Glossary — Google Ads, GEO and LLM concepts
Precise, field-tested definitions of the key concepts around Google Ads, Generative Engine Optimization (GEO) and LLM/agent development.
-
E-E-A-T
# - Google evaluation grid: Experience (first-hand), Expertise (technical), Authoritativeness (external authority), Trustworthiness. Applied by human Quality Raters and algorithmic signals.
-
GEO (Generative Engine Optimization)
# - Discipline of optimizing content so AI answer engines (ChatGPT, Perplexity, Gemini, Claude) cite it, rather than only ranking on Google. Relies on factual density, source traceability and chunk-friendly structure.
-
MCP (Model Context Protocol)
# - Open standard (Anthropic, adopted by Google, Microsoft, OpenAI) letting an LLM agent call tools exposed by a third-party server via JSON-RPC. Two main flavours: WebMCP (client-side, in-browser) and MCP Server (remote, SSE/HTTP).
-
Performance Max (PMax)
# - Automated Google Ads campaign type that serves across every Google inventory (Search, Display, YouTube, Gmail, Discover, Shopping) from one budget and one conversion goal. The algorithm decides bidding and creative mix without per-channel granular control.
Wikidata : Q113463424