Every day, countless new articles, videos and blogs are uploaded to the Internet. Most of these languish in relative obscurity. But a few go viral: they explode across the web, attracting the attention of hundreds of thousands (sometimes millions) of people in short order. Content that goes viral may remain in the public consciousness for days, weeks, or longer. So for those who want to reach a large audience—whether it’s for a corporate marketing campaign or to promote a personal cause or to show the world your cat’s amazing yodeling ability—the question of what makes content go viral online important.

In our research, two of us (Milkman and Berger) have explored what it is that makes some content spread like wildfire and whether it’s possible to deliberately achieve viral status. If so, what are the characteristics of viral content? 

Our work builds on previous research exploring questions about which urban legends catch on and why? And what drives the spread of gossip?

We examined what content on the New York Times’ homepage is most widely shared, and what scientific research summaries are most likely to be passed along and found that viral content tends to be characterized by certain, predictable qualities. While content may be shared for many reasons, overall, content that elicits an emotional reaction tends to be more widely shared. In addition, stories stimulating positive emotions are more widely shared than those eliciting negative feelings, and content that produces greater emotional arousal (making your heart race) is more likely to go viral. This means that content that makes readers or viewers feel a positive emotion like awe or wonder is more likely to take off online than content that makes people feel sad or angry, though causing some emotion is far better than inspiring none at all. Also, anger-inducing content is more likely to be shared than sadness-inducing content because it produces greater emotional arousal or activation.

Our first study together on this topic examined the virality of nearly 7,000 articles published online by the New York Times. We explored how an article’s valence (whether it was positive or negative), emotionality (affect laden-ness), and likelihood of producing specific emotions (e.g., anger, awe, anxiety, sadness) related to its likelihood of making the Times’ most e-mailed list. Then we conducted a series of laboratory experiments to validate our underlying theories of virality.

In our study of the Times, we hypothesized that a story’s valence, and the specific emotions it elicits, influence whether an article makes it onto the paper’s most e-mailed list. In testing our hypotheses, we controlled for a range of factors including where a story is featured in the printed paper or online, the time it spent in prominent positions on the home page, its release timing, author’s fame, writing complexity, author’s gender, and length. We also controlled for an article’s topic (e.g., sports, arts) and practical value, as well as how interesting and surprising readers thought it was. We collected information about all articles appearing on the Times’ home page over a three-month period. Of these articles, 20% made it to the website’s most e-mailed list.

Examples of articles that earned top scores on some of the dimensions we analyzed include a story called “Redefining Depression as Mere Sadness,” which was rated as highly emotional. An article entitled “Wide-Eyed New Arrivals Falling in Love with the City” was rated as very positive, while one called “Germany: Baby Polar Bear’s Feeder Dies” was rated as very negative. One particularly sad story was entitled “Maimed on 9/11, Trying to Be Whole Again,” while an article with the headline “Rare Treatment is Reported to Cure AIDS Patient” scored high for awe.

What our findings mean for practical purposes is that if you’re trying to create content that will make a big splash, making the message positive is likely to help, and emotionality is key. Of course, more interesting, practically useful and surprising content is also more likely to go viral.

In follow-up laboratory experiments, we asked participants how likely they would be to share a story about a recent advertising campaign or customer service experience and manipulated whether the story in question evoked more or less of a specific emotion. Participants were assigned to read high- or low-emotion inducing versions of the same stories and were then asked if they would share the story with others. Just as in the study of the Times’ most e-mailed list, we found that participants reported a higher likelihood of sharing the same basic content when it induced more activating emotions, like amusement and anger. On the flip side, content inducing higher levels of sadness—a de-activating emotion—reduced readers’ reported likelihood of sharing a story. These experimental results show that it’s not just the case that “better” underlying content is more emotional and more likely to shared—the same content can be reframed as more emotional and activating to increase its chances of going viral.

We recently conducted a second study together of what goes viral, looking at a very different sample of content. We asked more than 800 scientists (including hundreds of co-authors) to summarize a recent major finding from their work and then exposed over 7,000 laymen to one such summary each, asking them how likely they would be to share it with others. Our findings in this very different setting mirrored what we had observed before—that more positive and more emotional content is more likely to be widely shared, as is more useful and interesting content. In this research, we could even study subtle differences in the way two co-authors of the same finding described it and how those differences influenced its share-ability. We saw that describing the very same research in a way that enhances its emotionality, positivity, and usefulness increased its likelihood of being shared.

The take-away message from this work is that if you want your video, press release, news story, blog post or tweet to reach as many people as possible, there are specific things you can do to increase its chances of being widely shared. Make it emotional—ideally triggering emotions like anger, anxiety or awe that tend to make our hearts race; and if you can, make it positive. This may be more even effective than other methods that are currently in wide use like targeting “influentials,” or opinion leaders. Crafting contagious content, as this research suggests, may provide more bang for your buck and create more reliably viral content.

Liz Rees-Jones is a freelance writer and novelist. She studied journalism at Utah State University and is currently pursuing an MFA in Creative Writing at Lesley University. 

Katherine Milkman (@Katy_Milkman) is the James G. Campbell, Jr. assistant professor at The Wharton School at the University of Pennsylvania with a secondary appointment at the University’s Perelman School of Medicine. She has been recognized by Poets and Quants as one of the top 40 business school professors under 40 and voted Wharton’s “Iron Prof” by the school’s MBA students. In the last decade, she has published dozens of research articles about her work, which relies heavily on big data to gain insights about how to improve decision making.

Jonah Berger (@j1berger) is a Marketing professor at Wharton and author of the recent New York Times and Wall Street Journal bestseller Contagious: Why Things Catch On. He has spent over 15 years studying social influence, what drives word of mouth and virality, and how this leads products and ideas to catch on. For more on how to craft contagious content, see Jonah’s book.

Are you a scientist who specializes in neuroscience, cognitive science, or psychology? And have you read a recent peer-reviewed paper that you would like to write about? Please send suggestions to Mind Matters editor Gareth Cook. Gareth, a Pulitzer prize-winning journalist, is the series editor of Best American Infographics and can be reached at garethideas AT gmail.com or Twitter @garethideas.