Binge watching and advertising

David A. Schweidel & Wendy W. Moe

Binge Watching and Advertising
How users consume media has shifted dramatically as viewers migrate from traditional broadcast channels toward
online channels. Rather than following the schedule dictated by television networks and consuming one episode of a
series each week, many viewers now engage in binge watching, which involves consuming several episodes of the
same series in a condensed period of time. In this research, the authors decompose users’ viewing behavior into (1)
whether the user continues the viewing session after each episode viewed, (2) whether the next episode viewed is
from the same or a different series, and (3) the time elapsed between sessions. Applying this modeling framework to
data provided by Hulu.com, a popular online provider of broadcast and cable television shows, the authors examine
the drivers of binge watching behavior, distinguishing between user-level traits and states determined by previously
viewed content. The authors simultaneously investigate users’ response to advertisements. Many online video
providers support their services with advertising revenue; thus, understanding how users respond to advertisements
and how advertising affects subsequent viewing is of paramount importance to both advertisers and online video
providers. The results of the study reveal that advertising responsiveness differs between bingers and nonbingers and
that it changes over the course of online viewing sessions. The authors discuss the implications of their results for
advertisers and online video platforms.
Keywords: binge watching, online streaming video, digital advertising, digital media consumption

edia consumption has changed dramatically in recent
years. Viewers have been moving away from
watching traditional broadcast channels and toward
online video consumption to gain more control over their
media consumption. Traditional media consumption, whereby
viewers watch shows according to the schedule and sequence
in which the networks broadcast them, has gradually given
way to viewers determining their own viewing schedule
through digital video recorders or on-demand programming (Littleton 2014). As a result of these trends, new
patterns of media consumption have emerged.
Rather than consuming one episode of a series each week
in accordance with a typical television schedule, viewers may

opt to view several episodes of a single series in immediate
succession. Surveys have revealed that a majority of consumers prefer to watch multiple episodes of their favorite
programs in a single sitting (Pomerantz 2013). A Nielsen
(2013) study finds that 88% of Netflix users and 70% of Hulu
Plus users reported watching at least three episodes of the
same program in one day.
In addition to individuals consuming more content, they
report doing so in a condensed period of time. According to a
survey conducted by Netflix and Harris Interactive in 2013,
61% of adults who stream television shows at least once a
week reported that they regularly engage in “binge watching”
sessions that consist of two to three episodes of a single
television series in one sitting, with nearly three-quarters of

respondents having positive feelings about binge watching
(Netflix 2013). In its 2014 Digital Democracy Survey, Deloitte
reports that 31% of respondents engaged in binge watching at
least once a week, with more than 40% of respondents age
14–25 engaging in the behavior weekly (Deloitte 2015).
What is binge watching?1 The Digital Democracy Survey
defined the activity as “watching three or more episodes of a
TV series in one sitting” (Deloitte 2015). Meanwhile, in the
survey conducted by Netflix and Harris Interactive, nearly
three-quarters of respondents defined binge watching as
“watching between 2–6 episodes of the same TV show in one
sitting” (Netflix 2013). These studies focus on viewing
within a single viewing session, but Netflix further reports
that for a particular serialized drama, 25% of viewers finished
the 13-episode season within two days, and almost 50% did
so within one week (Jurgensen 2013). A similar pattern is also

reported for a sitcom. Across these reports, binge watching is
characterized by two common elements. First, there is a
heavy rate of consumption, which may occur within a single
session or across multiple sessions that occur within a short
period of time. Second, a key feature that distinguishes binge
watching from heavy usage is that binge watching is characterized by consuming multiple episodes of the same series.
These two characteristics are consistent with the definition for
“binge watching” that Oxford Dictionaries added to its online
version in 2014 (Oxford Dictionaries 2014, 2016): “watching
multiple episodes of (a television program) in rapid succession, typically by means of DVDs or digital streaming.” In
this research, we define “binge watching” as the consumption
of multiple episodes of a television series in a short period
of time.

M

David A. Schweidel is Associate Professor of Marketing, Goizueta Business
School, Emory University (e-mail: ). Wendy W. Moe
is Professor of Marketing, Robert H. Smith School of Business, University of
Maryland (e-mail: ). Rajkumar Venkatesan served
as area editor for this article.

© 2016, American Marketing Association
ISSN: 0022-2429 (print)
1547-7185 (electronic)

1We use the term “binge watching” to refer to the activity and
“bingers” to refer to those individuals who engage in the activity.

1

Journal of Marketing
Vol. 80 (September 2016), 1–19
DOI: 10.1509/jm.15.0258


Consistent with the reports noted previously, Google
Trends reveals a sharp increase in searches for “binge
watching” beginning in 2013, as illustrated in Figure 1.
Although reports have acknowledged the trend toward binge
watching, we have little understanding of its implications
for online video platforms. Many online video services are
supported by advertising revenues, as is the case in our
empirical context of Hulu.com. But if viewers are immersed
in binge watching, are advertisements still effective?
Moreover, do the advertisements affect users’ subsequent
viewing behaviors? Elberse and Gupta (2009) report that
advertising on Hulu.com was more effective than advertising
on broadcast or cable television. Yet the analysis does not
distinguish between users’ responsiveness to advertising
when users are engaged in binge watching and when they are
not. If users who are binge watching are less responsive to
advertising, this may give advertising-supported online video
platforms pause in terms of encouraging such behavior.
In this research, we propose a model of viewing behavior
and advertising response and apply it to data from Hulu.com,
a popular online video platform. Our modeling framework
decomposes users’ viewing behavior into (1) the decision to
continue the viewing session after each episode, (2) whether
the next episode viewed is from the same or a different series,

and (3) the time elapsed between sessions, in an effort to
identify binge watching behavior and factors that affect this
behavior, including advertisements. In addition to these three

components of viewing behavior, we simultaneously model
users’ responsiveness to advertisements shown during episodes. This modeling approach allows us to examine how
advertising is related to viewing behavior, in terms of the
effect that binge watching has on advertising response, as
well as how advertisements affect viewing behavior.
Our analysis provides empirical evidence in the context
of online video consumption that viewing begets more
viewing (Kubey and Csikszentmihalyi 2002), suggesting that
binge watching is at least in part a malleable behavior (in
addition to being a user-specific tendency or trait). We also
find that advertisements shown during a viewing session
can deter binge watching behavior and in fact shorten the
length of the viewing session. Finally, we find that when
viewers engage in binge watching, they are less responsive
to advertising. In particular, we show how advertising responsiveness differs between users who have a propensity to
engage in binge watching and those users who shift in and out
of binge watching states. These findings have significant
implications for advertisers and online video platforms
supported by advertising revenues.
The remainder of this manuscript proceeds as follows.
We next provide a review of related research. We then describe our data before presenting our modeling framework
and model specification. Finally, we present our results and
conclude with a discussion of implications and directions for
future work.

FIGURE 1

Google Trend’s Index for “Binge Watching”
100

Indexed Volume of Search Activity

90
80
70
60
50
40
30
20
10

Ja
n
Ma 2, 2
01
r
Ma 2, 2 1
y 2 011
Ju , 20
11
l
Se 2, 2
01
p
1
No 2, 2

v 2 011
Ja , 20
n
1
Ma 2, 2 1
r 2 012
Ma , 2
y 2 012
Ju , 20
12
l
Se 2, 2
p 2 012
No , 20
v 2 12
Ja , 20
n
1
Ma 2, 2 2
0
r
1
Ma 2, 2 3
y 2 013
Ju , 20
l 2 13
Se , 20
p
13
No 2, 2

0
v 2 13
Ja , 20
n 2 13
Ma , 2
01
r
Ma 2, 2 4
y 2 014
Ju , 20
l 2 14
Se , 20
p 2 14
No , 2
v 2 014
Ja , 20
n
1
Ma 2, 2 4
0
r
1
Ma 2, 2 5
y 2 015
Ju , 20
15
l
Se 2, 2
0
p

15
No 2, 2
v 2 015
Ja , 20
n
15
Ma 2, 2
0
r 2 16
,2
01
6

0

Week

2 / Journal of Marketing, September 2016


Related Research
In this section, we provide a brief review of the literature
related to binge watching. We draw on multiple streams of
literature from behavioral, economic, and medical research.
Our intention is not to engage in testing particular theories
using the observational data available to us. Rather, we
provide this discussion to motivate our analysis of binge
watching behavior and our expectations for how the behavior
relates to advertising.
What Is Binge Behavior?

The psychological and medical literature considers binge
behavior an addiction (e.g., Gold, Frost-Pineda, and Jacobs
2003), research into which has shown that individuals often
engage in such behaviors to escape reality. Binge behavior,
in general, has been defined by psychological researchers
as an “excessive amount in a short time,” such as binge eating
or binge drinking (e.g., Heatherton and Baumeister 1991;
Leon et al. 2007). This raises the question: Do such addictive
behaviors extend to television consumption?
Kubey and Csikszentmihalyi (2002) delve into this question by examining the addictive nature of television and
comparing it to substance dependence. The authors note that
electroencephalogram (EEG) studies of individuals watching
television have found that people “reported feeling relaxed
and passive” and reveal that they exhibited “less mental
stimulation” (p. 76). In addition to proposing the association
of a relaxed feeling with viewing that continues throughout a
viewing session, the authors also contend that this association
is negatively reinforced by the stress that viewers experience
when the viewing session ends. As Kubey and Csikszentmihalyi
(2002, p. 77) note, “Viewing begets more viewing,” suggesting
that viewers exhibit a tendency to continue the viewing session
to maintain their current state of mind.
In the online environment, this relates to the concept
of “flow” (e.g., Ghani and Deshpand´e 1994; Hoffman and
Novak 1996), which characterizes immersive experiences in
which the user is in a state of focused concentration, intrinsic
enjoyment, and time distortion. Researchers have also linked
experiencing flow to addictive behaviors. In the context of
video games, Chou and Ting (2003) find that individuals who
experience flow are more likely to become addicted. They

also find evidence to suggest that experiencing flow is an
intermediary step through which repetitive behaviors contribute to addictive behaviors.
As Chou and Ting (2003) note, addictive behaviors have
been viewed in various ways depending on the field of study.
Economists have proposed the theory of rational addiction,
which posits that individuals who exhibit addictive behaviors
may be maximizing their utility and that past consumption can have a substantial impact on the utility derived
from future consumption (e.g., Becker and Murphy 1988). In
the marketing literature, Gordon and Sun (2015) develop a
dynamic model of rational addiction to examine the impact
cigarette taxes on consumption behavior.
Under rational addiction theory, binges can arise from
cyclical behavior (e.g., Becker and Murphy 1988; Dockner
and Feichtinger 1993). Using eating as an example, Becker

and Murphy (1988) describe individuals alternating between
periods of overeating and dieting in order to enjoy consuming
food while also maintaining their weight. In the context of binge
watching, we may find that users take longer to initiate a new
viewing session after a binge experience because they may
derive more utility from other activities.
Advertising and Binge Watching
Advertisements shown during a viewing session can be seen
as an interruption to the experience. We can liken the effect to
that of advertising interruptions during an online browsing
session. Previous studies have shown that online browsers
frequently enter a state of flow (Hoffman and Novak 1996).
Advertisements shown during these sessions interrupt the
flow state and can adversely affect the browsing experience.
Along these lines, Moe (2006) finds that pop-up promotions

that interrupt an online shopping session shorten the duration
of the session and encourage users to exit the site. By the same
token, we would expect that advertisements shown during a
viewing session might interrupt the viewing experience and
consequently contribute to an increase in viewers’ tendencies
to end the session.
Binge watching behavior can also have an impact on
advertising responsiveness. As noted previously, research on
binge and addiction behavior outside the context of binge
watching has shown that users engage in addiction behaviors
as an escape from reality (e.g., Gold, Frost-Pineda, and
Jacobs 2003). In other words, individuals engaged in a binge
state are immersed in an alternate reality. In the context
of binge watching, this alternate reality is created by the
video content, and advertisements shown during these
sessions can be seen as unwelcome reminders of the viewer’s
true reality. Thus, our expectation is that viewers engaged in
binge watching will be less responsive to advertisements than
viewers not engaged in binge watching because they prefer to
remain immersed in the context of the series they are viewing.

Modeling Framework
In this section, we conceptually describe our modeling
framework before presenting the data and the methodological
details of the model. Advertising-supported platforms that
provide streaming video content have an interest in two types
of behaviors of their users: viewing behavior and advertising
responsiveness. In this article, we simultaneously model
viewing behavior and advertising responsiveness at the level
of the individual user while allowing the two to be related.

Our goal is to capture characteristics of viewing behavior that
may indicate binge watching behaviors and relate those
characteristics to how the user responds to advertising.
To model viewing behavior, we consider users’ viewing
decisions at the end of each episode they have viewed. Consistent
with prior research on live television viewing (e.g., Rust and
Alpert 1984; Rust, Kamakura and Alpert 1992; Shachar and
Emerson 2000), we decompose a viewing session into a
series of choices made by users. First, after each episode, we
model users’ decisions to continue their viewing session by
watching another episode (of any program). Second, we
model users’ decisions to watch another episode of the same

Binge Watching and Advertising / 3


program, an option not considered in studies of live television
viewing because this decision is facilitated by today’s
streaming video services. Third, if a user decides to conclude
the current viewing session, we consider the time until the
user returns to the platform to begin a new viewing session.
Finally, we simultaneously model the user’s response to any
advertising to which he or she is exposed. Specifically, we
examine the number of advertisements on which a user clicks,
out of the total number of advertisements to which the user
was exposed during an episode, as a binomial process.
Overall, our modeling framework allows us to examine
both how advertisements affect viewing and how viewing
behaviors affect the user’s response to advertisements. In
modeling both advertising and viewing decisions, we

allow for heterogeneity across users and recognize that users’
tendencies for each behavior may be correlated. In addition to
variation in users’ tendencies, we account for shifts in behavior
that reflect prior viewing and advertising responsiveness.

Data
Data Description
The data for our empirical analysis consist of the video
viewing behavior of 9,873 registered users of Hulu.com from
February 28, 2009, to June 29, 2009.2 While maintaining a
library including both movies and television programs, Hulu.
com “had a brand promise that was clear and distinctive:
Hulu is where you go for network TV” (Hansell 2009). This
served as a point of differentiation compared with other
popular online video portals, such as YouTube, that were
populated primarily with user-generated content (e.g.,
Elberse and Gupta 2009). Due to this positioning, discussions have occurred in the popular press about Hulu.
com’s impact on broadcast and cable television’s business
models (e.g., Learmonth 2009; Rose 2008, 2009). Following
its 2009 Super Bowl advertisement, online conversation
about Hulu increased more than 250% (Eshman 2009). In
April 2009, Hulu announced a deal to make content from
Disney available to Hulu users, including episodes of primetime hits such as Lost, Grey’s Anatomy, Desperate Housewives, Ugly Betty, Samantha Who?, Scrubs, and Private
Practice, as well as content from ABC Family and Disney
Channel (Kilar 2009a). The five most popular shows on Hulu
in 2009 were Saturday Night Live, Family Guy, The Office,
The Simpsons, and Naruto Shippuden (a Japanese anime
series), and an episode of Family Guy was the most played
full episode (Kilar 2009b).
The data for each individual user consist of an event log

that indicates the videos viewed.3 Each episode of a program
2The data set employed in this study was previously employed by
Schwartz et al. (2011) and Zhang, Bradlow, and Small (2013). We
refer interested readers to these studies for additional details of the
data. We excluded data from 164 users who had sessions that
exceeded 24 hours in length, due to concerns about the veracity of
the data.
3We use the terms “viewers” and “users” interchangeably.
Typical of many panel data sets, our data set does not allow us to
distinguish among multiple individuals in the same household.

4 / Journal of Marketing, September 2016

is divided into multiple segments. The event log records the
time at which each video segment began, as well as information about the video segment, including the title of the
television series and episode, the season of the series to which
the episode belongs, and the episode number within the season.
In addition to the series and episodes that users viewed,
the event log also contains information on the advertising to
which users were exposed. The advertising data include a
time-stamp at which the advertisement was served to the user,
as well as the program and episode in which the advertisement aired. Our advertising data also indicate whether an
individual took action and clicked on the advertisement. Over
1.1 million advertisement impressions were recorded in our
data, with users clicking on 9,317 advertisements (.84% of
the advertisements).4 In Figures 2–4, we provide histograms
that show the distribution of the number of episodes viewed
by users, the number of unique programs viewed by users,
and the number of viewing sessions conducted by users,
respectively.

Although most users viewed several episodes, we find
variation in the number of viewing sessions that users conducted. While 25% of users conducted just one viewing
session, approximately an equal proportion conducted ten or
more viewing sessions. We also see that the number of series
that users viewed follows a bimodal distribution. While 35%
of users viewed only one or two series, more than 25% of
users viewed ten or more different series.
We provide descriptive statistics based on users’ behaviors across all viewing sessions in our data in the upper
portion of Table 1. We define a viewing session as a period of
video viewing separated by one hour or more of inactivity. In
the data, 9,873 users were responsible for 104,414 viewing
sessions, an average of 10.58 sessions per user during the
four-month data period. We provide descriptive statistics
about the composition of these sessions in the lower portion
of Table 1.
Although Table 1 provides a summary of viewing
behavior at the level of the session, such statistics do not shed
light on the heterogeneity that exists across users. They also
do not provide insight into the relationship between the
volume of online video consumption and the content consumed. To investigate these factors at the session level, we
present the number of programs viewed conditional on the
length of the session in Table 2.
According to the session-level data presented in Table 1,
in at least 50% of sessions, viewers watched two or more
episodes, and in at least 50% of the sessions, viewers constrained their viewed episodes to a single series. Additionally,
the lower bound of the interquartile ranges presented in
Table 2 is equal to one series, irrespective of the number of
episodes viewed. In other words, for each session length
considered, at least 25% of sessions involved viewers watching episodes from a single series. These statistics suggest the
prevalence of binge watching behavior in the data.

4During the data period, all videos on the site could be viewed free
of charge, and the service was strictly ad-supported. In June 2010,
after the data period concluded, Hulu introduced the Hulu Plus
subscription service.


50%

50%

45%

45%

40%

40%

35%
30%
25%
20%
15%

35%
30%
25%
20%
15%

10%

10%

5%

5%
0%

0%
1

2

3

4

5

6

7

8

9

10+

To further illustrate the prevalence of binge watching behavior in our data, we show the joint distribution of the number
of episodes and unique series viewed in a session in Table 3. We
see that 63.3% of viewing sessions consisted of a single series,
and 18.5% consisted of multiple episodes of a single series. We
next consider the fraction of users who engaged in different types
of viewing sessions. We consider three types of viewing sessions: (1) single-episode sessions, (2) multiepisode sessions that
consist of episodes from a single series, and (3) multiseries
sessions that consist of episodes from multiple series. We find
that 81.1% of users conducted at least one single-episode session, 49.0% of users conducted at least one multiepisode viewing
session that consists of episodes from a single series, and 69.0%
users conducted at least one multiseries session.

FIGURE 3
Distribution of the Number of Series Viewed
50%
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
1

2

3

4

5

6

7

Series Viewed

1

2

3

4

5

6

7

8

9

10+

Viewing Sessions

Episodes Viewed

Fraction of Users

FIGURE 4
Distribution of the Number of Viewing Sessions

Fraction of Users

Fraction of Users

FIGURE 2
Distribution of the Number of Episodes Viewed

8

9

10+

Taken together with Table 3, with nearly half of users
viewing multiple episodes of a single program in a session,
these statistics provide model-free evidence for the presence
of binge watching in our data. While our exploratory analysis
suggests that binge watching does occur, it does not enable
us to discern whether such behavior is driven by user-specific
traits or by recent viewing behavior. To disentangle these

competing explanations, as well as to understand how this
behavior is related to users’ advertising responsiveness, we
develop a joint model of viewing behavior and advertising
response. We next describe the key variables in our empirical
analysis before presenting our modeling framework.
Variable Specification
For our model, we first create a set of dependent variables that
characterize the various components of users’ viewing and
advertising response decisions. These variables are user-level
and time-varying from episode to episode. The viewing
decisions we are interested in reflect the length, variety of
programming, and frequency of viewing sessions. Because a
viewing session is only observed when at least one episode is
viewed, we characterize the length of a session according to a
user’s decision of whether or not to continue the session after
viewing an episode. In other words, we construct a binary
variable equal to 1 if the user views another episode and
equal to 0 if the user chooses to end the viewing session
(CONTINUE). To represent the variety of the viewing session, we again consider the user’s decision after viewing an
episode. Conditional on the user viewing another episode, we
construct a binary variable equal to 1 if the next episode is
from the same series as the episode just completed and equal
to 0 if it is from a different series (SAME). Finally, if the user
chooses to end the viewing session, we then consider the
frequency of viewing sessions by computing the time (in
days) until the next viewing session (FREQUENCY). To
model advertising click-through behavior, for each episode, we count the number of advertisements on which a

Binge Watching and Advertising / 5

TABLE 1
Descriptive Statistics

By User
Viewing sessions
Episodes viewed
Programs viewed
Ads shown
Ads clicked
Season finales views
Series finale views
By Session
Episodes viewed
Programs viewed
Ads shown
Ads clicked
Intersession time (days)

M

SD

10.58
34.01
8.60
111.70
.94
2.16
.98

19.31
114.57
12.62
368.39
4.14
8.23
4.36

3.21
1.66
10.56
.09
3.73

10.67
1.18
39.68
.95
9.84

Mdn

95% Range

4
8
4
22
0

0
0

[1,
[1,
[1,
[0,
[0,
[0,
[0,

2
1
6
0
.86

90% Range

68]
229.68]
43]
797.03]
8]
17]
8]

[1,
[1,
[1,

[1,
[0,
[0,
[0,

[1, 12]
[1, 5]
[0, 40]
[0, 1]
[.05, 32.63]

IQR

27]
79]
21]
255.20]
2]
2]
2]

[2,
[3,
[2,
[7,
[0,
[0,
[0,

[1, 6]

[1, 3]
[0, 19]
[0, 0]
[.07, 7.93]

10]
27]
10]
79]
1]
1]
1]

[1, 3]
[1, 2]
[3, 11]
[0, 0]
[.33, 2.25]

Notes: IQR = interquartile range, representing the range that defines the middle 50% of observations.

user clicks (CLICKTHRU) out of the total number of advertisements to which the user is exposed. Taken together, this
yields the click-through rate for each episode.
In addition, we construct a number of covariates that are
expected to influence the viewing and advertising clickthrough decisions. These covariates are designed to capture
the effects of temporal factors (time of day, day of week, etc.),
program variation (e.g., genre), and content to which a user
was previously exposed, including program content and
advertising. We provide a list of the covariates employed in
our analysis in Table 4.

First, we consider the effects of viewing variables that
capture previous viewing behavior. Specifically, we consider a breadth variable, which is measured as the number of
different series viewed in a session up until the point of the
behavioral decision being considered, and a depth variable,
which is measured as the number of episodes from the current
series viewed thus far in the viewing session. We expect these
covariates to affect both the viewing decisions and advertising responsiveness. We construct episode-by-episode breadth
and depth measures (BREADTH_EPISODE and DEPTH_
EPISODE) that are used to model within-session behaviors
(CONTINUE and SAME), as well as measures that summarize

breadth and depth within the previous session (BREADTH_
SESSION and DEPTH_SESSION) and that are used to model
intersession durations (FREQUENCY). In addition to the
breadth and depth measures, we include an indicator variable
to account for whether a user has viewed the current series in an
earlier viewing session (PRIORVIEW).
Second, throughout these viewing sessions, users are
exposed to advertisements. Users’ exposures to advertisements
vary with the amount of content they consume; that is, longer
programs include more advertising than shorter programs.
We construct a time-varying covariate, EXPOSURES, that
represents the number of advertisements to which a user has
been exposed in each episode. Because viewers’ interactions
with advertisements may affect their subsequent viewing
decisions, we also consider the impact of the number
advertisements on which the user has clicked in each
episode, captured by the variable CLICKTHRU.
Finally, we specify a number of control variables to
capture the effects of content and temporal differences (time

of day, day of week, etc.) on both viewing and advertising click-through decisions. Drawing on the data provided
by Hulu.com, we identify 18 genres in our data and construct a
series of 17 indicator variables (one for each genre, with action/

TABLE 2
Unique Series Viewed, by Number of Episodes in the Session
Number of Episodes
1
2
3
4
5
6
7
8
9
10+

Percentage of Sessions

M

SD

Mdn

44.79
21.71
11.96
6.90

4.24
2.68
1.73
1.25
.89
3.89

1
1.6
2.04
2.39
2.66
2.92
3.06
3.29
3.35
3.51

0
.49
.84
1.14
1.42
1.65
1.87
2.04
2.25
2.76

1

2
2
2
2
3
3
3
3
3

95% Range
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,

Notes: IQR = interquartile range, representing the range that defines the middle 50% of observations.

6 / Journal of Marketing, September 2016

1]
2]
3]
4]

5]
6]
7]
8]
9]
11]

90% Range
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,

1]
2]
3]
4]
5]
6]
6]
6]
7]
7]

IQR
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,
[1,

1]
2]
3]
3]
4]
4]
4]
5]
5]
5]


TABLE 3
Numbers of Episodes and Series Viewed, Across Sessions
Number of Episodes
Number of Series
1
2

3
4
5+

1

2

3

4

51

44.76%

8.58%
13.13%

4.01%
3.43%
4.52%

2.06%
1.70%
1.53%
1.60%

3.88%
3.14%

2.53%
1.92%
3.21%

adventure being our baseline genre). We also construct a
covariate to represent whether the episode just viewed
was a season finale (SEASON_FINALE) or a series finale
(SERIES_FINALE). Finally, we define a set of variables to
capture month, day-of-week, and time-of-day effects. For day of
week, we differentiate between weekdays (Monday–Thursday),
Friday, Saturday, and Sunday, again constructing indicator
variables whereby weekdays are considered the baseline.
For time of day, we divide the day into parts according to
Headen, Klompmaker, and Rust (1979). Specifically, we
construct indicator variables for early morning (7 A.M.-10 A.M.),
daytime (10 A.M.-5 P.M.), early fringe (5 P.M.-8 P.M.), and prime
time (8 P.M.-11 P.M.). The late fringe period of 11 P.M.-7 A.M. is
set as our baseline.

Model
Model Specification
Figure 5 provides a depiction of our modeling framework,
including the user decisions we consider and the variables that
influence them. Our model is made up of four decisions: (1)
whether to continue the viewing session by viewing another
episode (CONTINUE); (2) if the session is continued, whether to
watch a video from the same or a different series (SAME); (3)
how much time elapses until the next viewing session begins
(FREQUENCY); and (4) whether to click on an advertisement
(CLICKTHRU). We develop our model by first considering the

intrasession viewing behaviors of whether to continue a viewing
session and whether to watch the same series. We then present

TABLE 4
Variable Descriptions
Variable Name

Description

BREADTH_EPISODEits

Number of series previously viewed in session s by user i prior to video t; operationalized as
log(1 + number of series previously viewed) and mean-centered

DEPTH_EPISODEits

Number of episodes of the current series previously viewed in session s by user i prior to video
t; operationalized as log(1 + number of episodes previously viewed) and mean-centered

BREADTH_SESSIONis

Number of series viewed in session s by user i; operationalized as log(1 + number of series
viewed) and mean-centered

DEPTH_SESSIONis

The maximum number of episodes of a single series viewed by user i in session s;
operationalized as log(1 + number of episodes viewed) and mean-centered

PRIORVIEWits

Indicator variable for whether or not user i has previously viewed the same series as video t of
session s in an earlier viewing session

EXPOSURESits

Number of advertisements shown to user i in video t of session s

CLICKTHRUits

Number of advertisements clicked by user i in video t of session s

SEASON_FINALEits

Indicator variable for whether video t in session s viewed by user i is a season finale

SERIES_FINALEits

Indicator variable for whether video t in session s viewed by user i is a series finale

MONTHits

Month in which video t viewed by user i in session s began (March, April, May, June)

DAYits

Day of the week on which video t viewed by user i in session s began (Monday–Thursday,
Friday, Saturday, Sunday)

DAYPARTits

Day part in which video t viewed by user i in session s began (early morning, daytime, early
fringe, prime time, late fringe)

GENREits

Program genre of video t viewed by user i in session s (action/adventure, animation and
cartoons, comedy, drama, family, food and leisure, home and garden, horror and suspense,
music, news and information, other, reality and game shows, science fiction, sports, talk and
interviews, unknown, video games, web)

Binge Watching and Advertising / 7


FIGURE 5
Depiction of Modeling Framework
Viewing Variables

Viewing Behavior
Continue

Same

Frequency

Do I continue or
stop watching?
(pits)

Do I watch the

same or a
different series?
(qits)

How long before
I watch again?
(h(t))

•
•
•
•
•
•
•
•
•

Control Variables
•
•
•
•
•
•

GENRE
DAY
MONTH
DAYPART

SEASON_FINALE
SERIES_FINALE

DEPTH_EPISODE
DEPTH_EPISODE2
BREADTH_EPISODE
BREADTH_EPISODE2
DEPTH_SESSION
DEPTH_SESSION2
BREADTH_SESSION
BREADTH_SESSION2
PRIORVIEW

Ad Variables

Advertising Exposures
•
•

EXPOSURES
CLICKTHRU

Advertising Clickthrough
Do I click on the ad? (rits)

the model component for intersession timing decisions. Finally,
we describe our model for viewers’ advertising responsiveness.
For user i who has viewed t videos in the current session s,
we model two binary decisions: (1) the decision to continue
session s by viewing another video (CONTINUE) and (2) conditional on viewing another video, the decision to watch the

same series or a switch to a different program (SAME). As shown
in Figure 5, these decisions are affected by viewing variables
based on previous viewing activity, ad variables, and control
variables. After user i views video t, we model the probability
with which the user continues the current session as pits:
(1) logitðpits Þ = ai + q1 DEPTH EPISODEits
+ q2 DEPTH EPISODE2its
+ q3 BREADTH EPISODEits
+ q4 BREADTH EPISODE2its
+ q5 PRIORVIEWits + q6 EXPOSURESits
+ q7 CLICKTHRUits + q8 SEASON FINALEits
26

+ q9 SERIES FINALEits +

å q GENRE
j

its

j = 10
29

+

å

32

qj DAYits +

j = 27

å q MONTH
j

its

j = 30

36

+

å q DAYPART
j

its ,

j = 33

where ai is an individual-level intercept such that
ai = a + g i1 , with g i1 a random effect with mean 0 that
captures variation across users. We allow for nonlinear
effects of the viewing variables DEPTH_EPISODE and

8 / Journal of Marketing, September 2016

BREADTH_EPISODE. If we consider binge watching an
addictive behavior whereby “viewing begets more viewing,”

as proposed by Kubey and Csikszentmihalyi (2002), we
should expect DEPTH_EPISODE (captured by q1 and q2)
to have a positive effect on pits. Likewise, the impact of
BREADTH_EPISODE (reflected by q3 and q4) is expected
to be negative because continued viewing of a given program is more likely in such an addictive state in which
BREADTH_EPISODE of viewing is low. The coefficient q5
accounts for the effect of user i having viewed the current
program prior to episode t of session s on the decision to
continue the current viewing session. The effects of ad variables (EXPOSURES and CLICKTHRU) to which user i is
exposed during episode t of session s on the decision to
continue viewing session s are captured by q6 and q7. If we
assume that advertising breaks up the flow of a binge watching
session and discourages further viewing, we should expect
that q6 < 0 and q7 < 0. The remaining coefficients (q8–q36)
capture variation in the decision to continue the session that is
related to control variables, including whether video t is a season
(q8) or series (q9) finale, the genre viewed in video t (q10–q26),
and day and time (q27–q36) at which video t is viewed.
In the special case where q = 0, then logit(pits) = ai and the
length of the viewing session (in episodes) follows a shifted
geometric distribution at the user level. Following those who
employ this individual-level model for discrete-time customer base
analysis (e.g., Fader and Hardie 2009), we accommodate heterogeneity across users. We also allow the likelihood of continuing the
viewing session to shift depending on the content that a user views.
We employ a similar binary logit model for a user’s
decision to continue viewing the same program, as opposed
to viewing a different program. Conditional on user i


continuing viewing session s, we specify the probability with

which video t + 1 is from the same series as video t as qits:
(2)

logitðqits Þ = bi + y1 DEPTH EPISODEits

(4)

+ y4 BREADTH
+ y5 PRIORVIEWits + y6 EXPOSURESits
+ y7 CLICKTHRUits
+ y8 SEASON FINALEits
+ y9 SERIES FINALEits

å y GENRE
j

its

å y DAY

+

j = 27

32

+

j

å y MONTH
j

its

+

å y DAYPART

j = 33

j

its ,

where the user-level intercept is specified as bi = b + g i2 . As
in our specification of pits, our specification of qits enables us
to distinguish the user’s general tendency to watch the same
program (through g i2) from the impact of previously viewed
content (through y1–y5). If we assume that consuming
multiple episodes of the same series increases addictive
behavior, then we should expect to observe a tendency to continue viewing the same series as depth increases. Likewise, if
limited breadth contributes to addictive viewing behavior focused
on a single series, then we are more likely to observe a tendency to
switch series as breadth increases. The impact of ad variables
on this component of viewing behavior is reflected in y6 and
y7. The coefficients y8-y36 capture the effects of the control
variables that account for content and temporal differences.
Whereas Equations 1 and 2 characterize the user decisions
within a single viewing session, our third model component

(FREQUENCY) looks across viewing sessions. At the conclusion of viewing session s, we model the time until the start
of the next viewing session using a proportional hazard
model (e.g., Seetharaman and Chintagunta 2003). We assume a Weibull distribution for the baseline hazard (e.g.,
Helsen and Schmittlein 1993; Schweidel, Fader, and Bradlow
2008; Seetharaman and Chintagunta 2003), which accommodates increasing or decreasing hazards and nests the constant
exponential hazard process as a special case. 5 The baseline Weibull hazard is given by
(3)

bðtÞ = luðltÞ

u-1

(5)

Xis = g i3 + w1 DEPTH SESSIONis
+ w2 DEPTH SESSION2is
+ w3 BREADTH SESSIONis

+ w5

its

36

j = 30

where

+ w4 BREADTH SESSION2is

29

j = 10

SðtÞ = exp 0

EPISODE2its

+

!
Â
Ã
u
bðuÞexpðXis Þdu = exp -ðltÞ expðXis Þ ,

ðt

+ y2 DEPTH EPISODE2its
+ y3 BREADTH EPISODEits

26

and control variables to affect how quickly a user returns to the
website to begin a new viewing session. The resulting survival
function is then given by

,

where l > 0 and u > 0. The survival function can then be

written as a function of the hazard rate h(t), which is given by
h(t) = b(t)exp(Xis), where b(t) is the baseline Weibull hazard
rate and Xis captures the impact of covariates corresponding
to session s for user i. As described in Equations 1 and 2 and
illustrated in Figure 5, we allow viewing variables, ad variables,
5We compared the proposed model that uses the Weibull distribution
as the baseline hazard with the model that employs the exponential
distribution as the baseline hazard. Although we did not find any substantive differences between the specifications, on the basis of the logmarginal density, the model that uses the Weibull hazard better fits the
data. We therefore present the results corresponding to this model.

å EXPOSURES
å CLICKTHRU

i$s

+ w6
i$s
+ w7 maxðSEASON FINALEi$s Þ
+ w8 maxðSERIES FINALEi$s Þ
25

+

+

28

å w GENRE
j

iNis s

+

å w DAY
j

j=9

j = 26

31

35

å w MONTH
j

j = 29

iNis s

+

iNis s

å w DAYPART
j

iNis s ,

j = 32

and Xis captures the impact of observed covariates and unobserved heterogeneity across users.
The user-specific random effect g i3 has a mean of 0 and
captures unobserved differences across users. As described
in Table 4, the measures of DEPTH_SESSIONis and BREADTH_
SESSIONis provide measures of viewing depth and breadth that are
calculated according to viewing behavior in session s. For example,
if a user has viewed three episodes of one series and one episode of
another, the maximum depth during the session is 3. If a user is
addicted to a series and has watched many episodes of a given
program in a session (i.e., DEPTH_SESSION is high), then we
might expect the user to return to view more episodes in a relatively
short amount of time, reflected by w1 and w2. Alternatively, if users
exhibiting binging behavior cycle between different activities (e.g.,
Becker and Murphy 1988), we might expect increased depth to
increase the time until the next viewing session begins. To account
for the potential impact of ad variables on the time until user i
begins session s + 1, we aggregate the advertising exposure (w5)
and advertisement clicks (w6) throughout session s by summing
these variables across all videos that comprise the viewing session.
We also account for the content viewed during session
s and the time of session s. The coefficients w7 and w8
account for the presence of a season or series finale, respectively, in session s. If an episode viewed in session s is
a season finale, max(SEASON_FINALEits) = 1; otherwise,
max(SEASON_FINALEits) = 0. We control for the genre of
the final episode viewed by user i in session s (which we
denote video Nis) to account for the most recently viewed
content. Similarly, we control for the time (day part, day of

week, and month) at which video Nis is viewed.
We use a binomial distribution to model the number of
advertisements on which user i clicks in episode t of session s,
according to the number of advertisements to which the user
is exposed. We specify user i’s probability of clicking on
an advertisement during video t of session s, rits, using a

Binge Watching and Advertising / 9


binary logit model (e.g., Chatterjee, Hoffman, and Novak
2003; Dr`eze and Hussherr 2003; Hoban and Bucklin 2015;
Urban et al. 2014), where rits is affected by the viewing
variables and control variables shown in Figure 5:
(6)

logitðrits Þ = d i + f1 DEPTH EPISODEits
+ f2 DEPTH EPISODE2its
+ f3 BREADTH EPISODEits
+ f4 BREADTH EPISODE2its
+ f5 PRIORVIEWits
+ f6 SEASON FINALEits
+ f7 SERIES FINALEits
24

27

å f GENRE

+

j=8

j

its

å f DAY

+

j

30

34

å f MONTH

+

j = 28

its

j = 25

j

its

+

å f DAYPART

j = 31

its ,

j

where d i = d + g i4 and g i4 is a user-specific intercept; f1 and
f2 capture the impact of the depth of viewing of the series
viewed in video t; f3 and f4 account for the breadth of viewing
that has occurred in the session prior to viewing video t; and f5
controls for the impact of prior exposure to the series. In addition
to viewing variables, we also include control variables that
account for whether or not video t is a season (f6) or series finale
(f7), the genre of video t (f8–f24), and when video t is viewed
(f25–f34).6 If we assume that binge watching reduces viewers’
responsiveness to advertisements, we should anticipate that the
impact of depth on the click-through probability (captured by f1
and f2) is negative.
To complete our model specification, Equation 7 provides the
joint likelihood function. Let Yits = 1 when the next video that user i
chooses to view is from the same series as video t of session s, let
Yits = 2 when user i chooses to view a different program, and let
Yits = 3 when user i decides to end session s after viewing video t.
Let dis denote the time between sessions s and s + 1 for user i.
Combining the intrasession viewing decisions (whether to continue

the session and whether to view the same series), episode-level
advertising response, and proportional hazard model of intersession
durations, the likelihood of user i’s behavior is given by
(7)
&

Si

∏

Li =

s=1

N
is
∏ ðpits Þ1ðYits <3Þ ðqits Þ1ðYits =1Þ
t=1

!

· ð1 - qits Þ1ðYits =2Þ ð1 - pits Þ1ðYits =3Þ
Si

Nis

· ∏ ∏
s=1 t=1



EXPOSURESits
CLICKTHRUits


!

CLICKTHRUits
ð1 - rits ÞEXPOSURESits -CLICKTHRUits
· rits

!!'
 

1
SðT - diSi Þ.
- Sðdis Þ
∏ S dis +
24
s=1

Si -1

·

6Note that DEPTH_EPISODE and BREADTH_EPISODE , as
its
its
described in Table 4, are calculated based on user i’s viewing behavior
in session s prior to video t. Thus, we avoid problems with simultaneity.

10 / Journal of Marketing, September 2016

The first line of Equation 7 captures intrasession viewing
decisions to continue the current viewing session (with
probability pits) and to view another episode of the same
series, conditional on continuing the session (with probability
qits). If user i continues the session (Yits < 3), we employ a
binary logit model for the decision to view the same or
different content. A user chooses to not continue the session
after viewing episode Nis with a probability of 1 - piNis s . This
episode-level behavior is modeled for each of the Si viewing
sessions from user i. The second line of Equation 7 accounts
for user i’s response to advertising in video t of session s,
using a binomial distribution with probability rits.
The last line of Equation 7 models the intersession durations. Because we do not know the exact time at which a
viewing session ends, we treat the observed intersession
times dis as interval-censored data in which the contribution
of dis to the likelihood Li is based on the difference between
the survival function evaluated one hour (the cutoff used
to define the ends of sessions) after the end of the last episode
of session s and the survival function evaluated at the end
of the last episode of session s. We also account for the rightcensored duration observed between the end of our observation period and the end of the final session observed from
user i.
We assume all users are heterogeneous across their
viewing behavior and advertising responsiveness, with users’
tendencies being correlated. We assume that users’ behavioral tendencies, reflected by the user-level random effects g i$, are distributed such that g i$ ~ MVN(0,S) where 0
is a 4 · 1 vector of zeros and S is the covariance matrix.
The inclusion of user-specific random effects enables us to
distinguish between unobserved heterogeneity that exists

across users and state dependence arising from previously
consumed content (e.g., breadth and depth of viewing).
Neglecting to account for heterogeneity has also been shown
to yield misleading inferences (e.g., Hutchinson, Kamakura,
and Lynch 2000; Roy, Chintagunta, and Haldar 1996). In
addition to capturing differences across users for each of the
behaviors we model, the correlated random effects also serve
to induce a correlation at the margin among the four components of our model in a manner consistent with prior
research on brand choice (e.g., Bucklin and Gupta 1992),
online behavior (e.g., Johnson et al. 2004; Sismeiro and
Bucklin 2004), and in-store behavior (e.g., Hui, Bradlow, and
Fader 2009).
We assume an inverse Wishart prior for S and normal
priors for the fixed effects.7 We obtain posterior draws from
the joint posterior using Markov chain Monte Carlo (MCMC)
methods. We ran two independent chains with different
starting values until convergence. For each chain, we removed
the first 10,000 iterations as a burn-in period and then used the
subsequent 5,000 iterations for inference.
Is Binge Watching a Trait or a State?
Popular reports of binge watching confound two distinct
explanations for the activity. One explanation assumes that
7Because l > 0 and u > 0, we assume diffuse normal priors for log
(l) and log(u).


some users are more inclined to have long viewing sessions
in which they watch multiple episodes of the same program
and return quickly to begin new sessions. In other words,
binge watching is a time-invariant, user-specific trait. Our

empirical model captures variation in this trait across users by
accounting for user heterogeneity with the vector of random
effects g i$. Those users who have a high tendency to continue
their viewing sessions, watch additional episodes of the same
series, and return quickly to begin new viewing sessions can
be considered to have a higher tendency to engage in binge
watching behavior. Moreover, our approach allows us to
distinguish those who are prone to binge watching from those
who are not, a distinction of managerial relevance that we will
discuss in our results.
Alternatively, binge watching might also be a malleable
behavior stimulated by the content consumed. Rather than
arising from a time-invariant, user-specific trait, binge
watching may come about from a time-varying behavioral
state; that is, users may engage in “situational” binge
watching. We account for shifts in users’ likelihood of engaging in binge watching that may occur over time through the
covariates in our analysis, specifically, our measures of series
breadth (BREADTH_EPISODE and BREADTH_SESSION)
and depth (DEPTH_EPISODE and DEPTH_SESSION).
The coefficients associated with these variables reflect how
watching more episodes of a single series can contribute to
(or distract from) a state of binge watching behavior. Likewise, the coefficients for advertising exposure and responsiveness reflect how these factors may affect subsequent
viewing behavior. Because these effects vary across time,
they represent situational factors that can influence whether
a user is in a binge watching state, as opposed to affecting a
user’s inherent viewing trait. Our approach to accommodating shifts in users’ behavioral tendencies over time is in
line with Gordon and Sun’s (2015) accommodation for state
dependence through the use of observed state variables to
investigate addiction in the context of cigarette purchasing.
It is also consistent with prior research into consumers’

brand choice behaviors (e.g., Erdem and Sun 2001; Roy,
Chintagunta, and Haldar 1996).

Results
In this section, we present our modeling results as they pertain
to (1) baseline behaviors, (2) the effects of control variables
(including temporal factors and the effects of finales), (3) the
effects of viewing variables and ad variables on subsequent
viewing behavior, and (4) the effects of viewing variables on
advertising responsiveness. We focus our discussion on the
results as they pertain to binge watching behavior and how
this behavior affects and is affected by advertising. We also
discuss binge watching as a malleable behavioral state versus a
user-specific trait.
The first three rows of Table 5 provide the baseline
parameters associated with each of the four component
models. Table 5 also provides the posterior means associated
with each of the temporal factors (i.e., month, day of the
week, and day part) and season/series finales on both viewing
behavior and advertising response. Looking at how viewing

behaviors vary over the course of the day, we observe that
users are less likely to continue viewing sessions in the early
morning and daytime day parts than they are during late
fringe. During the early morning day part, we also observe
that users are less responsive to advertisements than during
late fringe. Compared with sessions that end during late fringe,
when the current session ends in one of the other day parts,
users return sooner to begin new viewing sessions.
Turning our attention to how viewing behavior shifts

with finales reveals an interesting pattern of results. After
viewing a season or series finale, users are less likely to
continue their viewing session, and users take longer to return
for a new viewing session after season finales. In addition to
these effects on viewing behavior, our results show that
viewers are less responsive to advertisements during season
finales.
Table 6 presents the variation in viewing behavior and
advertising responsiveness across genres. We observe differences in users’ tendencies to continue their viewing sessions according to the genre of the current episode (column
1), as well as differences in the time until users begin
new sessions according to the genre of the final episode
viewed in the current session (column 3). We also see that
users’ likelihoods of continuing to view the same series differ according to the genre of the current program (column 2).
We observe that users are both more likely to continue
watching the same series and more likely to continue the
current viewing session (vs. the action/adventure genre) in
8 of the 17 genres presented in Table 6, suggesting that there
is variation by genre in the “binge-worthiness” of programming.
In addition to genre effects on different aspects of viewing
behavior, there are also differences in advertising responsiveness related to genre. Compared with advertisements aired
during action/adventure programs, users are more responsive to
advertisements during programs in the animation and cartoons,
reality and game shows, and video games genres and less
responsive in the food and leisure, home and garden, news
and information, sports, and unknown genres.
Effects of Previously Viewed Content and
Advertising on Viewing Behavior
In Table 7, we consider the effects of previously viewed
content, operationalized using breadth and depth measures
of prior viewing during a viewing session, on subsequent

viewing behavior and advertising responsiveness. We also
consider the effects of advertising exposure on viewing
behavior.
Our results indicate that increased viewing depth is associated with binge behavior, as indicated by the effects of
DEPTH_EPISODE on the choices of whether to continuing
the session and of whether to view the same series, and by the
effect of DEPTH_SESSION on the time until the next session. The more episodes within a given series a user views,
the more likely he or she will continue to view episodes (q1 =
.79, q2 = .07), will view another episode from the same series
(y1 = 3.22, y2 = -.83), and will return quickly for a future
viewing session (w1 = .12, w2 = -.03). Thus, our results are
consistent with Kubey and Csikszentmihalyi (2002) in that
we also find that “viewing begets more viewing” in the

Binge Watching and Advertising / 11


TABLE 5
Effects of Seasonality and Finales on Viewing Behavior and Advertising Response
Dependent Variable
CONTINUE
Baseline Behaviors
Intercept
Weibull parameter (log(l))
Weibull parameter (log(u))

SAME

FREQUENCY

CLICKTHRU
25.11 (.05)

.66 (.02)

1.77 (.06)

Month (Baseline: March)
April
May
June

2.13 (.01)
2.16 (.01)
2.14 (.01)

2.17 (.03)
2.16 (.03)
-.02 (.03)

2.16 (.01)
2.24 (.01)
2.27 (.01)

.02 (.03)
.08 (.04)
.05 (.04)

Day of Week (Baseline: Monday–Thursday)
Friday

Saturday
Sunday

.00 (.02)
-.03 (.02)
.04 (.03)

-.01 (.04)
.59 (.05)
-.07 (.06)

.02 (.02)
.05 (.02)
.19 (.02)

.19 (.05)
.18 (.06)
2.20 (.08)

Time of Day (Baseline: Late Fringe [11 P.M.–7 A.M.])
Early morning (7 A.M.–10 A.M.)
Daytime (10 A.M.–5 P.M.)
Early fringe (5 P.M.–8 P.M.)
Prime time (8 P.M.–11 P.M.)

2.29
2.06
.01
.02

.37
.01
2.08
.01

.05
.13
.13
.10

2.11
-.06
.10
.03

Finales
Season finale
Series finale

2.14 (.02)
2.18 (.03)

(.02)
(.01)
(.01)
(.02)

(.03)
(.03)
(.03)

(.03)

2.11 (.04)
2.29 (.06)

23.18 (.03)
2.49 (.00)

(.02)
(.01)
(.01)
(.01)

2.07 (.02)
-.03 (.03)

(.05)
(.04)
(.03)
(.04)

2.13 (.06)
-.06 (.09)

Notes: Standard deviations across MCMC iterations appear in parentheses. Boldface indicates that 0 does not fall within the 95% highest posterior
density (HPD) interval.

context of binge watching a particular series. We do, however,
find evidence of a nonlinear effect of depth on continuing the
view the same program and on how quickly users return

to begin new sessions, suggesting that the depth of viewing
contributes to viewing more of the same content, up to a point.
If users were addicted to Hulu.com itself, we might also
expect to find increased breadth associated with longer
sessions and shorter durations between sessions. In contrast
to the effect of depth on the time between sessions, increased
breadth is associated with users being slower to return to
begin a new viewing session (w3 = -.01, w4 = .01). We do find
that breadth exhibits a positive effect on the decision to
continue the viewing session (q3 = .06, q4 = .01). This
suggests a degree of stickiness with the video platform itself,
but the effect is considerably smaller than the impact of
depth on continuing the viewing session. We also find that
increased breadth adversely affects the decision to continue
viewing the same series, up to a point (y3 = -6.46, y4 = 3.07).
Our findings appear to suggest that users exhibit stronger
addictive behavior toward particular series and not necessarily toward the platform itself. This further differentiates
between the behavior of heavy users who consume multiple
programs and heavy users who consume multiple episodes
of a single program, the latter of which would be consistent
with users who engage in binge watching. We illustrate the
effects of breadth and depth on viewing behavior in Figures
6–8, respectively.
More interesting are the effects of advertising on viewing
behavior. We find that advertising exposures and viewers’
responses to advertising adversely affect the decisions to

12 / Journal of Marketing, September 2016

continue the viewing session (q6 = -.06, q7 = -.04) and to

view the same series (y6 = -.12, y7 = -.11). In other words,
advertisements appear to discourage binge watching
behaviors. This is consistent with the research on binge and
addiction behavior that has shown addiction is often an
escape from reality (Heatherton and Baumeister 1991). In
our context, advertisements may serve as a reminder of the
viewer’s reality and draw them out of the alternate reality
created by the programming content.
Effects of Viewing Behavior on Advertising
Responsiveness
Finally, we turn to the impacts of viewing breadth and depth
on advertising responsiveness. We see that depth (f1 = .00,
f2 = -.05) and breadth (f3 = .07, f4 = -.20) both exhibit
nonlinear effects on advertising responsiveness that ultimately decline with viewing more episodes. We illustrate the
effects of breadth and depth on advertising responsiveness in
Figure 9.
As Figure 9 illustrates, advertising responsiveness peaks
when users have previously viewed three episodes of a single series and two different series in a viewing session. The
inverted U shape observed in both panels of Figure 9 may
stem from shifts in users’ viewing states. Whereas users
may initially become more responsive to advertising as they
become more engaged with the video content they consume
(e.g., Calder, Malthouse, and Schaedel 2009), this state may
give way to a flow state in which users tune out advertising
messages that are seen as disruptions (e.g., Hoffman and
Novak 1996).


TABLE 6
Effects of Genre on Viewing Behavior and Advertising Response

Dependent Variable
Genre (Baseline: Action/Adventure)

CONTINUE

Animation and cartoons
Comedy
Drama
Family
Food and leisure
Home and garden
Horror and suspense
Music
News and information
Other
Reality and game shows
Science fiction
Sports
Talk and interview
Unknown
Video games
Web

.49
.33
.08
.27
.52
.32
.05

-.07
.07
.45
.40
.11
.32
.19
.15
.19
.72

SAME

(.03)
(.02)
(.02)
(.07)
(.05)
(.07)
(.03)
(.10)
(.03)
(.16)
(.03)
(.02)
(.07)
(.03)
(.06)
(.07)
(.11)

.92
.51
2.11
2.71
.63
1.00
.18
22.05
.42
21.60
.62
.67
-.07
.45
2.84
22.92
2.64

FREQUENCY

(.06)
(.05)
(.05)
(.15)
(.08)
(.14)
(.07)
(.37)
(.06)

(.39)
(.06)
(.05)
(.13)
(.07)
(.16)
(.31)
(.15)

.01
2.06
2.04
.08
-.07
.21
-.01
-.14
2.06
-.19
-.04
-.02
-.12
2.11
2.12
2.19
.00

(.03)
(.02)
(.02)

(.06)
(.05)
(.07)
(.03)
(.08)
(.02)
(.15)
(.02)
(.02)
(.06)
(.03)
(.05)
(.06)
(.11)

CLICKTHRU
.16
-.06
.01
-.28
2.71
2.59
-.09
-.03
2.37
-.03
.15
-.09
2.57
-.14

2.90
.52
-.58

(.07)
(.04)
(.04)
(.19)
(.23)
(.28)
(.07)
(.24)
(.08)
(.57)
(.06)
(.05)
(.23)
(.10)
(.21)
(.22)
(.37)

Notes: Standard deviations across MCMC iterations appear in parentheses. Boldface indicates that 0 does not fall within the 95% HPD interval.

Although advertising response eventually declines with
respect to both viewing depth and breadth, the impact of
increased breadth is stronger than that of increased depth.
Because interacting with advertisements may provide users
with an opportunity to increase the variety in their viewing
activity, viewing episodes from multiple series may fulfill

users’ desire for variety in their viewing and therefore result in
their being less responsive to advertisements as breadth
increases (e.g., Menon and Kahn 1995). With additional
viewing, whether of different series or of additional episodes
of the same series, advertising responsiveness declines. This
creates a challenge for advertising-supported streaming video
platforms. Although longer viewing sessions enable these
platforms to show users more advertisements, which results
in additional advertising revenue under an impression-based
revenue model (e.g., cost-per-thousand pricing), users become

less responsive to advertisements aired later in their viewing
sessions.
Trait Versus State
The preceding discussion focuses on changes in a user’s
viewing behavior as a function of previously viewed content
(e.g., breadth and depth of viewing). Users may, for instance,
move in and out of binge watching states, depending on the
content they have already viewed in the session, which may
in turn affect subsequent viewing decisions and how they
react to advertisements to which they are exposed. However,
in addition to these situational factors, some users may have
a higher tendency for binge watching behavior; that is,
binge watching may be a user-level trait. To examine
binge watching and advertising response behavior as traits,
we consider the user-level random effects (g i$) for each

TABLE 7
Effects of Viewing Breadth and Depth and Advertising Exposure on Subsequent Viewing Behavior and
Advertising Response

Dependent Variable
Independent Variable
DEPTH_EPISODE
DEPTH_EPISODE2
BREADTH_EPISODE
BREADTH_EPISODE2
DEPTH_SESSION
DEPTH_SESSION2
BREADTH_SESSION
BREADTH_SESSION2
Prior viewing of program
Advertising exposures
Advertising clicks

CONTINUE
.79
.07
.06
.01

(.01)
(.01)
(.01)
(.01)

2.06 (.01)
2.06 (.00)
-.04 (.02)

SAME

3.22
2.83
26.46
3.07

FREQUENCY

(.02)
(.01)
(.04)
(.03)

.00
2.05
.07
2.20
.12
2.03
2.01
.01

.04 (.02)
2.12 (.00)
2.11 (.05)

CLICKTHRU

(.01)
(.01)
(.01)

(.01)

(.02)
(.01)
(.03)
(.04)

2.18 (.03)

.00 (.00)
.00 (.00)

Notes: Standard deviations across MCMC iterations appear in parentheses. Boldface indicates that 0 does not fall within the 95% HPD interval.

Binge Watching and Advertising / 13


FIGURE 6
Effects of Depth and Breadth on
Continuing the Viewing Session
A: Impact of Depth on Continuing Session

FIGURE 7
Effects of Depth and Breadth on
Viewing the Same Series
A: Impact of Depth on Viewing the Same Series

1

4

.8

3

.4
.2
0
–.2

1

2

3

4

5

6

7

8

9 10

–.4

Impact on logit(qits)

Impact on logit(pits)

.6

–.6

2
1
0
1

2

3

4

5

6

7

8

9

10

–1
–2
–3

–.8
–1

–4

Episodes Viewed

B: Impact of Breadth on Continuing Session

Episodes Viewed

B: Impact of Breadth on Viewing the Same Series

1

4

.8

3

.4
.2
0

–.2

1

2

3

4

5

6

7

8

9 10

–.4
–.6

2
1
0
1

2

3

4

5

6

7

8

9

10

–1
–2
–3

–.8
–1

Impact on logit(qits)

Impact on logit(pits)

.6

Series Viewed

component of our model. In Table 8, we present the posterior
means (and standard deviations) of the correlations between
the random effects.
The results indicate a significant relationship between the
different components of users’ viewing behavior. When all
else is held constant, users who are prone to have shorter
durations between viewing sessions (higher values of g i3) are
also prone to have longer viewing sessions (higher values of
g i1), as revealed by the positive pairwise correlation among
the random effects (posterior mean = .32). Users with higher
values of g i3 are also more likely to watch the same series
during their sessions (higher values of g i2), as suggested
by the positive correlation (posterior mean = .14). This set
of behaviors collectively describes binge watching. These
behaviors are also related to viewers’ tendencies to engage
with advertisements. In particular, viewers who tend to have
shorter durations between sessions are less responsive to
advertising (posterior mean correlation = -.05), as are those

14 / Journal of Marketing, September 2016

–4

Series Viewed

who have longer sessions (posterior mean correlation = -.12).
By allowing for both correlations in users’ tendencies as well
as the effects of time-varying covariates that capture previously viewed content, we can distinguish between binge
watching behavior as a behavioral state versus a user-specific

trait.
To further investigate the tendency to engage in binge
watching as a user-specific trait, we identify those users
who are more prone to engage in binge watching behavior
according to the posterior means of their user-specific random
effects. Because we define binge watching behavior as the
consumption of multiple episodes of a television program in a
short period of time, we identify users who (1) are prone to
watch additional episodes from the same series (vs. a different
series) and (2) are prone to consume more content in a period
of time. For the first criterion, we identify bingers as those
users who have a high tendency for the next episode viewed
to be from the same series (g i2). For the second, we classify


FIGURE 8
Effects of Depth and Breadth on
Intersession Time

A: Impact of Depth on Click-Through Behavior

.15

.1

.1

0

.05

0
1

2

3

4

5

6

7

8

9 10

–.05

Impact on logit(rits)

Impact on logit(λits)

A: Impact of Depth on Intersession Time

FIGURE 9
Effects of Viewing Breadth and Depth on
Advertising Responsiveness

–.1

3

4

5

6

7

8

9

10

–.2
–.3
–.4

–.6

Episodes Viewed

B: Impact of Breadth on Intersession Time

Episodes Viewed

B: Impact of Breadth on Click-Through Behavior
.1

.15

0

.05
0
1

2

3

4

5

6

7

8

9 10

–.05
–.1

Impact on logit(rits)

.1

Impact on logit(λits)

2

–.5

–.15

–.15

1
–.1

1

2

3

4

5

6

7

8

9

10

–.1
–.2
–.3
–.4
–.5
–.6

Series Viewed

Series Viewed

users as bingers if they have a high tendency to continue their
viewing sessions by viewing another episode (g i1) or if they
have a high tendency to return quickly to begin the next
viewing session (g i3).
Taking these criteria together, we classify users as bingers
if they have a posterior mean value of g i2 in the top quartile of
users and a posterior mean value of either g i1 or g i3 in the top
quartile of users. This definition results in bingers accounting
for 16.91% of users and viewing 44.28% of episodes in our
data. In Table 9, we compare the advertising responsiveness
of bingers and nonbingers. We do so for all advertisements

to which users are exposed, as well as separately for advertisements that are shown in the first two episodes of the
viewing session and advertisements shown later in the viewing session.
The results presented in Table 9 are consistent with our
interpretation of the relationship among user-specific random
effects shown in Table 8, with bingers being less likely to
respond to advertisements than nonbingers. Additionally, both
bingers and nonbingers are less responsive to advertisements to which they are exposed late in viewing sessions
compared with early in viewing sessions, consistent with

the impact of depth (f1 and f2) and breadth (f3 and f4) on
advertising responsiveness discussed in Table 7. Bingers,
who are more likely to have increased depth, are 10.6% less
likely to respond to advertising later in the session compared with early in the session; nonbingers, who may have
higher values of depth or breadth, are 20.3% less likely.8
This result is consistent with the relative impacts of depth
and breadth on advertising responsiveness illustrated in
Figure 9.
Summary of Results
Overall, our results provide important insights about viewing
behaviors and how these behaviors relate to advertising
response:
8This pattern of results holds when bingers are defined according
to median splits of the user-specific random effects, as well as when
we focus on users whose posterior means fall in the top 15% of users.
It also holds if bingers are defined as having high values of the
posterior means for g i1, g i2, and g i3, rather than for g i2 and either g i1
or g i3..

Binge Watching and Advertising / 15

TABLE 8
Correlations Among User-Level Random Effects

CONTINUE (g $1)
SAME (g $2)
FREQUENCY (g $3)
CLICKTHRU (g $4)

CONTINUE

SAME

FREQUENCY

CLICKTHRU

1
-.01 (.02)
.32 (.02)
2.12 (.02)

1
.14 (.02)
.01 (.02)

1
2.05 (.02)

1

Note: Each cell contains the posterior mean (and standard deviation in parentheses) of the correlation between a pair of user-level random effects,
across MCMC iterations. Boldface indicates that 0 does not fall within the 95% HPD interval.

1. Our viewing model provides evidence that viewing begets
more viewing. This is consistent with previous research on
binge and addiction behaviors.
2. The viewing model also shows that advertising exposures
during a viewing session discourage binge watching behavior
during the same session because users are less likely to continue the viewing session and less likely to view the same series
when they are exposed to advertising. However, advertising
does not prolong the time until users return to begin their next
viewing session.
3. Binge watching arises from both user-specific traits and
situational factors based on the content previously consumed.
Users with an inherent tendency to engage in binge watching
are less responsive to advertising, whereas those who do not
have such a tendency but engage in binge watching due to
content previously consumed are more likely to respond to
advertising.
4. The advertising response model, coupled with our post hoc
comparison of bingers and nonbingers, reveals that users
become less responsive to advertising later in viewing
sessions.

Conclusion
To the best of our knowledge, our research is the first to
investigate the binge watching phenomenon using observational data from an online video platform and tie this
behavior to advertising responsiveness. Consistent with popular reports, we characterize binge watching in terms of three
aspects of viewing behavior: continuing a viewing session,

consuming content from the same series, and returning
quickly to begin a new viewing session. Our analysis reveals
that binge watching stems from both time-invariant user traits
and situational factors related to the content previously
consumed in a viewing session. Our results show that userspecific traits for engaging in these behaviors, as inferred
from the user-specific random effects, are correlated. Beyond

user-specific tendencies, our analysis reveals that viewing
more episodes of a program increases the tendency to engage in these behaviors, consistent with the notions of flow
and viewers’ working toward a goal of completing a season
or series.
We find that users are more responsive to advertising
early in their viewing sessions. This is consistent with experimental research into the link between online engagement and advertising effectiveness (Calder, Malthouse, and
Schaedel 2009). As our comparison between bingers and
nonbingers reveals, the difference in advertising responsiveness between early in the viewing session and later in the
viewing session is more pronounced for nonbingers. This
pattern has not been documented in the context of online
viewing behavior and presents a challenge for advertisingsupported streaming video platforms.
Users who have long viewing sessions and visit a viewing
platform often can be seen as being engaged with the platform. As users view more episodes, particularly as they view
more episodes from the same series, they become more prone
to continue their viewing sessions. This provides the platform
with additional opportunities to show advertisements. Yet,
later in viewing sessions we see that users are less responsive
to advertisements. As a result, advertising-supported video
platforms must balance the increased opportunities to show
advertisements with the reduced efficacy of advertising (as
inferred from the click-through probability).
Our modeling approach provides a means of identifying
users who are prone to binge watching behavior and distinguish them from users who do not have an inherent tendency toward binge watching. While bingers make up less

than one-fifth of the users in our data set, they account for
nearly half of the episodes viewed. Advertising-supported
streaming video platforms might examine the composition of
each series’ viewer bases because series whose viewers tend

TABLE 9
Comparison of Advertising Responsiveness Between Bingers and Nonbingers

Fraction of users (n = 9,873)
Fraction of episodes (n = 355,766)
Ad response, overall
Ad response, early videos in session (first and
second videos)
Ad response, later videos in session (third and
higher videos)

16 / Journal of Marketing, September 2016

Bingers

Nonbingers

16.91%
44.28%
.78%
.83%

83.09%
55.72%
.89%

1.00%

.74%

.79%


to be bingers can be expected to have lower advertising response
rates than other series. This information could affect the rates
that advertisers would be willing to pay for advertisements
airing in such programs and, consequently, the value of such
content to the platform. Our analysis also suggests that it may
benefit advertisers to pay a premium to advertise early in users’
viewing sessions, given the decline in advertising response rates
observed later in viewing sessions. As Table 9 reveals, bingers
experience a smaller decline in advertising responsiveness than
nonbingers, so the extent of such a premium would depend on
the composition of the series’ viewer base.
Given the increased prevalence of binge watching that
has been reported, developing a richer understanding of
binge watching behavior is of interest. New technologies have
provided consumers with increased flexibility for consuming
media. Content providers have begun to facilitate and even
encourage binge watching with features such as automatically
beginning the next episode of a series and releasing all episodes of a season at once. Despite the control over the timing
of consumption that this provides users, it can give way to
consumption patterns akin to addictive behaviors that have
been linked to problems with impulse control, an inability to
consistently abstain, and a desire for high-sensation experiences (Sarramon et al. 1999). In the context of binge watching,
we may see viewers struggle with some of these same issues.

Investigating the physiological impact of repeated binge
watching activity may provide insight into any long-term
health consequences (e.g., Grøntved and Hu 2011).
Given viewers’ shifting media habits, we believe that
there are many opportunities for future research in this
space. On the methodological front, while we account for
state dependence through observed covariates based on prior
consumption in a manner akin to research into brand choice
(e.g., Erdem and Sun 2001; Keane 1997; Roy, Chintagunta,
and Haldar 1996), future research may develop sophisticated
approaches for examining binging behavior from observational data. One possibility would be to employ a hidden
Markov model (HMM; e.g., Netzer, Lattin, and Srinivasan
2008; Schwartz, Bradlow, and Fader 2014) to more flexibly
capture shifts in viewers’ consumption patterns and advertising
responsiveness.
Zhang, Bradlow, and Small (2013) propose the use of an
HMM to accommodate clumpiness in incidence data, such as
whether an individual visits a website each day. But, as we
discuss in this research, characterizing binge watching and
examining its impact on advertising require looking at not
only the frequency with which users begin viewing
sessions but also the length of sessions and the content
consumed during these sessions. A multivariate HMM (e.g.,
Kumar et al. 2011; Schweidel, Bradlow, and Fader 2010)

may be suited to this task, which could facilitate the incorporation of unobservable temporal shocks to the different
components of viewing behavior. Data from a longer observational period could also allow for the development of a
structural model to examine how the amount and timing of
advertisements jointly affects users’ online video consumption
behavior and platforms’ decisions about different pricing

models for consumers (e.g., subscription-based, advertisingsupported, or hybrid models). Such an approach could also
enable the incorporation of forward-looking consumers, who
may make viewing decisions about the content they will view
and the amount they will view in advance (e.g., “I’m going to
watch the full season of House of Cards this weekend”).
Although our research furthers understanding of binge
watching and how it affects advertising response, our data do
not shed light on the cognitive mechanisms contributing
to this behavior. Future research may employ additional
measures of advertising effectiveness to gain a more detailed
understanding of the impact of binge watching on advertising, such as ad recall and persuasion measures. Coupled with
additional research into viewers’ motivations for binge watching,
this might enable online video platforms and advertisers to
identify ways to encourage consumption (thereby allowing
more advertisements to be shown) without adversely affecting
viewers’ response to advertising.
Today’s media consumption by consumers takes place
across multiple platforms (e.g., Feit et al. 2013; Ghose,
Goldfarb, and Han 2013). Although we have investigated
binge watching behavior that occurs on computers, viewers
may also consume content through mobile devices (a
behavior that was limited in 2009 when our data were collected) and on live television. Future work may shed light on
how users’ tendencies vary according to the set of devices on
which content is consumed. For example, viewers who use
online video platforms to catch up on missed episodes of a
series they typically watch live may differ from those who
binge watch series online exclusively in terms of their
responsiveness to advertising. If viewers split their media
consumption between live television and online video platforms, research will be needed to provide guidance to
marketers about how such behavior may affect advertising

reach and frequency. The behavior might also affect networks’ decisions regarding selling advertising, as well as
determining when to release which programs on which
platforms. Much as ignoring the relationships that exist in
visitors’ website browsing behavior may result in misleading
inferences for advertisers (Danaher 2007), it is likely that
looking at viewing activity only platform by platform will no
longer be sufficient, with a cross-platform perspective being
necessary for both networks and marketers.

REFERENCES

Becker, Gary S. and Kevin M. Murphy (1988), “A Theory of Rational
Addiction,” Journal of Political Economy, 96 (4), 675–700.
Bucklin, Randolph E. and Sunil Gupta (1992), “Brand Choice,
Purchase Incidence, and Segmentation: An Integrated Modeling Approach,” Journal of Marketing Research, 29 (May),
201–15.

Calder, Bobby J., Edward C. Malthouse, and Ute Schaedel (2009),
“An Experimental Study of the Relationship Between Online
Engagement and Advertising Effectiveness,” Journal of Interactive Marketing, 23 (4), 321–31.
Chatterjee, Patrali, Donna L. Hoffman, and Thomas P.
Novak (2003), “Model the Clickstream: Implications for

Binge Watching and Advertising / 17


Web-Based Advertising Efforts,” Marketing Science, 22 (4),
520–41.
Chou, Ting-Jui and Chih-Chen Ting (2003), “The Role of Flow
Experience in Cyber-Game Addiction,” Cyberpsychology &

Behavior, 6 (6), 663–75.
Danaher, Peter J. (2007), “Modeling Page Views Across Multiple
Websites with an Application to Internet Reach and Frequency
Prediction,” Marketing Science, 26 (3), 422–37.
Deloitte (2015), “Digital Democracy Survey,” 9th ed., [available at
/>technology-media-telecommunications/us-tmt-DDS_Executive_
Summary_Report_Final_2015-04-20.pdf].
Dockner, Engelbert J. and Gustav Feichtinger (1993), “Cyclical
Consumption Patterns and Rational Addiction,” American Economic Review, 83 (1), 256–63.
Dr`eze, Xavier and Francois-Xavier Hussherr (2003), “Internet
Advertising: Is Anybody Watching?” Journal of Interactive
Marketing, 17 (4), 8–23.
Elberse, Anita and Sunil Gupta (2009), “Hulu: An Evil Plot to
Destroy the World?” HBS No. 510-005, (October). Boston:
Harvard Business School.
Erdem, T¨ulin and Baohong Sun (2001), “Testing for Choice
Dynamics in Panel Data,” Journal of Business & Economic
Statistics, 19 (2), 142–52.
Eshman, Brandon (2009), “Hulu Builds Brand Awareness in Crowded
Space,” Nielsen Online, (February 13), [available at http://www.
nielsen.com/us/en/insights/news/2009/hulu-builds-brand-awarenessin-crowded-space.html].
Fader, Peter S. and Bruce G.S. Hardie (2009), “Customer-Base
Valuation in a Contractual Setting: The Perils of Ignoring
Heterogeneity,” Marketing Science, 29 (1), 85–93.
Feit, Eleanor McDonnell, Pengyuan Wang, Eric T. Bradlow, and
Peter S. Fader (2013), “Fusing Aggregate and Disaggregate Data
with an Application to Multiplatform Media Consumption,”
Journal of Marketing Research, 50 (June), 348–64.
Ghani, Jawaid A. and Satish P. Deshpand´e (1994), “Task Characteristics and the Experience of Optimal Flow in HumanComputer Interaction,” Journal of Psychology, 128 (4), 381–91.
Ghose, Anindya, Avi Goldfarb, and Sang Pil Han (2013), “How Is

the Mobile Internet Different? Search Costs and Local Activities,” Information Systems Research, 24 (3), 613–31.
Gold, Mark S., Kimberly Frost-Pineda, and William S. Jacobs (2003),
“Overeating, Binge Eating, and Eating Disorders as Addictions,”
Psychiatric Annals, 33 (2), 117–22.
Gordon, Brett R. and Baohong Sun (2015), “A Dynamic Model of
Rational Addiction: Evaluating Cigarette Taxes,” Marketing Science, 34 (3), 452–70.
Grøntved, Anders and Frank B. Hu (2011), “Television Viewing and
Risk of Type 2 Diabetes, Cardiovascular Disease, and All-Cause
Mortality: A Meta-Analysis,” Journal of the American Medical
Association, 305 (23), 2448–55.
Hansell, Saul (2009), “Why Hulu Succeeded as Other Video Sites
Failed,” New York Times, (July 8), [available at gs.
nytimes.com/2009/07/08/why-hulu-succeeded-as-other-videosites-failed/].
Headen, Robert S., Jay E. Klompmaker, and Roland T. Rust (1979),
“The Duplication of Viewing Law and Television Media Schedule
Evaluation,” Journal of Marketing Research, 16 (August), 333–40.
Heatherton, Todd F. and Roy F. Baumeister (1991), “Binge Eating
as Escape from Self-Awareness,” Psychological Bulletin, 111 (1),
86–108.
Helsen, Kristiaan and David C. Schmittlein (1993), “Analyzing
Duration Times in Marketing: Evidence for the Effectiveness of
Hazard Rate Models,” Marketing Science, 12 (4), 395–414.

18 / Journal of Marketing, September 2016

Hoban, Paul R. and Randolph E. Bucklin (2015), “Effects of Internet
Display Advertising in the Purchase Funnel: Model-Based Insights
from a Randomized Field Experiment,” Journal of Marketing
Research, 52 (June), 375–93.
Hoffman, Donna L. and Thomas P. Novak (1996), “Marketing in

Hypermedia Computer-Mediated Environments: Conceptual
Foundations,” Journal of Marketing, 60 (July), 50–68.
Hui, Sam K., Eric T. Bradlow, and Peter S. Fader (2009), “Testing
Behavioral Hypotheses Using an Integrated Model of Grocery
Store Shopping Path and Purchase Behavior,” Journal of Consumer Research, 36 (3), 478–93.
Hutchinson, J. Wesley, Wagner A. Kamakura, and John G. Lynch Jr.
(2000), “Unobserved Heterogeneity as an Alternative Explanation for ‘Reversal’ Effects in Behavioral Research,” Journal of
Consumer Research, 27 (3), 324–44.
Johnson, Eric J., Wendy W. Moe, Peter S. Fader, Steven Bellman,
and Gerald L. Lohse (2004), “On the Depth and Dynamics of
Online Search Behavior,” Management Science, 50 (3), 299–308.
Jurgensen, John (2013), “Netflix Says Binge Viewing Is
No ‘House of Cards,’” The Wall Street Journal, (December 12),
[available at />Keane, Michael P. (1997), “Modeling Heterogeneity and State
Dependence in Consumer Choice Behavior,” Journal of Business & Economic Statistics, 15 (3), 310–27.
Kilar, Jason (2009a), “Like Minds,” (April 30), [available at http://
blog.hulu.com/2009/04/30/like-minds/].
——— (2009b), “Thank You for a Very Big 2009,” (December 30),
[available at />Kubey, Robert and Mihaly Csikszentmihalyi (2002), “Television
Addiction Is No Mere Metaphor,” Scientific American, 286 (2),
74–80.
Kumar, V., S. Sriram, Anita Luo, and Pradeep K. Chintagunta (2011),
“Assessing the Effect of Marketing Investments in a Business
Marketing Context,” Marketing Science, 30 (5), 924–40.
Learmonth, Michael (2009), “Maybe Hulu Really Wants to Take
over the (TV) World,” Advertising Age, (June 8), [available at
/>Leon, Gloria R., Kathleen Carroll, Benita Chernyk, and Stephen
Finn (2007), “Binge Eating and Associated Habit Patterns with
College Student and Identified Bulimic Populations,” International Journal of Eating Disorders, 4 (1), 43–57.
Littleton, Cynthia (2014), “Linear TV Watching Down, Digital

Viewing Up in Nielsen’s Q3 Report,” Variety, (December 3),
[available at />Menon, Satya and Barbara E. Kahn (1995), “The Impact of Context
on Variety Seeking in Product Choices,” Journal of Consumer
Research, 22 (3), 285–95.
Moe, Wendy (2006), “A Field Experiment Assessing the Interruption Effect of Pop-Up Promotions,” Journal of Interactive
Marketing, 20 (1), 34–44.
Netflix (2013), “Netflix Declares Binge Watching Is the New
Normal,” PR Newswire, (December 13), [available at http://
www.prnewswire.com/news-releases/netflix-declares-bingewatching-is-the-new-normal-235713431.html].
Netzer, Oded, James M. Lattin, and Vikram Srinivasan (2008), “A
Hidden Markov Model of Customer Relationship Dynamics,”
Marketing Science, 27 (2), 185–204.
Nielsen (2013), “‘Binging’ Is the New Viewing for Over-the-Top
Streamers,” Nielsen Online, (September 18), [available at http://
www.nielsen.com/us/en/insights/news/2013/binging-is-the-newviewing-for-over-the-top-streamers.html].


Oxford Dictionaries (2014), “New Words Added to OxfordDictionaries.
com Today Include Binge-Watch, Cray, and Vape,” [available at
/>——— (2016), “Binge Watching,” (accessed June 17, 2016),
[available at />american_english/binge-watching?q=binge+watching/].
Pomerantz, Dorothy (2013), “Binge Watching Is Our Future,”
Forbes, (May 29), [available at />dorothypomerantz/2013/05/29/binge-watching-is-our-future/].
Rose, Frank (2008), “Free, Legal and Online: Why Hulu Is the New
Way to Watch TV,” Wired, (September 22), [available at http://
www.wired.com/2008/09/mf-hulu/.
——— (2009), “Hulu, a Victim of Its Own Success?” Wired, (May
12), [available at />Roy, Rishin, Pradeep K. Chintagunta, and Sudeep Haldar (1996), “A
Framework for Investigating Habits, ‘The Hand of the Past,’ and
Heterogeneity in Dynamic Brand Choice,” Marketing Science,

15 (3), 280–99.
Rust, Roland T. and Mark I. Alpert (1984), “An Audience Flow Model
of Television Viewing Choice,” Marketing Science, 3 (2), 113–24.
———, Wagner A. Kamakura, and Mark I. Alpert (1992), “Viewer
Preference Segmentation and Viewing Choice Models for
Network Television,” Journal of Advertising, 21 (1), 1–18.
Sarramon, C., H. Verdoux, L. Schmitt, and M. Bourgeois (1999),
“Addiction and Personality Traits: Sensation Seeking, Anhedonia, Impulsivity,” L’Enc´ephale, 25, 569–75.
Schwartz, Eric M., Eric T. Bradlow, and Peter S. Fader (2014),
“Model Selection Using Database Characteristics: Developing a

Classification Tree for Longitudinal Incidence Data,” Marketing
Science, 33 (2), 188–205.
———, ———, ———, and Yao Zhang (2011), “‘Children of
the HMM’: Modeling Longitudinal Customer Behavior at
Hulu.com,” working paper, [available at />abstract=1904562].
Schweidel, David A., Eric T. Bradlow, and Peter S. Fader (2010),
“Portfolio Dynamics for Customers of a Multiservice Provider,”
Management Science, 57 (3), 471–86.
———, Peter S. Fader, and Eric T. Bradlow (2008), “Understanding
Service Retention Within and Across Cohorts Using Limited
Information,” Journal of Marketing, 72 (January), 82–94.
Seetharaman, P.B. and Pradeep K. Chintagunta (2003), “The Proportional Hazard Model for Purchase Timing: A Comparison of
Alternative Specifications,” Journal of Business & Economic
Statistics, 21 (3), 368–82.
Shachar, Ron and John W. Emerson (2000), “Cast Demographics,
Unobserved Segments, and Heterogeneous Switching Costs in
a Television Viewing Choice Model,” Journal of Marketing
Research, 37 (May), 173–86.
Sismeiro, Catarina and Randolph E. Bucklin (2004), “Modeling

Purchase Behavior at an E-Commerce Web Site: A Task-Completion
Approach,” Journal of Marketing Research, 41 (August),
306–23.
Urban, Glen L., Guilherme (Gui) Liberali, Erin MacDonald, Robert
Bordley, and John R. Hauser (2014), “Morphing Banner
Advertising,” Marketing Science, 33 (1), 27–46.
Zhang, Yao, Eric T. Bradlow, and Dylan S. Small (2013), “New
Measures of Clumpiness for Incidence Data,” Journal of Applied
Statistics, 40 (11), 2533–48.

Binge Watching and Advertising / 19


Copyright of Journal of Marketing is the property of American Marketing Association and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.