Oshi-Katsu and Online Backlash: Why It Looks Like Everyone Is Angry When About 1 in 70,000 Posts

When your oshi is caught in an online firestorm (enjou) on social media, it feels as though everyone in the world beyond the screen is angry. Harsh criticism piles up, quote posts multiply, and the same words meet your eyes again and again. But does that scene really reflect the fanbase as a whole, or society as a whole? Our starting point this time isn't to judge a firestorm as right or wrong, but to separate “who is angry” from “how many times you saw the anger”.

A private survey in 2026 estimated the oshi-katsu population at about 19.4 million people and the market size at about ¥4.1 trillion (roughly $27 billion). Meanwhile, in Japanese research on online firestorms, the number of people who actually write in about a single incident is estimated at roughly 1 in 70,000. The survey periods and targets differ, so the two figures can't be divided directly. Still, they are an important clue for not lumping a huge fan market and the small number who post into one “everyone.”

When reading figures, we also want to check the counting period, in addition to the number of people and the number of posts. The impression of seeing the same post many times in a day doesn't match the number of people who actually wrote an opinion in a week. When evaluating a voice that looked big, you can go back and confirm what the figure was counting.

On social media, it isn't only the number of people who posted but the reactions gathered by a post that shape how widely it is shown. A quote post that rebuts angry words, and a share that warns “this is awful,” are both reactions as far as the system is concerned. Even the actions of people who don't agree with the criticism can give that post more power to stand out. So if you measure the scale of a firestorm by how much of the screen it occupies, you end up counting the same small group's voices many times.

In a survey of Twitter users overseas, about 80 percent of posts were produced by the most active 10 percent. Even for political posts, a lively minority accounted for the overwhelming share, while ordinary users posted only a few times a year. This isn't a statistic that explains Japanese fandom as it is. But it's material for understanding the bias that silent people do not appear on the screen, and only those who post look like the whole population.

When firestorms seem to have increased, it's tempting to explain it as “Japanese people's reading comprehension has fallen.” But in PISA 2022, Japan's reading literacy ranked second among OECD member countries. In the adult skills survey too, Japan's average literacy is above the OECD average. At least, the explanation that the whole country's ability to read has collapsed doesn't fit these data.

On the other hand, in a survey by the Agency for Cultural Affairs, more than 40 percent each said that on social media it is “hard to tell how the other person feels” and “hard to judge whether information is true.” Even as people who read not a single book a month increase, most of them read text online almost every day. Rather than having moved away from text, it seems closer to reality to say that we have moved into an environment where short, fragmented information must be judged in large volume and at speed.

Classic research on rumor explains that the more important the topic and the more ambiguous the information, the more easily rumors spread. Announcements such as an oshi's sudden hiatus or a “serious violation of the rules” are typical cases that matter a lot to fans yet come with little information. Waiting with a blank is painful, and we want to combine past posts and someone's guesses to build a coherent answer quickly.

What's needed here isn't to force every circumstance into the open. There are areas that must be protected: romance and private life, health, contracts. What we need is a culture that treats putting off judgment not as indifference but as sincerity: being able to say “we don't know” when little is confirmed, putting primary information and speculation in separate columns, and being able to correct afterward.

Fans know their oshi's face, voice, words and long history of activity. But in most cases, the oshi doesn't know each fan personally. This one-way intimacy is called a parasocial relationship, and the loss felt when the relationship ends has been studied as a “parasocial breakup.” The more deeply someone has supported, the more they may experience an end of activities or a scandal not as mere news but as pain close to a real heartbreak.

Feeling pain is nothing to be ashamed of. The problem is whom that pain is directed at. The more precisely we build an ideal image from limited information, such as “they must be sincere” or “they would never betray us,” the more sharply differing information cuts as betrayal. And when “my oshi” becomes “our oshi,” criticism of the oshi is received as an attack on us, and it easily moves toward counterattacks between groups.

What we can confirm this time isn't a causal link in which oshi-katsu produces firestorms. It's this: only a portion of people post in a firestorm, and the mechanism of reactions shows their voices over and over; the more important and ambiguous the information, the more easily people fill the blank with a story; and the deeper the relationship with an oshi, the greater the pain of loss or betrayal can be. The overall picture that combines these is treated as a hypothesis from the show.

Companies and content operators can't judge success by the volume of engagement alone. They need to separate whether reactions are affection, rivalry or anger, and check the voices of the quiet majority by other means such as surveys and retention rates. Fans, too, can avoid taking the timeline for a miniature of society, and have the options of closing the screen, waiting for primary information, and holding off without knowing.

When you measure again, don't follow post counts as one number. Check separately whether the same people are posting repeatedly, whether quotes and rebuttals are increasing the exposure of the original post, and whether inquiries and cancellations outside social media are really increasing. The observation period also changes the meaning: the first few hours after an incident are different from a few days later. Don't replace momentary display counts with “the evaluation of the whole fanbase”, and don't force different sources into the same conclusion. That leads to calm judgment.

When an operator issues an explanation, too, there is no need to compete only on speed. If you separate the facts confirmed at present, the matters still being checked, the range you can't explain for privacy reasons, and when you will next report, you can reduce the room for speculation to spread as fact, even if you can't fill the blank completely. What you ask of fans is easier to act on if you go beyond “please stay calm” and show concrete lines: where to check official information, who to contact with inquiries, and not spreading personal information.

The timeline of a firestorm isn't society itself. What's there is one screen where strong emotion is made visible and reflected over and over. Is the visible society the same as the real society? Just asking this again creates room to check before reacting. Among those who don't speak up are people watching how things go, people who keep supporting somewhere else, and people who think they still lack the material to judge. That silence, too, can't be assumed to be indifference. What's needed to protect love for your oshi over the long term is not to reach the right answer quickly, but to be able to hold the time of not knowing in safety.