TFS-18501; No of Pages 12

Technological Forecasting & Social Change xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Technological Forecasting & Social Change

Effect of perceived value and social influences on mobile app stickiness and in-app
purchase intention
Chin-Lung Hsu a, Judy Chuan-Chuan Lin b,⁎
a
Department of Information Management, National Taipei University of Business, No. 321, Sec. 1, Jinan Rd., Zhongzheng District, Taipei 100, Taiwan, ROC
b
Department of Computer Science and Information Management, Soochow University, 56 Kueiyang Street, Section 1, Taipei 100, Taiwan, ROC

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

Article history: Recently, there has been a dramatic proliferation of mobile apps, many of which allow for in-app purchases;
Received 10 March 2015 however, little research has focused on what motivates a user to make such purchases. Based on the Affect–
Received in revised form 15 March 2016 Behavior–Cognition model (ABC model) of attitudes, we developed a model involving perceived value and
Accepted 3 April 2016
social influences. The model is then empirically evaluated using survey data collected from 485 users
Available online xxxx
regarding their perception of mobile apps. The findings indicate that stickiness and social identification
Keywords:
significantly influence a user's intention to make in-app purchases. Specifically, significant differences
Mobile app were found between users and potential users in terms of antecedents to forming stickiness and making
In-app purchases in-app purchases. The results may provide further insights into the development of effective mobile app
Perceived value business models and adoption strategies.
Social influences © 2016 Elsevier Inc. All rights reserved.
Stickiness

1. Introduction Such features/content can be categorized as: 1) permanent enhance-

ments to the app usage experience (e.g. removal of advertisements or
Global mobile device penetration continues to proliferate at a unlocking more options); 2) expendable updates such as buying extra
phenomenal rate, driving the robust growth of mobile commerce ap- lives, coins, and food in games; and 3) temporary subscriptions to con-
plications. According to comScore (Siwicki, 2013), as of June 2013 tent or services (MacFarlane, 2015).
mobile devices accounted for 55% of total time spent interacting Though in theory in-app purchase seems a sensible method to mon-
with online retail outlets, with desktop and laptop computers ac- etize free apps, implementation has proved challenging. Unlike in con-
counting for the remaining 45%. Most mobile interactions take ventional shopping contexts, in-app purchases require apps to first
place via mobile apps designed to fulfill specific user demands for attract users to download the trial version before subsequent in-app
shopping, entertainment, information, and social interaction. To purchase can take place, and only regular (sticky) users are likely to
achieve viability, the developers and publishers of mobile apps make such purchases (Yaloz, 2015). Therefore, the key element to the
must create suitable business models. success of mobile app is stickiness (Drell, 2013). However, few previous
The “freemium” approach has transformed the mobile apps sales studies have investigated the antecedents and effects of stickiness in the
market. More than 90% of apps offered in the Apple App Store in 2013 context of mobile apps. Therefore, this study examines factors that
were free (Golden, 2013), prompting initial downloads which paved influence app user behavior (i.e. stickiness), attitudes, and in-app
the way to in-app purchases later. Earlier generation apps depended purchase intention. These are key considerations for app developers
mostly on advertisements to produce revenue; however more and given that, in 2013, in-app purchases accounted for more than 70% of
more publishers are building in-app purchase functions as a primary US iPhone app revenue and as much as 90% of revenues in Asian mar-
means of monetizing their work. In-app purchases are defined as the kets (Koekkoek, 2013) and is expected to grow steadily (eMarketer,
purchase of digital products or services within a specific app via a 2015).
mobile device. The usual scenario is that consumers download and try Past studies have shown that product value is one of the important
out the basic, free version of the app, and then become willing or devel- factors in promoting app usage and purchases (e.g. Lin and Wang,
op the need to pay for premium access to enhanced features/contents. 2006; Turel et al., 2007a, b; Chang and Tseng, 2013). According to
motivation theory, both utilitarian and hedonic values have a strong in-
⁎ Corresponding author.
fluence on purchasing behaviors in both conventional and e-commerce
E-mail addresses: chinlung@ntub.edu.tw (C.-L. Hsu), jclin@csim.scu.edu.tw shopping settings. In addition, prior studies have confirmed that con-
(J.C.-C. Lin). sumer purchasing decisions may be affected by the opinions of his/her

http://dx.doi.org/10.1016/j.techfore.2016.04.012
0040-1625/© 2016 Elsevier Inc. All rights reserved.

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
2 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx

peers (Bearden and Etzel, 1982; Engel et al., 1993; Yang et al., 2007). Han, 2009). Different users value mobile apps for different reasons,
How this actually influences purchase decisions in the context of in- such as productivity enhancement or entertainment. Therefore, per-
app consumption is still unknown. Therefore, this study will address ceived value, including utilitarian value and hedonic value, is proposed
the following research questions: as a motivation for app usage. In addition, app users can interact to cre-
ate online communities and thus enhance their social relationships. For
1. To what extent does perceived value affect user's affective responses
example, social networking apps such as Facebook and the LINE SMS
(i.e. attitude and satisfaction) toward a mobile app?
platform (Naver, Inc.) allow users to exchange information through
2. How do social influence and an app user's preference/dependence on
text, images, video and audio via mobile devices. Therefore, social influ-
a mobile app affect his/her willingness to pay for extra functions?
ences, including social norms and social identification, are included as
3. How does purchase experience affect in-app purchasing behaviors in
additional beliefs and play a significant role as direct determinants for
this new mobile market?
a user's intention to use apps and make in-app purchases. Fig. 1 shows
To achieve the research goals, this study seeks to examine in-app the network of relationships in the model. The remaining section ex-
purchase intention by proposing and empirically testing a behavioral plains the theoretical foundations and rationale of the proposed links
model based on theories of perceived value and social influence. More- along with definitions of each construct.
over, given that in-app purchases occur after a user is already familiar
with the specific app, the user's prior experiences–attitudes, satisfac- 2.1. ABC model of attitudes
tion, and likelihood to continue using the app–play an important role
in user's willingness to pay for extra functions. Previous studies have According to the ABC model of attitudes, attitude is made up of three
verified that the determining factors of information technology and in- components: affect, behavior, and cognition. As shown in Fig. 2, cogni-
formation system (IT/IS) adoption differ between potential users and tive response influences affective response, which in turn shapes behav-
experienced users (Dwivedi and Irani, 2009; Hsu et al., 2007). Hence, ioral response. Cognitive response refers to the beliefs a person has
this study also aims to identify factors that influence in-app purchase in- about performing the target behavior. Affective response is categorized
tention for experienced users (i.e., users who have previously made in- into attitude and satisfaction (Al-Gahtani and King, 1999). Attitude is
app purchases) and potential users (users who have no prior in-app defined as an individual's positive or negative feelings about performing
purchase experience). the target behavior whereas satisfaction is defined as the degree to
This study used a structural equation model (SEM) to assess the em- which a user favorably perceives the overall assessment of performing
pirical strength of the relationships in the proposed model. In-app pur- the target behavior. Behavior is a measure of the strength of one's inten-
chases have proven to be an effective monetization strategy for tion to perform a specified behavior. In the past decades, many theories
freemium apps. Examining antecedents and the effects of stickiness on such as the theory of reasoned action (TRA), the technology acceptance
mobile app users' in-app purchase intention from the perspectives of model (TAM) and the theory of planned behavior (TPB) have adapted
perceived value and social influences will provide valuable insights the cognition–affect–behavior causal chain to predict user behavior
into in-app purchasing behavior as well as additional implications for (Fishbein and Ajzen, 1975; Davis, 1989; Ajzen, 1991).
theory of stickiness in the mobile app context. In addition, the results The ABC model of attitudes is a general model and does not specify
of this research can also help app developers devise effective mobile the beliefs that are operative for a particular behavior. Moreover, atti-
apps to improve their competitiveness. tude formation is closely related to the specific characteristics of a par-
ticular product and service. Therefore, while researchers have adapted
2. Conceptual model and research hypotheses the ABC model of attitudes to explain user behavior, salient beliefs
need to be considered in a specific context. For example, in the TAM
Fig. 1 illustrates the research model which is adapted from the ABC model, Davis (1989) proposed that user beliefs about perceived useful-
model of attitudes. It asserts that user intention to make in-app ness and ease of use had an impact on user attitudes toward IT/IS accep-
purchases is determined by attitude, satisfaction, social norms, social tance behavior within an organization. In the TPB model, the social
identification and stickiness. Furthermore, attitude and satisfaction norms and perceived behavioral control are both types of beliefs
mediated the impact of beliefs about perceived value (i.e., hedonic and which influence user behavior. Prior studies based on these theories
utilitarian values) on stickiness. Perceived value has been identified as also proposed specific beliefs to improve understanding of user behav-
a possible measure of experience of mobile technology usage (Turel ior for specific contexts, such as WWW, online games, electronic
et al., 2007a,b; Minna, 2005; Kim, 2010; Kuo et al., 2009; Kim and commerce and mobile commerce (Moon and Kim, 2001; Hsu and Lu,

Fig. 1. Research model.

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx 3

Fig. 2. Basic model (adapted from the ABC model of attitudes).

2004; Lee, 2008; Green and Pearson, 2011; Leong et al., 2013). However, Hypothesis 2a. Utilitarian value will positively affect user attitude.
in the context of mobile apps, the present study proposes that belief var-
iables include perceived value and social influences. Findings are ex- Hypothesis 2b. Utilitarian value will positively affect user satisfaction.
pected to shed light on the development of strategies to understand
and promote app usages and purchase.
2.2.2. Social influence perspectives
2.2. Hypotheses In addition to perceived values, many studies also have found that
social influence positively influences an individual's IT usage (Luarn
2.2.1. Beliefs regarding hedonic value and utilitarian value and Lin, 2005; Cheung et al., 2011; Ting et al., 2014). Social influence
A user's utilitarian and hedonic values are seen to influence his/her is the extent to which consumers perceive that their peers (e.g., family
IT/IS usage behavior. While utilitarian IT/IS focuses on increasing the and friends) believe that they should use a particular technology
user's task performance, hedonic values focus on increasing the user's (Venkatesh et al., 2012). Reference group theory also emphasizes that
pleasurable experience (van der Heijden, 2004). Studies have empirical- a consumer's behavior (e.g. purchasing decision) can be affected by
ly verified that both values have an influence on IT/IS perceptions and the opinions of his/her peers (Brown and Reingen, 1987; Kotler,
usage behavior (Kim and Han, 2009; Kim and Hwang, 2012; Chun 1999). This reference group influence has a similar impact to other so-
et al., 2012; Pöyry et al., 2013). According to motivation theory, percep- cial factors and can be regarded as a direct determinant of behavioral in-
tions of utility represent extrinsic motivation while the perceptions of tention (Thompson et al., 1991).
hedonic value represent intrinsic motivation (Zanna and Rempel, Venkatesh et al. (2003) observed that various theories use different
1988). Extrinsic motivation emphasizes performing a behavior to constructs to describe social influence. For example, TRA, TPB, and
achieve specific goals/rewards (Vellerand, 1997), while intrinsic moti- TAM2 propose “subjective norms” as a type of social influence factor. Sub-
vation refers to the pleasure and satisfaction derived from performing jective norms are defined as the individual's perception that his/her peers
a certain behavior (Deci and Ryan, 1987). Previous research in the IT/ would approve or disapprove of a particular behavior. In addition, innova-
IS domain suggests that high extrinsic and intrinsic motivation results tion diffusion theory (IDT) uses the term “image” to refer to the degree to
in increased intention to adopt a given IT/IS (Davis et al., 1992; Teo which use of an innovation is perceived to enhance one's image or social
et al., 1999; Lin and Lu, 2011; Yoo et al., 2012). The present study as- status. Although different terms have been used to describe social influ-
sumes that people use apps not only to achieve specific goals or improve ence, it generally consists of two distinct influences — informational influ-
task performance and productivity, but also for leisure and entertain- ence and normative influence. Moreover, social influence operates
ment. Therefore, utilitarian value and hedonic value are proposed as fac- through three distinct mechanisms — internalization, identification, and
tors which impact user's beliefs regarding app usage. compliance. First, informational influence is an internalization process
In the current study, we define hedonic value as the degree to which occurs when a user perceives information as enhancing existing
which a user derives pleasure from using an app. Utilitarian value knowledge acquired from reference groups. As the number of internet
is defined as the degree to which a person believes that using an users increases, information influences such as electronic word-of-
app enhances his or her task performance. Oliver (1993) suggested mouth (eWOM) recommendations play an increasingly important role
that perceived value should be conceptualized and measured as a in shaping customer behavior (Hennig-Thurau and Walsh, 2003). Hsu
cognitive construct, while customer satisfaction is conceptualized and Lin (2010) empirically verified that information influence has a sig-
as an affective construct. Lin and Wang (2006) empirically found nificant impact on adoption behavior for In-Car GPS navigation systems.
that perceived value had a significant effect on satisfaction in mobile Second, normative influence is a form of identification and compliance.
commerce contexts, while Kuo et al. (2009) showed that the per- Identification occurs when a user adopts an opinion held by others out
ceived value for mobile value-added services had a positive influence of concern with self-definition in relation to the group. Compliance occurs
on user satisfaction. In addition, when consumers view a target be- when a user conforms to the expectations of another to receive a reward
havior such as online shopping as being more valuable (e.g. more or avoid rejection and hostility. Social exchange theory supports this per-
convenient and cost effective), they will develop a more positive at- spective in saying that people perform specific behaviors because they ex-
titude toward that behavior. For example, Overby and Lee (2006) pect reciprocal benefits such as enhanced reputation, trust, power,
verified that preference (i.e., positive attitude) was influenced by he- dependence, promise and economic return (Blau, 1964; Emerson, 1981;
donic and utilitarian values in the context of online shopping. More- Donath, 1999; Liang et al., 2008).
over, from the perspective of motivational factors, previous studies In addition to social norms, roles are another common standard for
have identified positive relationships between utilitarian motivation the behavior of group members (Ellis and Fisher, 1994). Roles are
and attitude as well as between hedonic motivation and attitude “sets of behaviors that are considered appropriate for persons holding
(Davis et al., 1992; Childers et al., 2001; Monsuwé et al., 2004). Ac- particular positions in a group” (Triandis, 1977). People develop a
cordingly, we propose the following hypotheses: sense of membership and belonging in particular groups. Thus, their
perception of social identification increases when they identify with
Hypothesis 1a. Hedonic value will positively affect user attitude. and assume a role in a group through participation in group action.
Social identity theory suggests that people classify themselves into
Hypothesis 1b. Hedonic value will positively affect user satisfaction. social categories. This categorization changes the way people see

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
4 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx

themselves in the sense that it activates one's self-concept which is 2.2.4. Behavioral responses regarding stickiness and intention to make in-
related to the formation of emotions and norms relating to member- app purchases
ship in affiliated groups. Furthermore, self-categorization theory also In recent years, the typical digital business strategy for mobile app
emphasizes that beliefs about appropriate behavior follow directly publishers has been to offer apps on a free or trial basis (Singer-
from one's self-concept as a group member (Turner et al., 1987). Oestreicher and Zalmanson, 2013), in an attempt to retain and prolong
Self-concept mostly comprises people social identification, defined the user's stay. In the context of non-mobile websites, Demers and Lev
as the perception of belonging to a group, team or community. In (2000) proposed the concept of “stickiness” to describe the ability of a
many instances, social identification has been linked to a number of site to capture and retain the user's attention. Similarly, app stickiness
positive outcomes, such as enhanced behavioral intention, life satis- refers to the user's behavioral intention to re-use and prolong the dura-
faction, loyalty and group performance (Alvesson, 2000; Ellemers tion of each usage on an app. By increasing the number of times a user
et al., 2004; Bagozzi and Dholakia, 2006; Hsu and Lin, 2008; visits an app and the duration of each visit, increased stickiness is seen
Haslam et al., 2009). Therefore, this study uses social norms and so- as increasing the potential for in-app purchases. Both stickiness and
cial identification as beliefs of social influences to explore the impact in-app purchase intention reflect important dimensions of the app's
on stickiness and intention to make in-app purchases. Social norms value and are expected to be positively associated with the app's market
are defined as the degree to which a user perceives others as approv- value.
ing of his/her using an particular app. Social identification is defined We define stickiness as the degree to which a user re-uses a given
as the degree to which a user perceives that using a particular app app and prolongs the duration of each usage. Intention to make in-app
grants him/her membership in a particular community. Accordingly, purchases was defined as the degree to which a user would like to pur-
we propose the following hypotheses: chase products and/or services in the future within a given app. The
present study proposes that satisfaction had a negative effect on inten-
Hypothesis 3a. Social norms will positively affect user stickiness. tion to make in-app purchases because users who are satisfied with the
Hypothesis 3b. Social norms will positively affect user intention to functionality of a free or trial version app are less likely to purchase en-
make in-app purchases. hanced functionality. In addition, increased stickiness was found to
leave users more likely to make in-app purchases. Empirically, Lin
Hypothesis 4a. Social identification will positively affect user (2007) confirmed that a Web user's willingness to stick with a given
stickiness. website is a strong predictor of his/her intention to transact. According-
ly, we propose the following hypotheses:
Hypothesis 4b. Social identification will positively affect user intention
to make in-app purchases. Hypothesis 5b. Attitude will positively affect user stickiness.

Hypothesis 5c. Attitude will positively affect user intention to make

2.2.3. Affective responses regarding satisfaction and attitude in-app purchases.
Satisfaction and attitude are conceptualized as an individual's affec-
tive reactions toward system usage. User satisfaction is defined as the Hypothesis 6a. Satisfaction will positively affect user stickiness.
degree to which a user perceives overall system usage favorably. It is
Hypothesis 6b. Satisfaction will negatively affect user intention to
primarily an affective evaluative response. Attitude is defined as an
make in-app purchases.
individual's positive or negative feelings (i.e., evaluative affect) about
using the system (Davis et al., 1989). Al-Gahtani and King (1999) sug- Hypothesis 7. Stickiness will positively affect user intention to make
gested that satisfaction be used in parallel with the attitude construct in-app purchases.
to explaining IT/IS usage behavior. Furthermore, satisfaction is concep-
tually distinct from attitude in that satisfaction is a transient and
experience-specific affect, while attitude is relatively more enduring 3. Methodology
(Oliver, 1980). Therefore, a user may have a positive attitude (with a
pleasant experience) but still feel dissatisfied if his/her actual experi- 3.1. Sample
ence does not meet expectations (Bhattacherjee, 2001). In the context
of mobile apps, most users adopt try-first-and-purchase-later behavior In recent years, web-based surveys have emerged as an effective
since many app providers offer free versions for the users to download. means of collecting data for academic research. The empirical data for
Before downloading the apps, a user may develop expectations about the present study were collected through an online survey of app
app performance. Depending on how well the trial usage compares to users in Taiwan. An announcement of the survey goals was posted for
expectations, the user gains actual experience from using the trial app 30 days on heavily-trafficked web sites such as Sogi.com (http://www.
and develops perceptions and attitudes regarding the app's value, sogi.com.tw) and Facebook (http://wwww.facebook.com) as well as
along with his/her usage satisfaction. Therefore, this study uses attitude online communities such as Mobile01 (http://www.mobile01.com)
and satisfaction as affective factors and examines their impact on usage and campus BBS message boards devoted to issues related to mobile ap-
behavior. plications. The announcement included a hyperlink to the survey form.
We define attitude as the degree to which the user derives positive Participants who submitted valid questionnaires were entered into a
feelings from using a given app. Satisfaction was defined as the extent drawing to receive one of twenty NT$200 (US$6) convenience store
to which the user gives the app an overall favorable assessment. Previ- gift certificates. Duplicate responses were eliminated by filtering for
ous studies have found a significant relationship between these two fac- multiple uses of a single IP address or email account. The questionnaire
tors (Hellier et al., 2003; Abdul-Muhmin. 2011). For example, Shih included prompts to ensure that the respondents completed all survey
(2004) confirmed that a user's satisfaction with using the Internet items, thus eliminating invalid responses. The final sample included
played an important role in influencing user's attitude toward using 485 valid responses.
the Internet. In addition, Hung et al. (2003) found that user satisfaction Among the respondents, 56% were male, 63% were under the age of
positively influences an individual's attitude toward using Wireless 25, 61% had a bachelor's degree and 53% had experience making in-app
Application Protocol (WAP) services. Accordingly, we propose the purchases. Table 1 summarizes the demographics of the respondents.
following hypothesis: The demographic profile showed that users are relatively young and
generally well educated. This trend is somewhat similar to results re-
Hypothesis 5a. Satisfaction will positively affect user attitude. ported by FIND (2014). Generally, younger users are more accepting

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx 5

Table 1 Table 2
Demographic profile. Descriptive statistics (means and S.D.).

Measure Items Frequency Percent n = 485 Means S.D.

Gender Male 271 56 Hedonic value 4.10 0.55

Female 214 44 Utilitarian value 3.83 0.70
Age Under 20 76 16 Social norms 3.71 0.69
21–25 228 47 Social identification 3.57 0.61
26–30 93 19 Attitude 4.05 0.54
Over 30 88 18 Satisfaction 3.96 0.52
Education High school or less 30 6 Stickiness 3.73 0.65
Some college 18 4 Intention to make in-app purchases 3.14 0.88
Bachelor's degree 298 61
Graduate degree 139 29
Experience in making in-app Yes 255 53
purchases No 230 47
4. Results
Price range of in- apps purchases Under NT50 84 18
NT$30–NT$150 125 26 4.1. Descriptive statistics
NT$51–NT$300 28 6
OVER$300 18 4
Table 2 lists the means and standard deviations of the constructs. On
None 230 47
Monthly expenditure for in-app Under NT$50 71 15 average, the participants responded positively to the research con-
purchases NT$51–NT$100 75 16 structs (the averages all being greater than 3). However, on average,
NT$101–NT$500 72 15 the surveyed participants seemed to be slightly less concerned with in-
NT$501–NT$1000 21 4 tention to make in-app purchases, suggesting that users may have a re-
Over NT$1000 16 3
duced tendency to purchase in-app products and services.
None 230 47
Number of in-app purchases in the Under 6 173 35
past year 7–12 63 13 4.2. Analytical strategy for model assessment
Over 12 19 4
None 230 47
To test the proposed hypotheses, data were collected and analyzed
Credit card Yes 280 58
No 205 42 using structural equation modeling (SEM), a second-generation multi-
Preferred means of payment for Credit card 144 30 variate technique for analyzing causal models by estimating the mea-
in-app purchases payment surement and structural models using AMOS 21.0. We also tested the
Telecommunication 106 22 univariate normality of the collected data by skewness and kurtosis
bill
tests. The results showed that absolute skew and kurtosis values for
Convenience store 166 34
Stored value card 57 12 most of variables are less than 2 (see Table 3), indicating that the data
Other 12 3 in our study are close to the univariate normality (Kline, 2005). We
also assessed multivariate normality by using Mardia's method
(Mardia, 1970). According to Bollen (1989), if the Mardia's coefficient
of innovations, and are likely become the most active apps users and the is lower than p(p + 2) (where p is the number of observed variables),
most influential consumers in the mobile apps field in the future. Real- then the sample exhibits multivariate normality. In this study, the
izing the preferences and behavior of these future opinion leaders Mardia's coefficient was 191.406, which was lower than 675 (the
would be of great value to mobile app publishers.

Table 3
Item reliability, composite reliability and average variance extracted (AVE).
3.2. Measurement development
Item Skewness Kurtosis Item reliability Composite reliability AVE
The questionnaire included demographic information and con- HV1 −0.59 2.92 0.86 0.89 0.72
struct items. In addition, measures were adapted from previous HV2 −0.44 1.66 0.84
studies (Lin, 2007; Hsu and Lin, 2008; Davis et al., 1989; Kim et al., HV3 −0.32 0.94 0.85
2007; Bhattacherjee, 2001) to develop scales for measuring UV1 −0.52 0.45 0.84 0.88 0.71
UV2 −0.44 0.15 0.86
constructs such as perceived value (i.e., utilitarian value and hedonic UV4 −0.38 0.10 0.83
value), attitude, satisfaction, social norms and social identification, SN1 −0.39 0.57 0.74 0.86 0.68
stickiness and intention in the context of mobile apps (see Appendix SN2 −0.11 −0.46 0.88
for the complete questionnaire). Each item was measured on a five- SN3 −0.15 0.04 0.85
SI1 −0.31 0.20 0.79 0.75 0.50
point Likert scale, ranging from “strongly disagree” (1) to “strongly
SI2 0.10 −0.19 0.74
agree” (5). Each item was translated into Chinese following Brislin's SI3 −0.40 0.31 0.57
(1980) translation — back translation method. In the survey, respon- AT1 −0.42 1.59 0.76 0.87 0.70
dents were asked to specify an app that he/she uses most frequently, AT2 −0.31 1.10 0.87
and subsequently answered the rest of questionnaire with respect to AT3 −0.63 2.59 0.87
SA1 −0.54 2.00 0.81 0.90 0.68
the usage behaviors of that app.
SA2 −0.33 1.77 0.85
Before conducting the main survey, the instrument was validated SA3 −0.16 0.43 0.79
using both a pre-test and pilot. The pre-test involved 10 respondents SA4 −0.36 1.11 0.85
who were experienced smartphone app users. The respondents were ST2 0.03 0.11 0.72 0.84 0.64
ST3 −0.23 0.02 0.86
asked to comment on a list of items related to the constructs, including
ST4 −0.36 0.08 0.82
the wording of the scales, the length of the instrument, and the ques- IN1 −0.26 −0.50 0.89 0.90 0.74
tionnaire format. Finally, to reduce possible ambiguity, a pilot test was IN2 −0.16 −0.38 0.82
administered to 50 respondents selected from a population of app IN4 −0.30 −0.15 0.87
users. The results of the pilot test showed acceptable reliability and HV: Hedonic value; UV: Utilitarian value; SN: Social norms; SI: Social identification; AT:
validity. Attitude; SA: Satisfaction; ST: Stickiness; IN: Intention to make in-app purchases.

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
6 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx

value of p(p + 2), where p = 25). Therefore, the data in our study are Table 5
assumed to exhibit multivariate normality. After accounting for sample Fit indices for the structural model.

size, data normality and the characteristics of evaluation methods Measures Recommended Results References
(Olsson et al., 2000), we then adopted the method of maximum likeli- criteria
hood estimation (MLE) for the analysis. X2/d.f. b3 2.12 Bentler and Bonnet (1980), Hair et al.
(p-value) (0.000)a (2006), Schumacker and Lomax (2004),
4.3. Measurement model GFI N0.9 0.92 Seyal et al. (2002), Scott (1994), Ullman
AGFI N0.9 0.89 (2006)
CFI N0.9 0.96
A confirmatory factor analysis using AMOS was conducted to test the NFI N0.8 0.93
measurement model. The initial assessment indicated that items with a RMSEA b0.08 0.04
factor loading less than 0.5 should be removed (Hair et al., 1992). The a
For a larger sample size, significant p-values are expected (Hair et al. 2006).
final instrument included 25 items, as shown in Table 3. The data indi-
cate that the reliability of the individual items ranged from 0.57 to
0.89, exceeding the minimum acceptable value of 0.50 (Hair et al., respectively), thus supporting H1a, H2a and H5a. Together, these
1992). The internal consistency of the measurement model was three paths explain 71% of the observed variance in attitude. Hedonic
assessed by computing the composite reliability. Consistent with the and utilitarian value also affect satisfaction (β = 0.74, p b 0.001; β =
recommendations of Fornell (1982), composite reliability of all items 0.18, p b 0.001, respectively), providing support for H1b and H2b. The
exceeded the benchmark of 0.60. The average variance extracted for model accounted for 68% of the variance in satisfaction. Attitude (β =
all constructs exceeded the threshold value of 0.5 recommended by 0.29, p b 0.01), satisfaction (β = 0.22, p b 0.05) and social identification
Fornell and Larcker (1981). Since the three reliability values were all (β = 0.33, p b 0.001) had significant effects on stickiness, accounting for
above the recommended thresholds (Table 3), the scales for evaluating 49% of the variance. Therefore, H4a, H5b and H6a were supported. Final-
these constructs were deemed to exhibit adequate convergence ly, social identification and stickiness had a significantly positive effect
reliability. on intention to make in-app purchases (β = 0.34, p b 0.001; β =
Table 4 indicates that the variances extracted for each construct 0.34, p b 0.001), while satisfaction had significantly negative effect
were greater than any squared correlation among the constructs, there- (β = −0.23, p b 0.05), accounting for 32% of the variance. Therefore,
by implying that the constructs are empirically distinct (Fornell and H4b, H6b and H7 were supported. Contrary to expectations, social
Larcker, 1981). In summary, measurement model testing, including norms had no significant effect on stickiness or intention to make in-
convergent and discriminant validity measures, showed satisfactory re- app purchases. Hence, H3a and H3b were not supported. Furthermore,
sults. In addition, as shown in Table 5, the fitness measures for the mea- attitude had no direct influence on user intention to make in-app pur-
surement models are all acceptable. Consequently, all measures taken in chases, thus H5c was not supported.
this study show that the model provides a good fit to the data. To compare the effects of prior experience making in-app purchases
To test for potential common method bias, we adopted Harman's on intention to make further in-app purchases, the sample was further
one-factor test (Podsakoff et al., 2003) and ran an unrotated principle categorized into potential users and actual users. Among the 485 re-
component factor analysis for all constructs. The results generated six spondents, 256 had prior experience making in-app purchases. Table 6
factors (with an eigenvalue greater than one) with one factor account- lists descriptive statistics and t-test test results for potential and actual
ing for 39% of covariance among the measures. The other five contribut- users. On average, both groups responded positively to perceived attri-
ed 32% of the remaining variance, each accounting for 3%–9%. Thus, we butes such as hedonic value, utilitarian value, social norms and social
detect no apparent common method bias. identification as well as attitude, satisfaction, stickiness and intention
In addition, variance inflation factors (VIFs) was used to assess to make in-app purchases. Meanwhile, the difference between the
the degree of multicollinearity. We conducted a regression analysis two groups is statistically significant for all items except hedonic
by modeling intention to purchase as the dependent variable and value, attitude and satisfaction. The average ratings for utilitarian
the other nine variables as independent variables. The VIF ranges value, social norms, social identification, stickiness and intention were
from 1.453 to 2.891, which is below the suggested threshold of 3.3 significantly higher among actual users than for potential users.
(Diamantopoulos and Siguaw, 2006). Therefore, no significant Two separate structural modeling analyses were also conducted for
multicollinearity problem exists with regard to our data. the two groups. As shown in Fig. 4, for potential users, hedonic value,
utilitarian value and satisfaction were found to influence attitude
4.4. Structural model (β = 0.26, p b 0.05; β = 0.19, p b 0.001; β = 0.54, p b 0.001, respective-
ly). The model accounted for 74% of the variance in attitude. Moreover,
We analyzed the structural model by testing the hypothesized rela- hedonic value and utilitarian value directly influenced on satisfaction
tionships among the various constructs. As illustrated in Fig. 3, hedonic (β = 0.80, p b 0.001; β = 0.10, p b 0.05, respectively). Together, these
and utilitarian values, along with satisfaction significantly influence at- two paths accounted for 71% of the variance in satisfaction. The results
titude (β = 0.19, p b 0.01; β = 0.15, p b 0.001; β = 0.60, p b 0.001, show that an app's stickiness for potential users is related to the user's
satisfaction (β = 0.36, p b 0.01) and social identification (β = 0.29,
p b 0.01), accounting for 48% of the variance. Surprisingly, no significant
Table 4 path was found from positive attitude and social norms to stickiness.
Discriminant validity of users.
Finally, stickiness had a significant effect on intention to make in-app
HV UV SN SI AT SA ST IN purchases (β = 0.30, p b 0.01), accounting for 17% of the variance. Un-
HV 0.72 expectedly, attitude, satisfaction, social norms and social identification
UV 0.11 0.71 had no direct effect on intention to make in-app purchases.
SN 0.11 0.23 0.68 Results for actual users (Fig. 5) indicate that hedonic value and satis-
SI 0.20 0.19 0.26 0.50
faction influence attitude (β = 0.18, p b 0.05; β = 0.66, p b 0.001, re-
AT 0.39 0.18 0.20 0.24 0.70
SA 0.48 0.16 0.19 0.29 0.53 0.68 spectively). The model accounted for 71% of the variance in attitude.
ST 0.26 0.09 0.11 0.24 0.31 0.31 0.64 However, utilitarian value had no direct influence on attitude. More-
IN 0.04 0.08 0.08 0.15 0.11 0.09 0.18 0.74 over, hedonic value and utilitarian value directly influenced satisfaction
Diagonals (in bolds and italic) represent the average variance extracted (AVE), while the (β = 0.64, p b 0.001; β = 0.29, p b 0.001; respectively). Together, these
other matrix entries represent the shared variance (the squared correlations). two paths accounted for 65% of the variance in satisfaction. The results

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx 7

Fig. 3. Results of structural modeling analysis (n = 485).

show that attitude and social identification had a significant effect on finding is that users are willing to make purchases due to concern for so-
stickiness (β = 0.45, p b 0.001; β = 0.24, p b 0.05, respectively), ac- cial identification, which also directly impacts stickiness. Use of social-
counting for 54% of the variance. However, social norms and satisfaction and communication-oriented apps may provide users with a sense
had no direct influence on stickiness. Finally, stickiness and social iden- of belonging. This perception will impact the app's perceived sticki-
tification had a significant effect on the intention to in-app purchase ness and the user's intention to make in-app purchases. Surprisingly,
(β = 0.22, p b 0.05; β = 0.28, p b 0.01), accounting for 42% of the vari- social norms are found to have no significant effect on stickiness and
ance. However, the paths from attitude, satisfaction and social norms to intention. This finding differs from theories such as TRA, TPB and
intention to make in-app purchases were insignificant. TAM2 (Fishbein and Ajzen, 1975; Ajzen, 1991; Venkatesh and
Our study also adopted the approach proposed by Keil et al. (2000) Davis, 2000). However, it is consistent with previous studies (Chau
to test the significance of the coefficient differences. Table 7 lists the and Hu, 2002; Lewis et al., 2003; Hsu and Lin, 2008; Tsai and
comparison results, which show significant differences in most path co- Bagozzi, 2014), which found that social norms have no significant ef-
efficients between potential and actual groups. fect on user's behavioral intention. Indeed, users may develop values
and standards for their intention to use by referring to normative
5. Discussion and conclusion practices of a group or other users. However, such normative effects
mostly occur in compulsory usage situations. Generally, users do not
In-app purchases are one of the effective monetization strategies for consider the approval of others when choosing to adopt an app since
freemium apps, and the commercial success of an app depends largely they generally download apps on a voluntary basis to fulfill their pri-
on its “stickiness” (Yaloz, 2015). This study examines the antecedents vate needs, thus limiting the impact of social norms. In addition,
and effects of stickiness in the context of mobile in-app purchases. Find- users regard apps installed on their smartphones as a type of private-
ings and implications are discussed as follows. ly consumed necessity and, according to Bearden and Etzel (1982),
the reference group influence for such a product and brand is rela-
5.1. Discussion tively weak. In-app purchase behavior is largely dominated by app
attributes (i.e., hedonic value and utilitarian value) rather than by
First, the proposed model is found to accurately predict user inten- the influences of others.
tion to make in-app purchases. Stickiness is found to significantly and Though attitude has no significant influence on intention to make in-
positively influence intention, thus emphasizing that apps users are un- app purchases, satisfaction has a significant and negative effect on in-
likely to make in-app purchases for apps which they don't find engag- tention, apparently because users who are satisfied with a free or trial
ing. This shows that stickiness is not only important in web-based app have limited incentive to pay to upgrade. However, attitude and sat-
environments (Zott et al. 2000; Lin, 2007), but also is the key to increas- isfaction appear to have significant and positive effects on stickiness. In
ing online transactions in the context of mobile apps. Another plausible addition, satisfaction positively impacts attitude. In summary, the
model provides a conceptual depiction of the critical role played by af-
Table 6 fective factors in determining stickiness and thus increase intention to
Descriptive statistics and ANOVA. make in-app purchases.
Scale items Potential users Users p-Value
Second, both utilitarian value and hedonic value turn out to have a
(n = 229) (n = 256) significant influence on affective factors, though the hedonic value is a
stronger predictor than utilitarian value (β = 0. 74 and β = 0. 19).
Hedonic value 4.06 4.14 0.14
Utilitarian value 3.76 3.89 0.03* This reveals that utilitarian and hedonic value dimensions play impor-
Social norms 3.64 3.76 0.05* tant roles in predicting affective outcomes such as satisfaction and atti-
Social identification 3.48 3.65 0.00*** tude. Moreover, this study emphasizes entertainment and leisure-
Satisfaction 3.95 3.97 0.79
oriented apps, as opposed to productivity apps. While using such
Attitude 4.00 4.08 0.10
Stickiness 3.59 3.84 0.00*** apps, the hedonic value is the primary driver of user affect, thus signif-
Intention to make in-app purchases 2.60 3.62 0.00*** icantly decreasing the relative impact of utilitarian value.
⁎ p b 0.05. Finally, the findings reveal that determinants influencing intention
⁎⁎ p b 0.01. to make in-app purchases differ substantially between potential and ex-
⁎⁎⁎ p b 0.001. perienced users. As depicted in Figs. 4 and 5, the potential user's

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
8 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx

Fig. 4. Results of structural modeling analysis for potential users (n = 229).

intention is determined by stickiness alone while that of the experi- feelings (i.e. improved attitude) which meet or exceed their expecta-
enced user is determined by both stickiness and social identification. tions, rather than a transient emotional reaction (i.e. satisfaction).
Stickiness is an important factor in motivating the majority of current Therefore, for the experienced user, stickiness is directly influenced by
and potential users to make in-app purchases. We also found that social positive attitude. Satisfaction influences attitude, and thus indirectly in-
identification plays a key role in stickiness for both groups. Social iden- fluences stickiness. For both groups, hedonic value and utilitarian value
tification results in a significant difference between these two groups in seem to play an influential role in shaping user affect. Notably, utilitari-
terms of intention to make in-app purchases (17% and 42% for potential an value is found to not directly impact attitude in experienced users,
users and experienced users, respectively). but it does directly affect satisfaction. Intuitively, users have positive at-
Interestingly, the effects of affective factors on stickiness for both titudes toward apps they found to be useful. However, the analytical re-
groups yield different results. For potential users, satisfaction directly sults indicate that utilitarian value does not appear to drive user
impacts stickiness, while attitude has no significant effect on stickiness attitude. Instead, utilitarian value forms satisfaction, which in turn
but significantly influences satisfaction. This underscores the impor- shapes user attitudes. This implies that users will have a positive atti-
tance of satisfaction for potential users. Satisfaction is a positive tude toward apps they perceive as providing utilitarian value.
emotional state related to expectation performance discrepancy (con-
firmation). For potential users, the expectation is typically based on 5.2. Implications for researchers
others' opinions or information disseminated through mass media.
When potential users feel satisfaction (i.e. app performance exceeds As smartphones and mobile apps become more indispensable in our
expectations), they will use the apps more frequently and for longer du- daily lives, researchers working in the fields of marketing and technolo-
rations, thus enhancing their intention to make in-app purchases. For gy adoption need a deeper understanding of consumer behaviors in mo-
experienced users, satisfaction influences attitude, but no significant bile platforms. The results of this study provide a better theoretical
relationship was found between satisfaction and stickiness. One possi- understanding of mobile app user behavior in regard to in-app pur-
ble explanation is that experienced users tend to have a greater subjec- chases. Several implications are as follows: First, this study expands
tive expectation for making future in-app purchase because of their past our comprehension of the theory of stickiness and provides empirical
experience in making such in-app purchases. However, to make in-app evidence that stickiness is a significant driver for purchase intention,
purchases, such users may require an app to provide enduring positive even in the context of mobile app commerce. This study further

Fig. 5. Results of structural modeling analysis for actual users (n = 256).

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx 9

Table 7 value to be important in their attitude toward and satisfaction with a

Coefficient differences between groups. given app. In addition, hedonic value had a more significant impact on
Path coefficients affection than did utilitarian value. Therefore, app developers should
Hypothesis Potential users Actual users Difference
seek to maximize hedonic motivation such as enjoyment, fun and
pleasure.
H1a 0.26⁎ 0.18⁎ 0.08⁎⁎⁎
Finally, potential users and actual users should be targeted using
H1b 0.80⁎⁎⁎ 0.64⁎⁎⁎ 0.16⁎⁎⁎
H2a 0.19⁎⁎⁎ 0.07 0.12⁎⁎⁎ different marketing strategies. Marking to potential users should focus
H2b 0.10⁎ 0.29⁎⁎⁎ −0.09⁎⁎⁎ on stickiness, satisfaction and social identification. For current users,
H3a −0.14 0.04 0.18⁎⁎⁎ marketing should focus on attitude, social identification and stickiness.
H3b 0.08 0.10 −0.02 Perception of social identification has significant effects for both groups.
H4a 0.29⁎⁎ 0.24⁎ 0.05⁎⁎⁎
H4b 0.18 0.28⁎⁎ −0.10⁎⁎⁎
This finding shows that users stick with apps because of social identifi-
H5a 0.54⁎⁎⁎ 0.66⁎⁎⁎ −0.12⁎⁎⁎ cation. Thus, app providers should incorporate community and forum
H5b 0.21 0.45⁎⁎⁎ −0.24⁎⁎⁎ functions to encourage users to share their social lives and opinions,
H5c 0.02 0.04 −0.02 and thus prompt them to use the app more frequently and for longer
H6a 0.36⁎⁎ 0.14 0.22⁎⁎⁎
periods. This will, in turn, increase their intention to make in-app pur-
H6b −0.11 0.16 −0.27⁎⁎⁎
H7 0.30⁎⁎ 0.22⁎ 0.08⁎⁎⁎ chases. In addition, the results present the following implications for
app developers and publishers.
⁎ Significant at the p b 0.05 level.
⁎⁎ Significant at the p b 0.01 level.
⁎⁎⁎ Significant at the p b 0.001 level. 1. For potential users, app developers should be aware of the impor-
tance of satisfaction. Given the abundance of similar (and similar-
ly free) apps, satisfaction can increase stickiness and thus the
examines the effects of attitude and satisfaction toward stickiness and user's intention to make in-app purchases. As a result, app devel-
results indicate that both attitudes and satisfaction positively enhance opers should invest more effort in managing user expectations,
app user's intention to continue to use a particular app. which can be increased through mass advertisement and elec-
Second, while studies on extrinsic motivation such as utilitarian tronic word-of-mouth (eWOM). When a user's initial actual expe-
value for IT/IS use have received more attention in the IS literature, rience with an app (i.e., perceived hedonic value and utilitarian
our model implies that intrinsic motivation, such as hedonic value, value) meets such increased expectations, the user will feel satis-
can be an influential factor in affective response of app use. Third, fied. Therefore, app developers should seek to drive increased
given the multi-purpose characteristics (e.g. entertainment, social expectations through early adopters and opinion leaders.
and sharing) of apps, social identification is also verified to be anoth- 2. App developers should seek to improve the attitudes of current users
er important factor in driving in-app purchase intention. Social to increase stickiness for that demographic. Though satisfaction was
norms, on the other hand, do not have a significant effect on either not found to have a significant effect on stickiness, satisfaction was
stickiness or in-app purchase intention. This finding is in line with found to have a significant effect on attitude. For users who had pre-
prior related studies which indicated that social norms have a very viously made in-app purchases, satisfaction was found to have an im-
limited impact on voluntary-use application. pact on attitude, but not stickiness. Hence, app developers should
Finally, the results of this study demonstrate that past adoption seek to improve user attitude, and thereby increase stickiness and in-
experiences have different effects on mobile app usage behaviors. tention to make in-app purchases.
The findings are consistent with prior studies in various contexts
(e.g. Karahanna et al., 1999; Teo and Ranganathan, 2004; Crabbe
et al., 2009). Therefore, this study extends the understanding of IS 5.4. Limitations and future research
adoption in the mobile context and future studies should seek to fur-
ther expand our knowledge in this fast growing field. The results of the present study need to be interpreted with cau-
tion for several reasons. First, statistical analysis provides only nu-
merical relationships, and the results are interpreted subject to the
5.3. Implications for practice author's subjective appraisal. However, the presented results are
consistent with established theories and other studies, thus enhanc-
Our results appear to have important implications for app devel- ing confidence in the findings. Second, respondents were current
opers and publishers seeking to promote their in-app products and mobile app users, thus introducing potential self-selection bias.
services. First, stickiness and social identification had been empiri- Third, the level of analysis is a limiting factor. Since this study was
cally confirmed to have a significant influence on users' intention based on the ABC Model of attitudes, using perceived value and so-
to make in-app purchases. Therefore, app developers should put cial influences as additional antecedents of the behavioral intention
more emphasis on encouraging reuse and increasing the amount of to make in-app purchases, it is impossible to generalize the findings
time users spend on each session. Developers should also incorpo- to other IT/IS purchase behavior. Conceivably, intention to make in-
rate functions which allow and encourage users to share their social app purchases may be subject to additional factors such as personal-
lives through text, pictures, images and video. This sharing behavior ity and price. Moreover, the subjects were app users in Taiwan and
will attract increased feedback from peers, allowing the user to de- many of them were young and well-educated mobile users. Culture
velop a strong sense of social identification through group action. and lifestyles may differ among countries. Future research may
This will, in turn, enhance app stickiness and intention to make in- investigate the influence of these additional variables to produce a
app purchases. According to Find (2014), five of the ten most popular better understanding of in-app purchase behavior.
apps in Taiwan are social and communication apps (e.g., LINE,
Facebook, WhatsApp and WeChat). Such apps help users enhance
their social relationships and their self-concept within a community. Acknowledgments
This reflects the importance of social identification in increasing
stickiness. This study was supported in part by grants from the Ministry of
Second, this study also provides some guidelines to enhance app Science and Technology of the Republic of China under Contract
stickiness. Users were found to perceive utilitarian value and hedonic Number MOST-101-2410-H-031-025-MY2.

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
10 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx

Appendix A. List of items by construct References

A.1. Hedonic value Abdul-Muhmin, A.G., 2011. Repeat purchase intentions in online shopping: the role of
satisfaction, attitude, and online retailers' performance. J. Int. Consum. Mark. 23 (1),
5–20.
1. Using the app is fun for me.⁎ Al-Gahtani, S.S., King, M., 1999. Attitudes, satisfaction and usage: factors contributing to
each in the acceptance of information technology. Behav. Inform. Technol. 18 (4),
2. Using the app gives me pleasure*
277–297.
3. I enjoy using the app.* Alvesson, M., 2000. Social identity and the problem of loyalty in knowledge-intensive
companies. J. Manag. Stud. 37 (8), 1101–1123.
Ajzen, I., 1991. The theory of planned behaviour. Organ. Behav. Hum. Decis. Process. 50,
A.2. Utilitarian value 179–211.
Bagozzi, R.P., Dholakia, U.M., 2006. Open source software user communities: a study of
1. Using the app enables me to accomplish work, learning, communica- participation in Linux user groups. Manag. Sci. 52 (7), 1099–1115.
Bearden, W.O., Etzel, M.J., 1982. Reference group influence on product and brand pur-
tion and transactions more quickly.* chase decisions. J. Consum. Res. 9 (2), 183–194.
2. Using the app enables me to accomplish work, learning, communica- Bentler, P.M., Bonnet, D.C., 1980. Significance tests and goodness of fit in the analysis of
tion and transactions more effectively.* covariance structures. Psychol. Bull. 88 (3), 588–606.
3. Using the app enhances my effectiveness in work, learning, commu- Bhattacherjee, A., 2001. Understanding information systems continuance: an expecta-
tion–confirmation model. MIS Q. 25, 351–370.
nication and transactions. Blau, P., 1964. Exchange and Power in Social Life. Wiley, New York.
4. Using the app improves the quality my work, learning, communica- Bollen, K.A., 1989. Structural Equations With Latent Variables. John Willey, New
tion and transactions.* York.
Brislin, R.W., 1980. Translation and content analysis of oral and written material. In:
Triandis, H.C., Berry, J.W. (Eds.), Handbook of Cross-cultural Psychology Vol. 2.
A.3. Social norms Allyn & Bacon, Boston, MA, pp. 2389–2444.
Brown, J.J., Reingen, P.H., 1987. Social ties and world-of-mouse referral behavior.
J. Consum. Res. 14 (3), 350–362.
1. People who are important to me would approve of me using the Chang, E.C., Tseng, Y.F., 2013. Research note: E-store image, perceived value and
app.* perceived risk. J. Bus. Res. 66 (7), 864–870.
2. People who are important to me think that I should use the app.* Chau, P., Hu, P., 2002. Investigating healthcare professionals' decisions to accept tele-
medicine technology: an empirical test of competing theories. Inf. Manag. 39,
3. People who influence my behavior encourage me to use the app.* 297–311.
Cheung, C.M.K., Chiu, P.Y., Lee, M.K.O., 2011. Online social networks: why do students use
A.4. Social identification facebook? Comput. Hum. Behav. 27 (4), 1337–1343.
Childers, L.T., Carr, L.C., Peck, J., Carson, S., 2001. Hedonic and utilitarian motivations for
online retail shopping behavior. J. Retail. 77, 511–535.
1. Using the app would enhance my chance to meet people with whom Chun, H., Lee, H., Kim, D., 2012. The integrated model of smartphone adoption: he-
donic and utilitarian value perceptions of smartphones among Korean college
I share common interests.*
students. Cyberpsychol. Behav. Soc. Netw. 15 (9), 473–479.
2. I am proud to be a member of this app community.* Crabbe, M., Standing, C., Standing, S., Karjaluoto, H., 2009. An adoption model for mobile
3. Using the app gives me a strong feeling of belonging to a group.* banking in Ghana. Int. J. Mob. Commun. 7 (5), 515–543.
4. This app is a communication channel, allowing users to maintain Davis, F.D., 1989. Perceived usefulness, perceived ease of use, and user acceptance of in-
formation technology. MIS Q. 13, 319–339.
close ties and share information.
Davis, F.D., Bagozzi, R.P., Warshaw, P.R., 1989. User acceptance of computer technology: a
comparison of two theoretical models. Manag. Sci. 35 (8), 982–1003.
A.5. Attitude Davis, F.D., Bagozzi, R.P., Warshaw, P.R., 1992. Extrinsic and intrinsic motivation to
use computers in the workplace. J. Appl. Soc. Psychol. 22 (14), 1111–1132.
Diamantopoulos, A., Siguaw, J.A., 2006. Formative versus reflective indicators in organiza-
1. My attitude toward using the app is favorable.* tional measure development: a comparison and empirical illustration. Br. J. Manag.
2. I like using the app.* 17, 263–282.
Deci, E.L., Ryan, R.M., 1987. Accessibility and stability of predictors in the theory of
3. I feel good about using the app.* planned behavior. J. Pers. Soc. Psychol. 63 (5), 754–765.
Demers, Elisabeth, Lev, Baruch, 2000. A rude awakening: internet shakeout in 2000.
A.6. Satisfaction Working Paper FR 00-13, Bradley Policy Research Center Financial Research and
Policy. William E. Simon Graduate School of Business Administration, University of
Rochester.
1. Using the app makes me feel very satisfied.* Donath, J.S., 1999. Identity and deception in the virtual community. In: Smith,
M.A., Kollock, P. (Eds.), Communities in Cyberspace. Routledge, New York,
2. Using the app gives me a sense of enjoyment.*
pp. 29–59.
3. Using the app makes me feel very contented.* Drell, L., 2013. 9 mobile app KPIs to know. http://mashable.com/2013/09/04/mobile-app-
4. Using the app makes me feel very delighted.* metrics/#bZ.a4FYByiqY (accessed on 02/28/2016).
Dwivedi, Y., Irani, Z., 2009. Understanding the adopters and non-adopters of broadband.
Commun. ACM 52 (1), 122–125.
A.7. Stickiness eMarketer, 2015. http://www.emarketer.com/Article/Mobile-Game-Revenues-Grow-
165-2015-Surpassing-3-Billion/1012063 (accessed on 02/19/2016).
Ellemers, N.De., Gilder, D., Haslam, S.A., 2004. Motivating individuals and groups at work:
1. I would stay longer on this app than other apps.
a social identity perspective on leadership and group perspective. Acad. Manag. Rev.
2. I intend to spend more time on this app.* 29 (3), 459–478.
3. I use this app as often as I can.* Ellis, D.G., Fisher, B.A., 1994. Small Group Decision Making: Communication and the
4. I use this app every time I am online.* Group Process. McGraw-Hill.
Emerson, Richard.M., 1981. Social exchange theory. In: Rosenberg, Morris, Turner,
Ralph H. (Eds.), Social Psychology: Sociological Perspectives. Basic Books, New
A.8. Intention to make in-app purchases York, pp. 30–65.
Engel, J.F., Blackwell, R.D., Miniard, P.W., 1993. Consumer Behavior. seventh ed. Dryden
Press, New York.
1. I intend to continue purchasing in-app products and services.* Find, 2014. http://www.iii.org.tw/Service/3_1_1_c.aspx?id=1356 (accessed on 11/21/
2. I strongly recommend others to purchase in-app products and 2014).
services.* Fishbein, M., Ajzen, I., 1975. Belief, Attitude, Intention and Behavior: An Introduction to
3. I find purchasing in-app products and services to be worthwhile. Theory and Research. Addison-Wesley, Reading, MA.
Fornell, C.R., 1982. A Second Generation of Multivariate Analysis Methods: Vols. I and II
4. I will frequently purchase in-app products and services in the future.*
Praeger Special Studies, New York.
Fornell, C.R., Larcker, D.F., 1981. Structural equation models with unobservable variables
and measurement error. J. Mark. Res. 18 (1), 39–50.
Golden, M.E., 2013. http://flurrymobile.tumblr.com/post/115189750715/the-history-of-
⁎
Denotes items retained for data analysis. app-pricing-and-why-most-apps-are#.Vlk5edKUdqU (accessed on 12/11/2014).

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx 11

Green, D.T., Pearson, J.M., 2011. Integrating website usability with the electronic com- Olsson, U.H., Foss, T., Troye, S.V., Howell, R.d., 2000. The performance of ML, GLS,
merce acceptance model. Behav. Inform. Technol. 30 (2), 181–199. and WLS estimation in structural equation modeling under conditions of
Hair, J.F., Anderson, R.L., Tatham, W.C., 1992. Black, Multivariate Data Analysis With Read- misspecification and nonnormality. Struct. Equ. Model. 7 (4), 557–595.
ings. MacMillan, New York. Overby, J.W., Lee, E., 2006. The effects of utilitarian and hedonic online shopping
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., Tatham, R.L., 2006. Multivariate Data Anal- value on consumer preference and intentions. J. Bus. Res. 59, 1160–1166.
ysis. Pearson University Press, New Jersey. Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y., Podsakoff, N.P., 2003. Common method biases in
Hellier, P.K., Geursen, G.M., Carr, R.A., Rickard, J.A., 2003. Customer repurchase intention: a behavioral research: a critical review of the literature and recommended remedies.
general structural equation model. Eur. J. Mark. 37, 1762–1800. J. Appl. Psychol. 88 (5), 879–903.
Hennig-Thurau, T., Walsh, G., 2003. Electronic word-of-mouth: motives for and conse- Teo, T., Ranganathan, C., 2004. Adopters and non-adopters of business-to-business elec-
quence of reading customer articulation on the internet. Int. J. Electron. Commer. 8 tronic commerce in Singapore. Inf. Manag. 42, 89–102.
(2), 51–74. Ting, K.C., Ting, P.H., Hsiao, P.W., 2014. Why are bloggers willing to share their thoughts
Hsu, C.L., Lu, H.P., 2004. Why do people play on-line games? An extended TAM with social via travel blogs? Int. J. Technol. Manag. 61 (1), 89–108.
influences and flow experience. Inf. Manag. 41 (7), 853–868. Turel, O., Serenko, A., Bontis, N., 2007a. User acceptance of wireless short messaging ser-
Hsu, C.L., Lin, J.C., 2008. Acceptance of blog usage: the roles of technology accep- vices: deconstructing perceived value. Inf. Manag. 44, 63–73.
tance, social influence and knowledge sharing motivation. Inf. Manag. 45, Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D., Wetherell, M.S., 1987. Rediscovering the
65–74. Social Group: A Self-categorization Theory. Basil Blackwell, Oxford, UK.
Hsu, C.L., Lin, J.C., 2010. A study of the adoption behavior for in-car GPS navigation sys- Pöyry, E., Parvinen, P., Malmivaara, T., 2013. Can we get from liking to buying? Behavioral
tems. Int. J. Mob. Commun. 8 (6), 603–624. differences in hedonic and utilitarian Facebook usage. Electron. Commer. Res. Appl.
Hsu, C.L., Lu, H.P., Hsu, H.H., 2007. Adoption of the mobile internet: an empirical 12, 224–235.
study of multimedia message service (MMS). OMEGA: Int. J. Manag. Sci. 35 Scott, J., 1994. The measurement of information systems effectiveness: evaluating
(6), 715–726. a measuring instrument. Proceedings of the Fifteenth International Confer-
Haslam, S.A., Jetten, J., Postmes, T., Haslam, C., 2009. Social Identity, Health and Well- ence on Information Systems, Vancouver, BC, pp. 111–128.
being: An Emerging Agenda for Applied Psychology 58. International Association of Schumacker, R.E., Lomax, R.G., 2004. A Beginner's Guide to Structural Equation Modeling.
Applied Psychology, pp. 1–23. second ed. Lawrence Erlbaum Associates, Inc., Mahwah, NJ.
Hung, S., Ku, C., Chang, C., 2003. Critical factors of WAP services adoption: an empirical Seyal, A., Rahman, M., Rahim, M., 2002. Determinants of academic use of the Internet: a
study. Electron. Commer. Res. Appl. 2 (1), 42–60. structural equation model. Behav. Inform. Technol. 1 (21), 71–86.
Karahanna, E., Straub, D.W., Chervany, N., 1999. Information technology adoption Shih, H.P., 2004. An empirical study on predicting user acceptance of E-shopping on the
across time: a cross-sectional comparison of pre-adoption and post-adoption be- web. Inf. Manag. 41, 351–368.
liefs. MIS Q. 23 (2), 183–213. Singer-Oestreicher, G., Zalmanson, L., 2013. Content or community? A digital business
Keil, M., Tan, B.C.Y., Wei, K.K., Saarinen, T., 2000. A cross-cultural study on escalation of strategy for content providers in the social age. MIS Q. 37 (2), 591–616.
commitment behavior in software projects. MIS Q. 24 (2), 299–325. Siwicki, B., 2013. It's official: Mobile devices surpass PCs in online retail. https://www.
Kim, B., 2010. An empirical investigation of mobile data service continuance: incorporat- internetretailer.com/2013/10/01/its-official-mobile-devices-surpass-pcs-online-
ing the theory of planned behavior into the expectation–confirmation model. Expert retail (accessed on 12/10/2014).
Syst. Appl. 37 (10), 7033–7039. Teo, T.S.H., Lim, V.K.G., Lai, R.Y.C., 1999. Intrinsic and extrinsic motivation in Internet
Kim, H., Chan, H., Gupta, S., 2007. Value-based adoption of mobile internet: an empirical usage. OMEGA Int. J. Manag. Sci. 27 (1), 25–37.
investigation. Decis. Support. Syst. 43 (1), 111–126. Tsai, Hsien, Tung, Bagozzi, Richard, P., 2014. Contribution behavior in virtual communi-
Kim, B., Han, I., 2009. What drives the adoption of mobile data services? An approach ties: cognitive, emotional, and social influences. MIS Q. 38 (1), 143–163.
from a value perspective. J. Inf. Technol. 24 (1), 35–45. Thompson, R., Thompson, C., Higgins, J., 1991. Howell personal computing: toward a con-
Kim, D.J., Hwang, Y., 2012. A study of mobile internet user's service quality perceptions ceptual model of utilization. MIS Q. 15 (1), 125–143.
from a user's utilitarian and hedonic value tendency perspectives. Inf. Syst. Front. Triandis, H.C., 1977. Interpersonal Behavior. Brooke/Cole, Monterey, CA.
14 (2), 409–421. Turel, O., Serenko, A., Bontis, N., 2007b. User acceptance of wireless short messaging
Kline, R.B., 2005. Principles and Practice of Structural Equation Modeling. second ed. The services: deconstructing perceived value. Inf. Manag. 44, 63–73.
Guilford Press, New York. Ullman, J.B., 2006. Structural equation modeling: reviewing the basics and moving for-
Koekkoek, H., 2013. How the Most Successful Apps Monetize Their User Base. Distimo ward. J. Pers. Assess. 87, 35–50.
Publication (file:///C:/Users/user/Downloads/Distimo%20Publication%20- Heijden, H., 2004. User acceptance of hedonic information systems. MIS Q. 28 (4),
%20March%202013%20-%20EN.pdf, accessed on 2/5/2015. March). 695–704.
Kotler, P., 1999. Kotler on Marketing — How to Create Win, and Dominate Markets. Simon Vellerand, R.J., 1997. Toward a hierarchical model of intrinsic and extrinsic motivation.
& Schuster Inc. Adv. Exp. Soc. Psychol. 29, 271–360.
Kuo, Y., Wu, C., Deng, W., 2009. The relationships among service quality, perceived value, Venkatesh, V., Davis, F.D., 2000. A theoretical extension of the technology accep-
customer satisfaction, and post-purchase intention in mobile value-added services. tance model: four longitudinal field studies. Manag. Sci. 46 (2), 186–204.
Comput. Hum. Behav. 25, 887–896. Venkatesh, V., Morris, M.G., Davis, G.B., Davis, F.D., 2003. User acceptance of information
Lee, M.C., 2008. Factors influencing the adoption of internet banking: an integration of technology: toward a unified view. MIS Q. 425–478.
TAM and TPB with perceived risk and perceived benefit. Electron. Commer. Res. Venkatesh, V., Thong, J.Y.L., Xu, Xin, 2012. Consumer acceptance and use of information
Appl. 8, 130–141. technology: extending the unified theory of acceptance and use of technology. MIS
Leong, L.Y., Hew, T.S., Tan, G.W.H., Ooi, K.B., 2013. Predicting the determinants of the NFC- Q. 36 (1), 157–178.
enabled mobile credit card acceptance: a neural networks approach. Expert Syst. Yaloz, B., 2015. Making it stick: why app stickiness matters and top tips to keep your
Appl. 40 (14), 5604–5620. users. http://www.growmobile.com/blog/making-it-stick-why-app-stickiness-
Liang, T.P., Liu, C.C., Wu, C.H., 2008. Can social exchange theory explain individual matters-and-top-tips-to-keep-your-users/ (accessed on 02/28/2016).
knowledge sharing behavior? A meta analysis. Proceedings of the 29th Inter- Yang, J., He, X., Lee, H., 2007. Social reference group influence on mobile phone purchasing
national Conference on Information Systems (ICIS 2008), Paris, France. behaviour: a cross-nation comparative study. Int. J. Mob. Commun. 5 (3), 319–338.
Lin, J.C., 2007. Online stickiness: its antecedents and effect on purchasing intention. Yoo, S.J., Han, S., Huang, W., 2012. The roles of intrinsic motivators and extrinsic motiva-
Behav. Inform. Technol. 26 (6), 507–516. tors in promoting e-learning in the workplace: a case from South Korea. Comput.
Lin, K.Y., Lu, H.P., 2011. Why people use social networking sites: an empirical Hum. Behav. 28, 942–950.
study integrating network externalities and motivation theory. Comput. Zanna, M.P., Rempel, J.K., 1988. In: Bar-Tal, D., Kruglanski, A.W. (Eds.), Attitudes: A New
Hum. Behav. 27, 1152–1161. Look at an Old Concept, in The Social Psychology of Knowledge. Cambridge University
Lin, H.H., Wang, Y.H., 2006. An examination of the determinants of customer loyalty in Press, pp. 315–334.
mobile commerce contexts. Inf. Manag. 43, 271–282. Zott, C., Amit, R., Donlevy, J., 2000. Strategies for value creation in e-commerce: best prac-
Lewis, W., Agarwal, R., Sambamurthy, V., 2003. Sources of influence on beliefs about infor- tice in Europe. Eur. Manag. J. 18, 463–475.
mation technology use: an empirical study of knowledge workers. MIS Q. 27 (4),
657–678. Chin-Lung Hsu is an associate professor of Information Management at National
Luarn, P., Lin, P.H.H., 2005. Toward an understanding of the behavioral to use mobile Taipei University of Business, Taipei, Taiwan. He received MBA and Ph.D. degrees
banking. Comput. Hum. Behav. 21, 873–891. from National Taiwan University of Science and Technology, Taipei, Taiwan, in
MacFarlane, C., 2015. What the app? In-app purchases explained. http://loudcrow.com/ 1999 and 2004, respectively. Dr. Hsu's articles have appeared in Computers in
what-the-app-in-app-purchases-explained/ (accessed on 2/6/2015). Human Behavior, Information & Management, Omega, Internet Research, Electronic
Mardia, K.V., 1970. Measures of multivariate skewness and kurtosis with applications. Commerce Research and Applications, Information Systems and e-Business Manage-
Biometrika 57, 519–530. ment, International Journal of Mobile Communications, International Journal of
Minna, P., 2005. Linking perceived value and loyalty in location-based mobile services. Computer Application in Technology, International Journal of Technology and Human
Manag. Serv. Qual. 15 (6), 509–538. Interaction, International Journal of Mobile Human Computer Interaction, and
Moon, J.W., Kim, Y.G., 2001. Extending the TAM for a world-wide-web context. Inf. Information Management & Computer Security. His research interests include
Manag. 38 (4), 217–230. electronic commerce and Internet marketing.
Monsuwé, T.P., 2004. Dellaert, B.G.C. and de Ruyter, K., what drives consumers to shop
online? Int. J. Serv. Ind. Manag. 15, 102–121. Judy Chuan-Chuan Lin is a professor of Computer Science and Information Management at
Oliver, R., 1993. Cognitive, affective and attribute bases of the satisfaction response. Soochow University, Taipei Taiwan. She received the BSEE from Columbia University, New
J. Consum. Res. 20, 418–430. York; the MS from Polytechnic School of Engineering at New York University, New York;
Oliver, R.L., 1980. A cognitive model of the antecedents and consequences of satisfaction the PhD in MIS from National Taiwan University of Science and Technology. Her papers have
decision. J. Mark. Res. 17, 460–469. appeared in Information and Management, International Journal of Information Management,

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012
12 C.-L. Hsu, J.C.-C. Lin / Technological Forecasting & Social Change xxx (2016) xxx–xxx

Behaviour and Information Technology, Electronic Commerce Research and Applications, Infor- of Mobile Human Computer Interaction, International Journal of Technology and Human Interac-
mation Systems and e-Business Management, Online Information Review, Journal of Information tion, and various international conference proceedings and journals. Her research interests
Science, Internet Research, International Journal of Mobile Communications, International Journal include internet marketing as well as electronic commerce and innovation adoption.

Please cite this article as: Hsu, C.-L., Lin, J.C.-C., Effect of perceived value and social influences on mobile app stickiness and in-app purchase
intention, Technol. Forecast. Soc. Change (2016), http://dx.doi.org/10.1016/j.techfore.2016.04.012