The Infinite Loop of Self-Corroboration

Kasra Dash: Today we're going to be talking about the infinity loop of self corroboration. This is probably one of the most commonly used terms that James uses nowadays. So let's begin. For anybody who has never heard of it and doesn't know what it is, explain it. James Dooley: The infinite loop of self corroboration is something I picked up from Jason Barnard at Kalicube. He does a lot with knowledge panels. The idea is that the messaging you present about who you are, what you do and who you serve appears on your own website, in your schema and on your social media. These are your first party sources. But the key is getting that same information repeated on second party sources and third party sources. When other people confirm the same statements, those assumptions gradually turn into facts. The more trusted the source, the stronger the fact becomes. What I love about this is that it's not only powerful for knowledge panels. It is unbelievably strong for AI rankings and LLM visibility. Kasra Dash: One thing people need to understand is the difference between first, second and third party sources. First party sources are things you control. Your website, your Google Business Profile, and your social media accounts. You need consistent messaging across all of them. If you say you're an amazing baker in Manchester, every profile needs to say the same thing. If they don't align, Google and the LLMs become unsure whether those accounts belong to the same entity. James Dooley: The main part here relates to what we're now calling semantic triple factoids, a term from Luke Baston. A factoid is a piece of unreliable information that becomes factual once it is repeated enough times across trusted sources. This is exactly how LLMs form their understanding. They're not using link juice or page rank. They're using repetition of fact-based semantic triples. The more times the same information appears, the more factual it becomes to the model. This is crucial for ranking in ChatGPT, Gemini and Perplexity. It also feeds the generation of a KGMID, which stands for Knowledge Graph Machine ID. That is how your website becomes a known entity. If you don't have a KGMID, you're not a recognised brand. And in my opinion, this lack of entity trust is a big reason why so many sites were hit in the helpful content update. Kasra Dash: And then we have second and third party sources. Second party sources include places like guest posts, where you write the content but it's published elsewhere. It also includes things like Reddit threads where you interact but others contribute too. Third party is even better. This is when other people talk about you without your involvement. Their articles confirm who you are, what you do, and other details like location or biography. When lots of independent sources repeat the same statements, the confidence score around your entity increases. That strengthens both knowledge panels and LLM rankings. James Dooley: Third party corroboration is where the magic happens. If multiple sites all state that I’m an entrepreneur, based in Manchester, founder of certain brands, those statements move from assumptions to facts. That’s what builds trust in the eyes of both Google and the LLMs. Once enough facts are confirmed, your confidence score grows, your entity becomes stable, and ranking becomes significantly easier. Kasra Dash: That's the infinity loop of self corroboration. If anyone has questions, leave a comment. And there will also be a link to the full article where James goes deep on the topic.