Anda di halaman 1dari 8

Impact of Message Design on Online Interactions: An

Empirical Investigation

Tianshu Sun

R.H.Smith School of Business
University of Maryland

Siva Viswanathan

R.H.Smith School of Business
University of Maryland

Elena Zheleva

Washington, D.C., USA

Recent advances in information technologies have provided firms
unprecedented ability to manage online social interactions among
their customers. Despite their growing popularity and importance,
there are very few studies that have examined the optimal design
choices for a firm seeking to maximize returns from firm-
mediated social interactions. Our study seeks to examine whether
and how a firm can enhance effectiveness of online social
interactions, by simply varying the message shared by referrers
with their social connections. More specifically, in this study we
focus on two key pieces of information contained in the messages
information about senders purchase status prior to referral, and
information about the existence of referral rewards and their
impacts on the recipients purchase decision. We design and
conduct a randomized field experiment testing the effectiveness of
different versions of firm-mediated messages in creating social
contagion, in collaboration with an online daily deal platform. Our
preliminary findings suggest that information about senders
purchase leads to more purchases by recipients. The findings of
the study will add to our understanding of the effect of different
messages on social interaction outcomes, and provide guidelines
for optimal design of messages at aggregate level as well as the
individual level.

General Terms
Management, Economics, Experimentation,Theory
Sharing, Referral, Randomized Field Experiment, Peer Effect
Social interactions have an important impact on consumer
choices and have been studied extensively (Godes,
Trusov, Bucklin, and Pauwels 2009 , Schmitt, Skiera, and Van
den Bulte 2011, Berger 2011, Moe 2014, Pfeiffer and Zheleva
2012). With the rapid advances in information technologies, a
large volume of social interactions and referrals are now created
and disseminated through online channels in the form of website-
mediated emails, social media posts, and mobile messages. An
interesting aspect of the online setting is that while consumers are
able to quickly disseminate online word of mouth about firms and
products, firms are also increasingly able to mediate these
interactions among customers. Firms have transitioned from being
passive observers and moderators of online social interactions to
becoming more active mediators (Godes et al 2005) of online
social interactions and referrals. Todays technologies provide
firms the ability to mediate and manage various granular aspects
of the social interaction process including - the motivations of the
referrers, the channels of referral, and the message, among others.
As noted by Godes et al (2005), despite the importance of social
interactions, little research has been done on how firms may
potentially use and impact them. This represents an exciting
research opportunity. In particular, there are very few studies that
have examined the optimal design choices for a firm seeking to
maximize returns from firm-mediated social interactions. Our
study seeks to fill this gap.
An important aspect of online social interactions is the
message that is shared between senders and recipients. In the case
of firm-mediated messaging among users, while the sender can
choose the recipients with whom they share the messages (online
word of mouth or referrals), the firm nevertheless, has the ability
to control several aspects of the message and its content. Such
firm-mediated messaging is increasingly the norm in a large
number of online websites, retailers, and platforms. Despite, the
increasing use of such mechanisms by firms online, there is very
little understanding of how different messages impact social
interactions and their outcomes. Given the ability of the firm to
partially control the content of the message that is shared between
the sender and the receiver, our study seeks to examine whether
and how a firm can enhance effectiveness of online social
interactions, by simply varying the message shared by referrers
with their social connections. More specifically, in this study we
focus on two key pieces of information contained in these
messages information about whether the sender has purchased
the product (i.e., the deal) prior to referral, and information about
the existence of monetary rewards for referrals and their impacts
Permission to make digital or hard copies of all or part of this work
for personal or classroom use is granted without fee provided that
copies are not made or distributed for profit or commercial
advantage and that copies bear this notice and the full citation on the
first page. Copyrights for components of this work owned by others
than the author(s) must be honored. Abstracting with credit is
permitted. To copy otherwise, or republish, to post on servers or to
redistribute to lists, requires prior specific permission and/or a fee.
Request permissions from

ICEC '14, August 05 - 06 2014, Philadelphia, PA, USA
Copyright is held by the owner/author(s).Publication rights licensed to ACM.
ACM 978-1-4503-2618-6/14/08 $15.00.
on the recipients purchase decision. We are especially interested
in the following questions:
1) Will knowledge of senders purchase affect recipients
purchase decision? How does this effect vary for products with
different characteristics and social tie of different type and
2) Will knowledge of the referral reward program affect the
recipients purchase decision as well as her decision to share the
deal with others?
3) What is the incremental effect when information about the
referral reward program is added into the message with senders
purchase information?
Despite the widespread popularity of referral rewards as well
as pre- versus post- purchase programs, there is limited research
on their impacts on social interactions. While a few prior studies
have highlighted the importance of pre- versus post-purchase
context (for instance see, Jing and Xie, 2011), as well as the
nuances of referral rewards in stimulating referrals (for instance,
see Ryu and Feick, 2007; Biyalogorsky et al 2001), our focus is
markedly different. We examine equally important yet,
unanswered questions given a (pre- or post-purchase) referral by
a sender, whether and how, the knowledge of a referrers purchase
(or the lack thereof) of the deal, as well as the knowledge of the
existence of monetary rewards for the referrer, impacts the
behaviors of the receivers.
Estimating the effect of message content of the sender on
social contagion (recipients purchase and sharing) has been
traditionally difficult for two reasons: first, the content of the
message in interpersonal communications is usually unobservable
to researchers; second, and probably more important, are issues of
endogeneity (Manski 1993), i.e. content of the message may be
correlated with the tie strength or the characteristics of the
recommended product. Several approaches for identifying peer
effects have been proposed, including dynamic matched sampling
( 2009), structural models (Ghose and Han 2010), ad
hoc approaches (Christakis and Fowler 2007) and instrumental
variables (Tucker 2008). However, most of above methods are not
applicable to our study because of unobserved data and potential
endogeneity. Thus, a randomized experiment is the ideal approach
to obtain unbiased estimates of causal effect of message on social
interactions. We therefore design and conduct a randomized field
experiment testing the effectiveness of different versions of firm-
mediated messages in creating social contagion among the users
of an online daily deals platform. We investigate the underlying
mechanisms of any observed differences in effectiveness by
analyzing field experiment results in detail.
Preliminary results show that information about the senders
purchase has a sizable and statistically significant effect on
recipients purchase. Interestingly, once information about referral
reward program is added, the effect of senders purchase is
attenuated and the marginal effect of the purchase information
becomes less significant.
This study is among the first to analyze the potential of firm-
mediated messaging and the findings of the study will not only
add to our understanding of the role of different messages on
referral outcomes, but also provide valuable guidelines for
optimal design of such information sharing mechanisms at group
level and at individual level.

There is a growing literature on social interactions (see Godes
et al (2005) and Libai for excellent reviews) and our
study is closely related to four streams of research.
The first stream is on the causal effect of peer influence.
Numerous studies have established the existence of social
influence through observational studies (Iyengar et al. 2011) and
randomized field experiments (Aral and Walker 2012, Bursztyn 2012). The message is widely considered as the most
fundamental factor driving social influence (Berger 2012). Our
study, with its primary focus on firm-mediated messaging,
extends the literature on peer influence by identifying the
incremental contribution of different components of a message
(e.g. purchase status of the sender) on recipient outcomes by
enabling or disenabling specific features in the message. While
prior studies (Aral 2011, 2012) have treated messages in
online social interactions as a whole, we are able to decompose
social influence at component level by varying different features
of a message. In addition, the wide range of product
characteristics included in our study will also enable us to
differentiate between the two mechanisms of social contagion:
social learning (recipient infers the high quality of products from
senders purchase) and network effects (recipient gets additional
utility from senders adoption of product, e.g. social events).
Our work also draws upon prior research on referral rewards.
There have been a number of analytical models examining the
optimal design of referral rewards from a firms perspective (for
instance see, Biyalogorsky et al 2001, Kornish and Li, 2010, Xia
et al 2011). A few experimental studies (Wirtz and Chew, 2002;
Ryu and Feick, 2007) have examined the impact of referral
rewards on the likelihood of referrals. A couple of studies (Tuk et
al 2009; Verlegh et al 2013) that have focused on the role of
rewarded referrals on recipient responses have been small scale
lab experiments involving students. Ours is the first large scale
field experiment to examine the role of the knowledge of
monetary rewards for the referrer on the recipients outcomes. Our
study also extends current work by analyzing the interaction
effects between referral awards and recipients perception of the
sender by manipulating features in the message.
Our work is also related to the literature on Group-Buying
specifically, the social network marketing mechanism embedded
in the Group-Buying selling format (Jing and Xie, 2011). Existing
research on Group-Buying largely focuses on the power of buying
groups on suppliers (Che and Gale, 1997; Chen and Li, 2011). An
exception is Jing and Xie (2011), who develop an analytical
model to compare and contrast Group-Buying - a mechanism to
motivate interpersonal information and knowledge sharing, with
Referral Rewards program a mechanism which uses monetary
rewards to motivate existing buyers to spread product information
to their peers. A key assumption underlying their model is that
while Group-Buying requires the customer to make referrals
before any transaction is completed, Referral Rewards programs
allow the customer to make a referral after she has purchased the
product. However, in a number of real-world Group-Buying
situations, including ours, customers have the ability to share
information about the deals both prior to as well as after their
purchase. The separation of the senders purchase status and her
sharing behaviors provides us an unprecedented opportunity to
examine the impact of the message content relating to the
purchase (or the lack thereof) by the sender on recipients
Finally, our proposed study (especially the personalization of
message design) is related to studies of factors moderating the
effectiveness of WOM, including the sender and the recipient
(Iyengar et al. 2011), tie strength, similarity between sender and
recipients tastes or opinions, and network characteristics
(Stephen and Berger 2012, Katona 2011). For instance,
Naylor, Lamberton, and Norton 2011 found that word-of-mouth
from similar others may have a more positive effect. Goldenberg,
Libai, and Muller 2001 found that the influence of weak ties is at
least as strong as the influence of strong ties for a given individual
in social network. Stephen and Lehman 2009 found the
transmitter and transmitter-recipient relationship characteristics as
potential drivers of reception/listening. We complement this
stream of research by examining the interaction effects of the
above factors with two key components of the message.

We collaborate with a daily deal platform in the US and propose a
randomized field experiment to study the causal impact of firm-
mediated message on recipients behavior. The platform offers a
wide range of daily deals for local services and standard products
at a high discount and has a large customer base. On each deal
page on the firms website, the platform provides channels
through which customers (senders) can share these deals with
their social connections. Customers (senders) can share deals with
their friends by clicking specific channel buttons which are
prominently displayed. Specifically, senders who wish to share
through email can add a recipients email address and click send.
For email referrals, the platform will then automatically deliver
emails to each recipients email address separately
. Every day, a
large volume of shares are made by customers through the firms
The platform uses a referral program to encourage social
interactions. To participate in this program, a user is first required
to purchase a particular deal. Then, the user is given the option to
share the deal with as many friends as desired. When three (or
more) of the senders referrals purchase the deal, the sender gets a
full refund of price she paid for the deal.
. Typically, a user shares the firm-created message
pertaining to the deal with her friends.
In this study, we aim to identify the effect of message design, i.e.,
effect of information about the 1) senders purchase status and 2)
referral program, on recipients purchase and sharing behaviors.
Using a 2*2 design (see Figure 7 on next page), we can identify
main effect as well as the interaction effects of information 1) and

Even though the sender may specify multiple recipients in a
single send, each email is sent separately and hence, each
recipient receives the email as a one-to-one personal share.
Hence, we define each sender-recipient pair in a multi-recipient
share as an independent share.
A fraction of the senders also share deals through their own
channels (e.g. copy-paste the deal URL into their own social
media or email account), leading to successful referrals. The
firm has no control on the message content of such social
interactions. Our field experiment focuses only on senders using
the firms platform/website for sharing/referrals.
2) for both pre- and post-purchase sharing/referrals. After the
sender confirms her share by clicking the send button, she will
be randomly assigned to one of the four test groups based on her
purchase status (Figure 7). There are four groups for senders in
post-purchase sharing test (1 control (C), and 3 treatments (T1-
T3)) and four groups (1 control (EC), and 3 treatments (ET1-ET3))
for senders in pre-purchase sharing test. The pre- and post-
purchase sharing test have two templates in common (control (T2
& ET2) and referral reward program (C&EC)) since no
information on purchase status is involved. Thus, we only need to
create six versions of the message templates by varying the
visibility of senders purchase status and referral reward program,
as illustrated in Figure 1-6.

Figure 1: Message Template for Control Group (i.e. C and EC)

Figure 2: Message Template for post-purchase treatment (T1)
Figure 3-6 (counterclockwise): Template for referral reward
treatment T2/ET2 (right-up), post-purchase & referral
treatment (T3) (left up), pre-purchase treatment ET1 (left
bottom) and pre-purchase & referral ET3 (bottom right)

Figure 7: Illustration
of random assignment
of sender into test
groups based on
purchase status
4.1 Level of Randomization
The randomization takes place at the level of the sender, i.e., the
recipients of a sender for a specific deal see the same template.
Randomization at the level of the sender (rather than at the level
of recipients) allows for better control of potential spillovers
between control and treatment groups and helps ensure that the
stable unit treatment value assumption (SUTVA) is not violated.
Such spillovers are more likely to happen within the local network
of a sender as compared to across senders networks (Aral
2011). For instance, two friends of sender S are more likely to
communicate about a specific concert deal (through sharing the
deal or through other modes of communication) and influence
each others decision as compared to recipients of two different
senders. (however, this is still a possibility and we will discuss
approaches to mitigate this concern). In addition, when the
purchase status of a sender changes (i.e. the sender shares the
same deal both before and after purchase), she may be randomly
assigned to a new template. Also, when a sender shares a different
deal, she may be randomly assigned to a new template.
4.2 Length of Experiment
The optimal sample size in field experiment has been well
discussed in previous literature (List 2008). Longer
experiment period or larger sample size will allow researchers to
identify smaller differences between control and treatment group.
In this study, we aim to identify both the average treatment effect
and heterogeneous treatment effect (for message personalization).
The latter imposes higher requirement on the sample size as it
further limits eligible sample to specific stratification. Based on
our calculations of the sample size vs. minimum detectable
difference we plan to run the experiment for 5 consecutive weeks.
The data for our study comes from customer- to-customer email
shares/referrals through the platform. For every firm-mediated
email share, we record the unique hashed identifier of the sender
(customer ID), the recipient (hashed email address), the shared
deal, as well as the assigned test group. We record the purchase
status of the sender (pre- or post-purchase share), the number of
recipients she specifies in the batch of sent messages, the
timestamp of share. Purchase decisions of the recipient are also
included. Finally, the refund status of the sender for her referrals
is also recorded. We further augment the above main dataset with
the historical data on sender and recipient's purchase history
before experiment as well as price and subcategory of deals. The
resulting dataset enables us to analyze the impact of message
design at a granular level (i.e. heterogeneous treatment effect or
moderating effect of sender, recipient, product characteristics)

We focus our discussion on preliminary results for post-purchase
share. Consistent with the level of randomization in experiment
design, we present our results at individual sender level (instead
of recipient level). As a check of the randomization, we present in
Table 1 the tests of equality of the sender-level covariates across
four test groups in post-purchase share. As expected from the
random assignment, the sample is well balanced across the

Table 1: Randomization Check
P-value (C=T1=T2=T3)
(Post-purchase Share)
Total Number of Past Purchases 0.69
Days after Creating Account 0.24
Average Number of Recipients
in each "Send"
Category of Deals in Share
(in total 8 category dummies)
all >0.2

6.1. Referral Purchases by Sender

We first present three preliminary findings for post-purchase
share using t-test (see Table 2 on next page).
1) (Effect of information about senders purchase) There is
an increase on average number of referral purchases in post-
purchase treatment over control. The increase is sizable and
statistically significant at 0.01 level, which indicates a strong peer
effect between sender and recipient.
2) (Incremental Effect) Interestingly, once information about
referral reward is added in the post-purchase treatment, the effect
on recipients purchase is attenuated and the increase over control
becomes insignificant. This drop is equivalent to half of the
difference between control group and post-purchase treatment
group (with p=0.12), indicating the negative incremental effect of
information about referral reward with information about senders
3) (Effect of information about referral reward) Finally, the
difference between reward treatment and control is relatively
small and not statistically significant. Adding senders purchase
status in the reward treatment greatly increases the take-up rate of
The regression results for post-purchase share also confirm
the insights of t-test statistics. We present OLS results in table 3
controlling senders characteristics and deal category dummies.
The results match the statistics in table 2: the coefficient on
treatment group 2 (post-purchase template) is highly significant
(p-value 0.001). In other words, on average, the sender in the
post-purchase treatment generates more purchases. This effect is
economically significant considering the large number of
customers who share through the platform.
In addition, test shows that the difference in referral purchase
between post-purchase treatment and post-purchase & reward
treatment is also sizable, though not statistically significant. The
drop can be attributed to the interaction effect of senders
purchase and the existence of referral reward.
Finally, we observe a small increase in conversion in the reward
treatment over the control treatment. However, such difference is
not statistically significant.

Table 2: Main Effect- t test ( for post-purchase share)
Dependent Variables P-value in t-test
Effect of Sender's Purchase
Incremental Effect of Referral
Reward Program when info on
Sender's Purchase is provided
Effect of Information about
Referral Reward Program
Joint Effect of Info on Sender's
Purchase and Info on Referral
Reward Program

Table 3: Main Effect- Regression(for post-purchase share)
Dependent Variables
P-value of t-test on regression
(with control)
Effect of Sender's Purchase
Incremental Effect of Referral
Reward Program when info on
Sender's Purchase is provided
Effect of Information about
Referral Reward Program
Joint Effect of Info on Sender's
Purchase and Info on Referral
Reward Program

6.2. Heterogeneity in treatment effect (for post-share)
We further explore the heterogeneity in the treatment effect for
post-purchase share, which may help us better understand the
underlying mechanism that leads to the effect. There are several
important sources that would result in heterogeneity in treatment
effect: number of recipients sender specified in sharing,
characteristics of deals in sharing, senders past experience,
recipients past experience and social tie. In table 4, we present
preliminary results to show heterogeneity in treatment effect for
the first source. In table 4, we examine the effect of the different
messages on senders who share with a different number of
recipients. We find that the treatment effect on senders who are
share with 1-2 recipients is not statistically significant. However,
the effect of post-purchase information for senders who are
sharing with 3-5 recipients is sizable and significant , which
indicates the major increase in senders average referral purchases
comes from the increase of referral purchases for those senders
who are sharing with 3-5 recipients.

Table 4: Heterogeneity in Treatment Effect (treatment effect for
sender with different number of recipients)
Treatment Effect
Significance of regression
coefficients (p<0.01)
(with control)
(T1-C) for sender
who only share with 1 recipient
(T1-C) for sender
who only share with 2 recipient
(T1-C) for sender
who only share with 3 recipient
(T1-C) for sender
who only share with 4 recipient
(T1-C) for sender
who only share with 5 recipient
(T2-C) for sender
who only share with 1 recipient
(T2-C) for sender
who only share with 2 recipient
(T2-C) for sender
who only share with 3 recipient
(T2-C) for sender
who only share with 4 recipient
(T2-C) for sender
who only share with 5 recipient
(T3-C) for sender
who only share with 1 recipient
(T3-C) for sender
who only share with 2 recipient
(T3-C) for sender
who only share with 3 recipient
(T3-C) for sender
who only share with 4 recipient
(T3-C) for sender
who only share with 5 recipient

Our study has a number of implications for practice. As
noted earlier, firm-mediated online social interactions are growing
in popularity. Our study will provide valuable guideline for firms
seeking to manage such online social interactions through
message design. The quantitative estimates and qualitative
understanding gained from this series of component-level studies
can guide the optimal design of messages for improving the
effectiveness of social interactions. While this study focuses on
two key components of the message, future studies can extend this
agenda by examining other features.
Identifying optimal design of firm-mediated message at a
group level is a valuable endeavor, but not the end in itself. With
the availability of large amount of data on the behaviors of
senders and recipients and their historical interactions, as well as
the ability to process requests in real time, firms can actually
personalize firm-mediated messages at an individual level. While
personalization is a common practice in the context of firm-
customer interactions, personalization of firm-mediated customer-
customer social interaction is still in its infancy. To investigate its
potential, we will examine heterogeneous treatment effects and
understand the heterogeneous treatment effect of message design
for different types of senders, recipients, strength of ties, and
products. We envision that in the near future when a firm gets a
request of email share from a sender, it would extract product
characteristics, sender and recipients purchase and interaction
histories, calculate optimal content and message design, and
deliver the content in real time in a personalized fashion. Our
work is a step in that direction.

[1] Aral, S. 2011. Identifying social influence: A comment on
opinion leadership and social contagion in new product diffusion.
Mar- keting Sci. 30(2) 217223.
[2] S. Aral, D. Walker, Manage. Sci. 57, 16231639 (2011).
[3] S. Aral, D. Walker, Science 337, 337341 (2012).
[4] Berger, Jonah and Eric Schwartz (2011), What Drives
Immediate and Ongoing Word-of-Mouth? Journal of Marketing,
48(5), 869-880.
-- Berger, Jonah and Katherine Milkman (2012), What Makes
Online Content Viral? Journal of Marketing, 49(2), 192-205.
[5] Christakis, N. A., J. H. Fowler. 2007. The spread of obesity in
a large social network over 32 years. New England J. Medicine
357(4) 370379.
[6] Biyalogorsky, E., E. Gerstner, B. Libai. 2001. Customer
referral management: Optimal reward programs. Marketing Sci.
20(1) 8295.
[7] Chen RR, Li C, Zhang RQ (2010) Group buying mechanisms
under quantity discounts. Working paper, University of California,
Davis, Davis.
[8] Chen, Yubo, Qi Wang and Jinhong Xie (2011), "Online Social
Interactions: A Natural Experiment on Word of Mouth versus
Observational Learning," Journal of Marketing Research (April),
[9] Duflo, E., R. Glennerster, M. Kremer. 2008. Using
randomization in development economics research: A toolkit. T. P.
Schultz, J. Strauss, eds. Handbook of Development Economics,
Vol. 4, Chap. 61
[10] Ghose, A., S.-P. Han. 2010. An empirical analysis of user
content gen- eration and usage behavior in mobile media.
Working paper, Stern School of Business
[11] Godes, David, Dina Mayzlin, Yubo Chen, Sanjiv Das,
Chrysanthos Dellarocas, Bruce Pfeiffer, Barak Libai, Subrata Sen,
Mengze Shi and Peeter Verlegh (2005), The Firms Management
of Social Interactions, Marketing Letters, 16 (3/4), 415-428.
[12] Godes, David and Dina Mayzlin (2009), Firm-Created
Word-of-Mouth Communication: Evidence from a Field Test,
Marketing Science, 28(4), 721-739.
[13] Goldenberg, J., S. Han, D. R. Lehmann, J. W. Hong. 2009.
The role of hubs in the adoption process. J. Marketing 73(2) 113.
[14] Goldenberg, Jacob, Barak Libai and Eitan Muller (2001),
Talk of the Network: A Complex Systems Look at the
Underlying Process of Word-of-Mouth, Marketing Letters, 12(3),
[15] Hu, M., M. Shi, J. Wu. 2013. Simultaneous vs. Sequential
Group-Buying Mechanisms. Management Sci. forthcoming.
[16] Iyengar, Raghuram, Christophe Van den Bulte and Thomas
W. Valente (2010), Opinion Leadership and Social Contagion in
New Product Diffusion, Marketing Science, 30(2),195-212.
[17] Jing X, Xie J (2011) Group buying: A new mechanism for
selling through social interactions. Management Sci. 57(8): 1354
[18] Libai, Barak, Ruth Bolton, Marnix S. Buegel, Ko de Ruyter,
Oliver Goetz, Hans Risselada, and Andrew T. Stephen (2010),
Customer-to-customer interactions: Broadening the scope of
word of mouth research, Journal of Service Research, 13 (3),
[19] L. Kornish and Q. Li. Optimal referral bonuses with
asymmetric information: Firm-offered and interpersonal
incentives. Marketing Science, 29(1):108 121, 2010.
[20] Moe and Schiwedel 2014 Social Media Intelligence ,
Publisher: Cambridge University Press (February 24, 2014)
Language: English
[21] Naylor, Rebecca Walker, Cait Poynor Lamberton and David
A. Norton (2011), Seeing Ourselves in Others: Reviewer
Ambiguity, Egocentric Anchoring, and Persuasion, Journal of
Marketing Research, 48, 617-631
[22] J. J. Pfeiffer III and E. Zheleva. Incentivized sharing in social
networks. In (WOSS), 2012.
[23] Ryu, G., L. Feick. 2007. A penny for your thoughts: Referral
reward programs and referral likelihood. J. Marketing 71(1) 84
[24] Schmitt, Philipp, Bernd Skiera and Christophe Van den Bulte
(2011), Referral Programs and Customer Value, Journal of
Marketing, 75, 46-59.
[25] Stephen, Andrew and Jonah Berger (2012), Creating
Contagious: How Social Networks and Item Characteristics
Combine to Spur Ongoing Consumption and Drive Social
Epidemics,Working Paper.
[26] Stephen, Andrew T. and Donald R. Lehmann (2009), Is
Anyone Listening? Modeling the Impact of Word-of-Mouth at the
Individual Level, Working Paper.
[27] Trusov, Michael, Randolph E. Bucklin and Koen Pauwels
(2009), Effects of Word-of-Mouth Versus Traditional Marketing:
Findings from an Internet Social Networking Site,Journal of
Marketing, 73(5), 90 102
[28] Tucker, C. 2008. Identifying formal and informal influence in
tech- nology adoption with network externalities. Management
Sci. 54(12) 20242038