Semantic SEO

From strings to things

Google’s shift toward meaning began in 2012 with the Knowledge Graph, which it introduced as “things, not strings,” a move away from matching the words in a query toward understanding the real-world entities behind them and how they connect. The launch version already held more than 500 million objects and over 3.5 billion facts about them. Later systems (RankBrain, neural matching, BERT, MUM) pushed the same idea further: read intent and context, not just term overlap.

For a single page, semantic SEO means covering a topic the way an expert would. Instead of one thin page per keyword, you define the core entity, name the related entities a reader expects, and answer the follow-up questions the topic implies. A page on “cold brew coffee” that also covers grind size, steep time, ratio, and how its caffeine compares to hot coffee signals depth on the whole subject, not one phrase.

Practical building blocks:

• Topic clusters: a pillar page plus interlinked pages on each subtopic.
• Entities: name people, places, products, and concepts clearly, and reinforce them with schema.org structured data in JSON-LD.
• Intent coverage: answer the informational, comparison, and how-to angles a searcher actually has.

The payoff is topical authority. Cover a subject fully and you become eligible for dozens of related queries, not just the one phrase you started with.