Available online at www.sciencedirect.

com

ScienceDirect
Journal of Interactive Marketing 31 (2015) 28 – 41
www.elsevier.com/locate/intmar

The Effects of Adopting and Using a Brand's Mobile

Application on Customers' Subsequent Purchase Behavior☆
Su Jung Kim a,⁎& Rebecca Jen-Hui Wang b & Edward C. Malthouse c
a
Iowa State University, 116 Hamilton Hall, Ames, IA 50011, USA
b
Northwestern University, 2001 Sheridan Road, 4th floor, Evanston, IL 60208, USA
c
Northwestern University, 1845 Sheridan Road, Evanston, IL 60208, USA

Available online 11 August 2015

Abstract

Mobile applications (apps) have become an important platform for brands to interact with customers, but few studies have tested their effects on
app adopters’ subsequent brand purchase behavior. This paper investigates whether adopters’ spending levels will change after they use a brand’s
app. Using a unique dataset from a coalition loyalty program with implementations of propensity score matching and difference-in-difference-in-
difference methods, we compare the spending levels of app adopters with those of non-adopters. Specifically, we examine whether the use of the
app’s two main interactive features—information lookups and check-ins—influences adopters’ spending levels. We find that app adoption and
continued use of the branded app increase future spending. Furthermore, customers who adopt both features show the highest increase. However,
we also observe “the recency effect” – when customers discontinue using the app, their spending levels decrease. Our findings suggest that sticky
apps which attract continuing uses can be a persuasive marketing tool because they provide portable, convenient, and interactive engagement
opportunities, allowing customers to interact with the brand on a habitual basis. We recommend that firms should prioritize launching a mobile app
to communicate with their customers, but they should also keep in mind that a poorly designed app, which customers abandon after only a few
uses, may in fact hurt their brand experience and company revenues.
© 2015 Direct Marketing Educational Foundation, Inc., dba Marketing EDGE.

Keywords: Mobile app; Log data; Location check-ins; Interactivity; Stickiness; Purchase behavior; Propensity score matching; Difference-in-difference-in-difference
(DDD) model

Introduction mobile technologies comes with the proliferation of various types

of apps. Over 110 billion apps have been downloaded in 2013
The rapid adoption of smartphones and subsequent develop- (Ingraham 2013, October 22; Welch 2013, July 24). The number
ment of mobile applications (“app” or “apps” hereafter) have is expected to surpass 268 billion by 2017 (Fox 2013, Sep 19).
been changing the ways in which customers interact with a brand. Mobile apps account for more than 50% of time spent on digital
According to comScore (2014, October 7), smartphone penetra- media (Lipsman 2014, June 25), suggesting that apps have
tion in the United States reached 72% as of August 2014. The deeply penetrated into the daily lives of smartphone users.
switch from feature phones to smartphones is happening globally, Companies have welcomed mobile apps as an additional
and many parts of the world are embracing the adoption of communication channel to attract new customers and increase
smartphones (Meena 2014, August 8). The rapid growth of brand loyalty among existing ones (Wang, Kim, and Malthouse
2016). They realize that customers use a variety of app features
☆ We acknowledge support from Northwestern University’s IMC Spiegel Digital to perform diverse tasks such as searching, retrieving, and
and Database Research Center and its Executive Director Tom Collinger. We thank sharing information, passing time with entertainment content,
the Air Miles Rewards Program for providing access to their data.
⁎ Corresponding author. paying bills, and navigating maps. Therefore, companies have
E-mail addresses: sjkim@iastate.edu (S. Kim), started to use apps to increase brand awareness and enhance
r-wang@kellogg.northwestern.edu (R.J.-H. Wang), ecm@northwestern.edu brand experience. Given younger customers' high expectations
(E.C. Malthouse). regarding a brand's savvy use of mobile technology (Annalect

http://dx.doi.org/10.1016/j.intmar.2015.05.004
1094-9968/© 2015 Direct Marketing Educational Foundation, Inc., dba Marketing EDGE.
 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41 29

2015, January 28), it becomes increasingly important for devices are unique in that they are portable, personal, networked,
companies to offer branded apps that can provide a seamless, textual/visual and converged. They explain that these unique
convenient, and enjoyable customer experience. characteristics provide value to customers, including information
Despite the growing interest in apps and their potential (e.g., searching for information), identity (e.g., expressing
marketing impact, there is a dearth of research on the use of personality), social (e.g., sharing experience or granting/gaining
branded apps as a persuasive communication channel or loyalty social approval), entertainment and emotion (e.g., killing time
building platform that can influence future spending levels. This by streaming music/movies or playing games), convenience
paper addresses such shortcomings and examines the adoption (e.g., speedy multitasking such as paying bills), and monetary
(i.e., downloading and logging into the app at least once) and uses value (e.g., getting discount or promotion offers).
of two interactive features (i.e., information lookups and location Motivated by these types of value and convenience, customers
check-ins). We analyze the effect of app adoption on purchase may initially adopt an app, however, what makes a branded app
behavior by comparing the spending levels of adopters with those powerful is its interactive nature that allows them to experience
of matched non-adopters, whose pre-adoption demographic and the brand by using its advanced features (Bellman et al. 2011;
behavioral characteristics are similar. We also investigate the Calder and Malthouse 2008; Mollen and Wilson 2010; Novak,
impact of using different app features on purchase behavior. Hoffman, and Yung 2000; Yadav and Varadarajan 2005). Kim,
Lastly, we examine the impact of repeated and discontinued use of Lin, and Sung (2013) document interactive features that promote
a branded app. Most studies have examined the positive outcomes the persuasive effectiveness of branded apps. Their content
of adopting and/or using apps, but we know little about how their analysis of 106 apps of global brands finds that most employ
spending levels change if customers cease using them. About attributes such as vividness, novelty, motivation, control,
20% of the apps are used only once after being downloaded (Hosh customization and feedback. Their findings suggest that, unlike
2014, June 11). Nearly half of customers report that they would computer-based websites, branded apps provide “anytime,
delete an app if they find a single bug in it (SmartBear 2014). anywhere” interactivity with navigation and control features that
Thus, understanding how customers' purchase behavior changes customers can easily use in the mobile environment. Furthermore,
after abandoning a branded app has major marketing implications. these apps also increase customers' enjoyment and willingness to
This paper offers two substantive contributions. First, by continue the relationship with the brands by giving them a sense
linking customers' app use with their actual spending, this study of agency.
provides quantified evidence that shows how adoption of a Although there are numerous studies on mobile marketing,
branded app impacts customers' spending at its firm. Second, it few specifically examined the financial effect of using branded
discusses the mechanisms for how a branded app leads to an apps. One exception is a study by Bellman et al. (2011), which
increase in future spending, using the concepts of interactivity and conducts an experiment to see whether using popular branded
stickiness, which have been mostly discussed in the web context. apps impacts brand attitude and purchase intention. They
By applying these concepts to mobile apps, this study expands demonstrate that using branded apps increases interest in the
our understanding of how the use of interactive technology brand and product category. They also find that apps with an
influences purchase behavior. Lastly, we offer an empirical informational/user-centered style had larger effects on increasing
contribution by using improved measures for key variables—app purchase intention than those with an experiential style because
use and purchase behavior—taken from customers' app logs and the former focus on the users, not on the phone, thus encouraging
transaction data. Given the issues of using self-reported measures them to make personal connections with the brand.
of mobile phone use (Boase and Ling 2013) and purchase Reflecting on a limited number of previous studies, we see
intentions (van Noort, Voorveld, and van Reijmersdal 2012), our that the persuasive effectiveness of branded apps is attributable
study adds to the mobile marketing literature by incorporating to the rich user experience made possible by interacting with the
behavioral measures of app use and purchase histories. app and the brand. Branded apps allow customers to get easy
access to information, enjoy entertainment, receive customized
Literature Review and Hypotheses coupons, and experience the brand on the move. Also, the
interactive features deepen the customer–brand relationship and
Values Created by Using a Mobile App serve as antecedents of positive attitude toward the brand, purchase
intention, and ultimately purchase behavior. Based on the review
A growing body of literature has identified distinctive of existing literature, we propose the following hypothesis.
characteristics of mobile devices and discussed their implications.
Perceived ubiquity is one of the most important aspects of mobile H1. The adoption of a branded app increases subsequent
services (Balasubramanian, Peterson, and Jarvenpaa 2002; Watson spending with a brand.
et al. 2002). Okazaki and Mendez (2013) find that perceived
ubiquity is a multidimensional construct consisting of continuity Experiences of Using Interactive Features of a Mobile App
(i.e., “always on”), immediacy, portability, and searchability. In a
marketing context, mobile media are distinct from mass marketing Research on human–computer interaction provides theoretical
channels because the former allow for location-specific, wireless, explanations for how branded apps create awareness, attitudes,
and portable communication (Shankar and Balasubramanian intentions, and behavior. Previous studies have shown that there
2009). Similarly, Lariviere et al. (2013) contend that mobile are five mechanisms in which technology influences the process
30 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41

of persuasion: (1) triggering cognitive heuristics, (2) enabling the customers to navigate the app, which provides brand informa-
receiver to be the source, (3) creating greater user engagement tion according to their individual preferences through a high
through interactivity, (4) constructing alternative realities, and level of machine interactivity. Sponsor location check-in is
(5) affording easier access to information (Sundar et al. 2013). another feature that enables customers to find nearby sponsor
Among these, greater user engagement and easier access to storefronts using their phones and “check in” at the stores.
information through interactive features seem highly relevant to Customers can play a monthly “Check-In Challenge Game,”
mobile technologies and merit further explanation. make a list of the stores that they have visited, and/or share their
The concept of interactivity has long been discussed since the check-in locations on social media. This feature both provides
emergence of the Internet. Kiousis (2002) defines interactivity entertainment and location-based information as well as instills
after thoroughly reviewing previous literature on this concept as a sense of social connectedness by allowing virtual interactions
“the degree to which a communication technology can create a with other customers and/or brands (i.e., person interactivity).
mediated environment in which participants can communicate Thus, we propose our second hypothesis:
(one-to-one, one-to-many, and many-to-many), both synchro-
nously and asynchronously, and participate in reciprocal message H2. The more interactive features app adopters use, the higher
exchanges (third-order dependency). For human users, it the increase in their subsequent spending.
additionally refers to the ability to perceive the experience as a
simulation of interpersonal communication and increase their Mobile App Stickiness: Repeated vs. Discontinued Use
awareness of telepresence (p. 372).” In a mobile context, Lee
(2005) differentiates between mobile and online interactivity by Up to this point, we maintain that when customers adopt a
highlighting features only available on a mobile platform. In brand's app and use its interactive features, their subsequent
addition to the four components of online activity (user control, spending increases. Questions remain if and how subsequent
responsiveness, personalization, and connectedness), mobile spending changes when app-adopting customers develop a
interactivity provides ubiquitous connectivity (i.e., mobility and habit of using it repeatedly. We use the concept of “stickiness”
ubiquity) and contextual offers (i.e., personal identity and from extant literature to posit how repeated use of an app
localization). He finds that all the elements but personalization influences purchase behavior.
increase customers' trust in mobile commerce, which in turn Website stickiness is defined as “the ability of web sites to draw
improves attitudes toward, and behavioral intentions to, engage and retain customers” (Zott, Amit, and Donlevy 2000, p. 471)
in mobile commerce. and usually measured as the length and/or the frequency of website
Gu, Oh, and Wang (2013) discuss the natures and benefits of visits. Stickiness is regarded as the level of users' value-expectation
machine interactivity and person interactivity. Machine interac- when they visit a site. In other words, if website users decide that a
tivity refers to interactions between humans and the medium. It website meets their expectations and provides enjoyable experi-
provides users with access to dynamic information and the ability ences, they are more likely to come back until the visiting behavior
to customize the information they receive. Previous studies show becomes a routine. Li, Browne, and Wetherbe (2006) approach
that the different tools available to access embedded information stickiness from a relational perspective, arguing that users do not
such as hyperlinks and drag-and-zoom features produce positive perceive a website as being separate from an organization that
attitudes toward the interface and message content (Sundar et al. provides it and treat it as a representative of the organization. Thus,
2010). Person interactivity, on the other hand, is defined as the continued relationship with the website leads to commitment
interactions between humans through a medium. It offers a sense and trust toward the website as well as the organization that
of social connectedness and empowerment among users by provides it. Previous studies support this relational approach:
providing a channel for customers to communicate with the they find a positive association between website stickiness and
brand and/or other customers. Gu, Oh, and Wang (2013) show relational concepts such as trust, commitment, and loyalty (Li,
that person interactivity influences customers' perceived reliability Browne, and Wetherbe 2006), which positively influences
of a website, which in turn increases the financial performance of purchase intention (Lin 2007).
the given brands. More specifically, interactivity increases the Furner, Racherla, and Babb (2014) apply website stickiness
perception of usefulness, ease of use, and enjoyment among those to a mobile context and introduce the concept of mobile app
who adopt the technology (Coursaris and Sung 2012). A recent stickiness (MASS). According to their framework, the outcome
study on mobile location-based retail apps shows that interactivity of mobile app stickiness can manifest in the forms of trust in
has a positive influence on affective involvement, which in turn the provider, positive word-of-mouth behavior, and commer-
leads to downloading and using the apps (Kang, Mun, and cial transactions on the app. If customers find that a branded
Johnson 2015). app provides a novel brand experience that has not been
In this study, we focus on the use of two interactive features possible through other media channels or it fulfills customers'
that are available from a branded app of a coalition loyalty informational needs (e.g., product reviews, store locations,
company: information lookups and sponsor location check-ins. coupons) or entertainment needs (e.g., games, check-ins), this
Customers of the loyalty program can find information on their will build up trust in the value of the app and the provider,
loyalty point balances and transaction records. Additionally, which then increases the level of commitment and repeated
they can search for, and email the information about, reward use. As suggested by the literature on website stickiness, we
items that they are interested in. These features motivate the expect that stickier apps will eventually lead to an increase in
 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41 31

purchase behavior. We use continued patronage as an

indicator of app stickiness in this study, following the
approach in website stickiness research.
In contrast, if consumers find the app useless or irrelevant
(i.e., “less sticky”), it is likely that they will form a negative
attitude toward the brand for its lack of understanding of their
customers' needs. Hollebeek and Chen (2014) discuss how
consumers' perception of brands' actions, quality/performance,
innovativeness, responsiveness, and delivery of promise can
create both positively- or negatively-valenced brand engagement.
We adopt their conceptual model and predict that a branded app
that fails to meet consumers' expectations will create negative
brand attitudes, which will result in a decrease in purchase
intention or actual purchase behavior.
In sum, we expect that downloading and using a well-
designed branded app will improve the customers' attitude
toward, and increase the purchase of, the brand's products or
services. On the contrary, if the app does not satisfy their needs,
they may abandon it, develop negative attitudes toward the (a) (b)
brand, and even decrease future purchases. Thus, we posit:

H3. Repeated use of the app increases subsequent spending.

H4. Discontinuing the use of the app decreases subsequent

spending.

Research Design

Data

We have data from Air Miles Reward Program (AMRP) in

Canada, which has been operating since 1992 and is one of the
largest coalition loyalty programs in the world. Customers of (c)
the loyalty program earn reward points for purchasing at
partnering sponsors across various categories, such as grocer- Fig. 1. Screenshots of the brand's mobile app application.
ies, gas, banking, automobile repairs, and other types of stores.
They can redeem their points for various types of rewards,
including merchandise, gift cards, and flight tickets. and afterwards (from September 2012 to February 2013). The
In 2012, AMRP launched a mobile app on Android and iOS. company also provided a control group of another 10,766
After logging in, customers can check point balances, transaction customers who had never adopted the app during or prior to
histories and reward items as shown in Fig. 1a. Besides looking February 2013. To reduce potential selection bias, we match each
up information, one notable feature is sponsor location check-in, app adopter with a non-adopter using demographics and pre-
which allows customers to find nearby sponsors (e.g., gas, period point accruals, which serve as a proxy for measuring their
grocery, or toy stores) with GPS-enabled mobile devices and spending behavior. This unique dataset and quasi-experimental
“check in” at the stores (Fig. 1b). App users can play the design allow us to estimate the effects of adopting and using the
“Check-In Challenge,” a game that picks 50 customers with the branded app on customers' spending levels.
most check-ins each month and rewards them with double points
(Fig. 1c). The check-in feature is similar to other location-based
apps such as Foursquare (Frith 2014). Propensity Score Model and Matching
On August 30, 2012, AMRP released an app update and
advertised two particular features—balance checking and As with all observational studies, our research may be subject
location check-in. Using this release date to separate the to selection biases because customers are not randomly assigned
customers' point accruals before and after app adoptions, we to app adoption. Therefore, adopters and non-adopters may have
analyze 10,776 app adopters, who downloaded and logged in preexisting differences that also influence their post-period point
with the app in September 2012. Besides their app login and accruals, which is our dependent variable of interest and a proxy
location check-in histories, we also have their point accrual that measures their spending behavior. Given the potential of
histories six months prior (from March 2012 to August 2012) confounds, we employ propensity score matching, which is one
32 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41

of the most widely used methods to reduce selection bias twenty-one NDs are greater than 0.10.2 After matching, the NDs
(Rosenbaum and Rubin 1985; Rubin 2007; Stuart 2010). range from 0.0004 to 0.0282. Except for the three variables,
Rosenbaum and Rubin (1985) define matched samples as a ln(Points Accrued in March + 1), ln(Points Accrued in July + 1),
method for selecting untreated units to form a control group and ln(Points Accrued from Banking + 1), which already have
“that is similar to a treated group with respect to the distribution very small NDs to begin with, all other covariates' balances are
of observed covariates.” They show that, conditional on those improved after matching.
observed variables, treatment in the matched sample is, in effect, Whether matched sampling should be performed with or
randomized, thus reducing selection bias. However, matching on without replacement depends on the original dataset. Matched
every characteristic becomes challenging when the dataset has sampling with replacement yields better matches, because any
many dimensions. To overcome the curse of dimensionality, controls that are already matched with treated units can still be
extant literature suggests calculating, and matching on, the available for the next matching iteration. Obtaining closer
propensity scores (Rosenbaum and Rubin 1983). We first matches is especially important when the original reservoir of
estimate a propensity score model relating customers' propensity control units is small, as in our study. Our final matched sample
to become app users with their characteristics. We then calculate consists all app adopters (n = 10,776) and 5,127 distinct non-
the estimated propensity for each user and select a subset of the adopter customers. In the following regression models that test
control customers that have similar estimated values.1 our hypotheses, if a control customer is matched with two (or
We now specify the propensity score model that estimates more) mobile adopters, his records are duplicated in order to
the probability, or propensity, of app adoption. Denoted by Pi, maintain the one-to-one ratio. After obtaining a matched sample
it is defined as: and a propensity score for each of its customer, we include the
score as a covariate in the regressions that test our hypotheses.
Pi ≡ Pr½Adopt i ¼ 1jd i ; ln ðvi þ 1Þ; ð1Þ
Effects of App Adoption by Feature Usage (H1 and H2)
where Adopti is a binary variable that indicates whether
customer i adopts the mobile app in September 2012, vector di
Having obtained the propensity scores and a matched
contains demographics such as age and gender, and vector vi
sample, we proceed with a difference-in-difference-in-difference
includes behavioral characteristics exhibited during the six
(DDD) regression, which is an extension of the difference-in-
months prior to app adoption, i.e., total points accrued in each
difference (DD) model that is used when there is only one
month from March through August 2012, total points accrued
treatment of interest. The motivation behind a DD model is to
by spending on banking, retail, food, gas, and other miscella-
control for any a priori individual differences between treated and
neous charges, total number of sponsors visited, number of
control units. Even though we already account for observed
points used for reward redemptions, and customers' tenure with
characteristics using propensity score matching, the adopters and
the program in days. All numerical behavioral variables are
non-adopters may still have unobserved differences. By assuming
log-transformed to symmetrize the distributions and reduce the
that those unobserved characteristics for both the treated and the
influence of extreme observations. We use binary logistic
control groups remain unchanged before and after the treatment
regression to estimate Pi.
shock (i.e., app's update release), a DD model accounts for the
Having estimated the propensity score model, we compute the
^ i, groups' fixed effects and seasonality to derive the treatment effect
estimated probability to adopt the app, i.e., the propensity score P
by comparing the before-and-after changes in the dependent
for each customer. We then employ 1:1 matching (1 control to 1
variable of the treated group with that of the control group. A
app adopter) using the nearest-neighbor matching algorithm with
DDD model extends a DD model, and it is used when there are
replacement. We evaluate whether covariate balance improves by
two treatments present instead of one. In our study, we are
calculating the normalized differences (NDs) in means for all
interested in the effect of adopting two app features, namely,
covariates and compare the two sets of NDs before and after
information lookup and sponsor check-ins. We develop a
matching (Imbens and Rubin 2015). ND is defined as:
balanced, monthly-level, cross-sectional panel from the cus-




x jt − x jc
tomers' purchase histories and demographic profiles. The time
ND j ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð2Þ horizon includes both the pre- (March 2012 to August 2012) and
S 2jt þ S 2jc post-adoption periods (September 2012 through February 2013).
2 The DDD model is specified as:

where NDj is the normalized difference in means for covariate j, x jt

ln ðyit þ 1Þ ¼ α1 þ α1i þ α1p þ β1 0 ðLi  T Þ
is the mean of the covariate for the treated units, x jc is the mean of
þ β2 0 ðC i  T Þ þ β3 0 ðLi  C i  T Þ
the covariate for the control units, sjt2 is the variance of the covariate
þ β 4 Li þ β 5 C i þ β 6 ðLi  C i Þ þ β7 0 T
j for the treated units, and sjc2 is the variance of the covariate j for
^ i þ εit
þ β8 0 d i þ β9 0 xi þ β 10 P ð3Þ
the control units. Table 1 shows that before matching, nine out of

1
Readers can find the rationale for choosing propensity score matching over 2
Extant literature (e.g., Imbens and Rubin 2015, p. 277), suggest that as a
the two-stage-least-squares (2SLS) methods with instrumental variables in general rule of thumb, NDs should be close to zero, and if they are greater than
Appendix A. 0.25, regression analyses may become unreliable.
 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41 33

Table 1
Descriptive statistics of propensity score model variables, before and after matching.
App-adopter Non-adopter Normalized difference In means
Before matching After matching Before matching After matching
Mean Std. dev. Mean Std. dev. Mean Std. dev.
Demographics
Is aged 18 to 34 0.4410 0.4965 0.1814 0.3854 0.4408 0.4965 0.5841 0.0004
Is aged 35 to 44 0.2105 0.4077 0.1561 0.3630 0.2124 0.4090 0.1409 0.0047
Is aged 45 to 54 0.1281 0.3342 0.1853 0.3886 0.1334 0.3400 0.1578 0.0157
Is aged 55 to 64 0.0684 0.2524 0.1566 0.3634 0.0658 0.2479 0.2819 0.0104
Is aged 65 plus 0.0282 0.1656 0.1425 0.3496 0.0295 0.1692 0.4179 0.0078
Is female 0.3389 0.4734 0.3988 0.4897 0.3433 0.4748 0.1244 0.0093
Is male 0.3299 0.4702 0.2717 0.4449 0.341 0.4741 0.1272 0.0235
Pre-adoption purchase behavior
ln(Points Accrued in Mar + 1) 2.2653 1.71 2.2592 1.7035 2.2532 1.6913 0.0036 0.0071
ln(Points Accrued in Apr + 1) 2.2771 1.6952 2.3183 1.7014 2.2721 1.6884 0.0243 0.0030
ln(Points Accrued in May + 1) 2.5226 1.7533 2.4971 1.7463 2.5167 1.7493 0.0146 0.0034
ln(Points Accrued in Jun + 1) 2.5408 1.7597 2.5664 1.7562 2.5511 1.7612 0.0146 0.0059
ln(Points Accrued in Jul + 1) 2.4080 1.6857 2.4277 1.6916 2.3772 1.6978 0.0117 0.0182
ln(Points Accrued in Aug + 1) 2.4853 1.7151 2.4050 1.7268 2.4724 1.7178 0.0467 0.0075
ln(Points Accrued from Grocery + 1) 2.8969 1.9744 3.0000 2.0102 2.8811 1.9688 0.0517 0.0080
ln(Points Accrued from Retail + 1) 1.1762 1.5215 1.2067 1.4980 1.2049 1.5148 0.0202 0.0189
ln(Points Accrued from Gas + 1) 1.6599 1.8171 1.3167 1.7125 1.6548 1.7940 0.1944 0.0028
ln(Points Accrued from Banking + 1) 1.2967 2.3615 1.2906 2.3640 1.2808 2.3805 0.0026 0.0067
ln(Points Accrued from Other + 1) 1.0846 1.6352 0.9827 1.5916 1.0665 1.6013 0.0632 0.0112
ln(Number of Sponsors Visited + 1) 1.4177 0.5345 1.3791 0.5137 1.4208 0.5069 0.0736 0.0060
ln(Redeemed Points + 1) 0.4485 1.7383 0.5687 1.9397 0.4391 1.7123 0.0653 0.0054
ln(Tenure in Days + 1) 7.7595 0.9022 8.0336 0.8936 7.7331 0.9715 0.3053 0.0282
Note. After matching, 10,776 mobile app adopters and 10,776 paired non-adopters as a control group with replacement (5,127 distinct non-adopters).

where yit is the number of points accrued3 (1 point is earned by signs and magnitudes of vectors β1, β2, and β2, whose estimates
spending approximately $30 CAD) by customer i in month t, and give us the percent changes in yit after each type of app adoptions.
α1 is the fixed intercept. We classify each customer in the Finally, we control for any observed heterogeneity effects,
matched sample into one of the four categories: non-adopters, including customers' demographics (di); pre-adoption behavior
adopters who only do information lookups with the app, adopters (xi), which contains total number of sponsors visited, number of
who only do sponsor check-ins, and adopters who use both points used for reward redemptions and customers' tenure with the
features. The coefficients of Li and Ci account for the group fixed- program in days5; and propensity score ðP ^ i Þ. 6 We also control for
effects of being a customer that lookups information or one that customers' unobserved heterogeneity at both the matched-paired
checks in. If a customer does both, his fixed-effect, which is level, α1p, and at the individual level, α1i. Because we have done
assumed to be constant before and after the app update release, is propensity score matching, i.e., each adopter is paired with a
captured by the coefficient of Li × Ci in addition to those of main non-adopter, we have a multi-level structure, in which individuals
effects Li and Ci. In order to both control for seasonality and within a matched pair are correlated, and observations within an
observe how the influence of app use may vary over time, we use individual are also correlated. Thus, unobserved heterogeneity is
vector T, which contains six binary variables indicating if points accounted for both within a matched pair (α1p) and at the individual
have been accrued during a given month from September 2012 level (α1i). Each of α1p and α1i is assumed to be i.i.d. normally
through February 2013, e.g., T = [1 0 0 0 0 0] for t = distributed with mean 0 and variances of σp2 and σi2, respectively.
September.4 To test H1 and H2, therefore, we observe the
Quantifying the Effect of App Stickiness and Disengagement
(H3 and H4)
3
It has been shown that in the DD framework, estimation implementation that
uses a stack of observations with saturated explanatory variables (i.e., treatment, The aforementioned DDD model allows us to assess spending
time, and their interaction) is equivalent to taking the paired difference (Angrist changes after adopting the app. Now we proceed with examining
and Pischke 2009, pp. 233–234). Taking the paired difference has the the effect of repeated use as posited in H3, which is tested with
disadvantage of not being able to easily control for fixed effects due to
demographics and other observable heterogeneity. Thus, in our model, the
5
dependent variable of interest is log-transformed points accrued in a given Pre-adoption point accruals are excluded because they are the dependent
month, rather than a difference. variable in the model, per the DDD framework.
4 6
Similar to a traditional DD model, our model uses observations in any given The descriptive statistics and the correlation matrix of focal variables are
month prior to the app update release on August 30, 2012 as the baseline. available from the corresponding author upon request.
34 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41

the log–log model depicted below.7 Using data from the app the number of information lookups a customer did during the
adopters and their matched non-adopters, we regress each week (Look_Frequencyiw) and the number of sponsor check-ins
customer's weekly spending from September 2012 to February (Check_Frequencyiw). We also include two-way interaction
2013 on the causal variables that represent repeated use and terms, Liw × Agei and Ciw × Agei, indicating whether customer
discontinued use. We specify our model as follows: i belongs to a particular age group and whether he does
information lookups or sponsor check-ins during week w,
ln ðyiw þ 1Þ ¼ α2 þ a2i þ a2p þ γ1 ln ðCumulative Look iw−1 þ 1Þ respectively. As with the propensity score model, we use
þγ2 ln ðCumulative Check iw−1 þ 1Þ customers with unknown age as baseline, and categorize the
þγ3 Is Before First Look iw þ γ4 Is Before First Check iw rest of the customers into one of five age groups (aged 18–34;
þγ5 ln ðLook Recencyiw þ 1Þ 35–44; 45–54; 55–64; and 65 and older). Similarly, we include
þγ6 ln ðCheck Recencyiw þ 1Þ two-way interaction terms for gender, Liw × Genderi and Ciw ×
þγ7 ln ðLook Frequencyiw þ 1Þ ð4Þ Genderi, to indicate whether the customer belongs to a gender
þγ8 ln ðCheck Frequencyiw þ 1Þ þ γ9 0 ðLiw  Agei Þ
þ γ10 0 ðC iw  Agei Þ þ γ011 ðLiw  Genderi Þ
group and whether he does information lookups or sponsor check-
þ γ012 ðC iw  Genderi Þ þ γ13 Is iOS Useri ins during week w. Again, we use customers with unknown gender
þ γ14 Is Android Useri þ γ15 0 w þ γ16 0 d i þ γ17 0 vi as baseline, and categorize the rest as either male or female. We
þ γ18 P^ i þ εiw also control for the platform that operates the app since conven-
tional wisdom suggests that iOS users may be more affluent than
Android users, and thus customers may have different spending
where yiw is the number of points accrued in week w, and as we
levels depending on which mobile platforms they use. As with age
have already mentioned, is a close approximation of the customer's
and gender, we use unknown platform as the baseline, and denote
spent dollars, and α2 is the fixed intercept. We represent repeated
Is_iOS_Useri as the binary variable that indicates whether we
app use with two variables: cumulative number of information
observe customer i being an iOS user. Similarly, Is_Android_Useri
lookups (Cumulative_Lookiw − 1) and cumulative number of
indicates whether the customer used Android anytime during our
sponsor check-ins (Cumulative_Checkiw − 1) customer i has done
study period. Finally, we control for weekly seasonality for each of
by the end of the previous week, w − 1. To verify H3, we observe
the 25 weeks, as well as customers' observed heterogeneity (di
the signs and magnitudes of γ1 and γ2.
and vi) and unobserved heterogeneity (α2i and α2p).
H4 posits that discontinued use of the app will decrease
future purchases. We operationalize discontinued use with
Results
recency measures of lookups (Look_Recencyiw) and check-ins
(Check_Recencyiw), recording the days since the last app use for a
Propensity Score Model
given feature by the week's end. Prior to customers' app feature
adoptions, their recency measures are set to 0 day, and binary
First, we estimate the propensity score model, which predicts
variables, Is_Before_First_Lookiw and Is_Before_First_Checkiw,
the likelihood of a customer adopting the app. Table 2 shows the
are set to 1 indicating that they have yet to use the app feature
estimates. Younger customers are more likely to adopt than older
(Is_Before_First_Lookiw for not having adopted information
customers, and the oldest customers are the least likely to adopt.
lookups yet, and Is_Before_First_Checkiw for not having adopted
Males are more likely to adopt than females. Customers with
check-ins). After app adopters' first uses of an app feature, their
higher point accruals in August just before the adoption (β^ ¼ 0:
corresponding binary variables are set to 0. Note that smaller
078, p b .001) are more likely to adopt the app. Those with point
values of recency correspond to more recent events. Thus, we
accruals due to spending on gas ( β^ ¼ 0:09 , p b .001) and
expect negative estimates of γ5 and γ6.
miscellaneous charges (β^ ¼ 0:02, p b .05) are also more likely to
Besides capturing the causal variables of interest, namely, the
adopt. Customers who used more points to redeem for rewards
effect of repeated use and discontinued use of the app, we control
are less likely to adopt ( β^ ¼ −0:026, p b .01), suggesting that
for other behavioral variables and interaction terms. Since app
after a large reward redemption, they are less likely to use the app
use within the week might influence spending, yiw, we include
to learn more about the loyalty program or to check their point
7
The correlation table of main variables suggests that our dependent variables balances. After the propensity score calculations, we match up
of interest regarding H3 and H4 may be highly correlated. To ensure that our each app adopter with a control customer, i.e., non-adopter.
reported results for H3 and H4 are not influenced due to potential multi-
collinearity, we conduct four additional sets of random effect models as Effects on Point Accruals After App Adoption (H1 and H2)
robustness checks: (1) the first alternative model examines the effect of
cumulative use of lookups and check-ins while excluding all of the recency
variables, (2) the second alternative model examines the effect of app use
After obtaining the propensity scores and a matched sample,
recency while excluding the effects of cumulative app use, (3) the third we begin our regression analyses to test our hypotheses. We use
alternative model examines the effect of discontinued lookups without any of a random intercept DDD model that incorporates both observed
the check-in variables, and (4) the fourth alternative model examines the effect heterogeneity and unobserved heterogeneity. In addition to the
of discontinued check-ins without any of the look up variables. The results of DDD model specified in Eq. (3), as a robustness check, we also
these four alternative models remain similar to those from our original model,
and in some cases their standard errors are smaller, suggesting even higher run an alternative DDD model, where instead of T, we denote
levels of significance. In sum, our findings pertaining to H3 and H4 remain time-fixed effect before and after the app update release with a
unchanged after having conducted numerous robustness checks. binary variable ta, which is set to 1 if the given month is after the
 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41 35

Table 2 Results from the month-to-month model show how the effect
Propensity score model estimates. changes over time. The impact on subsequent spending is the
Dependent variable Is becoming a user of the app greatest immediately after the adoption: compared to the monthly
Estimate (Std. err.) spending average in the pre-adoption period, in September, being
Intercept –0.4041 ⁎⁎ (0.1534) a user who does information lookups leads to a exp(0.314) −
Is aged unknown (Baseline) 1 = 37% increase ( β^ ¼ 0:314, p b .001) in point accruals, while
Is aged 18–34 1.2291 ⁎⁎⁎ (0.0481) using sponsor check-ins gives exp(0.280) − 1 = 32% ( β^ ¼ 0:28
Is aged 35–44 0.6428 ⁎⁎⁎ (0.0512) 0, p b .001). The effects attenuate afterwards. In October, being a
–0.0171
Is aged 45–54 (0.0534)
user of information lookups leads to a 23% (β^ ¼ 0:207, p b .001)
–0.4597 ⁎⁎⁎
increase, while using sponsor check-ins yields 14% ( β^ ¼ 0:133,
Is aged 55–64 (0.0599)
Is aged 65 + –1.2260 ⁎⁎⁎ (0.0749)
Is gender unknown (Baseline) p b .001). By the end of our study period (February 2013), the
Is female –0.2091 ⁎⁎⁎ (0.0370) effects are 22% ( β^ ¼ 0:199 , p b .001) and 16% ( β^ ¼ 0:150 ,
Is male 0.1288 ⁎⁎⁎ (0.0387) p b .01) for information lookups and sponsor check-ins, respec-
ln(Points Accrued in Mar + 1) –0.0006 (0.0145)
tively. The interaction effect from being a user of both features
ln(Points Accrued in Apr + 1) –0.0207 (0.0153)
ln(Points Accrued in May + 1) –0.0034 (0.0152) is positive but insignificant for all months except for September,
ln(Points Accrued in Jun + 1) –0.0515 ⁎⁎⁎ (0.0151) right after the app adoption (β^ ¼ −0:122, p b .05). Therefore, for
ln(Points Accrued in Jul + 1) –0.0427 ⁎⁎ (0.0154) customers that use both features, their September spending
ln(Points Accrued in Aug + 1) 0.0776 ⁎⁎⁎ (0.0145) increases by exp(0.314 + 0.280 − 0.122) − 1 = 60%, and their
ln(Points Accrued from Grocery) 0.0166 (0.0125)
February spending increase by exp(0.199 + 0.150) −1 = 42%.
ln(Points Accrued from Retail) –0.0139 (0.0122)
ln(Points Accrued from Gas) 0.0934 ⁎⁎⁎ (0.0106) Without complex econometrics, Fig. 2 presents model-free
ln(Points Accrued from Banking) 0.0272 + (0.0109) line charts that show how different types of app users increase
ln(Points Accrued from Other) 0.0174 ⁎ (0.0102) their point accruals over time. As with the DDD model, we
ln(Number of Sponsors Visited + 1) 0.0786 (0.0488) classify all app adopters into one of three types: adopters who
ln(Redeemed Points + 1) –0.0262 ⁎⁎ (0.0084)
only do information lookups with the app during the post-
ln(Tenure in Days + 1) –0.0162 (0.0185)
adoption period of September 2012 to February 2013, adopters
Note. 10,776 app adopters and 10,776 non-adopters before matching. who only do check-ins, and adopters who use both features. The
Standard errors are in parentheses.
+
p b .10.
charts also show point accruals of each type's matched control
⁎ p b .05. group, i.e., non-adopters. The gap between the lines of adopters
⁎⁎ p b .01. and non-adopters shows a moderate increase in September for
⁎⁎⁎ p b .001. customers who do information lookups and a sharp increase
for customers who use both features, suggesting that the
app release, and 0 otherwise, while also controlling for seasonality more features the adopters use, the more they outspend the
month-by-month. Thus, Table 3 presents two sets of estimate non-adopters.
results—the overall effect on monthly spending after adoption,
and the effect broken down by month from September 2012 to
February 2013 per specifications in Eq. (3). The latter reveals Effects of Repeated App Use on Point Accruals (H3)
whether and by how much the effect of app adoption persists.
Any combination of feature use leads to an increase in Confirming H1 and H2 establishes that, using the branded app
subsequent point accruals (and therefore spending),8 which is associated with an increase of 19%–48% in point accruals (and
confirms H1. Overall, being an adopter who does information therefore spending) over the course of six months after adoption.
lookups leads to a exp(0.213) − 1 = 24% increase (β^ ¼ 0:213, Using a separate model of a balanced weekly panel, constructed
p b .001) in post-adoption spending, while using sponsor from customers' purchase histories and demographic profiles, we
check-ins gives an effect of exp(0.177) − 1 = 19% ( β^ ¼ 0:17 now seek to determine the effect of repeated use on spending (H3).
7 , p b .001). Thus, customers who use both features increase We operationalize repeated app use with two variables of interest:
their spending by exp(0.213 + 0.177) −1 = 48% after adopting the cumulative numbers of information lookups and sponsor
the app. We therefore confirm H2—using different combinations check-ins a customer has done by the end of the previous week. As
of app features leads to different effects on subsequent spending.9 shown in Table 4, both variables have positive and significant
influence on point accruals. If customers increase their cumulative
information lookups by 10%, their weekly point accruals increase
8
We also performed sensitivity analyses according to Rosenbaum bounds
by (1.100.0182) − 1 =0.17% ( β^ ¼ 0:0182, p b .001), and if they
(Keele 2010; Rosenbaum 2005, 2010). Results show that for customers that increase their cumulative check-ins by 10%, their weekly point
only do information lookups or check-ins, unobserved variables will have to accruals increase by (1.100.0181) − 1 =0.17% ( β^ ¼ 0:0181,
influence the odds of app adoption by 20% for the effect on point accruals to p b .01). Based on the matched sample, which spans from
become null. For customers that do both check-ins and information lookups, September 2012 to February 2013, on average, customers' weekly
they will have to change the odds of app adoption by 50% to nullify the effect.
9
As a robustness check, we estimate a third DDD model using the sample of point accruals is 11.9 and their cumulative information lookups
5,127 control customers and matching it to 5,127 app adopters. The findings and check-ins are 1.7 and 1.5, respectively. Thus, by increasing
remain the same. See Appendix B for details. their cumulative information lookups by 1 unit, their weekly point
36 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41

Table 3
Estimates of focal parameters, DDD model (H1 and H2).
Dependent variable ln(Points Accrued in Month t + 1)
Lookup Check-in Lookup × Check-in
Is Post-Adoption 0.2131 ⁎⁎⁎ (0.0128) 0.1770 ⁎⁎⁎ (0.0353) 0.0148 (0.0421)
Sep 0.3140 ⁎⁎⁎ (0.0169) 0.2804 ⁎⁎⁎ (0.0457) –0.1218 ⁎ (0.0541)
Oct 0.2069 ⁎⁎⁎ (0.0175) 0.1329 ⁎⁎ (0.0488) 0.0745 (0.0569)
Nov 0.2189 ⁎⁎⁎ (0.0188) 0.1443 ⁎⁎ (0.0528) 0.0754 (0.0614)
Dec 0.1805 ⁎⁎⁎ (0.0195) 0.1546 ⁎⁎ (0.0536) 0.0075 (0.0626)
Jan 0.1594 ⁎⁎⁎ (0.0188) 0.2001 ⁎⁎⁎ (0.0507) 0.0187 (0.0602)
Feb 0.1987 ⁎⁎⁎ (0.0186) 0.1498 ⁎⁎ (0.0500) 0.0347 (0.0592)
Demographics
Is aged 18 to 34 –1.1045 ⁎⁎⁎ (0.0641)
Is aged 35–44 –0.5035 ⁎⁎⁎ (0.0431)
Is aged 45–54 0.2314 ⁎⁎⁎ (0.0304)
Is aged 55–64 0.5321 ⁎⁎⁎ (0.0434)
Is aged 65 plus 0.9811 ⁎⁎⁎ (0.0694)
Is female 0.1604 ⁎⁎⁎ (0.0214)
Is male –0.0412 ⁎ (0.0197)
Pre-adoption behavioral characteristics and propensity score
ln(Number of Sponsors Visited + 1) 1.5273 ⁎⁎⁎ (0.0150)
ln(Redeemed Points + 1) 0.1343 ⁎⁎⁎ (0.0045)
ln(Tenure in Days + 1) 0.0864 ⁎⁎⁎ (0.0088)
Propensity score 3.5277 ⁎⁎⁎ (0.2010)
Fixed intercept –2.0356 ⁎⁎⁎ (0.1080)
Matched pair random intercept covariance 0.0062 (0.0108)
Individual random intercept covariance 1.0292 ⁎⁎⁎ (0.0152)
Note. The coefficients labeled under “Is Post-Adoption” are estimated from an alternative model that examines the overall treatment effect. Coefficients labeled by
month, i.e., Sep through Feb, are estimated using the model specified in Eq. (3), which breaks down the treatment effect by month after app adoption. Estimates for
demographics, pre-adoption behavioral characteristics, propensity score, fixed intercepts, and random intercept covariance are the same between the two models.
Estimates for seasonality effects are not reported. 10,776 app adopters and 10,776 paired non-adopters as a control group with replacement (5,127 distinct
non-adopters). Total number of observations = 10,776 × 2 × 12 months = 258,624.
Robust standard errors per White (1980) are in parentheses.
+
p b .10.
⁎ p b .05.
⁎⁎ p b .01.
⁎⁎⁎ p b .001.

0:0182
accruals increase by ð1þ1:7
1:7 Þ −1 ¼ 0:85% , or 0.10 points, Effects of Discontinued Use on Point Accruals (H4)
which is equivalent to $3.0 CAD a week. If the customers increase
their check-ins by 1 unit, their weekly point accruals increase by H4 predicts that discontinuing the use of the app has a
0:0181
ð1þ1:5
1:5 Þ −1 ¼ 0:93%, or 0.11 points, which is equivalent to negative influence on spending. We operationalize discontinued
$3.3 CAD. H3 is confirmed. use with recency, the number of days since using the feature.

Before Adoption After Adoption Before Adoption After Adoption Before Adoption After Adoption

Fig. 2. Average monthly point accruals by app adoption type. Note. One point is equivalent to approximately $30 CAD in spending. Number of adopters of both
features = 2,531. Number of adopters of information lookups = 7,652. Number of adopters of check-ins = 573. Total of 10,776 adopters and 10,776 paired
non-adopters as a control group with replacement (5,127 distinct non-adopters).
 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41 37

Table 4 Both variables' coefficients are negative and significant10

Estimates of focal parameters, log–log model (H3 and H4). (information lookup recency: β^ ¼ −0:012 , p b .05; sponsor
Variable Estimate (Std. err.) check-in recency: β^ ¼ −0:017, p b .01), suggesting that as the
App use number of days since the last information lookup or check-in
ln(Information Lookups in Week w + 1) 0.3314 ⁎⁎⁎ (0.0207) increases, spending levels drop. If customers increase their
ln(Check-ins in Week w + 1) 0.0364 ⁎ (0.0177) information lookup recency by 10%, their weekly point accruals
ln(Cumulative Lookups by Week w − 1 + 1) 0.0182 ⁎⁎⁎
decrease by 1 − (1.10− 0.012) = 0.11%. If they increase their
(0.0053)
ln(Cumulative Check-ins by Week w − 1 + 1) 0.0181 ⁎ (0.0090)
Has no lookups so far by week w –0.0765 ⁎⁎⁎ (0.0225) check-in recency by 10%, their weekly point accruals decrease by
Has no check-ins so far by week w –0.05114 + (0.0304) 1 − (1.10− 0.017) = 0.16%. On average, customers' information
ln(Information Lookup Recency in Days –0.0122 ⁎ (0.0058) lookup and check-in recencies are 3.8 and 8.5, respectively.
in Week w + 1) If customers lapse in their information lookups by one day, their
ln(Check-in Recency in Days in Week w + 1 ) –0.0170 ⁎ (0.0070) −0:012
Demographics and interactions 3:8 Þ
weekly point accruals drop by 1−ð3:8þ1 ¼ 0:28%, which is a
Is aged 18 to 34 0.0685 (0.1516) decrease of 0.03 points, or $1.0 CAD a week. If customers lapse in
Is aged 18 to 34 × has lookups in week w –0.0448 ⁎ (0.0210) their check-ins by one day, their weekly point accruals drop by
−0:017
0.1141 ⁎⁎
8:5 Þ
1−ð8:5þ1 ¼ 0:19%, which is a decrease of 0.02 points, or
Is aged 18 to 34 × has check-ins in week w (0.0429)
Is aged 35 to 44 0.0762 (0.0830) sixty-seven cents CAD a week. The results suggest that as
Is aged 35 to 44 × has lookups in week w –0.0810 ⁎⁎⁎ (0.0244)
customers cease using the app, the brand's revenues will decrease.
Is aged 35 to 44 × has check-ins in week w 0.0814 (0.0563)
Is aged 45 to 54 0.1032 ⁎⁎⁎ (0.0135) H4 is confirmed.
Is aged 45 to 54 × has lookups in week w –0.1326 ⁎⁎⁎ (0.0280)
Is aged 45 to 54 × has check-ins in week w 0.0460 (0.0617) Discussion
Is aged 55 to 64 0.0238 (0.0560)
Is aged 55 to 64 × has lookups in week w –0.0965 ⁎⁎ (0.0357)
Mobile apps are examples of “non-push marketing contacts”
Is aged 55 to 64 × has check-ins in week w –0.0853 (0.0750)
Is aged 65 plus –0.0510 (0.1241) (Shankar and Malthouse 2007, p. 3), with which customers
Is aged 65 plus × has lookups in week w –0.0873 (0.0541) decide to interact with a brand. It is important to understand their
Is aged 65 plus × has check-ins in week w 0.2185 (0.1546) effectiveness so that firms know whether to invest in mobile
Is female 0.0068 (0.0252) strategies. This study examines the effects of adopting and
Is male –0.0008 (0.0166)
Is female × has lookups in week w –0.0464 ⁎ (0.0189)
interacting with a brand's app on subsequent spending levels.
Is female × has check-ins in week w 0.0512 (0.0461) It also tests how the use of different interactive app features
Is male × has lookups in week w –0.0326 + (0.0188) influences purchases. Additionally, it explores the effects of app
Is male × has check-ins in week w –0.0131 (0.0441) stickiness (i.e., repeated app use) and discontinued app use. We
Is iOS user 0.0711 ⁎⁎⁎ (0.0082) find that (1) branded app adoption has a positive effect on
Is Android user 0.0494 ⁎⁎⁎ (0.0099)
Pre-period behavioral characteristics and propensity score
purchase behavior; (2) the positive effect on spending persists for
ln(Points Accrued in Mar + 1) 0.0481 ⁎⁎⁎ (0.0034) at least six months after adoption; (3) the positive effect of branded
ln(Points Accrued in Apr + 1) 0.0346 ⁎⁎⁎ (0.0042) app adoption is elevated when customers become more active and
ln(Points Accrued in May + 1) 0.0389 ⁎⁎⁎ (0.0035) use more features available on the branded app; (4) branded app
ln(Points Accrued in Jun + 1) 0.0342 ⁎⁎⁎ (0.0068) use has a cumulative effect: as customers repeatedly use the app,
ln(Points Accrued in Jul + 1) 0.0617 ⁎⁎⁎ (0.0061)
their spending levels increase even more; and (5) discontinuing
ln(Points Accrued in Aug + 1) 0.1148 ⁎⁎⁎ (0.0094)
ln(Points Accrued from Grocery + 1) 0.0926 ⁎⁎⁎ (0.0037) the use of the branded app is associated with reduced future
ln(Points Accrued from Retail + 1) 0.0295 ⁎⁎⁎ (0.0034) spending. These findings are consistent with previous research on
ln(Points Accrued from Gas + 1) 0.0467 ⁎⁎⁎ (0.0109) the effectiveness of branded apps (Bellman et al. 2011), mobile
ln(Points Accrued from Banking + 1) 0.0237 ⁎⁎⁎ (0.0033) interactivity (Gu, Oh, and Wang 2013), mobile advertising (Rettie,
ln(Points Accrued from Other + 1) 0.0270 ⁎⁎⁎ (0.0041)
ln(Number of Sponsors Visited + 1) –0.1559 ⁎⁎⁎ (0.0140)
Grandcolas, and Deakins 2005), and positively- and negatively-
ln(Redeemed Points + 1) 0.0172 ⁎⁎⁎ (0.0038) valenced brand engagement (Hollebeek and Chen 2014). We also
ln(Tenure in Days + 1) 0.0249 ⁎⁎⁎ (0.0042) find that the concept of stickiness, which has been mostly tested in
Propensity score –0.1708 (0.5209) a web context, can be applied to a mobile context. Consistent with
Note. 10,776 app adopters and 10,776 paired non-adopters as a control group with the framework of Furner, Racherla, and Babb (2014), we find that
replacement (5,127 distinct non-adopters). Total number of observations = the level of stickiness (i.e., repeated use of the app) increases
10,776 × 2 × 25 weeks = 538,800. Estimates for seasonality effects are not purchase behavior. To the best of our knowledge, this research is
reported.
also one of the first empirical studies that show an app's failure in
Robust standard errors per White (1980) are in parentheses.
+
p b .10. satisfying customers' needs hurts brand sales, as customers
⁎ p b .05. discontinue the use of it.
⁎⁎ p b .01.
⁎⁎⁎ p b .001. 10
To check whether propensity score matching yields different estimates from
a null model that potentially suffers from selection bias, we conduct a Durbin–
Wu–Hausman test, comparing estimates from the null model without
propensity score matching to one that incorporates it. Null hypothesis is
rejected, suggesting that the two models yield different coefficient estimates.
The test results are available from the corresponding author upon request.
38 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41

Our results provide substantive implications for the theory of Methodologically, while our research design uses propensity
interactivity and stickiness, as well as cross-channel advertising. score matching to reduce confounding due to non-random
First, this study expands the concept of interactivity and stickiness assignment to app adoption, our approach does not fully solve the
to a mobile context. Previous research has shown that greater issues of endogenous selection bias (Elwert and Winship 2014).
engagement through interactivity generates favorable attitudes Our models control for observed confounders, but any unob-
toward message creators (Sundar et al. 2013). Our study confirms served confounders, e.g., brand attitude or Internet experience,
that greater engagement with the interactive app features (and are assumed away. Ideally, we would incorporate instrumental
presumably with the brand) produces a higher level of subsequent variables and two-stage-least-square (2SLS) regressions in
spending. Customers who adopt a branded app and continue addition to propensity score matching. For instance, a potentially
to interact with it will deepen their relationships with the brand. suitable instrumental variable could be customers' participation
By logging in with the app and using its features, customers on social media sites or terms of their smartphone contracts, but
experience multiple touch points with the brand. If we consider the unfortunately we do not have this type of data. Another way to
continuous and immediate nature of mobile technologies, adopters validate the findings from our observational study would be
have more opportunities to be reminded of the values that the brand conducting experiments that allow for better control of app
offers and build a relationship of trust and commitment. Thus, use and measure pre- and post-adoption spending behavior
providing a stickier app that encourages adopters to use it (e.g., Armstrong 2012; Goldfarb and Tucker 2011).
repeatedly and frequently is a key mobile strategy. In addition, Despite these limitations, this paper offers important sugges-
habitual use of an app integrates the brand into the customer's life tions for future research. Few studies have examined the effect of
(Wang, Malthouse, and Krishnamurthi 2015). The optimization of app adoption and usage with actual feature usage data that are
cross-channel advertising can maximize the marketing potential of linked to purchase behaviors of the same individuals. Future
each communication channel. research should test the uses of other interactive features and see
However, the finding that abandoning an app decreases what specific ones are the most effective in changing brand
subsequent purchase behaviors warns against releasing an app attitude, purchase intention, or loyalty, and under which
whose value to customers is not clear, so that it seems unhelpful or conditions. Future studies should also test the effects of branded
irrelevant to them. Why customers decrease their spending levels apps in other industries.
after they stop using the branded app is beyond the scope of this Since this study concerns what happens after customers adopt
paper. However, we speculate that the decline may be attributable and repeatedly use (or stop using) a branded app, the process in
to negative or dissatisfying experiences that the customers develop which a branded app gets discovered in app store and downloaded
while exploring the app. Hollebeek and Chen (2014) identified is not investigated in this study, but merits attention from both
perceived brand value as one of the key antecedents of brand academics and professionals. Technological adoption model
engagement behavior. If consumers negatively assess the benefits (TAM) provides a conceptual framework for identifying factors
of a branded app, they are more likely to experience negatively- that motivate app adoption. TAM maintains that Perceived
valenced brand engagement such as negative brand attitude and Usefulness (PU) and Perceived Ease of Use (PEOU) are two
sharing negative word-of-mouth, which then becomes an fundamental determinants of technology adoption (Davis 1989).
antecedent of decreasing sales. If that is the case, companies Recent literature on TAM suggests that there are four types of
should be cautious about making apps available on the market- external factors that influence the level of PU and PEOU:
place before their functionalities are fully tested. It may be even individual differences, system characteristics, social influence,
more harmful than not providing any app. Also, app recency can and facilitating conditions (Venkatesh and Bala 2008). Given the
serve as an early indicator of customer attrition and should be information and hedonic nature of mobile apps and the influence of
monitored as a trigger event (Malthouse 2007). peer reviews on technology adoption, how system characteristics
This study has a few limitations. Data wise, we do not have and social influence, among the four external factors, play a role in
complete information on what specific functions the customers consumers' adoption of branded apps becomes an important area
used (e.g., when they looked up information, were they checking for future research. For example, what are the key design features
their point balances or browsing reward items), which makes it that affect the perceived level of interactivity and/or cognitive and
impossible to estimate the effect of each function beyond the affective involvement? How much do consumers rely on ratings or
broad categories of information lookups and check-ins. Our study reviews of apps when there is a need to download an app? How do
also assumes that customers did not look up information when these two external factors influence PU and PEOU? Are there
they did a check-in, which may lead to overestimating the latter's differences in motivating factors between branded app adoption
effect. Additionally, we do not have session times, so a ten-minute and generic app adoption?
session is assumed to impose the same effect as a one-minute From a practical standpoint, our findings suggest that
session. It is possible that accounting for usage time may yield companies should provide branded apps that can offer a unique
more nuanced insights. Finally, even though the sample size is brand experience without any technical hassles. A recent
large, this study uses data from only one brand that provides survey from Forrester shows that only 20% of digital business
services and offerings across multiple categories. Hence, our professionals update their apps frequently enough to fix bugs
findings are conditional on the fact that a brand's app successfully and handle OS updates, signaling the lack of mobile strategies
delivers value and interactivity to its customers, and that its in industry (Ask et al. 2015, March 23). Marketers should think
products are relevant to customers' needs. about what user interface and features can be implemented to
 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41 39

increase interactivity and user engagement (Kiousis 2002; Lee covariates exhibit similar distributional patterns between the
2005). Companies should also monitor the app marketplace for treated and the control groups. In contrast, conventional
customers' evaluations on their branded apps. Negative word- implementations of 2SLS regressions do not improve covariate
of-mouth on a branded app may signal the brand's lack of balance across all observed variables. Instead, they derive
readiness in the fast-changing mobile environment. With the causality of the endogenous variable of interest, i.e., variable x,
explosive growth of mobile technologies and app culture, by assuming that the instrumental variable can only influence the
customers' expectations of a useful and enjoyable mobile dependent variable through x. If x influences the independent
experience will become the norm. Whether companies can meet variable directly, then 2SLS also yields biased results. In some
those expectations is a determining factor in maximizing the cases 2SLS bias may be greater than the OLS bias because even
impact of mobile applications. with a small direct effect between the dependent variable and the
instrument z, the parameter estimate can be biased upward by
1/Corr(z,x), and standard error be biased downward, resulting in
Appendix A significant but incorrect results and conclusions.
Furthermore, an instrumental variable that is only weakly
Below we provide our rationale for choosing propensity score correlated with x can lead to inconsistent parameter estimations
matching over 2SLS with instrumental variables by providing a (Bound, Jaeger, and Baker 1995; Nelson and Startz 1990), and
comparison between the two methods. there are no statistical tests that can definitively show the
Propensity score matching differs from 2SLS in two ways. degree to which an instrumental variable is valid (Morgan and
First, after propensity scores are calculated, researchers will Winship 2015, p. 301).
perform the following procedures: check to ensure improved Finally, 2SLS only addresses confounding of one instru-
covariate balance, matching and/or weighting the sample with the mental variable at a time and does not account for covariate
propensity scores, and then perform regression analyses using the balance across all variables. Given the advantages and
matched or weighted sample. 2SLS, on the other hand, does not disadvantages of the methods, one would ideally have a good
drive at covariate balance or a “perfect stratification” (Morgan instrumental variable and apply both approaches. However, in
and Winship 2015, p. 291). our dataset, there are no variables that affect app adoption and
Second, propensity score matching improves a regression yet are not directly correlated with point accrual behavior.
sample's covariate balance by selecting a subset of control units Thus, instead of using an ill-suited instrumental variable, we
that are similar to the treated units based on observable variables. use propensity score matching to control for confounds due to
This assumes that all confounds are accounted for when observed customers' demographics and preexisting behavior.

Appendix B. Estimates of focal variables, DDD models, matching on non-adopters

Dependent variable ln(Points Accrued in Month t + 1)

Lookup Check-in Lookup × Check-in
Is Post-Adoption 0.2000*** (0.0111) 0.2378*** (0.0350) –0.0063 (0.0402)
Sep 0.2872*** (0.0208) 0.3028*** (0.0657) –0.1118 (0.0755)
Oct 0.2067*** (0.0208) 0.2235*** (0.0657) –0.0115 (0.0755)
Nov 0.2112*** (0.0208) 0.2381*** (0.0657) 0.0389 (0.0755)
Dec 0.1718*** (0.0208) 0.2082** (0.0657) 0.0233 (0.0755)
Jan 0.1463*** (0.0208) 0.2217*** (0.0657) 0.0024 (0.0755)
Feb 0.1770*** (0.0208) 0.2325*** (0.0657) 0.0211 (0.0755)
Demographics
Is aged 18 to 34 –2.3575*** (0.0601)
Is aged 35–44 –0.8697*** (0.0370)
Is aged 45–540 0.4036*** (0.0321)
Is aged 55–64 0.6834*** (0.0397)
Is aged 65 plus 1.5517*** (0.0591)
Is female 0.3491*** (0.0223)
Is male –0.1306*** (0.0212)
Pre-adoption behavioral characteristics and propensity score
ln(Number of Sponsors Visited + 1) 1.3564*** (0.0181)
ln(Redeemed Points + 1) 0.1707*** (0.0048)
ln(Tenure in Days + 1) 0.0089 (0.0098)
Propensity score (i.e., estimated probability to not adopt the app) –15.5305*** (0.3681)
Fixed intercept 6.0630*** (0.1828)
Matched pair random intercept covariance 0.0050 (0.0142)
Individual random intercept covariance 0.9336*** (0.0182)
40 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41

References Hosh, Dave (2014), “App Retention Improves — Apps Used Only Once Declines
to 20%,” http://info.localytics.com/blog/app-retention-improves (June 11).
Angrist, Joshua David and Jörn-Steffen Pischke (2009), Mostly Harmless Imbens, Guido W. and Donald B. Rubin (2015), Causal Inference in Statistics,
Econometrics: An Empiricist's Companion. Princeton, NJ: Princeton Social, and Biomedical Sciences: An Introduction. New York, NY:
University Press. Cambridge University Press.
Annalect (2015), “#GenerationTech: Millennials & Technology,” http://www. Ingraham, Nathan (2013), “Apple Announces 1 Million Apps in the App Store,
annalect.com/generationtech-millennials-technology/ (January 28). More Than 1 Billion Songs Played on Itunes Radio,” http://www.theverge.
Armstrong, J. Scott (2012), “Illusions in Regression Analysis,” International com/2013/10/22/4866302/apple-announces-1-million-apps-in-the-app-store
Journal of Forecasting, 28, 3, 689–94. (October 22).
Ask, Julie A., Jeffrey S. Hammond, Carrie Johnson, and Laura Naparstek Kang, Ju-Young M., Jung Mee Mun, and Kim K.P. Johnson (2015), “In-store
(2015), “The State of Mobile App Development: Few eBusiness Teams Mobile Usage: Downloading and Usage Intention Toward Mobile Location-
Keep Pace With Customer App Expectations,” https://www.forrester.com/ based Retail Apps,” Computers in Human Behavior, 46, 210–7.
The+State+Of+Mobile+App+Development/fulltext/-/E-res120267 (March 23). Keele, Luke (2010), An overview of rbounds: An R package for Rosenbaum
Balasubramanian, Sridhar, Robert A. Peterson, and Sirkka L. Jarvenpaa (2002), bounds sensitivity analysis with matched data. Columbus, OH: White Paper.
“Exploring the Implications of M-commerce for Markets and Marketing,” 1–15.
Journal of the Academy of Marketing Science, 30, 4, 348–61. Kim, Eunice, Jhih-Syuan Lin, and Yongjun Sung (2013), “To App or Not to App:
Bellman, Steven, Robert F. Potter, Shiree Treleaven-Hassard, Jennifer A. Engaging Consumers via Branded Mobile Apps,” Journal of Interactive
Robinson, and Duane Varan (2011), “The Effectiveness of Branded Mobile Advertising, 13, 1, 53–65.
Phone Apps,” Journal of Interactive Marketing, 25, 4, 191–200. Kiousis, Spiro (2002), “Interactivity: A Concept Explication,” New Media &
Boase, Jeffrey and Rich Ling (2013), “Measuring Mobile Phone Use: Self- Society, 4, 3, 355–83.
Report Versus Log Data,” Journal of Computer-Mediated Communication, Lariviere, Bart, Herm Joosten, Edward C. Malthouse, Marcel Van Birgelen,
18, 4, 508–19. Pelin Aksoy, Werner H. Kunz, and Ming-Hui Huang (2013), “Value Fusion:
Bound, John, David A. Jaeger, and Regina M. Baker (1995), “Problems with The Blending of Consumer and Firm Value in the Distinct Context of Mobile
Instrumental Variables Estimation when the Correlation Between the Technologies and Social Media,” Journal of Service Management, 24, 3,
Instruments and the Endogenous Explanatory Variable Is Weak,” Journal 268–93.
of the American Statistical Association, 90, 430, 443–50. Lee, ThaeMin (2005), “The Impact of Perceptions of Interactivity on Customer
Calder, Bobby J. and Edward C. Malthouse (2008), “Media Engagement and Trust and Transaction Intentions in Mobile Commerce,” Journal of Electronic
Advertising Effectiveness,” in Kellogg on Advertising and Media, Bobby J. Commerce Research, 6, 3, 165–80.
Calder, editor. Hoboken: NJ: Wiley. 1–36. Li, Dahui, Glenn J. Browne, and James C. Wetherbe (2006), “Why Do Internet
comScore (2014), ComScore Reports August 2014 U.S. Smartphone Subscriber Users Stick with a Specific Web Site? A Relationship Perspective,”
Market Share. (October 7). International Journal of Electronic Commerce, 10, 4, 105–41.
Coursaris, Constantinos K. and Jieun Sung (2012), “Antecedents and Consequents Lin, Judy Chuan-Chuan (2007), “Online Stickiness: Its Antecedents and Effect on
of a Mobile Website's Interactivity,” New Media & Society, 14, 7, 1128–46. Purchasing Intention,” Behaviour & Information Technology, 26, 6, 507–16.
Davis, Fred D. (1989), “Perceived Usefulness, Perceived Ease of Use, and User Lipsman, Andrew (2014), Major Mobile Milestones in May: Apps Now Drive
Acceptance of Information Technology,” MIS Quarterly, 13, 3, 319–40. Half of All Time Spent on Digital. (June 25).
Elwert, Felix and Christopher Winship (2014), “Endogenous Selection Bias: Malthouse, Edward C. (2007), “Mining for Trigger Events With Survival
The Problem of Conditioning on a Collider Variable,” Annual Review of Analysis,” Data Mining and Knowledge Discovery, 15, 3, 383–402.
Sociology, 40, 1, 31–53. Meena, Satish (2014), “Forrester Research World Mobile and Smartphone
Fox, Zoe (2013), “Global App Downloads to Pass 100 Billion This Year,” Adoption Forecast, 2014 to 2019 (Global),” https://www.forrester.com/
http://mashable.com/2013/09/19/app-downloads-100-billion/?utm_cid=mash- Forrester+Research+World+Mobile+And+Smartphone+Adoption+Forecast+
com-fb-main-photo (Sep 19). 2014+To+2019+Global/fulltext/-/E-RES118252 (August 8).
Frith, Jordan (2014), “Communicating Through Location: The Understood Meaning Mollen, Anne and Hugh Wilson (2010), “Engagement, Telepresence and
of the Foursquare Check-in,” Journal of Computer-Mediated Communication, Interactivity in Online Consumer Experience: Reconciling Scholastic and
19, 4, 890–905. Managerial Perspectives,” Journal of Business Research, 63, 9–10, 919–25.
Furner, Christopher P., Pradeep Racherla, and Jeffry S. Babb (2014), “Mobile Morgan, Stephen L. and Christopher Winship (2015), Counterfactuals and
App Stickiness (Mass) and Mobile Interactivity: A Conceptual Model,” The Causal Inference: Methods and Principles for Social Research. New York,
Marketing Review, 14, 2, 163–88. NY: Cambridge University Press.
Goldfarb, Avi and Catherine Tucker (2011), “Online Display Advertising: Nelson, Charles R. and Richard Startz (1990), “The Distribution of the
Targeting and Obtrusiveness,” Marketing Science, 30, 3, 389–404. Instrumental Variables Estimator and its T-ratio when the Instrument Is a
Gu, Rui, Lih-Bin Oh, and Kanliang Wang (2013), “Differential Impact of Web and Poor One,” Journal of Business, 63, 1, S125–40.
Mobile Interactivity on E-retailers' Performance,” Journal of Organizational Novak, Thomas P., Donna L. Hoffman, and Yiu-Fai Yung (2000), “Measuring
Computing and Electronic Commerce, 23, 4, 325–49. the Customer Experience in Online Environments: A Structural Modeling
Hollebeek, Linda D. and Tom Chen (2014), “Exploring Positively- Versus Approach,” Marketing Science, 19, 1, 22–42.
Negatively-valenced Brand Engagement: A Conceptual Model,” Journal of Okazaki, Shintaro and Felipe Mendez (2013), “Perceived Ubiquity in Mobile
Product & Brand Management, 23, 1, 62–74. Services,” Journal of Interactive Marketing, 27, 2, 98–111.

Notes to Appendix B
Note. As a robustness check, we use “no adoption” as treatment and first estimate a propensity score model of 5,127 non-adopters and 10,776 app adopters, and then
pair up each non-adopter with a distinct app adopter. Estimates from the propensity score model are not reported. After obtaining a non-adopter matched sample, we
perform the DDD models per our research design. Regression sample includes 5,127 non-adopters and 5,217 paired app adopters. Total number of observations =
5,127 × 2 × 12 months =190,836. Conclusions regarding H1 and H2 remain the same.
The coefficients labeled under “Is Post-Adoption” are estimated from an alternative model that examines the overall treatment effect. Coefficients labeled by month,
i.e., Sep through Feb, are estimated using the model specified in Eq. (3), which breaks down the treatment effect by month after app adoption. Estimates for
demographics, pre-adoption behavioral characteristics, propensity score, fixed intercepts, and random intercept covariance are the same between the two models.
Estimates for seasonality effects are not reported.
+
p b .10, *p b .05, **p b .01, ***p b .001. Standard errors are in parentheses.
 S.J. Kim et al. / Journal of Interactive Marketing 31 (2015) 28–41 41

Rettie, Ruth, Ursula Grandcolas, and Bethan Deakins (2005), “Text Message Theory and Practice, James P. Dillard, Lijiang Shen, editors. Thousand
Advertising: Response Rates and Branding Effects,” Journal of Targeting, Oaks, CA: Sage Publications. 388–404.
Measurement and Analysis for Marketing, 13, 4, 304–12. van Noort, Guda, Hilde A.M. Voorveld, and Eva A. van Reijmersdal (2012),
Rosenbaum, Paul R. and Donald B. Rubin (1983), “The Central Role of the “Interactivity in Brand Web Sites: Cognitive, Affective, and Behavioral
Propensity Score in Observational Studies for Causal Effects,” Biometrika, Responses Explained by Consumers' Online Flow Experience,” Journal of
70, 1, 41–55. Interactive Marketing, 26, 4, 223–34.
——— and ——— (1985), “Constructing a Control Group Using Multivariate Venkatesh, Viswanath and Hillol Bala (2008), “Technology Acceptance
Matched Sampling Methods That Incorporate the Propensity Score,” The Model 3 and a Research Agenda on Interventions,” Decision Sciences, 39,
American Statistician, 39, 1, 33–8. 2, 273–315.
——— (2005), “Sensitivity Analysis in Observational Studies,” in Wang, Rebecca Jen-Hui, Edward C. Malthouse, and Lakshman Krishnamurthi
Encyclopedia of Statistics in Behavioral Science, Brian S. Everitt, David (2015), “On the Go: How Mobile Shopping Affects Customer Purchase
C. Howell, editors. Behavior,” Journal of Retailing, 91, 2, 217–34.
——— (2010), Design of observational studies. ———, Su Kim, and Edward C. Malthouse (2016), “Branded Apps and Mobile
Rubin, Donald B. (2007), “The Design Versus the Analysis of Observational Platforms as New Tools for Advertising,” in The New Advertising: Branding,
Studies for Causal Effects: Parallels with the Design of Randomized Trials,” Content, and Consumer Relationships in the Data-driven Social Media Era,
Statistics in Medicine, 26, 1, 20–36. Ruth E. Brown, Valerie K. Jones, Ming Wang, editors. Santa Barbara, CA:
Shankar, Venkatesh and Sridhar Balasubramanian (2009), “Mobile Marketing: ABC-CLIO.
A Synthesis and Prognosis,” Journal of Interactive Marketing, 23, 2, Watson, Richard I., Leyland F. Dill, Pierre Berthon, and George M. Zinkhan
118–29. (2002), “U-commerce: Expanding the Universe of Marketing,” Journal of
——— and Edward C. Malthouse (2007), “The Growth of Interactions and the Academy of Marketing Science, 30, 4, 333–47.
Dialogs in Interactive Marketing,” Journal of Interactive Marketing, 21, 2, Welch, Chris (2013), “Google: Android App Downloads Have Crossed 50
2–4. Billion, over 1M Apps in Play,” http://www.theverge.com/2013/7/24/4553010/
SmartBear (2014), “2014 State of Mobile,” http://www2.smartbear.com/rs/ google-50-billion-android-app-downloads-1m-apps-available (July 24).
smartbear/images/SmartBear-2014-Software-Industry-Survey-Testing.pdf. White, Halbert (1980), “A Heteroskedasticity-consistent Covariance Matrix
Stuart, Elizabeth A. (2010), “Matching Methods for Causal Inference: A Estimator and a Direct Test for Heteroskedasticity,” Econometrica, 48, 4,
Review and a Look Forward,” Statistical Science, 1, 1–21. 817–38.
Sundar, Shyam S., Qian Xu, Saraswathi Bellur, Haiyan Jia, Jeeyun Oh, and Yadav, Manjit S. and Rajan Varadarajan (2005), “Interactivity in the Electronic
Guan-Soon Khoo (2010), “Click, Drag, Flip, and Mouse-over: Effects of Marketplace: An Exposition of the Concept and Implications for Research,”
Modality Interactivity on User Engagement with Web Content,” Suntec Journal of the Academy of Marketing Science, 33, 4, 585–603.
City, Singapore: International Communication Association. Zott, Christoph, Raphael Amit, and Jon Donlevy (2000), “Strategies for Value
———, Jeeyun Oh, Hyunjin Kang, and Akshaya Sreenivasan (2013), “How Creation in E-commerce: Best Practice in Europe,” European Management
Does Technology Persuade? Theoretical Mechanisms for Persuasive Journal, 18, 5, 463–75.
Technologies,” in The Sage Handbook of Persuasion: Developments in