seo

seo

Authoritative reference for search engine optimisation across traditional SERP, Generative Engine Optimisation (GEO), and Answer Engine Optimisation (AEO). Covers research, technical audits, and the full content-to-ranking pipeline.

1) Keyword research pipeline

# Seed → cluster → prioritise
def keyword_pipeline(seed: str, domain: str) -> list[dict]:
    # 1. Expand seed via SERP "People Also Ask" + autocomplete
    # 2. Cluster by semantic similarity (embeddings or manual)
    # 3. Score each cluster: volume × (1 - difficulty) × relevance
    pass

# Scoring formula
def priority_score(volume: int, difficulty: float, relevance: float) -> float:
    return volume * (1 - difficulty) * relevance

# Target: difficulty < 0.40, relevance > 0.70, volume > 500/mo

2) GEO — Generative Engine Optimisation

GEO targets AI-generated answers (ChatGPT, Perplexity, Gemini) rather than blue links.

# GEO citation requirements (per article)
- Cite 3+ primary sources (academic, government, or high-DA publications)
- Include a structured "Key Takeaways" section at H2 level
- Add FAQ schema with 5+ Q&A pairs that mirror natural language queries
- Use direct answer format: lead with the answer, then explain
- Entity density: 2-4 named entities per 200 words
- Include a data table or comparison table — AI engines frequently quote tables

# Detect citation gaps vs. top-3 SERP competitors
async def find_citation_gaps(slug: str, domain: str) -> list[str]:
    competitor_urls = await get_top3_serp(slug)
    their_citations = await extract_citations(competitor_urls)
    our_citations   = await extract_citations([f"/{slug}"])
    return [c for c in their_citations if c not in our_citations]

3) AEO — Answer Engine Optimisation

# AEO structural requirements
- H1 must contain the primary keyword verbatim
- First paragraph answers the query in ≤ 50 words (featured snippet target)
- Use question-format H2s: "What is X?", "How does X work?"
- Structured data: Article + FAQPage + BreadcrumbList schemas
- Reading level: Flesch-Kincaid grade 8-10 (accessible to AI parsers)