Episode 9: How Google Uses the Knowledge Graph in Its AE Algorithm
Episode 9: How Google Uses the Knowledge Graph in Its AE Algorithm
Google announced the Knowledge Graph in 2012, but the infrastructure behind it predates the announcement by years. Andrew Hogue’s team at Google was building structured data systems from around 2005 - what they called the fact repository - precisely because Google had recognised that returning a list of pages wasn’t the same as answering a question. To answer a question, you need to understand the entities involved, their attributes, and their relationships to each other. Pages are documents. The Knowledge Graph is a map of meaning.
Episode 9 of the #SEOisAEO series explored how that map feeds into Google’s answer engine algorithm, with Cindy Krum, Andrea Volpini, and Bill Slawski joining Jason Barnard to trace the Knowledge Graph from its origins to its operational role in 2018.
Slawski’s contribution was particularly valuable. His long-running analysis of Google patents gave the episode a level of technical grounding that most public discussions of the Knowledge Graph lack. The patents confirm what the AEO model predicts: Google’s entity-based systems don’t just store facts, they evaluate the confidence with which a fact can be asserted. A claim about an entity that appears in one source carries less weight than the same claim corroborated across multiple independent sources. That confidence scoring is the mechanism behind algorithmic credibility - and it’s why the work of building entity signals across authoritative third-party sources is not optional for brands that want to be recommended by answer engines.
Volpini brought the Knowledge Graph into the content context: how structured data markup gives publishers a direct channel to communicate entity information to Google, reducing the guesswork in Knowledge Graph construction and increasing the accuracy of the entity’s representation. Schema isn’t decoration. It’s the language a brand uses to tell an algorithm exactly what it is.
Krum’s framing connected the Knowledge Graph to mobile and voice: the same entity-based system that powers Knowledge Panels on desktop powers the answers that voice assistants return. The graph is the shared infrastructure across all of Google’s answer surfaces, which is why entity optimisation scales across platforms in a way that page optimisation never could.
The presentation deck used during Episode 9 of the #SEOisAEO series is preserved on SlideShare.
Published by: Semrush. Host: Jason Barnard. Guests: Cindy Krum, Andrea Volpini, Bill Slawski. October 30, 2018
