Virality is dead

When everything is viral, there is no virality

Welcome to the Sociology of Business. In my last analysis, Product as the acquisition channel, I explored how to turn product legibility into strategy. If you are on the Substack, join the chat. With one of the paid subscription options, join Paid Membership Chat, and with the free subscription, join The General Chat on The Sociology of Business WhatsApp group.

I can’s remember what went viral last December. I can, however, easily remember things that were viral in 2016 (hygge, Pokémon GO, the Snapchat Dog filter, carpool karaoke), 2014 (Pharrell’s big hat, ice bucket challenge), 2012 (Gangam Style), 2010 (planking), 2008 (Obama), 2006 (OK Go treadmill video).

There are two reasons for this.

First, the culture moved much slower, and social media were still relatively new. It wasn’t until 2016 that Instagram introduced algorithmic feed, which prioritized content based on affinity and engagement, rather than time and one’s immediate social graph. TikTok, based on similar algorithmic personalization, gained traction only in late 2018.

Things went viral not only because a lot of people shared them, but also because they stood out against a known, stable, relatively slow-moving pop-culture narrative. These things were cultural signals.

This signaling power is another reason for past viral moments’ memorability. Memes, content, fashion, lexicon, etc. all had cultural legibility, even when, as was the case with the Gangam Style video, they were sang entirely in Korean and contained intensely local cultural references. People shared them because they were new and different.

Comparatively, in early 2025, middle schoolers across the country began shouting “6-7” in classrooms—when a teacher turned to page 67, when lunch was six minutes away, or for no apparent reason at all. It was a joke without a punchline or setup. The phrase meant nothing, but using it made students feel like members of a bigger, cooler peer group.

For a few weeks, it spread rapidly. A kid in Ohio posted a video. The TikTok algorithm picked it up, tested it with similar users, measured engagement, and began feeding it to broader clusters of middle school accounts. Within days, it jumped from Ohio to California to Texas. Teachers started hearing it in hallways. Parents noticed their kids saying it at dinner. And that’s when it died.

The kids didn’t stop saying “6-7” because they got bored with it. They stopped because adults could now decode it. Teachers were using it. Parents were asking what it meant. The phrase wasn’t new anymore, and it couldn’t differentiate insiders from outsiders. It couldn’t signal that you were part of something special that not everyone understood. Within a week of adults catching on, the meme was buried in the graveyard alongside every other reference that had become too visible to the wrong audience.

This wasn’t fickleness. It was economics. And it reveals why virality—once the holy grail of marketing strategy—has become economically worthless.

Death of virality can be traced to social media’s algorithmic personalization, exponential growth of content (and content creators), and corresponding fragmentation of culture.

More people, more content, smaller audiences.

Small audiences are created through algorithmic personalization of the high volume of content. We see content, based on our taste profiles, regardless of who created it.

Algorithms’ role here is not to amplify culture the way mass media did—it is to sort it.

“6-7” spread by a kid posting a video. The algorithm shows it to a small test group of users with similar engagement patterns. Some laugh and share it. Some scroll past. Based on those reactions, the algorithm tags the video with metadata about what kind of content it is and which microsegments of users respond to it. Then it shows the video to more people in those microsegments. React positively, and you get tagged as someone who likes this type of humor. Scroll past, and you get tagged differently. Over time, you’re sorted into tighter and tighter clusters of users with similar preferences so that every piece of content can be optimized for conversion.

The economic incentive isn’t to create big, heterogeneous audiences anymore. It’s to create homogeneous micro-audiences where messages can be personalized and response rates maximized. Even when something like “6-7” appears to go viral—when it jumps from thousands to millions of impressions overnight—what’s actually happening is simultaneous micro-adoptions across fragmented taste communities. The algorithm isn’t broadcasting to everyone. It’s targeting hundreds of small clusters that happen to respond similarly to the same content.

Virality is mistakenly equated with popularity and cultural and economic value.

Pay to read the rest of this analysis, including why brands are still optimizing for the 1990s on the platforms designed for virality, and what kind of cultural output avoids virality trap.