What Makes Content Go Viral? What Science Says About Viral Content
- Feb 9
- 7 min read
Why do things go viral? Turns out there’s a pretty easy way to define a piece of content that has the potential to do so. It’s not random or mysterious, because there are real commonalities with posts that are shared widely across the internet. If you’re curious about what those commonalities are, then stick around and we’ll talk about what the actual research tells us. We’re all about science here, so let’s uncover the science of what makes things go viral. So I’m going to tell you about a few studies that show us exactly what viral content is made of, and what characteristics are more likely than anything else to get a lot of people to share it.
If you’re a content creator, you don’t have to brandish your pitchforks and take on the algorithms anymore, you can just listen in and incorporate these findings into your own work. And then….uh…..profit!?
Admittedly, I’m straddling a weird line here, because this is not a business or marketing channel with entrepreneurial advice. Even though I have fifteen years of experience in content, editorial, and marketing, what I do here is that I talk about the hard science and research behind things. And that’s what I’m going to do now. If you’re watching this for business-minded reasons, I’d encourage to continue, and I’d also argue that learning about the real data behind whatever it is you're working on will always get you better results than vibes and the so-called “common knowledge” of your industry.
Okay, so I came to this subject while researching my last video about viral posts and collective narcissism. If you haven’t seen that yet, I highly recommend you check it out if you’re interested in going viral, or you just want to know why human information spreads the way it does.
The thing that got me into that video was a study about the viral posts surrounding the stop the steal movement that trump launched into when he lost the 2020 presidential election. The main thing that will help us segue into today’s subject matter is the fact that those posts were obviously extremely negative. And for my entire career, I’ve been taught, and I’ve taught others the same, that content goes the most viral if it elicits strong negative emotions in the reader or viewer. This is the entry point into what makes content go viral. Everyone that I’ve worked with, everyone that I talk to who doesn’t work in the field professionally, all of us seem to have a pretty solidified opinion based on our own experience that creating very negative reactions in people is the way to make a lot of money making content.
So you can imagine my surprise when I found out that’s wrong. Yeah. Crazy.
We have this gigantic negative feedback loop of digital publishing right now, and it’s all based on the fact that outrage keeps us glued to our screens and scrolling our feeds. There are successful examples of people and organizations that have made a killing off of this, creating giant echo chambers based on in-group superiority and appealing to people whose ideologies they don’t even share, they just found it easy to piss them off repeatedly enough to make a bunch of ad dollars from their eyeballs. And this is a thing I’ve talked about before, and I have a whole video or even two about that.
The thing is, though, opinion is not data. Anecdotes are not data. Even if everyone believes something, that doesn’t make it real. I won’t go too far with this because strong negative emotions can obviously make content go viral, we see it happen every day, and it’s also backed up in the research. But it’s not the most effective way to make viral content. There’s a surprising characteristic of viral content that makes something more likely to be widespread on the internet, and far more likely than content that creates strong negative emotions. So what is it?
The short answer is that content that evokes positive emotions in people is more likely to go viral overall. This was surprising for me. The longer explanation for it is that researchers from UPenn did a comprehensive analysis of about seven thousand headlines over a three-month time period, and what they found was a hierarchy of virality.
Interestingly, the UPenn study found that positive emotional content was more likely to go viral overall. The results found that this is because strong emotional responses are what works, like we thought, just that what creates those strong emotions is not what we thought. The feeling of awe, being either positive or negative, was very powerful, and anger and frustration worked well, but positivity wins out above all.
The conclusions they drew from the data are partially that strong physiological arousal is what makes content go viral. Whereas activating emotions help content spread, deactivating emotions like anger and sadness actually hinder the spread of content and make people less inclined to share them. It didn’t matter what kind of content or how interesting, or helpful, or practical it was, the results held up across all different varieties of topics and fields.
But that’s not where the story ends either. Another study out of the University of Texas also surveyed a huge amount of shared articles from users. Researchers looked at 387,486 articles sent from 7 million unique users on the messaging app WeChat. They found that the most important driver of the spread of viral content were three emotional states, love, surprise, and anxiety. And this is pretty consistent with the results from UPenn. But Texas researchers even added a delineation by age and other social factors, and this is fascinating. Because older people were found to be more likely to share articles that elicited the emotional responses of anger and anxiety, and so were people without many friends. Younger people were more likely to share articles that elicited feelings of disgust. People with a lot of friends were found to share articles that showcased love, anxiety, anticipation, or disgust.
So the crazy thing here is that even emotions within emotional clusters can have totally opposite results. The lead researcher pointed out that both anger and anxiety are negative emotions, for instance, but anxiety far outspread anger when it came to analyzing the virality of the sample content. And the factors that make different people share different things are visible here, too, and the implications of that are held up in other research, too. Because love and anxiety are potentially linked to connection or the desire to support or warn other people. Other emotions like many negative ones, anger in particular, are isolating and probably shared less because they’re self-focused and not communal feelings. Interestingly enough, they found the same discrepancy with positive emotions, and joy has the same dampening effect on whether content is shared, also because researchers point out that it’s a self-focused and isolating emotion.
A final study I pulled for this analysis comes from the University of Michigan. Researchers here analyzed two giant data sets full of messages, both from really controversial and scandalous moments -- one from the time of the Brexit vote, and the other from the debate surrounding the Nord Stream 2 natural gas pipeline. The Michigan study found three things that made posts go viral during these two moments in time. The first was negativity bias, the second was causal arguments, and the third was threats to core values.
The negativity bias results make sense as we’ve seen before. I do wonder if the data sets and moments in time they picked maybe pushed the collective samples toward the negative in the first place, so I’m thinking about that. And also we know that different negative emotions have considerably different results in making content go viral.
A causal argument is a tool of language, obviously that can be used in debate at moments like the researchers analyzed. Things like, if they pipeline is built and turned on, it’ll impact the surrounding environment in a negative way, or if the pipeline is built, the economic impact will be really spiffy. The study authors pointed out the fact that causal language is simply one of the main ways that humans communicate overall, and apparently 33% of verb relations in the French language express causation.
But it's really the threats to core values that ties all this together for me. The researchers pointed out that this concept could be found across all other groupings. You can have positive values, negative values, causal arguments about values, and positive or negative causal arguments. Because with a super unofficial and high-concept survey of the data from all of these studies, it seems to me to say something about the way we connect with other and how we perceive ourselves. And in my last video about the viral negative posts from the stop the steal movement, the data told us that was driven by collective narcissism, which is a mechanism for feeling that your in-group is superior to other groups.
Once again, I am left with the conclusion that we are just cave people playing with talking robots and nuclear weapons. Physiologically, we still live in the world of hunter-gatherers and woolly mammoths. Things are scary inside our heads and scarcity is the governing concept of our collective actions.
Fortunately or well really unfortunately, for marketers, our human programming makes it easier to predict what kind of content is going to get traction. As we’ve seen today, things that go viral are more likely to be associated with positive emotions above all. Some negative emotions also have a chance to viral, some more than others, like anxiety over anger. What brings it all home is whether content expresses a values statement that others can readily identify with, and share as an aspect of declaring that identity to the world. That’s a big one.
Make sure to keep looking out for studies that further explore why information spreads on the internet, I know I definitely will. In order to figure out how to effectively communicate with one another and solve the big problems of our time, one of the best and most important things we can do is simply pin down exactly how to talk to each other. Sharing viral content might not seem like the best way to do that, especially when the data shows that only 25% of people who share articles actually read them. I talk about that in my video on headlines and the state of modern journalism. But crafting content that will actually see a wide audience is a good place to start anyway, and we have the tools now to find an audience for things we want to share.
I’ll see you next time. Don’t forget to like the video and subscribe to the channel so I can go viral. Happy viral content extravaganza to all who celebrate.
Sources:
Comments