Keyword research is where most SEO programs quietly leak time. A mid-sized site exports 4,000 queries from Search Console, and someone spends two days dragging rows around a spreadsheet, guessing which queries belong on the same page. The output is subjective, unrepeatable, and stale the moment traffic shifts. The fix isn’t a better spreadsheet —…
Detect keyword cannibalization programmatically from GSC API data, score each conflict by recoverable traffic, and decide what to consolidate or differentiate.
DuckDB, BigQuery, and pandas benchmarked on the same 25M-row GSC export. Which engine wins for which SEO workload, with working code and a decision framework.
Choosing an embedding model for SEO isn’t a leaderboard question. A practitioner’s framework for OpenAI vs Voyage vs Cohere vs BGE, with a reproducible AUC benchmark, the dimension trade-off, and the cost math at scale.
Spreadsheets break past a few hundred URLs. Use OpenAI embeddings, Search Console overlap, and an n8n workflow to flag cannibalization across thousands of pages, score severity, and pre-classify each pair as merge, differentiate, or canonical.