Universal Keyword Research

UNIVERSAL KEYWORD RESEARCH

A platform-agnostic keyword research methodology. Keyword research isn't a Pinterest thing or a blog thing: it's a "find the actual words real people use before writing anything" practice, and the same underlying process applies whether the deliverable is a pin, a blog post, an Etsy listing, an email subject line, a YouTube title, or an AI Overview answer block. This skill is the research methodology. It does not write the final content itself; it produces a validated, prioritized keyword list that feeds into whatever you write next.

WHEN TO USE THIS SKILL

• Any explicit keyword research or keyword strategy request, on any platform
• Planning a content pillar, content calendar, or topic cluster, not just a single quick post
• Competitor gap analysis ("what are competitors covering that this site/shop isn't")
• Questions about specific keyword tools (Ahrefs, Semrush, Keysearch, Ubersuggest, Pinterest Trends, Google Keyword Planner, KeywordToPin, or others)
• AI search / answer engine optimization (AEO) keyword and query questions (query fan-out, AI Overview targeting, "what would ChatGPT or Perplexity surface for this")
• "What should I title, name, or tag this" for any product, post, or listing

CORE WORKFLOW (APPLIES TO EVERY PLATFORM)

1. Clarify the actual goal: what is this keyword research for (a content pillar, a calendar, a single product, a gap check), and which platform(s) matter
2. Generate seed keywords: from the product or topic itself, stated in the words a customer or searcher would actually use, not internal/brand language (see references/core-keyword-fundamentals.md Section 1)
3. Expand the seed list: autocomplete, related-term tools, AI-assisted brainstorming. Treat this expanded list as a starting point only, never a finished one
4. Tag search intent: informational, navigational, commercial investigation, or transactional (Section 2 of the same file)
5. Validate against real data: autocomplete, an actual keyword tool, or platform-native trend data. Never hand over a brainstormed guess with the same confidence as verified data (Section 7)
6. Score and prioritize: relevance, volume, winnability, and business potential, each rated 1 to 5 (Section 6)
7. Hand off the output: to whatever content-creation step comes next (writing the post, building the pin, drafting the listing)

ROUTING TABLE

Read references/core-keyword-fundamentals.md first for any request. It's the shared foundation every platform file builds on. Then read whichever platform file(s) the request actually needs. There's no need to read all four for every request; read what the request calls for.

A NOTE ON FACTS AND TOOLS

Every statistic and tool claim in the reference files was checked against a current, named source as of the last_updated date in this file's frontmatter, not pulled from memory or assumed current. Tool pricing specifically changes often and is flagged with a "confirm before quoting" note wherever it appears. Treat any number in this skill as directional and confirm it on the vendor's own site before it goes into a budget conversation or a client-facing document. If a tool comes up that isn't covered in these files, research it the same way before recommending it: confirm what it actually does and what it costs from a current source, rather than assuming from the name or from general category knowledge.

OUTPUT FORMAT

A finished keyword research deliverable should include, at minimum:

• The seed keyword(s) and where they came from
• The validated keyword list, grouped by parent topic/cluster where relevant
• Search intent tagged per keyword
• A priority score, or at least a clear best-to-worst rank, per keyword
• A flag on anything still brainstormed-only and not yet validated against real data

QUALITY CHECKLIST

Before treating a keyword research pass as finished:

• Every keyword traces back to a seed and a reason it was included
• Search intent is tagged for every keyword
• Brainstormed-but-unvalidated keywords are clearly flagged as such, not presented as confirmed data
• Tool and pricing claims are current and sourced, not assumed
• Priority or ranking is present, not just a flat list
• Numbers attributed to a named source (a specific study, a specific tool's own page) rather than presented as general knowledge