Sunday, August 25, 2013

Unlocking the VAULT: Why things go viral...

Making a positive viral communication is one of the holy grails of marketing and branding. Tonight, my blog post is a synthesis of some of the best academic and practitioner work on what drives virality.

Let's start off with some questions.
  • Why are some pictures re-Tweeted and others don't get a single chirp?
  • How come some articles are "most emailed stories" and others collect cyber dust moments after they are posted?
  • Why are some YouTube videos passed on?
  • What drives online virality?
Numa Numa Kid: >55 million views

Springer Family Vacation: 13 views
A good starting point to answer this question is this video. It's a Ted Talk by Google-Guy Kevin Allocca whose job is to follow trends on YouTube. He has some cool stats (e.g. in 2011, more than 2 days of content were added every minute to YouTube) and a great point of view on the things that drive virality. He says there are 3 things: tastemakers, communities of participation, and unexpectedness. While Allocca is focusing on viral video to deliver his viral drivers, we're going to assume that his  framework also holds for images and articles. After all, videos, images and articles are doing the same thing: telling a story.

The starting point of any video, image, or article is the content itself, so let’s reorganize Allocca's framework into a more natural order:
  1. Unexpectedness - unexpected content
  2. Tastemakers - influencers who promote the content
  3. Communities of participation - content that people can join in the theme/message participate
It makes a lot of sense that tastemakers drive virality. A Tweet about “Call Me Maybe” from Justin and Selena got an army of teeny-boppers making Carly Rae’s video a top viewed within hours. Tosh's nod to Rebecca Black's Friday rocketed interest in her song. And, there can hardly be an argument that communities of participation enhance viral appeal too. Search "Harlem Shake" on You Tube you get more than 3,800,000 results. (Which doesn't say much for how people spend their time, if you ask me).  Participation enhances commitment to the original work and share appeal within social networks.
Harlem Shake: >3.8 million results
Harlem Shake Army: >92 million views
But now it’s time to dig a little deeper and make a challenge to Allocca’s framework.  It starts with a big, but reasonable assumption that the most watched videos have a viral component to them. I don't know share rates, but any video that has over 500,000,000 views has to have been passed along. With that assumption, check out the most popular videos on YouTube. Most of them are pop music videos with scripts out of a can. In other words, I'm arguing they do not score high on "unexpectedness".
Top 6 videos on YouTube:  Music out of a can?
Let’s look at the NY Times most emailed stories of today. Top shared articles are about how market fluctuations are just "a variant on the same old story", and how tablets can help kids learn to read. This is hardly unexpected content. It is also not unusual for the NY Times to carry these stories.

I decided to look for a better framework to answer tonight's questions, but to my surprise, there isn't a lot of meaty content out there. So I did what my academic training taught me to do - a lit review. I synthesized the best and current literature on word-of-mouth and online virality to unlock the vault of what goes viral. Not surprisingly, the starting point is the message.

Unlocking the V.A.U.L.T. of Virality

V. is for valence
The question related to valence is this: is a positive or negative piece of information more likely to be passed along?

Positive pass-along
The academic case for passing along positive news has been around since the 1950s. In 1958, Homans built the case that people like to pass along good news because it a social exchange value. This idea gets extended to this:  if I pass along some good information, my friends and acquaintances are happy- and I get associated to the news. So, if I pass along positive news to someone, it makes me look good and it bolsters my self-concept (Wojnicki and Godes 2008).

Negative pass-along
There is also a lot of evidence for a negativity bias (see: Bad is Stronger than Good by Baumeister, Bratslavski, Finkenauer, Vohs 2001). There is a prevailing attitude in the press that bad news sells and since negative info is more diagnostic and tends to be more useful for us we are more likely pass that along. Think about it: if there is a radiation leak in Japan that affects the Pacific salmon, you’ll want to share the news to your friends to keep them from harm. After all, it is better to be safe than sorry (Heath 1999).

So the answer is…
The reality is that people do pass along both positive and negative news. But both experiments and field data on the NY Times "most emailed stories" by Berger and Milman (2011) suggest there is a postivity bias. So if your intuition was that good news spreads better, you'd be right.

A is for activation
Activation relates to how emotionally charged up a person gets from the video/story/picture. There are lots of different emotions (some positive, some negative), but these emotions fire people up to varying degrees. Awe, amusement, anger and anxiety are examples of emotions where, when triggered, motivate people to share the news (Berger and Milman 2011). If the story content makes you content or sad, and  you’ll be a lot less willing to share the story. This explains Charlie Bit My Finger (high on the Awwwwwwe) and Kony (high on Anger)  got shared a lot. It also helps explain what drives people to share amusing ads like John West salmon and DAR pictures from the Chive.

Charlie: >540 million AWWWWEs
U is uniqueness
This one is pretty obvious. Would you rather pass on something everybody has seen or something creative. If information is novel or unique, it tends to be more interesting and shareworthy.

Level of Extreme
This one is more tricky. What is shared more: more extreme news or more moderate news? The case for both laid out below.

The case for extreme information
The Extremity theory posits that people pass on more extreme news. The justification is pretty simple:  extreme news (either good or bad) is more interesting (Heath 1999), more helpful for decision making (Skowronski and Carlston 1989), and that people are more likely to engage in word-of-mouth when they are extremely satisfied or dissatisfied (Anderson 1998). Beyond that, extremely good or bad emotionally charged info is often used to deepen social connection (Festinger, Riecken and Schachter 1956). Ask anyone who has experienced a trauma, and they tend to love sharing their story. Furthermore, if you are trying to raise money for your cause, or demonize your political opponent, there is incentive to exaggerate (Heath 1999).

The case for moderate information
The argument for propensity to pass along more moderate information is centered (pardon the pun) on the Centrality Theory.  This theory goes like this.   People are reluctant to pass on extreme news because the stories may be less believable (Grice 1975) and the news teller is concerned about how the receiver will thinks of him/her (Tecklock, Skitk, and Boettger 1989).

So the answer is…
There is a bias to pass over more moderate news (Heath 1999). There are exceptions to this of course. Extreme news will be passed on when it matches the emotional valence of the situation.

T is timeliness
This is pretty straightforward- timing matters. It’s no coincidence that Rebecca Black’s Friday song consistently spiked on Friday or that Santa images get passed along more at Christmas time.
Rebecca Black's Friday spikes on Fridays (from TedTalk)

So we can now marry our VAULT literature review with You Tube’s structure to have a more thorough perspective on what drives virality:
  1. VAULT
  2. Tastemakers
  3. Communities of Pariticipation
So now we have a more thorough understanding of what drives virality:  we have unlocked the vault to what makes a video, article, or image go viral.


  1. Thanks for the help on this. Much appreciated.