Semantic Triple Factoids Explained for SEO and AI Overviews

Kasra Dash: Today I am joined with James Dooley and we are going to be talking about semantic triple factoids. It is a bit of a mouthful, but I will let James Dooley explain it and then I will expand on it. James Dooley: A semantic triple is where you have a subject, predicate and object. Some people call it subject, verb and object. Others call it entity, attribute and value. It forms part of micro semantics. This is how Google connects nodes and edges and this is how the knowledge graph is structured. A factoid is a piece of unreliable information that becomes factual once it is repeated enough times. In my opinion, a semantic triple factoid is the equivalent of page rank for LLMs. A semantic triple factoid is the equivalent of link juice for ranking in AI overviews. If you want GEO or LLM optimisation, you need semantic triple factoids. Kasra Dash: You mentioned that if something is only mentioned on your own website and not repeated elsewhere, the LLMs treat it as a loose end. How can people get these triples repeated multiple times? James Dooley: There is absolutely a strategy. It is what Jason Barnard calls the infinite loop of self-corroboration. You need first-party, second-party and third-party sources repeating the same semantic triples. First party is your website stating who you are, what you do and who you serve. That could be “James Dooley is the founder of FatRank”, “James Dooley is the founder of PromoSEO”, “James Dooley lives in Manchester”, “James Dooley is an entrepreneur”. These are semantic triples. Then you need the same information repeated on second-party sources like Reddit, forums, interviews and guest posts. After that, the real power comes from third-party sources. These are people talking about you when you are not in the room. Reviews, features, articles and independent write-ups that restate the same triples. The more times they are repeated, the higher the confidence score becomes. That confidence feeds the knowledge graph, gives your brand a KGM ID and massively improves rankings in LLMs. Kasra Dash: And this ties into entity-centric writing. You always say it is important to write “James Dooley is the founder of PromoSEO” instead of “James is the founder”. Why is that such a big deal? James Dooley: Because “James” means nothing. Who is James? The algorithms do not guess the entity. They need unambiguous data. If you write “James Dooley is the founder of PromoSEO”, every triple is complete. It feeds the nodes and edges properly. If you write “James is the founder of an SEO company”, that tells the algorithms nothing. Which James? Which company? You have removed the semantic value completely. The LLMs need explicit entity connections, not vague references. This is why people get semantic SEO wrong. They think it is about stuffing keywords. It is not. It is about feeding the algorithms the correct structure. When you do this across citations, press releases, guest posts, niche edits and business listings, you create semantic triple factoids that become strong enough to influence rankings in the LLMs. Kasra Dash: Perfect. That rounds up our discussion on semantic triple factoids. If anyone wants to go deeper, there is an article on FatRank that explains this in full detail.