agentleFS
kastriasanikastriasani/google-ads-skills/skills/google-ads-keyword-research/SKILL.md

google-ads-keyword-research

Skill1 starsChanged 5 months agoMITView on GitHub

Turn seed terms into a structured Google Ads keyword plan with match types, themed ad groups, intent classification, and a negative-keyword list. Use when the user wants to build a keyword plan, expand seed keywords, generate keyword variants, group keywords into ad groups, classify search intent, choose match types, mine a search-terms report for waste, or find negative keywords. Trigger phrases include keyword research, keyword plan, expand keywords, ad group structure, match type strategy, negative keywords, search terms report, and reduce wasted spend on Google Ads.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run google-ads-research first if the market and ICP are not yet understood.

    line 26
  2. Distinctive:

    Search Console: query demand for the domain to seed expansion.

    line 75

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Google Ads Keyword Research

Turn a handful of seed terms into a build-ready keyword plan: expanded variants, intent classification, match-type formatting, tightly themed ad groups, and a negative-keyword list mined from real search-terms data. The plan decides where budget goes and which clicks you refuse to pay for.

When to use

  • The user has seed terms or a product and needs a structured keyword plan and ad-group structure.
  • The user wants to classify intent, choose match types, or organize keywords into themed ad groups.
  • The user has a search-terms report and wants to find wasted spend and recommend negatives.

Run google-ads-research first if the market and ICP are not yet understood. Hand off the ad-group structure to google-ads-campaign-setup and the conquest terms to google-ads-competitor-intel.

Workflow

  1. Expand seeds into a plan: python3 scripts/keyword_expander.py --seeds "crm software, project tool" --brand Acme This generates modifier variants (buy, best, near me, pricing, free, software, ...), formats exact/phrase/broad for each keyword, classifies intent (transactional / commercial / informational / navigational), clusters keywords into themed ad groups by token overlap, and dedupes. Use --geo to replace "near me" with a real location; --json for pipelines.
  2. Review the intent mix and ad-group clusters. Intent drives match type and campaign placement: keep informational terms out of conversion campaigns so they do not drain budget. See references/intent-and-grouping.md.
  3. Apply match-type strategy. Transactional and commercial terms favor exact and phrase for control; broad match is a discovery play that requires Smart Bidding plus a negative list. See references/match-types.md.
  4. Refine the ad groups by hand. The clustering is a starting skeleton; split any group whose keywords would need different ad copy, and attach one landing page per group so ads stay relevant.
  5. Once campaigns have run, mine the search-terms report for waste: python3 scripts/negative_keyword_miner.py --input search_terms.csv --min-cost 10 It flags zero-conversion terms over a cost/click threshold (term-level negatives) and recurring n-grams that waste spend systemically (high-leverage negatives), each with a recommended match type and list level (ad group / campaign / account). --json for pipelines; --demo for sample data.
  6. Apply negatives at the right level and cadence. Universal junk (jobs, salary, free for a paid product, crack, torrent) goes on a shared account-level list; ad-group conflicts get ad-group negatives. Sanity-check every n-gram against intent before applying. See references/negative-keywords.md.

Inputs and data sources

Both scripts run standalone on pasted input and stdlib only.

  • keyword_expander.py: needs only seed terms; optional brand and geo.
  • negative_keyword_miner.py: needs a search-terms report as CSV or JSON with fields term, clicks, cost, conversions.

Optional sources improve accuracy:

  • Google Ads Keyword Planner: real search volume, competition, and CPC.
  • Google Ads search-terms report: the source of truth for negatives and for graduating good queries to exact.
  • Search Console: query demand for the domain to seed expansion.
  • Google Sheets / BigQuery: store the plan and the running negative list.

Outputs

  • A themed keyword plan: ad groups, each keyword with intent and exact/phrase/ broad formatting (markdown or JSON).
  • A prioritized negative-keyword list: term-level and n-gram negatives with recommended match type and list level, plus a wasted-spend summary.

References

  • match-types.md: exact/phrase/broad mechanics, the reach-vs-control tradeoff, match type by intent, RLSA, and common mistakes.
  • intent-and-grouping.md: the four intent buckets, why theme-tight ad groups win Quality Score, and the grouping method.
  • negative-keywords.md: negative match types, list levels, mining the search-terms report, thresholds, and review cadence.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.