RESEARCH DESIGN REPORT

Virality

Burgundy School of Business

Data Science, Research Design
Table of Contents
Abstract.................................................................................................................................................1

Interest of the study..............................................................................................................................2

Literature review...................................................................................................................................2

Research question.................................................................................................................................5

Research design.....................................................................................................................................6

Expected results....................................................................................................................................7

Appendix...............................................................................................................................................8
1 - Definitions....................................................................................................................................8

References.............................................................................................................................................9
 Research Design Report Virality

Abstract

New means of communication brings a new form of communication through time.

Today, information is shared in colossal quantity and the boom of social media brings the
question does quantity takes over quality. Now, for a piece of information to be seen over
the massive second-to-second actualisation over dozens of social media, it must be viral and
take over as many media as possible in a shorter time. This is a new form of word-of-mouth,
quick and short. Understanding this new form of diffusion has important implications for
both consumers and brands, Word-of-mouth is known to boost diffusion and sales, and
today studies try to assess what leads people to talk about certain products or brands rather
than others and why a certain piece of information becomes viral over others.
A trend today that sum-up new implications of virality are probably internet memes.
Using ancient Greek etymology mimeme “imitated thing", “to imitate”1, the spread of ideas
and cultural phenomena, the concept applies today to “internet memes”, and describes a
piece of content that becomes viral by human creativity itself, using an idea and “alternate”
it, with use of humour, catchphrase, image to spread the idea widely (Dawkins, 2013). They
are new forms of communication, evolving fast as well (see Dank memes 2), and open new
questions on whether there is a sense for specific content to become viral or not because it
is the nonsense of memes which often makes their popularity 3.
We will examine through a literature review the extent to what virality is linked with
characteristics of the content itself4, then the characteristics of the people sharing5 it and
finally characteristics of the channels6 it gets shared on and how it influences whether it gets
massively shared or less. We will propose an experiment, in the end, to address that further.
KEYWORDS: Word-of-Mouth, Viral Marketing, Social Transmission, Online Content,
Diffusion, Networks, Online Networks, Social Media, Channel, meme (see Appendix)

1
‘Meme’, in Wikipedia, 31 March 2020, https://en.wikipedia.org/w/index.php?title=Meme&oldid=948329666.
2
‘Internet Meme’, in Wikipedia, 13 April 2020, https://en.wikipedia.org/w/index.php?
title=Internet_meme&oldid=950640798.
3
Annaliese Griffin, ‘“Dank” Is the New Umami’, Quartz, accessed 14 April 2020,
https://qz.com/quartzy/1221995/dank-is-the-new-umami/.
4
‘Social Transmission, Emotion, and the Virality of Online Content - MSI Web Site »’, Marketing Science
Institute, accessed 14 April 2020, https://www.msi.orghttps://www.msi.org/reports/social-transmission-
emotion-and-the-virality-of-online-content.
5
Andrew Stephen et al., ‘The Effect of Transmitter Activity on Information Dissemination Over Online Social
Networks’, 6 September 2012.
6
Jonah A. Berger and Raghuram Iyengar, ‘How Interest Shapes Word-of-Mouth Over Different Channels’, SSRN
Electronic Journal, 2012, https://doi.org/10.2139/ssrn.2013141.

Page | 1
 Research Design Report Virality

Interest of the study

Today’s business is social7 it is important at any steps of a marketing campaign or in

taking part of an event to think about information to share and its characteristics (choice of
words, target, appearance), it shapes the way it will be perceived and shared, it can become
viral but it should not be in a bad way and create a backlash. It could be first controlled by
the characteristic of words (example of semantic fields to take care of: Awe, Anger, Anxiety,
Sadness, Surprise, Practical Utility, Interest) and then additionally by appearance, release
timing, length, competition.
But it is not to be taken as granted as other as other processes shape the outcome of
the information “shareability”. Micro-level processes at an individual’s level give rise to
macro-level diffusion through online social networks. There is a part of behaviour-based
person characteristics. Channels are also important to consider as their continuity (the
speed at which the information is shared on it) will determine whether or not a product,
topic, or brand is interesting to talk about, therefore, whether it gets mentioned in certain
channels (e.g., online) rather than others (e.g., offline).
It is also important to understand, through these findings and actual research that
this subject of information sharing and virality is in constant evolution in pair with
technologies and means of communication. It is, above all, a constant market and trend
watch that is necessary.

Literature review

Three articles have been mainly studied here:

Social Transmission, Emotion, and the Virality of Online Content by Jonah Berger, Katherine
L. Milkman, April 1, 2012;
A comparison of the effect of transmitter activity and connectivity on the diffusion of
information over online social networks by Andrew T. Stephen, Yaniv Dover, Jacob
Goldenberg, April 29, 2010;
How Interest Shapes Word-of-Mouth Over Different Channels by Jonah Berger, Raghuram
Iyengar, March 1, 2012.

7
Andrew Stephen and Olivier Toubia, ‘Deriving Value from Social Commerce Networks’, Journal of Marketing
Research 47 (9 January 2009), https://doi.org/10.2139/ssrn.1150995.

Page | 2
 Research Design Report Virality

Jonah Berger and Katherine L. Milkman assess how individuals naturally communicate and
word to mouth place in buying decisions and behaviour towards a brand.

They start from the premise that word to mouth over social media is far more implicating
for customers and cheaper for brands than a traditional media campaign. But there is an
issue over possible backlash and negative contents.

Two experiments are conducted:

- A Field Study: Using NY Times data to assess how emotions shape virality. The Times
has been chosen for his variety of articles (world news, sports, travel…).
Using an automated web crawler with coding to look for full-text article in coding
related to anxiety, anger, awe, or sadness, articles were classified over the number
of positive words in opposition to “negative” one.
Other variables were taken into account: Appearance in the physical paper,
Appearance on the homepage, Release timing, Author fame, Writing complexity,
author gender, article length, competition

- Second experiment: Rather than seeing whether the content that evokes certain
emotions is more likely to be shared (as in the field study), participants have been
exposed to film clips that explicitly share certain specific emotions. The aim was to
“activate” a certain emotion to see if it influences their choices.

Main results:
Content that evokes either positive (awe) or negative (anger or anxiety) emotions
characterized by high arousal is more viral. Content that evokes low arousal emotion
(sadness) is less viral. Field study: 96% of articles that make the most emailed list do so only
once. The more positive the content is, the more likely it is to become viral.
Laboratory study: exposing participants to high arousal content (amusement) increased
their willingness to share content with others.

People share content regarding the way they want to be seen, some want rather be
spreading “positive” content to make others feel better than to make them feel angry or
upset.
External factors can affect too, separated from the content, functioning like advertising such
as appearance, the position of the content etc.

Page | 3
 Research Design Report Virality

In Andrew T. Stephen, Yaniv Dover, Jacob Goldenberg’s paper two characteristics are
compared: a transmitter’s activity (how frequently they transmit information on their social
media) and connectivity (how well connected they are on the network).
The premise is that sharing information (links) have become significant traffic sources for
many blogs and websites. In some cases, traffic to major websites (link sharing) coming from
social media (Facebook, Twitter) exceeds traffic from Google directly.

Retransmission is essential for information to diffuse over online social networks like
Twitter, Facebook but even through “old fashioned” email forwarding.

They studied it across three methods:

- The first experiment is behavioural and addresses micro-level psychological aspects.
It tests how people perceive information from transmitters who differs in term of
activity and connectivity and how likely they retransmit the information related to
that.

- The second experiment (simulation) is an agent-based model and test the sharing
and retransmission of information over a large macro-scale still regarding activity
and connectivity. The third empirical analysis of twitter will address that further.
Main results:
Retransmission is a micro individual-level action but can lead to collective and
aggregate-level retransmission which is macro-level diffusion.
A transmitter's activity influence diffusion and its connectivity have a large positive impact
on top.
The activity of the transmitter is more important than the content he is sharing itself.

Page | 4
 Research Design Report Virality

Jonah Berger with Raghuram Iyengar differentiates different means or “channels” for
people to share information, classified in continuous and discontinuous regarding the pace
at which the information is transmitted (pauses in between each information or not).
For them, channels and their continuity are shaping a new form of word of mouth.

People communicate information when they talk, but as with many types of consumer
behaviours, they also communicate things about themselves, this shows that consumer
propensities to talk about satisfying and dissatisfying experiences depend in part on their
desire to communicate certain expertise as well.
But people not only communicate through what they talk about, but they also communicate
through how they talk. Stylistic elements of conversation, such as rate of speech, speed of
each turn, and avoidance of pauses between conversational turns, all communicate things
about the speaker.

They Analyse that over 21,000 conversations (two datasets) and a laboratory experiment.

Main results:
Conversation channel continuity norms shape what gets discussed. In discontinuous
conversation channels (e.g., online posts or text), pauses between conversational turns are
expected, so people have time to select and craft what they say. Consequently, more
interesting products should be talked about more than boring ones.
In channels where conversations are expected to occur more continuously (e.g., face-to-face
or on the phone), however, there is less time to selectively pick what one talks about.
Consequently, how interesting products are to talk about should have less of an impact on
whether they get discussed.

Research question

There is plenty of content out or “memes” that seems to have the potential to go
viral but never do, “Is virality just random, as some have argued (Cashmore 2009; also see
Salganik, Dods, and Watts 2006), or might certain characteristics predict whether content
will be highly shared?”8.

8
‘Social Transmission, Emotion, and the Virality of Online Content - MSI Web Site »’.

Page | 5
 Research Design Report Virality

Research design
A meme is a carrier of information, it replicates itself across a population and
competes with other memes for survival9. Our experiment will try to assess why is it that
makes a meme “survive” among others, to estimate his potential of going viral.

We choose here a quantitative method to collect data. We will create a “meme

pool” consisting of images to be “shared” over an assessed group in different round, order
and picture shown change randomly at each turn, a set of four images are shown at each
turn (reflecting different emotions funny, serious, nonsensical, …), saving images that have
been chosen by participants for the first three turns. Images chosen during the first three
turns systematically appears as the first image for the following three turns, with the rest of
the set remaining random.
Images that have not been chosen at all during the first six turns are eliminated
from the pool. First six period got no variable except that.
Starting turn 7 and for three turns, some participants will be randomly assigned to
the role of “influencers”, they will have the opportunity to “build” the sets that will be
shown, from the entire pool, classified by “most shared in previous turns” and “less shared”
before the turn goes as usual for the others.
Turn 10 will mark the end of the experiment, in this turn participants will be shown
most successful memes and be asked on which media they will most likely share them (ex:
Facebook, Twitter, 9gag, traditional mail, Tumblr, Reddit,…).

Instructions (beginning of the experiment).

In this experiment, you are going to review a set of images. You will have 90 seconds to
choose a picture you will share to others in a random set of four. This process will be
repeated for three periods.  

Additional instructions – at the end of period 3

Participants are given this information:

“You might see pictures you have seen before; they represent the most shared in the
previous periods.”
Additional instructions at the end of period 6 – to participants assigned as “influencers”

“You are now influencer; you will have two minutes to build a set of four images that will be
shown to others, your pool is divided in most shared pictures so far and less shared, you can
from all of them to build your set.”

9
The Science of Dank Memes, accessed 14 April 2020, https://www.youtube.com/watch?
time_continue=683&v=HV9WEqLeBuo&feature=emb_title.

Page | 6
 Research Design Report Virality

Additional instructions at the end of period 6 – to all participants

Participants are given this information:

“You might see pictures you have seen before; sets are now built by other participants.”

Instructions (end of the experiment) – Turn 10

Participants are given this information:

“Experiment is about to end, you will now be shown the most successful pictures so far,
please select from the list the media in which you will be more likely to share any of those
pictures for real”.

Expected results

Attention span is limited, people have less and less time to review information today,
our experience mimic that by the timer for each turn, therefore they are more tended to
repost of “share” content that is already massively shared, just because it is what they are
seeing at first10.
The most successful memes are the one who survives the longest while mutating the
one who performs less. We give a chance to less successful memes to go out again when
sets are being made by participants themselves, but as time is short, successful memes
might get shared again, and just flourish further.
Not all media are suitable for memes, some social media are even considered as the
“place where the meme goes to die”11, we might see that in the choices the participants
made at the last turn.

Limits of the experiment

As expressed before, communication is in constant evolution, and virality is difficult
to estimate and assess, despite some clues that can explain it, memes are now part of
everyday social media life but are even more uncertain and hard to pin down. Memes most
often have life span on a few months, even a few weeks, and regarding their number today
no more than a few days sometimes.
Even following all “theory” and “clues” that might get a content to go viral, it can fail,
as there is still a part of uncertainty and “randomness”, or “chance”.

10
L. Weng et al., ‘Competition among Memes in a World with Limited Attention’, Scientific Reports 2, no. 1 (29
March 2012): 1–9, https://doi.org/10.1038/srep00335.
11
The Science of Dank Memes.

Page | 7
 Research Design Report Virality

Appendix

1 - Definitions
Word-of-Mouth
Communication from individuals to people. Ex: telling people you know about a
product or service, usually because you think it is good and want to encourage them
to try it.

Viral Marketing
Marketing activity in which information spreads between people quickly, mainly on
the internet.

Social Transmission
Social Transmission is the transfer of information or behaviours within a group of
individuals. It can include verbal and nonverbal communication, actions, knowledge,
and beliefs. It is, in other words, the spread of knowledge and cultures through social
interactions and the media.

Online Content
Any material that is available on the Web. Online content includes text, images,
animations, music and videos.

Diffusion
the spreading of something more widely.

Networks
A group or system of interconnected people or things.

Online Networks
A website that provides a virtual community for people interested in a subject.

Social Media
Websites and applications that enable users to create and share content or to
participate in social networking.

Channel
The mean in which a piece of information is transmitted.

Meme
A meme is an idea, behaviour, or style that spreads through imitation from person to
person within a culture, often to convey a phenomenon, theme, or meaning
represented by it.

Page | 8
 Research Design Report Virality

References

Berger, Jonah A., and Raghuram Iyengar. ‘How Interest Shapes Word-of-Mouth Over Different
Channels’. SSRN Electronic Journal, 2012. https://doi.org/10.2139/ssrn.2013141.

Griffin, Annaliese. ‘“Dank” Is the New Umami’. Quartz. https://qz.com/quartzy/1221995/dank-is-the-

new-umami/.

‘Internet Meme’. In Wikipedia, 13 April 2020. https://en.wikipedia.org/w/index.php?

title=Internet_meme&oldid=950640798.

‘Meme’. In Wikipedia, 31 March 2020. https://en.wikipedia.org/w/index.php?

title=Meme&oldid=948329666.

Marketing Science Institute. ‘Social Transmission, Emotion, and the Virality of Online Content - MSI
Web Site »’. https://www.msi.orghttps://www.msi.org/reports/social-transmission-emotion-
and-the-virality-of-online-content.

Stephen, Andrew, Yaniv Dover, Lev Muchnik, and Jacob Goldenberg. ‘The Effect of Transmitter
Activity on Information Dissemination Over Online Social Networks’, 6 September 2012.

Stephen, Andrew, and Olivier Toubia. ‘Deriving Value from Social Commerce Networks’. Journal of
Marketing Research 47 (9 January 2009). https://doi.org/10.2139/ssrn.1150995.

The Science of Dank Memes. https://www.youtube.com/watch?

time_continue=683&v=HV9WEqLeBuo&feature=emb_title.

Weng, L., A. Flammini, A. Vespignani, and F. Menczer. ‘Competition among Memes in a World with
Limited Attention’. Scientific Reports 2, no. 1 (29 March 2012): 1–9.
https://doi.org/10.1038/srep00335.

Page | 9