Available via license: CC BY-NC-ND 3.0
Content may be subject to copyright.
18
/ / /
Emotional aspects of
content impact on
whether it is shared.
19Insights / Vol. 5, No. 1, 2013, pp. 18 – 23 / GfK MIR
Is virality random? ///
One of the most popular online
videos of all time is “Charlie bit me”, a short film about two
little boys. They are sitting side by side in a chair when Charlie,
the younger brother, mischievously bites down rather hard
on his older brother Harry’s finger. Harry isn’t sure whether to
laugh or cry while baby Charlie is unmistakably delighted by
his little trick. Nothing much happens in the video, and yet the
clip had received more than 400 million views on YouTube by
the end of 2011.
But why did this video go viral? And more generally, why are
certain pieces of online content more viral than others? The
emergence of social media (e.g., Facebook and Twitter) has
boosted interest in word-of-mouth and viral marketing. But
while it is clear that consumers often share online content,
and that social transmission influences product adoption and
sales, less is known about why consumers share content or
why certain content becomes viral.
Why people share content ///
In fact, there are many
reasons why people enjoy exchanging content. One reason
why people share stories, news, and information is because of
the useful information contained. Coupons or articles about
good restaurants help people save money and eat more
healthily. Consumers may forward such practically useful con-
tent to help others or to appear knowledgeable and enhance
their self-image. Others might share practically useful content
because they hope to obtain equally useful information from
their friends in return.
Emotional aspects of content may also impact upon whether it
is shared. People discuss many of their emotional experiences
with others, and there is evidence that extremes of satisfac-
tion (highly satisfied or highly dissatisfied consumers) gener-
ate more word-of-mouth than average experiences.
keywords
Viral Marketing, Word-of-Mouth,
Social Transmission, Emotions, Online Content
•
the authors
Jonah Berger,
Joseph G. Campbell Assistant Professor of Marketing,
jberger@wharton.upenn.edu
Katherine L. Milkman,
Assistant Professor of Operations and
Information Management,
kmilkman@wharton.upenn.edu
Both at the Wharton School,
University of Pennsylvania,
Philadelphia, USA
Emotion and Virality:
What Makes Online Content Go Viral?
Jonah Berger and Katherine L. Milkman
— doi 10.2478 / gfkmir-2014-0022
OPEN
20 GfK MIR / Vol. 5, No. 1, 2013, pp. 18 – 23 / Insights
Content characteristics matter
Positive beats negative /// Common sense, as well as
simply considering the kind of information that is broad-
casted on news channels, suggest that people tend to pass
along negative news more than positive news. But the results
of a study conducted on the New York Times and its list of
most shared articles indicate that positive news actually tends
to be more viral (Box 1). Affect-laden content (independent
of valence) is more likely to make the most-emailed list than
content that does not evoke emotions. In fact, content is more
likely to become viral the more positive it is. In other words,
while either more positive or more negative content tends
to be more viral than content that does not evoke emotion,
positive content is generally more viral than negative content.
High-Arousal emotions favor sharing /// When taking
into account specific emotions, it becomes evident that the
role of emotion in transmission is more complex than mere
valence alone. While awe-inspiring (positive) content is more
viral and sadness-inducing (negative) content is less viral,
some negative emotions are positively associated with viral-
ity. Anxiety and anger-inducing stories are both more likely
to make the most-emailed list after controlling for an article’s
valence and emotional intensity. This suggests that trans-
mission is about more than just sharing positive things and
not sharing negative ones. Content that evokes high-arousal
emotions (i.e., awe, anger, and anxiety) after accounting for
valence is more viral.
{ Box 1 }
VIRalITy oF
New yoRK TIMes aRTIcles
•
The objective of this study was to investigate
which types of New York Times articles were highly
shared. A webcrawler visited the Times’ homepage
(www.nytimes.com) every 15 minutes from 30August
to 30 November 2008 and recorded information about
every article (approximately 7,000 articles in total) on
the homepage as well as on each article on the most-
emailed list (updated every 15 minutes). The dataset
captured information on several potentially relevant
characteristics such as topic area, location in the news-
paper and author fame. Of all of the articles, 20 %
earned a position on the most-emailed list.
To document the emotional content of each article,
automated sentiment analysis was used to quantify
the positivity (i.e., valence) and emotionality (i.e.,
affect-ladenness) of each article. A computer program
(LIWC) counted the number of positive and negative
words in each article using a list of 7,630 words that
were classified as either positive or negative. Positiv-
ity was quantified as the difference in percentage
between the positive and negative words in an arti
-
cle. Emotionality was quantified as the percentage of
words classified as either positive or negative.
Human coders were necessary to classify the extent to
which content exhibited more specific characteristics,
in particular anger, anxiety, awe or sadness. There was
a closer focus on negative emotions because they are
easier to differentiate and classify than positive emo-
tions. The coders also distinguished whether articles
contained practically useful information, or whether
they evoked interest or surprise (control variables).
Only a random subset (approx. 2,500) of the articles
were subject to human coding. A logistic regression
model was applied to estimate the influence of the
variables.
»
Content that evokes high-arousal
emotions like awe, anger or anxiety
is more viral.
«
21Insights / Vol. 5, No. 1, 2013, pp. 18 – 23 / GfK MIR
What else favors virality? /// Articles that are more
interesting, informative (practically useful) and surprising
are more likely to make the Times’ most-emailed list. Simi-
larly, being featured for longer in more prominent positions
on the Times homepage (e.g., the lead story as opposed
to at the bottom of the page) is positively associated with
making the list. However, the relationship between the
emotional characteristics of content and virality holds even
when these content characteristics are taken into account.
Therefore, the correlation between stories that evoke certain
emotions and their higher virality is not simply down to edi-
tors featuring those types of stories, thereby mechanically
increasing their virality.
The same is true for longer articles, articles by more famous
authors and articles written by women. These are also more
likely to make the most-emailed list, but the emotional effect
can still be observed above and beyond these other effects.
The results also hold true independent of an article’s general
topic (according to 20 areas classified by the Times such as
Science or Health). This indicates that they are not merely
driven by certain areas tending to be the ones to both evoke
certain emotions and be particularly likely to make the most-
emailed list. The observed relationships between emotion
and virality hold true not only across topics but also within
them. Even among opinion or health articles, for example,
awe-inspiring articles tend to be more viral.
Figure 1 shows how different characteristics shape the
virality of an article. Virality is driven by more than just
valence. Sadness, anger, and anxiety are all negative emo-
tions, but while sadder content is less viral, content that
evokes anxiety or anger is actually more viral. Positive and
negative emotions characterized by activation or arousal
(i.e., awe, anxiety, and anger) are positively linked to virality,
while emotions characterized by deactivation (i.e., sadness)
are negatively linked.
figure 1:
Percentage change in the probability of making the most-emailed list
due to a one standard deviation increase in various article traits
Anger
Awe
Practical Value
Interest
Anxiety
Emotionality
Surprise
Positivity
Sadness
Time at Top of Homepage
34 %
25 %
– 16 %
30 %
30 %
21 %
13 %
18 %
14 %
20 %
22 GfK MIR / Vol. 5, No. 1, 2013, pp. 18 – 23 / Insights
External drivers of attention (e.g., being prominently fea-
tured) shape what becomes viral, but content characteris-
tics are of similar importance. For instance, a one standard
deviation increase in the amount of anger an article evokes
increases the odds that it will make the most-emailed list
by 34 % (Figure 1). This increase is equivalent to spending
an additional 2.9 hours as the lead story on the Times web-
site, which is nearly four times the average number of hours
articles spend in that position. Similarly, a one standard
deviation increase in awe increases the odds of making the
most e-mailed list by 30 %.
Testing the high-arousal-emotions effect in a marketing
context /// Two follow-up experiments confi rmed the
results of the New York Times fi eld study in two different
and more marketing-related contexts. The experiments used
different versions of stories to test how different amounts
of amusement (positive emotion) or anger (negative emo-
tion) infl uenced arousal and sharing. Participants said they
would be more likely to share an advertising campaign when
it induced more amusement. They also said that they would
be more likely to share a customer service experience when
it induced more anger. In both cases the higher willingness to
share was driven by the arousal it evoked.
»
Rather than targeting “special” people,
it may be more beneficial
to focus on crafting contagious content.
«
Participants said they would be more likely to share an
advertisement if they faced the high as opposed to low
amusement/anger advertisement. The results were similar
for arousal: the high amusement/anger condition evoked
more arousal than the low amusement/anger advertisement.
Marketing Implications /// When looking to generate
word-of-mouth, marketers often try targeting “infl uentials”
or opinion leaders. But while this approach is pervasive, its
value and cost-effectiveness is doubtful. Rather than tar-
geting “special” people, it may be more benefi cial to focus
on crafting contagious content. The study results illuminate
how content characteristics can improve virality.
>
Amuse rather than relax your audience ///
The
fi ndings shed light on how to design successful viral
marketing campaigns and craft contagious content.
While marketers often produce content that paints their
product in a positive light, content is more likely to be
shared if it evokes high-arousal emotions. Ads which
make consumers content or relaxed, for example, will not
be as viral as those which amuse them.
> Negative emotions do no harm when they activate ///
Further, while some marketers might shy away from ads
that evoke unpleasant feelings, negative emotions can
actually increase transmission if they are characterized
by activation. BMW, for example, created a series of short
online fi lms called “The Hire” that they hoped would go
viral. The series included car chases and story lines that
often evoked anxiety (with titles such as “Ambush” and
“Hostage”). While one might be concerned that negative
emotion would hurt the brand, the study results suggested
that it would increase transmission because anxiety induces
arousal (incidentally, “The Hire” was highly successful, gen-
erating millions of views).
> Sadness hinders transmission /// According to the
study, sadness is not a good vehicle for viral communi-
cation initiatives. While emotions that generate activation
and arousal favor sharing, no matter whether they are pos-
itive or negative, sadness actually has a negative impact
on willingness to share, no matter whether the induced
condition of sadness is low or high. Therefore, public health
information, for example, is more likely to be passed on if
it is framed to evoke anger or anxiety rather than sadness,
which is more frequently encountered.
23Insights / Vol. 5, No. 1, 2013, pp. 18 – 23 / GfK MIR
Berger, Jonah (2011),
“Arousal Increases Social Transmission of Information”,
Psychological Science, 22(7), 891 – 893.
Berger, Jonah (2013):
“Contagious: Why Things Catch On”; Simon & Schuster
Cashmore, Pete (2009),
“YouTube: Why Do We Watch?”
http://www.cnn.com/2009/TECH/12/17/cashmore.
youtube/index.html
Godes, David and Dina Mayzlin (2009),
“Firm-Created Word-of-Mouth Communication:
Evidence from a Field Test”,
Marketing Science, Vol. 28, pp. 721 – 739.
FURTHER READING
> Angry consumers’ online actions might be critical
/// Similar points apply to managing online consumer
sentiment. While some consumer-generated content (e.g.,
reviews and blog posts) is positive, much is also negative,
and can lead to consumer backlash if it is not carefully
managed. Mothers who were offended by a Motrin ad
campaign, for example, banded together and began post-
ing negative YouTube videos and tweets. While it is impos-
sible to address all negative sentiment, certain types of
negativity may be more important to address because
they are more likely to be shared. Customer experiences
that evoke anxiety or anger, for example, are more likely
to be shared than those that evoke sadness (and textual
analysis can be used to distinguish different types of
posts). Consequently, it may be more important to rectify
experiences that make consumers anxious rather than
disappointed.
—
Managerial Summary of an article published in
the top academic journal “Journal of Marketing
Research”:
Berger, Jonah; Milkman, Katherine L. (2012):
“What Makes Online Content Viral?”, Journal of
Marketing Research, Vol. 49, No. 2 (April), pp. 192 – 205.