Professional Documents
Culture Documents
a
Griffith Business School, Griffith University, Australia
b
Business School, Sun Yat-sen University, Guangzhou, China
Keywords: Drawing on social identification theory and motivation of expectancy, the current study examines the re-
Social identification lationship between social identification, customer engagement and purchase intention in the context of online
Customer engagement communities. Social identification is operationalised into person-to-person and person-to-community identifi-
Purchase cations, whereas customer engagement into attitudinal and behavioural engagement. This study was undertaken
Online community
with members from two of the most popular online communities in China. The study confirms that identifying
with other members of the online community is significantly related to identification with the community; both
identifications have a significant impact on customer engagement, which can lead to purchase intention.
☆
Note: This project is funded by The National Natural Science Foundation of China (71272194, 71772186).
⁎
Corresponding authors.
E-mail addresses: cathyjournalarticles@gmail.com (C. Prentice), hanxy@mail.sysu.edu.cn (X.Y. Han).
https://doi.org/10.1016/j.jretconser.2018.12.014
Received 30 September 2018; Received in revised form 11 November 2018; Accepted 27 December 2018
0969-6989/ © 2018 Elsevier Ltd. All rights reserved.
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
2018; Prentice and Loureiro, 2018; Vivek et al., 2018; van Doorn et al., 2. Literature review
2010). The psychological conditions are more reflective of customer
discretion in engaging with the firm or brand, and hence require 2.1. Social identification in an online community
minimal or no financial or other business-oriented involvement from
the firm (Prentice and Loureiro, 2018). Prentice et al. (2018) refer to An online community is a social network which allows individuals
such discretionary engagement as an organic approach to customer with similar interests in a business, a brand or the brand associated
engagement. These organic antecedents provide businesses with cost- products to virtually interact online with one another through a specific
effective means to promote customer engagement and achieve financial social media platform (Rheingold, 1993). The community enables
advantages. members to post views and opinions, comment on discussions, and offer
The current study contends that customers’ social identification can or receive advice. An online community generally involves three parties
be an organic and customer-based antecedent and examines how such (the company, customers and other members) (Muniz and O'Guinn,
identification drives customers to volitionally engage with the business- 2001). However, the company often plays a minor role in members’
related communities and affect their subsequent purchase. This con- online activities. Members opt to interact with each other, instead of the
tention is consistent with Vroom's (1964) expectancy theory of moti- company which is motivated by commercial interests. Through regular
vation. The theory indicates that individuals are motivated to engage in online interactions, the community members become a virtual family or
certain behaviours or actions, for the prospect of generating desired invisible friends (Ridings et al., 2002). Gradually the member develops
outcomes. In this respect, social identity can be an intrinsic motivator. a sense of belongingness to, and social identification with, the online
Social identity theory indicates that individuals tend to classify them- community (Algesheimer et al., 2005a, 2005b).
selves and others into various social categories and associate them- Social identification is fundamentally reflective of one's desire for
selves with a group that they would like to be identified with (Ashforth social identity and affiliation with a consciously chosen social group.
and Mael, 1989). Individuals may have a high level of social identity Social identity refers to an individual's membership, emotional
when they are aware of their membership in a group and the mem- bonding, and perceived value in that social group (Tajfel, 1978). Social
bership is significant. In an online brand community, customers who identity theory indicates that individuals develop identification through
are identified with a brand and its associated communities are more classifying themselves and others into various social categories, thereby
likely to engage with the brand and community through interacting and engendering social comparison and positive distinction (Zhang and
communicating with other members of the community (Prentice et al., Zuo, 2006). Given this relational and comparative nature, individuals
2018) since social identity reinforces belongingness to and having af- seek to identify themselves in a social group to enhance their self-es-
filiation to a chosen group or community (Trepte, 2006). These custo- teem (Abrams and Hogg, 1988). An individual who identifies with a
mers are more likely to develop significant relationships with other social group tends to choose activities congruent with salient aspects of
customers and turn to those sites as a source of information and support their identities and perceive themselves as psychologically intertwined
(Hajli et al., 2014). Such engagement is likely to lead to brand purchase with the fate of the group (Ashforth and Mael, 1989).
and loyalty, as the process of receiving and offering information and Research on social identity is primarily focused on individual
support reinforces customers’ understanding and connection with the identification with an organization, a community, or any social group.
brand. The longer the customer participates in the community, the For instance, identification with an organization enhances support for
more likely he or she is identified with the brand and the community. and commitment to the organization. Strong identification motivates
Consistent with the foregoing discussion, this study draws upon van members to exert additional effort toward tasks that contribute to co-
Doorn et al. (2010) and Prentice and Loureiro's (2018) conceptual workers and to the organization (Ashforth and Mael, 1989; Dutton
framework with a focus on the customer-based engagement model, and et al., 1994). Very few studies have attempted to understand how an
adopts Prentice et al. (2018) organic engagement concept to investigate individual's identification with other members of community may affect
how customers’ social identification with an online brand community, their identification with the community. In this study we refer to the
influences their engagement with the brand and affects subsequent former as customer-to-customer identification (CCI); the latter as cus-
brand purchases. In particular, the study examines how the length of a tomer-to-community identification (CMI). Most individuals are in-
customer’ membership with a chosen online brand community affects volved with the online community to communicate with other members
their identification and engagement with the community. The proposed and to seek or offer information and support (Armstrong and Hagel,
relationships are shown in Fig. 1. Understanding these relationships 1996; Prentice et al., 2018). Information from other members are per-
provides researchers and business practitioners with a cost-effective ceived to be more objective and helpful than information provided by
means to address customer engagement and product sales, since the the company (Kozinets, 2002). Those who genuinely and constantly
customer-based social identification is self-driven precondition, rather share information and offer support are naturally recognised and
than expensive firm-based antecedents (such as incentives and re- identified by other members who seek and receive such information
wards). The following section reviews the relevant literature on the and support. Likewise, the latter tend to initiate their identification to
proposed relationships and forms hypotheses accordingly. Methodology interact with others. The receiving and giving processes reinforce an
for testing these hypotheses is outlined, followed by a presentation of individual's identification with other members of the community, and
the results. Discussion of the findings and implications are provided and subsequently with the community. This discussion informs the fol-
conclude the paper. lowing hypothesis:
H1. Customer-to-customer identification (CCI) is positively related to
customer-to-community identification (CMI).
Customer Purchase
2.2. Customer engagement
Social
idenficaon engagement intenon
340
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
multidimensional perspective seems to be dominant in the literature their organizational membership and foster belongingness to the group
with a focus on the behavioural dimension (e.g. Hollebeek, 2011; (Kohler et al., 2011; Kim et al., 2011). If the belongingness to the group
Mollen and Wilson, 2010; Patterson et al., 2006; So et al., 2016a, is strong, the identification with the community would be enhanced
2016b; Vivek et al., 2010). Customer engagement behaviours refer to a accordingly. This will likely intensify members’ willingness to com-
customer's behavioural manifestation toward an object (for example a municate and build friendships with others, and amplifies their at-
brand, a firm or other customers) beyond purchase, including writing tachment with the community (Nambisan and Baron, 2009). These
reviews, helping others, blogging, providing referrals, spreading word- social benefits can be crucial motivators to members’ participation in
of-mouth (WOM) recommendations for a brand or product, and co- the online community (Wang and Fesenmaier, 2004). Subsequently, the
creating values for the firm or the brand they opt to be engaged with members tend to share their knowledge and willingly engage with the
(Brakus et al., 2009; Brodie et al., 2011; Doorn, et al., 2010; Hoyer community (Haeyoung et al., 2014; Qu and Lee, 2011). This view is
et al., 2010; Nambisan and Nambisan, 2008). Some researchers consistent with social exchange theory (Emerson, 1976), which ex-
(Hollebeek et al., 2014; Mollen and Wilson, 2010; Vivek et al., 2014) plains reciprocal relationships between engaging parties. Identifying
refer to customer engagement as a psychological state, which derives with the online brand community is perceived to be a type of social
from a customer's emotional and cognitive connection with a brand benefit (sharing information and receiving support). Reciprocally, the
within a certain context. community members would intend to engage with the brand. This
Customer engagement should be manifested in cognitive, emo- discussion informs the following hypothesis:
tional, and behavioural involvement with a business or a brand. Brodie
H4a. Customer-community identification is positively related to
et al. (2011) concur with this view and define CE as a dynamic psy-
customer attitudinal engagement.
chological process, based on the existence of a customer's interactive,
co-creative experiences with a specific object in a specific set of con- H4b. Customer-community identification is positively related to
ditions. Consistent with this view, the current study contends that a customer behavioural engagement.
genuinely engaged customer must demonstrate both attitudinal and
behavioural manifestations. Psychological engagement implicates cus-
tomer attitude and may have long-term benefits for the firm (Prentice, 2.4. Customer engagement and purchase
2013). The attitudinal aspect conveys a psychological state with cus-
tomers’ high involvement, belongingness, and connections toward Online brand communities are intended for firms and customers to
firms, and comprises of cognitive and affective components referred to engage with each other and to strengthen the ties between company
as customer attitudinal engagement (CAE). The behavioural aspect re- and customers, and ultimately promote sales and enhance customer
flects a behavioural manifestation beyond transactions, which is driven loyalty. Prior research has shown how customer engagement has a di-
by one's cognition and affection (Porter et al., 2011) and referred to as rect positive relationship with customer purchase and loyalty (Brodie
customer behavioural engagement (CBE). Individual behaviour is lar- et al., 2011; Gummerus et al., 2012; Hollebeek, 2011; Prentice et al.,
gely determined by behavioural intention, and the intention is largely 2018; Vivek et al., 2012). The level of customer engagement influences
governed by attitude (Fishbein and Ajzen, 1975). The attitude – in- the level of customer loyalty and success of the brand (So et al., 2014).
tention – behaviour relationship has been extensively discussed in the Indeed, customers’ intention to engage with the brand community
literature (Dong et al., 2010a, 2010b; Han et al., 2008; Li et al., 2009). (ACE) can be transferred to their intention to buy products or services
Attitudinal customer engagement serves as the psychological founda- from the brand. Hence:
tion of behavioural customer engagement, and hence, affect engage- H5. Customer attitudinal engagement is positively related to customer
ment behaviours. purchase.
H2. Customer attitudinal engagement is positively related to customer Behavioural customer engagement refers to a customer's beha-
behavioural engagement. vioural manifestations toward a brand or firm beyond purchase, and
derived from their interactive experiences with the firm, brand, or other
customers. During the interactive process, customers will develop close
2.3. Social identification and customer engagement connections with the brand associated communities. This connection
prompts customers to purchase brand related products or services
Social identification with a brand community augments the sig- (Prentice et al., 2018). Behavioural engagement facilitates customers
nificance of customer-firm relationship and promotes the development transiting on the belief-attitude-behaviour continuum (Vivek et al.,
of positive attitudes such as trust, commitment, or involvement (Brodie 2012) and influences customer loyalty toward the firm and repeat
et al., 2011). CCI, as previously discussed, reflects the fulfilment of purchase behaviours (Harrigan et al., 2017; So et al., 2014; Van Doorn
customers’ requirements for relationships, which involve associating et al., 2010). Accordingly, the following hypothesis is offered:
with other members of the online community, feeling understood,
supported, and admired, as well as understanding, supporting, and H6. Customer behavioural engagement is positively related to customer
admiring others. Such identification is manifested in the primary psy- purchase.
chological substrate for deep, committed, and meaningful relationships
with the organization (Bhattacharya and Sen, 2003). Customer en-
2.5. The moderating role of membership
gagement describes precisely, the connectedness between customers
and the firm, and requires customers to devote extra efforts beyond
Given an online community generally enables customers or any
purchase toward the firm.
individuals to join, participate, or leave the community, the length of
H3a. Customer-customer identification is positively related to customer membership or involvement with the community may be indicative of
attitudinal engagement. the level of engagement with the brand or the business, and embodies
the members’ commitment and the nature of their activities. Wang and
H3b. Customer-customer identification is positively related to customer
Fesenmaier (2004) found that membership status had a significant in-
behavioural engagement.
fluence on the level of participation in online travel communities.
A virtual community is a combination of human relationships and Members with different demographic backgrounds vary in the degree of
comprises of members sharing mutual interests or hobbies. Through identification with the organization (Dutton et al., 1994; Wang and
interaction with like-minded individuals, community members perceive Fesenmaier, 2004), and then perform differently within the community.
341
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
New members might have a lower identification and remain in the refers to a customer's perception of oneness and affinity with a brand
process of understanding the community, attracted by the values re- community, this identification reflects a sense of belongingness to a
ceived from the community (Prentice et al., 2018). New members’ in- community. Accordingly, the items that were used to measure CMI
teractions with existing members may facilitate or speed up the process were adapted from Mael and Ashforth (1992). This scale has four items,
of identification with other members as well as with the community. On including “This community's success is my success”. The Cronbach
the other hand, a long-time member may identify strongly with the alpha for this scale was 0.92. Items used to measure CCI were adapted
community, and then act as a leader in community interaction from Algesheimer et al. (2005a, 2005b) and refer to the identification
(Thompson and Sinha, 2008), recommending the community to more with group members, whereby a customer construes himself or herself
people and helping more entrants to blend in with the community. as group members. Three items were selected for this study and include
Accordingly, the member's duration of membership may affect cus- “The friendships I have with other community members mean a lot to
tomer engagement and purchase through its influence on their identi- me”. The Cronbach alpha for this scale is 0.84.
fication. The longer a member is involved with the community, the Customer attitudinal engagement was measured by adapting Vivek
more he or she would identify with the brand and the community, et al. (2014). Customer attitudinal engagement indicates the extent to
therefore the more participative and engaged the member is, with re- which an individual connects with community or other members. Six
gard to, both attitudes and behaviours (Wang and Fesenmaier, 2004). items were included in this study and the Cronbach alpha for this scale
Therefore, the length of membership is likely associated with product is 0.92. Customer behavioural engagement was measured by adapting
purchase. This discussion informs the following hypothesis: items from Verleye et al. (2014). Customer behavioural engagement
represents a customer's behavioural manifestation toward a virtual
H7a. Duration of membership has a significant moderating effect on the
community beyond purchase, such as compliance, cooperation, feed-
relationship between social identification and customer attitudinal
back, and helping others. Four items were included in this study and the
engagement.
Cronbach alpha for this scale is 0.88.
H7b. Duration of membership has a significant moderating effect on the Customer purchase intention refers to customers’ intention to buy
relationship between social identification and customer behavioural products or services through links in the community. Three items were
engagement. adapted from Zeithaml et al. (1996) in this study and the Cronbach
alpha for this scale is 0.91. Duration of membership was measured
H7c. Duration of membership has a significant moderating effect on the
using a single question: “How long have you been a member of the
relationship between customer attitudinal engagement and purchase.
community?” The factor loadings, composite reliability, and validity
H7d. Duration of membership has a significant moderating effect on the are reported in the next section.
relationship between customer behavioural engagement and purchase.
3.3. Procedures
3. Methods Prior to the survey, the study instrument was validated by two
marketing Professors who have published widely in similar research
3.1. Sample areas. The survey was then pilot tested through sojump.com, a profes-
sional and popular survey website in China. Ten members from one of
The sample in this study was drawn from two major and largest the selected online tourism communities located in different regions of
online tourism brand communities in China: mafengwo.cn and qyer.com. China were targeted to complete questionnaires. As a result, items were
Online tourism communities in China are primarily classified into two revised for better understanding and clarity. The revised questionnaires
categories. One is the vertical enterprise category, focusing on specific were subsequently assessed by the primary researcher and colleagues
areas or particular needs, such as www.mafengwo.cn and www.qyer. who specialise in customer engagement research or teach research
com; the other is a comprehensive enterprise category that provides all- methodology. Relevant modifications were applied with a view to im-
around strategy guides to tourists, such as www.ctrip.com. An in- proving face validity and readability.
tegrated online tourism brand community provides a one-stop tourism The survey was conducted online. By searching keywords on
service platform and has a requirement for financial performance. This Microblog such as “qyer.com”, “mafengwo.cn”, and “travel tips”, we
type of brand community aims to enhance users’ attachment to the identified members who participated online tourism communities reg-
brand. However, a vertical online tourism brand community offers free ularly and sent them the questionnaire link with an invitation letter.
services to users and members, and its development pattern is widely The invitation email contained a brief description of the project and a
divergent from an integrated community. Given its free service and hyperlink that directed participants to the survey. Two mechanisms in
informative nature, this type of brand community is more popular and the survey tool were deployed: a social media collector to distribute the
attracts more consumers to engage within the community. According to survey, and a Web-Link collector with a fixed URL link that enabled the
an iResearch report (2014), there were about 55 million users of ma- dissemination of the survey through different channels. These user-
fengwo.cn, and about 50 million users of qyer.com.1 The two platforms friendly features assisted in accessing a wider population. Participants
were ranked the top two online brand communities in China and sub- were informed the completion of the questionnaire was taken as im-
sequently make these two communities ideal choices for the purpose of plied consent. Participants remained anonymous and the data was kept
this study. de-identifiable. An incentive of a raffle draw was offered to encourage
participation. The survey's welcome page contained external links to
3.2. Measures the ‘Project's Description’ as well as the draw's “Terms and Conditions”,
which were accessible to participants at any point in time.
All study variables were measured by adapting scales applied within As a result, 515 usable responses were generated. Of the total
previous research, with items being assessed using a 7-point Likert sample, 66.5% of respondents were female and 33.5% were male (see
scale. (1 = strongly disagree, 7 = strongly agree). Within this study, Table 1). The majority of respondents fell in the age group of 18 – 25
customer identification is classified into “customer-community identi- (63.7%), followed by the 26 – 35 group (26.1%). At least 76.6% of
fication” (CMI) and “customer-customer identification” (CCI). CMI respondents had acquired a diploma, certificate or bachelor's degree. Of
the total sample, 309 respondents, or (60.2%), had been within a
community for less than 2 years with 54% of respondents coming from
1
Source: iResearch: China Online Travel UGC Report in 2014. mafengwo.com.
342
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
Table 1 validity. The square root of average variance extracted of each con-
Profile of respondents. struct exceeds the correlation between constructs, indicating dis-
Variables n % criminant validity. The results for correlations, means and SD among
study variables are provided in Table 3.
Gender
Female 342 66.5
4.2. Hypotheses testing
Male 171 33.5
Age
<18 years 8 1.6 The hypotheses testing was estimated through SEM with Lisrel 8.72
18–25 years 327 63.7 using the maximum likelihood estimation method. The results demon-
26–35 years 156 30.4 strated a very good fit with the model, with χ2 = 408.17, df = 145, χ2/
36 – 55 years 21 4.1
df = 2.81, p = .00, NFI = 0.98, CFI = 0.99, IFI = 0.99, RFI = 0.98,
> 55 years 1 0.2
Education GFI = 0.92, AGFI = 0.89, RMSEA = 0.062, and SRMR = 0.029. The
Postgraduate degree 2 0.4 results show that CCI is significantly related to CMI (ß = .73,
Graduate Diploma 92 17.9 p < .001), supporting H1. Both CCI and CMI have significant effects on
Bachelor's degree 393 76.6
customer attitudinal engagement: ß = 0.46, p < .001 for CCI and
Diploma/certificate 23 4.5
Secondary school 3 0.6
ß = 0.40, p < 0.001 for CMI. Attitudinal engagement also sig-
Duration of membership nificantly relates to behavioural engagement (ß = 0.51, p < .001).
< 1 year 134 26.1 Both engagement dimensions exert a significant impact on purchase
1–2 year 175 34.1 intention: ß = 0.45, p < .001 for CAE and ß = 0.35, p < .001 for
2–3 years 123 24.0
ABE. The results are show in Table 4.
3 – 4 years 48 9.4
> 4 years 33 6.4 H7 proposed the moderating role of membership duration. A multi-
Online community group analysis was performed to test moderation (Wulf et al., 2001;
Mafengwo 277 54 Palmatier et al., 2007; Han et al., 2008). First, the total sample was split
Qyer 197 38.4 into two sub-groups based on the duration of being a member of the
Others 39 7.6
community, with one group being a member ≥ 2 years (referred to as
senior group for the purpose of this study, or SG), the other < 2 years
3.4. Common method variance (referred to as junior group or JG), with 204 respondents falling into
the SG group and 309 into the JG group. The analyses demonstrates
In addition to the ex-ante procedure described above to minimise acceptable model fit indices for both sub-samples:
common method bias, ex-post statistical remedies were implemented to
examine common method variance, including Harman's single factor SG: χ2 = 267.84, df = 145, χ2/df = 1.85, NFI = 0.97, CFI = 0.99,
test, partial correlation procedure, and controlling for the effects of an IFI = 0.99, RFI = 0.96, RMSEA = 0.061, SRMR = 0.038,
unmeasured latent methods factor (Podsakoff et al., 2003). First, Ex- GFI = 0.89, AGFI = 0.84;
ploratory Factor Analysis (EFA) was performed by loading all study JG: χ2 = 347.68, df = 145, χ2/df = 2.40, NFI = 0.98, CFI = 0.99,
variables to examine the number of factors and respective variance IFI = 0.99, RFI = 0.97, RMSEA = 0.069, SRMR = 0.033,
explained. Results from this procedure revealed five factors and the first GFI = 0.90, AGFI = 0.85
factor explains about 20% of variance. Second, a partial correlation
procedure by including a marker variable (perceived usefulness) was The path coefficients in both sub-samples were first assumed equal,
performed. Inclusion of the marker variable did not alter the postulated then, a comparison of the constrained (constraining the equality of path
relationships and significance of these relationships. Finally, following coefficients in both sub-samples) and unconstrained models were per-
the recommendation of Podsakoff et al. (2003), the last procedure was formed through a chi-square difference test. The results (see Table 5)
to control for the effect of an unmeasured latent factor and to compare revealed that all paths are significantly different between two samples,
the item loadings with and without adding an unmeasured latent indicating existence of a significant moderating effect. Specifically, the
methods factor. This procedure resulted in some adjustments in the paths CCI→CMI, CMI→CAE, CBE→PI, CBE→PI are greater in JG than
measurement model. Finally, diagnosis of multicollinearity using the in SG; however, the paths CCI→CAE, CAE→CBE, CAE→PI are smaller
Variance Inflation Factor (VIF) demonstrated no issue of multi- in JG than in SG. The relationship between CMI and CBE is not sig-
collinearity, as all VIFs were below 5.0 (Neter et al., 1985). nificant in JG, nevertheless significant in SG.
The study hypotheses imply a few mediation paths, with customer
engagement dimensions mediating between customer identification
4. Analysis and results and purchase intention. Further analysis was performed to understand
whether customer engagement played a mediation role in purchase. We
4.1. Measurement model assessed mediation effect with Monte Carlo Method (see Table 6). The
mediation effect is significant when zero is not included in the 95%
As existing scales were used to measure the study variables, confidence interval (Selig and Preacher, 2008). According to the results,
Confirmatory Factor Analysis (CFA) with maximum likelihood estima- 95% confidence internal for all mediations are not including zero, in-
tion was performed to assess their reliability and validity (Hu and dicating the existence of mediation effect. The bias corrected boot-
Bentler, 1999). The result of model fit indices were acceptable: (χ2 strapping p value tested mediating effects (Bollen and Stine, 1990;
(140) = 401.06, p < .0005, CFI = 0.99; NFI = 0.98; GFI = 0.93, Shrout and Bolger, 2002). Table 7 displays the bias corrected boot-
RMSEA = 0.062, SRMR = 0.027). The results of standardized residual strapping p values, and the path coefficient estimates indicating the
co-variances and modification index values also indicated no con- magnitude of the effect between two variables with or without in-
spicuously significant changes to the model. The average variance ex- cluding a mediator.
tracted for each variable was over .50, indicative of adequate con-
vergence (Fornell and Larcker, 1981). The composite reliability was 5. Discussion
acceptable for each of the factors. Factor loadings were positive and
statistically significant (See Table 2). All items have significant loadings Drawing on social identification theory and motivation of ex-
on their corresponding constructs, demonstrating adequate convergent pectancy, the current study examines the relationship between social
343
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
Table 2
Confirmatory factor analyses results.
Item FL α CR AVE
Customer-customer identification The friendships I have with other community members mean a lot to me 0.71 0.84 0.74 0.50
If community members planned something, I’d think of it as something “we” would do rather than something 0.80
“they” would do
I see myself as a part of this community 0.87
Customer-community identification When someone criticizes this community, it feels like a personal insult 0.82 0.92 0.86 0.61
This community's success is my success 0.89
When someone praises this community, it feels like a personal compliment 0.92
When a story in the media criticized this community, I would feel embarrassed 0.82
Customer Attitudinal engagement I am heavily into this community 0.79 0.92 0.87 0.87
I am passionate about participating in this community 0.84
I pay a lot of attention to anything about this community 0.87
Anything related to this community grabs my attention 0.87
I enjoy this community more when I am with others 0.74
This community is more fun when other people around me join it too 0.74
Customer behavioural engagement I adequately complete all expected behaviours in this community 0.84 0.88 0.82 0.53
I do things to make the community's job easier 0.84
I provide the community some useful suggestions to improve services 0.80
I help other members get what they need within this community 0.74
Purchase intention I would purchase products or services through this community 0.85 0.91 0.87 0.69
The community is my first choice to buy relative products or services 0.88
I would do more businesses with this community in the next few years 0.92
Note: α = Cronbach's alpha, FL = factor loadings, CR = composite reliability, AVE = average variance extracted.
344
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
Table 5
The results for moderation testing.
Proposed model JG SG Δχ2(Δdf = 1)
Table 7
Results for Mediation test.
Variables Direct effect Mediation mediation Direct effect Mediation Mediation
345
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
that in Prentice (2013) and Hawkins and Vel (2013). This finding study context. Third, the current study only included one customer-
provides insights into the dimensionality of customer engagement and based antecedent. Other customer factors proposed in van Doorn et al.’s
reveals their relationship sequence. (2010) conceptual framework and in Prentice and Loureiro (2018) are
worth looking into in future studies. Fourth, this study only focused on
7. Practical implications positive engagement, extending the investigation to examine negative
engagement would provide more insights for researchers and practi-
The study also has practical implications. Firms today understand tioners to better understand the impact of customer engagement. Fi-
the value of engaging customer while in reality, customers are often nally, future research shall extend the outcome variable by including
disengaged from the firm or brand associated community. The study other customer loyalty behaviours.
offers a cost-effective means for businesses to engage their customers.
The cost-effectiveness is reflected in little involvement by the firm and Acknowledgement
leaving customers to communicate at their own discretion and engage
freely with other members who have similar interests. The study un- This study was funded by National Natural Science Foundation of
earths that customers’ social identity in a brand community can be a China (71272194, 71772186) and Natural Science Foundation of
contributory factor of customer engagement in online communities and Guangdong Province (2018A030313862).
subsequent purchase. These results offer companies insights into the
implementation of customer engagement strategies and the enhance- References
ment of customers’ connection with organizations.
Specifically, the firm should find the means to boost customer Abrams, D., Hogg, M., 1998. Social Identifications. Routledge, London.
identification with the brand community and improve the quality of Algesheimer, R., Dholakia, U.M., Herrmann, A., 2005a. The social influence of brand
community: evidence from European car clubs. J. Mark. 69 (3), 19–34.
community relationships. In particular, the firm could endeavour to Algesheimer, R., Dholakia, U.M., Herrmann, A., 2005b. The social influence of brand
facilitate their identification process such as organizing charitable community: evidence from European car clubs. J. Mark. 69 (3), 19–34.
programs, conducting environmental activities, and actively under- Armstrong, A., Hagel, J., 1996. The real value of communities. Harv. Bus. Rev. 74 (3),
134–140.
taking social responsibilities. On the other hand, online communities Ashforth, B.E., Mael, F., 1989. Social identity theory and the organization. Acad. Manag.
could help customers find like-minded members and immerse them- Rev. 14 (1), 20–39.
selves in interactive experiences with others. Interactions by commu- Bandyopadhyay, S., Martell, M., 2007a. Does attitudinal loyalty influence behavioral
loyalty? A theoretical and empirical study. J. Retail. Consum. Serv. 14 (1), 35–44.
nity members can occasionally stimulate innovative ideas, which be- Bandyopadhyay, S., Martell, M., 2007b. Does attitudinal loyalty influence behavioral
come increasingly important for companies to leverage user generated loyalty? A theoretical and empirical study. J. Retail. Consum. Serv. 14 (1), 35–44.
content. Collectively, companies should endeavour to enhance cus- Bhattacharya, C.B., Sen, S., 2003. Consumer-company identification: a framework for
understanding consumers' relationships with companies. J. Mark. 67 (2), 76–88.
tomer-customer interactions, and seek the means to empower custo-
Bilro, R.G., Loureiro, S.M.C., Guerreiro, J., 2018a. Exploring online customer engagement
mers to participate in marketing activities and product development. with hospitality products and its relationship with involvement, emotional states,
Given the positive relationship between customer engagement and experience and brand advocacy. J. Hosp. Mark. Manag. https://doi.org/10.1080/
purchase intentions, the firm should develop an appropriate business 19368623.2018.1506375.
Bilro, R.G., Loureiro, S.M.C., Ali, F., 2018b. The role of website stimuli of experience on
intelligence system and motivate customers to engage with the brand engagement and brand advocacy (forthcoming). J. Hosp. Tour. Technol. https://doi.
by developing a reward system and providing incentives for actively org/10.1108/JHTT-12-2017-0136.
engaged customers, for instance, motivating customers to become Bollen, K.A., Stine, R., 1990. Direct and indirect effects: classical and bootstrap estimates
of variability. Sociol. Methodol. 20 (1), 115–140.
opinion leaders or knowledge contributors; launching interesting topics Brakus, J. Jo.ško, Schmitt, B.H., Zarantonello, L., 2009. Brand experience: what is it? How
that lead them to build or merge into additional social networks and is it measured? Does it affect loyalty? J. Mark. 73 (3), 52–68.
connect with other social mediums. These activities can enforce inter- Brodie, R.J., Hollebeek, L.D., Jurić, B., Ilić, A., 2011. Customer engagement: conceptual
domain, fundamental propositions, and implications for research. J. Serv. Res. 14 (3),
actions between customers, creating referral value, influencer value, 252–271.
and knowledge value, and may even prompt the purchase of products Dong, D.H., Liu, S.N., Jin, Y.F., Qi, H., 2010a. An empirical research on e-commerce
or services. customer website loyalty. In: Proceedings of International Conference on
Management and Service Science. pp. 1–4. IEEE.
The significant moderating role of membership duration indicates
Dong, Xiaosong, Wang, Chengzhang, Zhao, Xing, 2010b. Research on customer satisfac-
that customers with different membership lengths should be classified tion, attitude loyalty, the behavior honest's dissymmetry antecedent. East China
into different tiers with different reward systems. The result shows that Econ. Manag. 24 (1), 154–160.
Dutton, J.E., Dukerich, J.M., Harquail, C.V., 1994. Organizational images and member
new entrants are more likely to be influenced by values produced by
identification. Adm. Sci. Q. 39 (2), 239–263.
communities and perceived similarities between themselves and com- Einwiller, S.A., Fedorikhin, A., Johnson, A.R., Kamins, M.A., 2006. Enough is enough!
munities, which sparks their engagement intention with communities. When identification no longer prevents negative corporate associations. J. Acad.
On the other hand, regular customers focus more on the similarities and Mark. Sci. 34 (2), 185–194.
Emerson, R.M., 1976. Social exchange theory. Annu. Rev. Sociol. 2 (1), 335–362.
interactions between themselves and other members. Therefore, for Fishbein, M., Ajzen, I., 1975. Belief, attitude, intention and behaviour: an introduction to
new entrants, it would be helpful for the online brand community to theory and research. Philos. Rhetor. 41 (4), 842–844.
encourage members to provide informative content to improve their Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with unobservable
variables and measurement error. J. Mark. Res. 18 (1), 39–50.
engagement. For regular or senior members, the community should Guesalaga, R., 2016. The use of social media in sales: individual and organizational
create more opportunities to promote the interactions among members. antecedents, and the role of customer engagement in social media. Ind. Mark. Manag.
54 (1), 71–79.
Gummerus, J., Liljander, V., Weman, E., Pihlström, M., 2012. Customer engagement in a
7.1. Limitations and future research Facebook brand community. Manag. Res. Rev. 857–877.
Haeyoung, L., Reid, E., Kim, W.G., 2014. Understanding knowledge sharing in online
Although this study provides some interesting insights, some lim- travel communities: antecedents and the moderating effects of interaction modes. J.
Hosp. Tour. Res. 38 (2), 222–242.
itations are acknowledged. First, this study undertook a cross-sectional
Hajli, N., Lin, X., Featherman, M., Wang, Y., 2014. Social word of mouth: how trust
design, and causal relationships between variables need to be tested develops in the market. Soc. Sci. Electron. Publ. 56 (5), 673–689.
with longitudinal research. Second, the sample was only comprised of Han, Xiaoyun, Robert Jr., J. Kwortnik, Wang, Chunxiao, 2008. Service loyalty: an in-
tegrative model and examination across service contexts. J. Serv. Res. 11 (1), 22–42.
online brand community members in China, the findings may not be
Harrigan, P., Evers, U., Miles, M., Daly, T., 2017. Customer engagement with tourism
generalizable in other study contexts. Including more brand commu- social media brands. Tour. Manag. 59, 597–609.
nities from different industries may offer different insights. Future re- Hawkins, K., Vel, P., 2013. Attitudinal loyalty, behavioural loyalty and social media: an
search should also cross-validate the findings in other regions where the introspection. Mark. Rev. 13 (2), 125–141.
Hollebeek, L., 2011. Exploring customer brand engagement: definition and themes. J.
culture and macro environmental factors are different from the current
346
C. Prentice et al. Journal of Retailing and Consumer Services 47 (2019) 339–347
347