Modeling ReTweet Dynamics

Earlier this year I read a paper called “Modeling Blog Dynamics” in which they propose a method of modeling the spread of links through the blogosphere using zero-crossing random walks and exploitation vs. exploration applied to a logical flowchart model:

The authors suggested that the model could be used in influence maximization algorithms which aim to identify key, influential individuals in a given social network for the purposes of viral marketing. I was intrigued by the possibilities and have been tossing around a possible flowchart model of how individuals decide to ReTweet specific Tweets since reading that paper. Here’s my first attempt:

There are three steps in the process where a marketer can increase the chances of a specific Tweet being ReTweeted. The first step indicates that a user must be following the sender of the target Tweet; the second step means that they must actually see the Tweet in question (try to imagine what percentage of your friend’s timeline you actually see). Step three is where the user must find some motivation to ReTweet it.

Maximizing the number of followers the Tweet’s original sender has is fairly straightforward, and most of my Science of ReTweets data has explored the ReTweet motivation percentage. I had not put much effort into analyzing statistics around the attention problem, but I’ve begun to.

Because there is no way to exactly measure what percentage of followers will actually read a given Tweet, the next best metric we have is click through percentages, so that is what I’ve been working with. You can expect to see more work to that end in the next few weeks.

My work has been concentrated on maximizing the contagiousness of ideas, whereas much of the aforementioned academic work focuses on the people involved in spreading ideas. So you can also expect to see me advance the concepts of “ReTweetability” I began a few months ago with the purpose of identifying influential users.

Read More

Does Social Media Accelerate the Spread of Dangerous Ideas?

Is the social web becoming a dangerous platform for contagious, destructive ideas? As social media usage grows and becomes a hive mind of collective consciousness, it enables a number of positive things to happen, but it also presents a grave danger in the form of dangerous memes.

Dan Dennet gave a great TED talk that I’ve mentioned before where he explores dangerous memes. He defines these as parasitic ideas that subordinate genetic interests, in that they can flourish and spread even when they cause harm to the people who contract them. Examples of these are “ideas to die for” like communism, capitalism, religion, fascism and contagious suicide.

Memes are ideas that act as viruses and spread from person to person. In biological infection extreme dense populations often form worst breeding grounds. Many of history’s deadliest outbreaks started in the extremely dense populations of Asia. Cholera started in Bengal and spread across India in the early eighteen hundreds. The black death is widely believed to have begun in central Asia and the third bubonic plague pandemic began in the Yunnan province of China in 1855

If memes are idea viruses, population density can be compared to technologies that bring minds closer together. Social media not only does this, but it also increases the reach available to a single infected person and the frequency of contact that other minds have with new ideas.

The old, industrial media regime had several buffering factors that hindered the spread of contagious ideas. The gatekeepers of broadcast media companies often did extensive fact checking on new stories. The speed and frequency of idea transmission under the old media was also much less than that presented by social media.

We’ve already begun to see the beginnings of dangerous meme outbreaks in social media. Many are relatively benign like the celebrity death hoaxes of stars like Tila Tequila, Britney Spears and Zach Braff. We’ve all seen examples of incorrect “facts” spreading across Twitter at lightening speed through ReTweets and stock prices have felt the pain of a rumor posted to a social site.

More sinister variations on this theme have also begun to emerge including a suspected “web-based Suicide Cults” in England and Japan, a “flash mob riot” in Philadelphia, online gang recruitment, and racist and neo-Nazi social networking. The giant Unification Church “cult” also has strong presences on Facebook and Youtube.

These phenomenon are likely only the tip of the iceberg. How long will it be before a dangerous cult, racist faction or mass-panic inducing hoax emerges that has been specifically designed for social media contagiousness?

Read More

Designing for Contagiousness

When marketers think about “going viral,” they think about creating infectious content and getting it in front of the right people. And while that’s a great place to start, contagious design can also go a long way to help. Here’s 7 things to think about when designing a blog or site for maximum viral effect.

Timeliness & Urgency

Breaking news is one of the most contagious types of content, in every form I’ve studied. From social news and voting to Twitter, if you are the first to cover some important information, you’ll almost assuredly go viral. The reasons for this include the importance of information scarcity and the reputation-boost that comes from being the first to know about something cool.

  • Color: Use hot colors to emphasize new content– think about how highlights breaking news in yellow at the top of the page.
  • Time and Date Stamps: Most blog themes include time and date of posting, but you should work to make this information as prominent as possible on fresh content. Perhaps even go so far as to remove, or downplay, this element on old content to reduce a reader’s resistance to sharing “stale” information.
  • Positioning: Be sure to place links to the newest content on a site in prominent locations, even on older content. Again, notice’s breaking news block.

Social Proof

People are much more likely to spread content if other people have also spread it. Think about long email chain letter forward headers, or ReTweets with multiple usernames included. Informational cascades are a strong example of this kind of behavior from economics and game theory. Signs of activity and positive feedback from other users can also function as a signal of quality and authority, as well as implicit call to actions; “I loved this post so much I Tweeted about it” subtly says “and you should too” to readers.

  • Comments: Once a post has a couple of comments on it, display the number of comments in a visible location.
  • Social Media Reactions: Using a plugin like Disqus allows you to display comments about a piece of content left in places other than the blog itself.
  • Subscribers: Any blog that is using Feedburner can include a small graphic that shows the number of users subscribed. If that number is high (above a few thousand), consider including it in the theme.
  • Number of ReTweets: The TweetMeme plugin makes it very easy to display the number of times a piece of content has been ReTweeted.
  • Testimonials: Readers, users or customers might have said nice things about your site (either in email or on a social site); ask them for permission, and then feature their feedback on your site.


Nobody wants to be the one who told all their friends about something, only to have it turn out to be a hoax, so communicating a sense of authority is a key task for a designer interested in stimulating viral content sharing. In his work on applied memetics, Francis Heylighen specifically mentions authority as a required criteria for an idea to spread, saying:

…hosts or vehicles that are held in high regard or considered to represent expertise in the domain, will be more easily noticed and accepted

  • Custom Design: One of the most obvious signals of a low-authority site is a commonly reused theme. Don’t use default templates; invest in designing a look and feel that is unique to your site and communicates your brand well.
  • Professionalism: Your design can convey a level of professionalism through sophisticated use of fonts (only one or two at most), whitespace, grid layouts and mature color schemes. Don’t use too many animated gifs and don’t have auto-play music. Avoid “MySpace-esque” design.


Huge blocks of shapeless text are daunting, boring and unlikely to be spread. Social media users have demonstrated a well-known preference for content that has the feeling of chunked, easy-to-scan information.

  • Chunking: Make sure your designs allow long blocks of text to be broken into short, scannable chunks. Focus on “top ten list” types of content.

Viral Calls to Action

Most successful contagious ideas contain some element of evangelism, where the meme itself contains instructions to spread it. On Twitter, including the phrase “please ReTweet” dramatically increases the likelihood that you’ll get ReTweeted. Most religions teach followers of the need to convert non-believers, and most chain letters explicitly ask recievers to send them on.

  • Social Media Buttons: The easiest viral calls to action are buttons or badges from various social media sites, including Digg, Reddit, Twitter (“Tweet This” buttons) and Facebook. Place these buttons in as many prominent locations as you can.
  • Triggers and Motivations: My viral content sharing survey uncovered a handful of motivations that can be used to persuade users to share your content. Include these triggers near your calls to action.


People don’t spread content they’ve seen or heard a million times before. For an idea to be contagious, it has to be new in some way. In fact, my research has shown that ReTweets contain more uncommon words than regular Tweets do.

  • Uniqueness: As noted in the section above about authority, your designs should stand out from the pack. Don’t make your site look like the default WordPress template, and replace the default Drupal favicon.


Once you’ve started to get traffic to a site from your viral efforts, you’ll want to convince visitors to stick around. Users of social media sites are the most prolific content sharers, so you should especially concentrate on getting new users from social sites engaged and committed.

  • Subscriptions: Feature your RSS and email subscription options everywhere you can, and as prominently as makes sense in your design.
  • Twitter: Direct as many of your readers as you can to follow you on Twitter. This includes links in your sidebar, as well as at the end of each piece of content.
  • First Time Visitors: Consider using a plugin like WWSGD to show a special message just to new site visitors that encourages them to subscribe or follow you. This way, you can really drive the point home with new users, but avoid annoying your repeat readers.
Read More

The Science of ReTweets Report

After putting together the most recent version of my “Science of ReTweets” presentation and putting it up on Slideshare, I got a lot of great feedback, including that it’s a little hard to understand without my explanations along with each slide.

So I pulled all the data together (including some I’ve never published on this blog) with the basic transcript of the talk I give for each slide into one 22 page PDF. That report has already been featured on Fast Company and if you want to get a copy of it, all you have to do is subscribe to my blog, either by RSS or email:

Once you’ve subscribed, you’ll see a link at the end of each post (including this one) that you can click to download the report.

And if you liked this, don’t forget to buy my book.

Read More

Why Jokes Go Viral

Jokes, despite their popularity and widespread sharing across the Internet, are not a new concept. From a very young age we learn the setup of a joke, and very quickly catch on to the pattern of joke telling: someone shares a joke, I find it funny, and days later retell the joke to a group of friends, family or coworkers. Maybe some time later one of them will tell it at a party and the chain goes on. Jokes can play an integral role in socializing; in fact, certain people seem like nothing more than joke perpetuation machines and would be lost without their repeatable nature.

How do they work? Why do we spread them on? And how can I engineer more infectious jokes?


Two fish in a tank.
One turns to the other and says: “Do you know how to drive this?”

The human experience is full of cognitive dissonance and disconnects; incidents where our perception and our reality clash. Evolutionarily, these episodes can be scary at best, and deadly at worst. If you’re not expecting a certain type of berry to be poisonous and you go out and forage them for dinner, perception and reality can very quickly clash in a toxic fashion. Our ancestors lived in constant fear of being the last to know some vital piece of information, and it is scarcity that makes knowledge valuable and contagious; you’re not doing your tribal duty if you don’t tell everyone which berries will kill you.

Urban legends demonstrate a similar trait called delayed orientation. The protagonist is operating under commonly held assumptions: her perception of the scratching on the roof of her boyfriend’s car on that darkened lovers’ lane tells her not to go check it out. In the morning, it turns out it was her boyfriend, hung upside down by the serial killer and she could have saved him–if only that information wasn’t so scarce.

Perception colliding with reality causes goosebumps. And we love sharing goosebumps–they’re contagious.

Jokes address that tension and resolve it in a nonthreatening manner. Remember that swine flu joke graphic with the little kid licking the pig’s nose? Catharsis. See the 1963 Tanzanian contagious laughter outbreak as another example.

A patient says: “Doctor, last night I made a Freudian slip, I was having dinner with my mother-in-law and wanted to say: “Could you please pass the butter.” But instead I said: “You silly cow, you have completely ruined my life.”

If you want to mastermind your own humor pandemic, play on existing incongruities.

Joke Cycles

A joke cycle is a collection of jokes that revolve around a single event, idea or person. They tend to start quickly, in memetic waves, and die unfunny deaths just as soon as they began. Cycle-based jokes often recycle old structures with the new topic; how many times have we heard the same joke, just different names? Cycles commonly address topics of great societal unease like swine flu, celebrity deaths, and the Challenger disaster; they’re like catharsis epidemics.

In 1993’s Healing with Humor, Dr Arthur Asa Berge says “whenever there is a popular joke cycle, there generally is some widespread kind of social and cultural anxiety, lingering below the surface, that the joke cycle helps people deal with.”

In your humor laboratory, base your jokes on those moments of cultural tension.


Before you can retell a joke, you must be able to remember it. Specific details in a joke can help create a mental image of the scene, aiding in recall. However, these minutiae are easy to lose or mutate between tellings–successful jokes don’t depend on fine-grain elements, they are enhanced by them.

The unexpectedness that makes a joke funny can also make it hard to recall, as the human mind finds it easier to store information types we’ve dealt with before. This is why the most clichéd jokes are often the easiest to remember. Contagious and funny jokes frequently make use of a new-old model, where either new content is shoehorned into an old structure, or a old content is reworked into a new structure.

Why did the chicken cross the road?

DR. SEUSS: Did the chicken cross the road? Did he cross it with a toad? Yes, The chicken crossed the road, But why it crossed, I’ve not been told!

KARL MARX: It was an historical inevitability.

BILL CLINTON: I did not cross the road with THAT chicken. What do you mean by chicken? Could you define chicken, please?

ERNEST HEMINGWAY: To die. In the rain. Alone.

State-dependent memory and sensorial triggers can also aid the recall process. Some people have a joke they tell every single time they’re drunk. And some people have jokes they tell every time someone says a certain word or phrase.

Make sure your viral joke is easy to remember.

Social Setting

I always look for a woman who has a tattoo. I see a woman with a tattoo, and I’m thinking, okay, here’s a gal who’s capable of making a decision she’ll regret in the future.

What’s the first thing you do before telling a joke? You look around you. Once you remember a joke, the social setting you find yourself decides if you will tell it. Off-color jokes are the most obvious example, but there are other, more subtle variations exist. People in “high-culture” environments often refrain from telling jokes at all, preferring witty retorts to canned material. Gender often plays a roll as in the stereotype of “ladies who don’t tell jokes and sluts who laugh at dirty jokes.”

Be aware of your target demographics’ social considerations when constructing quips.


Twitter is the perfect petri dish to test out your jokes. Throw a few ideas against the wall and see which are ReTweeted.

If you liked this post, check out the rest of the ProtoViral Posts.

Read More

Get More ReTweets With

I’ve been working on a variety of tools that use my ReTweet Mapper as a foundation, like The ReTweetability Index. But the coolest one of them all will be

I’ve had most of this code written for quite a while, but I’ve been too busy to finish it enough to release it. So I pulled out some of the basic functionality and made a little teaser version just to see what everyone thinks about it.

This version of the tool allows you to enter the URL of a page or blog post you’d like to have ReTweeted and receive suggestions of highly ReTweetable related words to add to your content and use when Tweeting said content.

The full version will show you those users who are the most ReTweetable when it comes to content related to yours. It will also allow you to schedule Tweets to be posted during the most ReTweetable times and days, as well as “seeding” your content via DM to those ReTweetable users who have given you permission to DM them. Once your Tweet goes out, the system tracks ReTweets and clicks it gets, thereby allowing you to test and refine your content further.

So go check it out and let me know what you think.

Read More

What is Viral Marketing Science?

I tend to look at social and viral marketing on a campaign level, evaluating viral marketing campaigns as a whole instead of individual components. For me, viral marketing science is all about figuring out what and how things spread, as opposed to the more general “how communities interact online,” and so the science comes in when various elements are interacting with each other and with the audience.

It is important to note that this does not mean that viral marketing is purely tactical; on the contrary, there is a great deal of strategy present in how these campaigns fit into a brand’s overall marketing mix. The science is in hitting the sweet spot between viral tactical elements and overarching marketing strategy.

The fields I draw from commonly include sociology, neurology, statistics, history, psychology (especially of the evolutionary type), economics, biology and memetics. I also use metaphors, terms and models from epidemiology as tools to help communicate about viral marketing, as these are much more commonly understood.

I see much of the information currently available about social and viral marketing as being comprised of two distinct types: conjecture-driven and data-driven. The former is the majority, a formulation of advice based on anecdotal evidence and “what seems right.” My work with multivariate testing, combined with research from The Tipping Point and Freakonomics, has shown me that the actual data often disproves the conclusions drawn purely from gut-feelings. My efforts have focused on creating content that is backed by facts, not feelings, and falls into the data-driven bucket. I call it viral marketing science.

The first thing that got me thinking about the potential power of scientific viral marketing was, surprisingly, a work of fiction: Neal Stephenson’s Snow Crash. In it, the villain creates a biolingusitic virus based on a prototypical, brain-stem related Sumerian language. He uses the virus to basically enslave a whole bunch of people in a world domination plot.

I also believe that there is plenty of room for art in viral marketing; the creativity, intuition and improvisation involved in a successful campaign often come from a deep understanding of the data involved. But the brute creative genius most people assume is the core of contagious campaigns can make the entire exercise seem like black magic and entirely unpredictable. However, using scientific methods, it is possible for mere mortals to create repeatably viral campaigns.

Read More

Alex Bogusky: What’s Wrong With ReTweets

Alex Bogusky is co-chairman of Crispin Porter + Bogusky, or CP+B as those in the know like to say, a large advertising agency best known for edge-pushing viral marketing. As such, they’re also one of the few ad agencies whose work I admire. CP+B has won a slew of awards; they were recently named Creativity’s Agency of the Year and have collected a handful of One Club Pencils in just a few years.

After a short stint on Twitter, Alex publicly quit the micro-blogging service, saying it “wasn’t for him.” Intrigued by his thoughts on ReTweeting, I asked him to do an email interview.

Dan Zarrella: I’m a huge fan of the viral work CP+B has done, especially for Burger King (most famously the Subservient Chicken and most recently the Whopper Sacrifice). You guys seem to be using new mediums and platforms to build these campaigns. Have you used Twitter like this yet, or do you have any plans to?

Alex Bogusky: That’s really kind of you to say. We just love interacting with consumers instead of talking at them. We love the back and forth. We’re playing around a little bit with the Old Navy Supermannequins. Just getting going but it should get interesting soon.

DZ: Do you think Twitter has reached a critical mass yet where big brands are well served by engaging the audience there?

AB: I don’t know the raw numbers for twitter but there are certainly some big wins to be had. But we can’t really look at any social media in isolation because they’re all bouncing off one another and influencing each other in real time.

DZ: To me ReTweeting is an incredibly open and powerful viral messaging mechanism. Do you think ReTweeting is going to be an important tactic for viral marketing in the future?

AB: It could be. It’s certainly faster at garnering more eyeballs than e-mail, IM or chat. The question will be does it amplify or simply speed up the process. Right now we have no way of knowing because it hasn’t got the adoption yet.

DZ: You and Chris Anderson (among others) have recently expressed concern with the carefree and often intent-changing way in which the tweets you post are ReTweeted. My understand is that the core problem is that if someone else ReTweets your content, but rewrites it, it may appear that you’ve said something you haven’t. Is this a big enough problem to make you weary of ReTweeting as a marketing form?

AB: This is my personal issue with retweeting and it makes me personally uncomfortable. As a marketer I don’t see it as a concern. You want people to put their stamp on marketing even if it seems negative. Consumers play rough and you have to let them play. But as a consumer I would be concerned that somebody can appear to be putting your thoughts forward but in fact they have changed them. Maybe even reversed them. It seems like it would benefit the service to somehow lock the original message if it is presented as a RT.

DZ: What do you think is the most exciting and/or important thing about Twitter for viral marketing going forward? (Is there anything exciting or important?)

AB: I don’t like to approach new media as a marketer. I prefer to approach it as a consumer and that appreciation is more likely to inspire thinking that might help our clients. Like any new media there will be significant resistance to marketing/advertising. That was a huge them with Subservient Chicken. It was actually the first hyper successful marketing foray into the web and it pissed a lot of people off. We can expect some of the same here. Twitter will also put off integrating marketing into the service for as long as possible to get the raw numbers up and the behavior fully adopted. But at some point to monetize they will probably look to advertisers. So I think we can expect the tools that marketer have at their disposal to increase in the same way it did with Google. It may all be fairly invisible to the user ala adwords but the ability to monitor and jump into real time conversations will have tremendous appeal. I’m a little skeptical of its actual value or to put it more accurately I think this opportunity will be overvalued in the beginning in much the same way banners were overvalued in the beginning and then saw that value plummet. But I’m not a futurist. Specifically because I’ve been around long enough now to see that most predictions don’t pan out. What I like about twitter is what’s happening right now and what will be happening in the next fifteen minutes. It’s extremely exciting.

Read More

How ReTweets Spread: The Epidemiology of Viral Messaging on Twitter

Now that my ReTweet mapping system is functioning, I’m able to start compiling more granular data on the actual dynamics of the spread of ReTweet streams.

First, I’ll start with some simple averages. For the first 3 numbers –depth, users and Tweets, I’m looking at entire ReTweet streams, that is the whole tree, starting with the original Tweet and all of the subsequent ReTweets. What we see here is that the large majority of ReTweet streams only contain 2 levels of depth, that is ReTweets of the initial Tweet do not themselves produce further ReTweets. They also tend to only include two participating users (the original Tweeter and the ReTweeter) and two individual Tweets. From this, we begin to understand that most RT streams are merely one user ReTweeting another, and never go any further.


Average Value
Depth 2.09
Users 2.41
Tweets 2.44

The next data point I’m looking at is the ReTweets per Follower (RTpF) ratio for the users involved in streams I’ve indexed (just under 20,000 users). The graph below shows the distribution of RTpF in the top 9000 most followed users in my database, I’ve graphed the actual distribution line in blue, with a 30-point moving average over it in black.

Here we see that while most users had an RTpF of under 1% in my dataset, some users showed much larger ratios, possibly indicating that there are a class of users who are more “ReTweetable” than others. In the future, as I have more data indexed, I plan to release a list of those users with the highest RTpF ratios.

To explore the followers to ReTweet issue a bit more, I then analyzed correlations between followers and stream depth and total Tweets. I used two follower numbers, the first is the number of followers the “root” level of the stream had, that is how many people were potentially exposed to the “seed” Tweet. The second follower number is the combined total of followers of every user who participated in each stream.

While we must remain cautious not to assume a causal relationship between these numbers, it does become clear that there is no significant correlation between either follower number and the depth of a stream. There is on the other hand a significant, though weak, positive correlation between the number of users exposed to a Tweet and the number of times it was ReTweeted. What this (and the distribution graph above) tells me is that while users who have more followers get ReTweeted more often, the number of followers plays a less-than-expected role in predicting how widely something is ReTweeted. I expect to find that the actual content of Tweets explains more of its “ReTweetability”.


Values Correlation
Seed Followers to Total Tweets .226
Seed Followers to Depth .029
Total Followers to Total Tweets .383
Total Followers to Depth .132

The last data point I looked at in this stage of research are average reproduction rates, that is how many ReTweets in turn triggered further ReTweeting. This is comparable to the biological Reproduction rate (R0) concept in that it represents the average number of additional infections a single case of infection results in.

Of those streams with 2 or more levels, only 7.57% eventually gain an additional level, yet, of those streams with 3 or more levels, nearly 11.5% grow another level. This trend continues out to the 5th level (I did index some streams with more than 5 levels, but not enough to generate any significant data). The more levels a ReTweet stream has, the more it is likely to accumulate.

What this may indicate is that social proof (or imitation more specifically) plays a role in a user’s decision to ReTweet. The more users a Twitterer sees ReTweet something, the more likely they are themselves to ReTweet it. Another factor in the decision to ReTweet that this data point (as well as the previously noted higher occurrence rate of the “please” call-to-action) may be highlighting is that when the act ReTweeting is called to a user’s attention, they may be more likely to ReTweet.

Reproduction Rates by Depth

Depth Reproduction Rate
2 7.57%
3 11.47%
4 22.31%
5 48.44%

Look for more posts on this subject, as I’ll be developing more functionality into my ReTweet tools and I’ll also start investigating content-based correlations, that is what are the factors of the content of a Tweet that make it more or less likely to be ReTweeted.

Read More

What’s in a Retweet? The Data Behind Viral Messaging on Twitter

I started collecting ReTweets a few weeks ago and have collected just over 84,000. I’m working on a system that will allow for mapping and analysis of ReTweet streams (sneak peak below), but in building that, I’ve already uncovered some interesting data.

Contrary to what I initially thought, “RT” is used more than 4 times more often than the full word “retweet”.

ReTweets occur at an average rate of around 258 per hour, and show a distinct increase during the work day and early evening.

Retweets contain the word please over 5 times more often than most tweets.

Retweets are generally longer than other tweets.

Almost 70% of ReTweets contain a link.

Tinyurl is overwhelmingly preferred as the URL shortener to use in ReTweets.

Let me know what other data points you’d like to see and I’ll see what I can do.

And here’s a very simple, very rough preview of the mapping tool:

Read More