Breakout Scale

On the internet, some groups deliberately try to influence public opinion using fabricated accounts and false stories. Such campaigns run covertly, often on behalf of governments or political groups. The crucial question is not how many accounts are involved. What matters is whether the message actually reached anyone at all. That is exactly what the Breakout Scale measures: a scale with six levels describing how far such a campaign has broken out of its own small bubble. Researcher Ben Nimmo proposed it in 2020, and it has since become a standard tool in disinformation reports.

Why reach matters more than account count

When a company like Meta or OpenAI reports that it has removed 500 fake accounts, that sounds dramatic. But the number says almost nothing about how dangerous the campaign actually was. 500 accounts that only like each other’s posts cause no harm. A single post, on the other hand, picked up uncritically by a major newspaper, can reach millions of people.

The Breakout Scale therefore separates two things that are often confused: the effort attackers put in and their actual success. An elaborate network can remain completely ineffective. A poorly made fake can go viral by sheer chance. Only the scale makes this difference visible and comparable.

For platforms and authorities, this is a matter of priorities. Not every campaign can be pursued with the same amount of effort. The scale helps decide which cases are urgent and which can be monitored but set aside. It is thus less a measuring instrument than a decision-making aid.

The six levels from whisper to harm

At level one, a message stays within a single community online, such as a forum, and is not carried further. At level two, it jumps to multiple platforms but remains confined within the campaign itself. Only at level three do real people pick up the content and voluntarily spread it further, without knowing where it originated. This is the actual breakout that gives the scale its name.

Level four means that traditional media pick up the claim, or that it circulates simultaneously across many different communities. At level five, well-known figures adopt the message: politicians, celebrities, large accounts. Level six, finally, is reached when the campaign leads to real-world actions, such as violence or a political decision.

The scale is useful because its levels can be verified. One can show that a TV station picked up a report. One can hardly prove, on the other hand, that someone changed their mind. The Breakout Scale therefore deliberately measures spread rather than impact on people’s minds. This is its greatest strength and, at the same time, its most important limitation.

Where the scale appears in reports

The Breakout Scale is most commonly encountered in the transparency reports of major tech companies. OpenAI has used it since 2024 to assess cases in which someone misused language models for propaganda. Notably, most of these campaigns remained at level one or two. AI made writing text cheaper, but hardly gave the messages any additional reach.

Research institutes and fact-checkers also draw on the scale when analyzing campaigns from Russia, China, or Iran. In news articles, it usually appears as a subordinate clause, such as: The campaign achieved no significant reach. Behind that statement often lies an assessment based on this scale.

A common misconception is that the Breakout Scale measures how dangerous a piece of false information is in terms of content. It does not. A harmless advertising message can reach level five, while a dangerous conspiracy narrative can fizzle out at level one. The scale describes only the path through the public sphere, not the content.