,547 17,318
SG bank 2 Singapore Singapore 1932 467.5 billion SGD 656,834 4514 18.3K 2.477K
SG bank 3 Singapore Singapore 1935 388 billion SGD 520,622 null null 6595
SG bank 4 Singapore London 1865 (Established operation 2558.1 billion USD 2605,810 24,602 8320 22,499
in Singapore in 1877)
SG bank 5 Singapore New York 1812 (Established operation 1917 billion USD 138,555 50.6K 389.5K 1.752K
in Singapore in 1902)
AU Bank 1 Australia Sydney 1911 975 billion AUD 750,250 19,097 87,546 12,740
AU Bank 2 Australia Dockland 1982 806.5 billion AUD 227,639 4623 67,311 7656
AU Bank 3 Australia Melbourne 1835 889.9 billion AUD 357,191 14,244 127,394 10,217
AU Bank 4 Australia Sydney 1817 879.6 billion AUD 242,472 10.5k 57.3k Not disclose
AU Bank 5 Australia Brisbane 1863 50.212 billion AUD 39,674 1618 10.2k 576
a
as of 2018.
b
as of Jan 1st 2019.

Appendix 3

Classification of Post Content.

Level 1
categories Level 2 categories Level 3 categories Examples

Informative Product 1. Product information 10% discount on purchase by using a banking service
content
Incentive 2. Lucky draw Congratulations post for a lucky winner
3. Contest Ask me anything with the ‘Selena Ling contest’
4. Share-based benefits (share a comment) Win $10 – the person who gets the highest number of
likes by sharing a post
Information 5. Basic financial information Exchange rates and interest rates
6. System-related information Apologies to customers who are not able to connect with
the call centres or perform mobile banking tasks
Persuasive Pathos: Philanthropic 7. Simply introducing banks’ involvement into CSR Homage, a social enterprise supported by a Singaporean
content efforts and corporate campaign to support local community bank
social responsibility (CSR)
strategies.
8. Simply introducing banks’ involvement into CSR Staff training to ensure banking staff are ready for the
campaign about employees jobs of the future
9. Inspirational stories / emotion evoking stories Interview with a successful employee / community
about employees, community, etc. It could arise member who talks about the keys to success and how
from banks’ CSR campaign (as in sub-category 7 or they handled the challenges they experienced in business
8), but presented as a touching story)
10. Inspirational quotes Quote: “Money can’t buy happiness”
Logos: General facts not 11. General facts not directly related to the General facts related to finance: Tips on how to avoid
directly related to the company/products scams or spot a hoax.
company/products Tips for better financial habits, such as saving for the
future
US-China trade talks
General facts not related to finance: zero waste life:
upcycle watches made from soda can bottoms to natural
cleaner made from citrus peel waste (Recycle more Waste
less)
Dining Advisor
Rise in Popularity of Home Renovations In Australia
8 things to know about Ang Baos (red packet)
Seasonal 12. Seasonal greetings Happy New Year
greetings/Entertainment
13. Puzzles, games Game: find words

References
American Bankers Association (2017). The State of Social Media in Bank-
ing: Results of an American Bankers Association Research Study. Washing-
Abdul-Ghani, E., Hyde, K.F., Marshall, R., 2019. Conceptualising engagement in a con-
ton, D.C.Retrieved 21 July 2019 fromhttps://www.aba.com/Products/Endorsed/
sumer-to-consumer context. Aust. Market. J. (AMJ) 27 (1), 2–13.
Documents/ABASocialMedia_Report.pdf.
Ajzen, I., 1991. The theory of planned behavior. Organ. Behav. Hum. Decis. Process.
Anderson, S.P., Renault, R., 2006. Advertising content. Am. Econ. Rev. 96 (1), 93–113.
50 (2), 179–211.
Areni, C.S., 2003. The effects of structural and grammatical variables on persuasion:
Alalwan, A.A., Rana, N.P., Dwivedi, Y.K., Algharabat, R., 2017. Social media in mar-
an elaboration likelihood model perspective. Psychol. Market. 20 (4), 349–375.
keting: a review and analysis of the existing literature. Telem. Inf. 34 (7),
Ashey, C., Tuton, T., 2014. Creative strategies in social media marketing: an ex-
1177–1190.
ploratory study of branded social content and consumer engagement. Psychol.
Aldlaigan, A.H., Buttle, F.A., 2001. Consumer involvement in financial services: an
Market. 32 (1), 15–27.
empirical test of two measures. Int. J.Bank Market. 19 (6), 232–245.

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

16 P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx

Ashley, C., Tuten, T., 2015. Creative strategies in social media marketing: an ex- Fredrickson, B.L., 2001. The role of positive emotions in positive psychology: the
ploratory study of branded social content and consumer engagement. Psychol. broaden-and-build theory of positive emotions. Am. Psychol. 56 (3), 218.
Market. 32 (1), 15–27. Friestad, M., Wright, P., 1999. Everyday persuasion knowledge. Psychol. Market. 16
Austrade - Australian Trade and Investment Commission. (2016). In- (2), 185–194.
vestment opportunities in digital technologies in Australia. Re- Gamboa, A.M., Goncalves, H.M., 2014. Customer loyalty through social networks:
trieved 29 September 2019 from file:///C:/Users/jc203156/Downloads/ lessons from Zara on Facebook. Bus. Horiz. 57 (6), 709–717.
Investment- opportunities- in- Australian- Digital- technologies.pdf. Godin, S., 1999. Permission Marketing: Turning strangers Into Friends and Friends
Australian Competition & Consumer Commission (ACCC) (n.d). Protecting yourself Into Customers. Simon and Schuster, New York.
from scams. Retrieved 20 July 2019 fromhttps://www.accc.gov.au/consumers/ Grott, E.M., Cambra-Fierro, J., Perez, L., Yani-de-Soriano, M., 2019. How cross-culture
consumer-protection/protecting- yourself- from- scams. affects the outcomes of co-creation. Eur. Bus. Rev. 31 (4), 544–566.
Babin, B.J., Darden, W.R., Griffin, M., 1994. Work and/or fun: measuring hedonic and Harris, J. (2017). Cover All the Bases with 21 Winning Content Marketing Tech-
utilitarian shopping value. J. Consum. Res. 20, 644–656. niques,
Bakker, A.B., 1999. Persuasive communication about AIDS prevention: need for Hashmeta (2018). Social Media Landscape in Singapore. Retrieved 19 July 2019 from
cognition determines the impact of message format. AIDS Educ. Prev. 11 (2), https://hashmeta.com/blog/social- media- landscape- in- singapore- 2019/.
150–162. Heller Baird, C., Parasnis, G., 2011. From social media to social customer relationship
Banerjee, S., Chai, L., 2019. Effect of Individualism on Online User Ratings: theory management. Strategy & Leadership 39 (5), 30–37.
and Evidence. J. Global Market. doi:10.1080/08911762.2018.1549690. Heong, S.K. (2018). Discussion: How Much Can Singaporeans Really Trust Their Fi-
Berger, S.C., Messerschmidt, C.M., 2009. Babbling before banking? Online communi- nancial Adviser? Retrieved 23 February 2020 from https://dollarsandsense.sg/
ties and pre-purchase information seeking. Int. J. Bank Market. 27 (6), 446–466. discussion- can- singaporeans- really- trust- financial- adviser/.
Bokek, E., Tversky, B., 2016. Creating visual explanations improves learning. Cognit. Hirschmann, E.C., Holbrook, M.B., 1982. Hedonic consumption: emerging concepts,
Res. Princip. Implic. 1 (27). https://doi.org/10.1186/s41235- 016- 0031- 6. methods and propositions. J. Mark. 46 (3), 92–101.
Brech, F., Messer, U., Schee, B.V., Rauschnabel, P.A., Ivens, B., 2017. Engaging fans and Hodis, M.A., Sriramachandramurthy, R., Sashittal, H.C., 2015. Interact with me on
the community: examining factors related to interaction with institutions of my terms: a four segment Facebook engagement framework for marketers. J.
higher education. J. Market. High. Educ.. https://doi.org/10.1080/08841241.2016. Market. Manag. 31 (11–12), 1255–1284.
1219803. Hofstede Insights, n.d. Retrieved 23 March 2020 from https://www.hofstede-
Bu, Y., Parkinson, J., Thaichon, P., 2020. Digital content marketing as a catalyst for insights.com/country- comparison/australia, singapore/.
e-WOM in food tourism. Aust. Market. J. (AMJ). https://doi.org/10.1016/j.ausmj. Hofstede, G., Hofstede, G.J., Minkov, M., 2010. Cultures and Organizations: Software
2020.01.001. of the Mind, 3rd Ed. McGraw-Hill, New York.
Bullas (n.d). 10 Powerful Tips to Increase Fan Engagement on Facebook. Re- Hollebeek, L.D., Macky, K., 2019. Digital content marketing’s role in fostering con-
trieved March 1 st, 2020 from https://www.jeffbullas.com/10- powerful- tips- sumer engagement, trust, and value: framework, fundamental propositions, and
to-increase-fan-engagement-on-facebook/. implications. J. Interact. Market. 45, 27–41.
Burton, S., Dadich, A., Soboleva, A, 2013. Competing voices: marketing and coun- Hollebeek, L.D., 2013. The customer engagement/value interface: an exploratory in-
ter-marketing alcohol on twitter. J. Non-Profit Public Sector Market. 25 (2), vestigation. Aust. Market. J. 21 (1), 17–24.
186–209. Holliman, G., Rowley, J., 2014. Business to business digital content marketing: mar-
Calder, B.J., Malthouse, E.C., Schaedel, U., 2009. An experimental study of the rela- keters’ perceptions of best practice. J. Res. Interact. Market. 8 (4), 269–293.
tionship between online engagement and advertising effectiveness. J. Interact. Houk, K.M., Thornhill, K., 2013. Using Facebook page insights data to determine
Market. 23 (4), 321–331. posting best practices in an academic health sciences library. J. Web Librari-
Chaffey, D., Ellis-Chadwick, F., 2019. Digital Marketing. Pearson, UK. anship 7 (4), 372–388.
Chaffey, D., Smith, P.R., 2013. Emarketing Excellence: Planning and Optimizing Your House, R.J., Hanges, R.J., Javidan, M., Dorfman, R.W., Gupta, V., 2004. Culture, Lead-
Digital Marketing, 4 th ed Routledge, London. ership, and Organisations: The GLOBE Study of 62 Societies. Sage, CA Thousand
Chang, Y.T., Yu, H., Lu, H.P., 2015. Persuasive messages, popularity cohesion, and Oaks.
message diffusion in social media marketing. J. Bus. Res. 68 (4), 777–782. Hovland, C.I., Weiss, W., 1951. The Influence of Source Credibility on Communication
Charello, M., 2017. Essentials of Social Media Marketing. Idaho Falls, Stukent. Effectiveness. Public Opin. Q. 15 (4), 635–650.
Cheong, H.J., Morrison, M.A., 2008. Consumers’ reliance on product information and Johnson, T.J., Kaye, B.K., 2004. Wag the blog: how reliance on traditional media and
recommendations found in UGC. J. Interact. Advert. 8 (2), 38–49. the internet influence credibility perceptions of weblogs among blog users. J.
Chiu, Y.L., Chen, K.H., Wang, J.N., Hsu, Y.T., 2019. The impact of online movie word- Mass Commun. Q. 81 (3), 622–642.
of-mouth on consumer choice. Int. Market. Rev. doi:10.1108/IMR- 06- 2018- 0190. Jones, M.A., Reynolds, K., Arnold, M.J., 2006. Hedonic and utilitarian shopping value:
Chu, S.C., Kamal, S., 2008. The effect of perceived blogger credibility and argument investigating differential effects on retail outcomes. J. Bus. Res. 59 (9), 974–981.
quality on message elaboration and brand attitudes: an exploratory study. J. In- Kelman, H.C., 1961. Processes of opinion change. Public Opin. Q. 25, 57–78.
teract. Advert. 8 (2), 26–37. Klassen, K.M., Borleis, E.S., Brennan, L., Reid, M., McCaffrey, T.A., Lim, M.S., 2018.
Colgate, M., Stewart, K., 1998. The Challenge of Relationship in Services - A New What people “like”: analysis of social media strategies used by food industry
Zealand Study. Int. J. Serv. Ind. Manag. 9 (5), 454–468. brands, lifestyle brands, and health promotion organizations on Facebook and
Cummins, S., Peltier, J.W., Dixon, A., Niedermeier, K.E., Wang, E., Zhang, X., 2016. The Instagram. J. Med. Internet Res. 20 (6), e10227.
use of social media among business-to-business sales professionals in China. J. Koiso-Kanttila, N., 2004. Digital content marketing. J. Market. Manag. 20 (1–2),
Res. Interact.Market. 10 (1), 33–49. 45–65.
Dahl, S., 2018. Social Media Marketing: Theories and Applications. Sage, Los Angeles. Kumar, A., Bezawada, R., Rishika, R., Janakiraman, R., Kannan, P., 2016. From social
Dainton, M., 2005. Applying Communication Theory For Professional Life: A Practi- to sale: the effects of firm-generated content in social media on customer be-
cal Introduction. SAGE Publications, Thousand Oaks, Calif. havior. J. Mark. 80, 7–25.
de Mooij, M., 2017. Comparing dimensions of national culture for secondary analy- Laksamana, P., 2018. Impact of social media marketing on purchase intention and
sis of consumer behaviour data of different countries. Int. Market. Rev. 34 (3), brand loyalty: evidence from Indonesia’s banking industry. Int. Rev. Manag. Mar-
444–456. ket. 8 (1), 13–18.
de Ruyter, K., Keeling, D.I., Ngo, L.V., 2018. When nothing is what it seems: a digital Lee, D., Hosanagar, K., Nair, H.S., 2018. Advertising content and consumer engage-
marketing research agenda. Aust. Market. J. (AMJ) 26 (3), 199–203. ment on social media: evidence from Facebook. Manage. Sci. 64 (11), 5105–5131.
de Vries, L., Gensler, S., Leeflan, P.S., 2012. Popularity of brand posts on brand fan Malthouse, E.C., Haenlein, M., Skiera, B., Wege, E., Zhang, M., 2013. Managing cus-
pages: an investigations of the effects of social media marketing. J. Interact. tomer relationships in the social media era: introducing the social CRM house.
Market. . 26 (2), 83–91. J. Interact. Market. 27 (4), 270–280.
Deighton, John, Kornfeld, L, 2009. Interactivity’s unanticipated consequences for McCroskey, J.C., Teven, J.J., 1999. Goodwill: a reexamination of the construct and its
marketers and marketing. J. Interact. Market. 23 (1), 4–10. measurement. Commun. Monogr. 66 (1), 90–103.
Deloitte (2018). Flying through Turbulence. Retrieved 29 September 2019, McGuire, W.J., 1969. The nature of attitudes and attitude change. In: Lindzey, G.,
fromhttps://www2.deloitte.com/au/en/pages/financial-services/articles/ Aronson, E. (Eds.), The Handbook of Social Psychology, 3. Addison-Wesley, Read-
flying- through- turbulence.html#. ing, MA, pp. 136–314.
Dhar, R., Wertenbroch, K., 20 0 0. Consumer choice between hedonic and utilitarian McMillan, S.J., Hwang, J.S., Lee, G., 2003. Effects of structural and perceptual factors
goods. J. Market. Res. 37 (1), 60–71. on attitudes toward the website. J. Advert. Res. 43 (4), 400–409.
Dolan, R., Conduit, J., Frethey-Bentham, C., Fahy, J., Goodman, S., 2019. Social me- Merz, M.A., He, Y., Alden, D.L., 2008. A categorization approach to analysing the
dia engagement behavior: a framework for engaging customers through social global consumer culture debate. Int. Market. Rev. 25 (2), 166–182.
media content. Eur. J. Mark. 53 (10), 2213–2243. Mitic, M., Kapoulas, A., 2012. Understanding the role of social media in bank mar-
du Plessis, C., 2015. An exploratory analysis of essential elements of content mar- keting. Market. Intell. Plann. 30 (7), 668–686.
keting. In: Proceedings of the Second European Conference on Social Media, Mogaji, E., Danbury, A., 2017. Making the brand appealing: advertising strategies and
pp. 122–129. consumers’ attitude towards UK retail bank brands. J. Prod. Brand Manag. 26 (6),
Durkin, M., Mulholland, G., McCartan, A., 2015. A socio-technical perspective on 531–544.
social media adoption: a case from retail banking. Int. J. Bank Market. 33 (7), Muntinga, D.G., Moorman, M., Smit, E.G., 2011. Introducing COBRAs. Int. J. Advert.
944–962. 30 (1), 13–46.
Festinger, L., 1957. A Theory of Cognitive Dissonance. Stanford University Press. Nguyen, B., Chang, K., Simkin, L., 2014. Customer engagement planning emerg-
Fisher, W.R., 1984. The Narrative Paradigm: in the Beginning. J. Commun. 34 (1), ing from the “individualist-collectivist”-framework. Market.Intell. Plann. 32 (1),
74–89. 41–65.

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008
JID: AMJ
ARTICLE IN PRESS [m5G;June 7, 2020;15:16]

P. Wang and B. McCarthy / Australasian Marketing Journal xxx (xxxx) xxx 17

O’Keefe, D.J., 1990. Persuasion: Theory and Research. Sage Publications, Newbury Schulze, C., Schöler, L., Skiera, B., 2014. Not all fun and games: viral marketing for
Park, CA. utilitarian products. J. Mark. 78 (1), 1–19.
Overby, J.W., Lee, E.J., 2006. The effects of utilitarian and hedonic online shopping Sensis (2018). Yellow Social Media Report 2018. Melbourne, Victoria. Re-
value on consumer preference and intentions. J. Bus. Res. 59 (10–11), 1160–1166. trieved 23 July 2019 fromhttps://www.sensis.com.au/about/our-reports/
Ozdora-Aksak, E., Atakan-Duman, S., 2015. The online presence of Turkish banks: sensis- social- media- report.
communicating the softer side of corporate identity. Public Relat. Rev. 41 (1), Sherif, C.W., Sherif, M., Nebergall, R.E., 1965. Attitude and Attitude Change: The So-
119–128. cial Judgment-involvement Approach rpt. 1981. Greenwood Press, Westport, CT.
Park, H.H., Jeon, J.O., 2018. The impact of mixed eWOM sequence on brand attitude Sherif, M., Hovland, C.I., 1961. Social Judgment: Assimilation and Contrast Effects in
change: cross-cultural differences. Int. Market. Rev. 35 (3), 390–411. Communication and Attitude Change. Yale University Press.
Peppers, D., Rogers, M., 2011. Managing Customer Relationships: A Strategic Frame- Sitta, D., Faulkner, M., Stern, P., 2018. What can the brand manager expect from
work, 2 nd ed Wiley, Hoboken, NJ. Facebook? Aust. Market. J. (AMJ) 26 (1), 17–22.
Petty, R.E., Cacioppo, J.T., 2012. Communication and Persuasion: Central and Periph- Smart Nation (2018). Transforming Singapore. Retrieved 29 September 2019
eral Routes to Attitude Change. Springer Science & Business Media. fromhttps://www.smartnation.sg/why- Smart- Nation/transforming- singapore.
Petty, R.E., Cacioppo, J.T., Goldman, R., 1981. Personal involvement as a determinant Smieško, I., 2016. Criminal liability of Facebook reaction buttons in environment of
of argument-based persuasion. J. Pers. Soc. Psychol. 41 (5), 847–855. https://doi. the Slovak Republic as a form of hate speech. Societas Et Iurisprudentia 4, 213.
org/10.1037/0022-3514.41.5.847. Sparks, B.A., Perkins, H., Buckley, R., 2013. Online travel reviews as persuasive com-
Petty, R.E., Cacioppo, J.T., 1986. The elaboration likelihood model of persuasion. Adv. munication: the effects of content type, source, and certification logos on con-
Exp. Soc. Psychol. 19, 123–205. sumer behaviour. Tourism Manag. 39, 1–9.
Pinto, M.B., Yagnik, A., 2017. Fit for life: a content analysis of fitness tracker brands StatCounter (2019a). Social Media Stats Australia. Retrieved 22 July 2019 fromhttp:
use of Facebook in social media marketing. J. Brand Manag. 24 (1), 49–67. //gs.statcounter.com/social- media- stats/all/australia.
Pulizzi, J., Barrett, T., 2010. Get Content, Get Customers. McGraw-Hill, Columbus, OH. StatCounter (2019b) Social Media Stats Singapore. Retrieved 22 July 2019 fromhttp:
Rancati, E., Gordini, N., Capatina, A., 2016. Conceptualizing and measuring content //gs.statcounter.com/social- media- stats/all/singapore.
marketing in luxury firms: an exploratory analysis. In: Global Marketing Strate- Tsiotsou, R.H., 2019. Rate my firm: cultural differences in service evaluations. J. Serv.
gies For the Promotion of Luxury Goods. IGI Global, pp. 109–132. Mark. 33 (7), 815–836.
Resnik, A., Stern, B.L., 1977. An analysis of information content in television ad- Veale, H.J., Sacks-Davis, R., Weaver, E.R., Pedrana, A.E., Stoové, M.A., Hellard, M.E.,
vertising. J. Mark. 41 (1), 50–53. Retrieved 28 September2019 from http: 2015. The use of social networking platforms for sexual health promotion: iden-
//contentmarketinginstitute.com/2017/05/winning-content-marketing-tactics/. tifying key strategies for successful user engagement. BMC Public Health 15 (1),
Rinaldi, C., 2015. Financial Marketer’s Guide to the Top 5 social Media platforms. 85.
The Financial Brand Retrieved 2 April 2020 from https://thefinancialbrand.com/ Vejačka, M., 2017. Social Media Marketing in Comparison With Other Forms of Mar-
51381/guide- to- social- media- in- banking/. keting in the Slovak Banking Sector. Market/Tržište. 29 (1), 23–38.
Rissanen, H., Luoma-Aho, V., 2016. (Un) willing to engage? First look at the engage- Villarroel Ordenes, F., Grewal, D., Ludwig, S., Ruyter, K.D., Mahr, D., Wetzels, M.,
ment types of millennials. Corpor. Commun. Int. J. 21 (4), 500–515. 2018. Cutting through content clutter: how speech and image acts drive
Roig, J.C.F., Garcia, J.S., Tena, M.A.M., Monzonis, L.J., 2006. Customer perceived value consumer sharing of social media brand messages. J. Consum. Res. 45 (5),
in banking services. Int. J. Bank Market. 24 (5), 266–283. 988–1012.
Rowley, J., 2008. Understanding digital content marketing. J. Market. Manag. 24 World Bank (n.d). Gross savings (% of GDP). Retrieved 3 April 2020 from https://
(5–6), 517–540. data.worldbank.org/indicator/ny.gns.ictr.zs.

Please cite this article as: P. Wang and B. McCarthy, What do people “like” on Facebook? Content marketing strategies used by retail
bank brands in Australia and Singapore, Australasian Marketing Journal, https://doi.org/10.1016/j.ausmj.2020.04.008