Keyword research

Keyword research is the discovery of the demand you can plausibly capture. The naive version is a volume list; the elite version is an opportunity-scored map. Volume alone is a vanity metric: a 50,000-volume head term you cannot rank for is worth less than a 300-volume term with clear intent, low competition, and a buyer at the end of it.

Good research starts from the customer, not the tool. The highest-ROI source is customer-question mining: scraping Reddit and Quora, reading support tickets, and pulling the top queries from Search Console to find the exact language customers use. That surfaces long-tail, high-intent questions that keyword tools, which lean on autosuggest and aggregate volume, often miss entirely.

From there you classify each candidate by intent, estimate difficulty against the actual competitors ranking for it, and score opportunity as a function of volume, intent value, difficulty, and the SERP feature mix (a query with an AI Overview and a featured snippet behaves differently than ten plain blue links). For AI search there is a new layer: the conversational, fuller-sentence queries people type into ChatGPT and Perplexity, which favor question-shaped, passage-structured content. We ground difficulty in the real SERP, not a black-box score.