skill-018-duplicate-keyword-detection

Skill 018: Duplicate Keyword Detection

Purpose

Find duplicate keywords across ad groups and campaigns, considering match type and Quality Score to determine which instances to consolidate. Duplicates fragment data, create internal competition, and waste budget.

Data Requirements

Data Source: Standard

Standard Data:

• data/account/campaigns/*/*/keywords.md - Keyword settings with match type and Quality Score
• data/performance/campaigns/*/*/keywords_metrics_30_days.md - Keyword performance

Reference GAQL:

SELECT
  campaign.name,
  ad_group.name,
  ad_group_criterion.keyword.text,
  ad_group_criterion.keyword.match_type,
  ad_group_criterion.quality_info.quality_score,
  metrics.impressions,
  metrics.clicks,
  metrics.cost_micros,
  metrics.conversions
FROM keyword_view
WHERE ad_group_criterion.status = 'ENABLED'
  AND segments.date DURING LAST_30_DAYS

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

Analysis Steps

1. Normalize and index: Extract keywords with text (lowercase, trimmed), match type, campaign, ad group, QS
2. Find exact duplicates: Same keyword text + same match type in different locations
3. Find match type variants: Same keyword text with different match types (exact + phrase + broad)
4. Compare performance: For each duplicate set, compare QS, CTR, CPA, conversions
5. Calculate waste: Sum cost across duplicate instances; identify budget fragmentation

Thresholds

Condition	Severity
Same keyword+match in 3+ locations	Critical
Duplicate across campaigns (internal competition)	Critical
Exact duplicate with 3+ QS difference	Warning
Low-QS duplicate has more spend	Warning
All three match types for same term	Info

Output

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

Short:

## Duplicate Keyword Audit
**Account:** [Name] | **Duplicates:** [X] sets | **Instances:** [Y]

### Critical ([Count])
- **[keyword]** ([match]): [X] instances, QS range [Y]-[Z] → Keep highest QS, pause others
- **Cross-campaign duplicate**: "[keyword]" in [Campaign A] and [Campaign B] → Consolidate

### Warnings ([Count])
- **[keyword]**: Low-QS instance getting [X]% of spend → Redirect to high-QS instance

### Recommendations
1. Consolidate [X] duplicate sets
2. Estimated efficiency gain: +[Y] avg QS, -$[Z] fragmented spend

Detailed adds:

• What Was Checked (normalization, match type comparison)
• Exact duplicates table (Keyword, Match, Instance 1, QS, Instance 2, QS, Cost Split)
• Match type variants table (Keyword, Exact QS, Phrase QS, Broad QS, Locations)
• Consolidation recommendations (Keyword, Current State, Action, Expected Impact)