skill-014-search-term-mining-negatives

Skill 014: Search Term Mining for Negative Keywords

Purpose

Identify irrelevant, wasteful, or poor-performing search queries that should be added as negative keywords to prevent future spend waste. Proactive negative keyword management reclaims budget for better-performing queries.

Data Requirements

Data Source: Standard

Standard Data:

• data/performance/campaigns/*/*/search_terms_metrics_30_days.md - Search query performance
• data/account/campaigns/*/negative_keywords.md - Existing campaign negatives
• data/account/campaigns/*/*/negative_keywords.md - Existing ad group negatives
• data/account/shared_negative_lists.md - Shared negative keyword lists
• data/account/campaigns/*/*/keywords.md - Active keywords (to avoid conflicts)

Reference GAQL:

SELECT
  campaign.name,
  ad_group.name,
  search_term_view.search_term,
  search_term_view.status,
  metrics.impressions,
  metrics.clicks,
  metrics.cost_micros,
  metrics.conversions
FROM search_term_view
WHERE segments.date DURING LAST_30_DAYS
  AND metrics.cost_micros > 0

Use /google-ads:get-custom if you need different date ranges.

Analysis Steps

1. Find high-spend zero-conversion terms: Filter Conversions = 0, Cost >= 2x target CPA or $20 minimum
2. Identify low-engagement terms: CTR < 1% AND Impressions >= 100 OR Clicks >= 10 AND Conv = 0
3. Detect waste patterns: Scan for "jobs", "free", "how to", "login", "support", competitor research terms
4. Check against existing negatives: Verify candidates not already blocked
5. Recommend negative level: Account-wide patterns → shared list; campaign-specific → campaign negative

Thresholds

Condition	Severity
Cost >= 3x Target CPA AND Conv = 0	Critical
Cost >= 2x Target CPA AND Conv = 0	Warning
Clicks >= 20 AND Conv = 0 AND CTR < 1%	Warning
Contains waste pattern ("free", "jobs") AND Conv = 0	Warning

Output

Use Short format by default. Use Detailed if user requests comprehensive analysis.