Researcher says viral metrics can mislead how we read online trends

In short: A NYU researcher says we often overestimate what is “viral” because social media is fragmented and views can be manufactured.

What's going on

Ruby J. Thelot, a cyber-ethnographer and media theory researcher at New York University, told WIRED that “virality” is no longer a reliable way to tell if something will last as a real trend.

He says one reason is that the internet is now split into many separate “digital islands.” In other words, people see very different feeds, even on the same app. Something can look huge inside one corner of the internet, but barely exist outside it.

Thelot also points to Goodhart’s law, which he explains as a problem that happens when a measurement becomes the goal. It is like judging restaurants only by Instagram photos, then watching restaurants focus on photo spots instead of food. He argues that creators and marketers can “engineer” lots of views without the idea actually sticking in everyday culture.

He gave an example from his own research on dating negativity online. After reviewing about 1,000 Instagram and TikTok videos with over 1 million views, he says roughly 25 percent were negative. His takeaway is that culture writers may be overstating “dating burnout” based on what looks loud online.

Thelot also discussed prediction markets, where people bet on real-world outcomes. He compares them to earlier meme coin trading and says the pull is often social. Group chats, shared information, and even losing together can become a form of bonding.

What to watch

Expect more debates about whether big view counts mean real adoption, especially as platforms promote content that is designed to spread fast. For people tracking trends, tools that focus on “what is actually lasting” may matter more than raw reach, including products that help teams monitor online chatter like /viral-curation.

Source: Wired