Negative Keyword Miner

Systematic approach to extracting, categorizing, and managing negative keywords across Google Ads and Microsoft Ads campaigns. Covers search term report analysis, AI-assisted categorization, match type strategy, shared lists, cross-campaign conflict detection, and industry-specific templates.

Capabilities

1. Search Term Report Analysis - Pull and analyze search queries that triggered ads
2. AI Categorization - Automatically classify irrelevant queries by category
3. Match Type Analysis - Detect broad match bleeding and close variant issues
4. Negative List Management - Build campaign-level and shared negative keyword lists
5. Cross-Campaign Conflict Detection - Find negatives that block desired traffic
6. Industry Templates - Pre-built negative keyword lists by industry vertical

Workflows

Workflow 1: Search Term Report Mining

Step 1: Pull search term report via GAQL

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

Step 2: Calculate efficiency metrics

For each search term:

CPA = Cost / Conversions
ROAS = Conversion Value / Cost
Waste Score = Cost × (1 - Conversion Rate / Campaign Avg Conversion Rate)

Step 3: Flag candidates for negatives

Step 4: Categorize flagged terms

Group into categories for systematic addition:

Workflow 2: Broad Match Bleeding Analysis

What is broad match bleeding? Broad match keywords trigger on loosely related searches that waste budget. This has intensified with Google's expanded close variants and broad match modifier deprecation.

Step 1: Identify broad match keywords

SELECT
  ad_group_criterion.keyword.text,
  ad_group_criterion.keyword.match_type,
  campaign.name,
  ad_group.name,
  metrics.impressions,
  metrics.clicks,
  metrics.cost_micros,
  metrics.conversions
FROM keyword_view
WHERE
  ad_group_criterion.keyword.match_type = 'BROAD'
  AND segments.date DURING LAST_30_DAYS
ORDER BY metrics.cost_micros DESC

Step 2: Cross-reference with search terms

For each broad match keyword, compare triggered search terms:

Broad Keyword: "running shoes"
Search Terms Triggered:
  ✅ "buy running shoes online" — relevant
  ✅ "best running shoes 2026" — relevant
  ⚠️ "running shoe repair near me" — tangential
  ❌ "running shoes drawing" — irrelevant
  ❌ "free running shoes giveaway" — irrelevant
  ❌ "running shoe factory jobs" — irrelevant

Step 3: Calculate bleed rate

Bleed Rate = Irrelevant Search Term Impressions / Total Search Term Impressions × 100

Interpretation:
  < 10%: Healthy — normal broad match behavior
  10-25%: Moderate — add negatives for top offenders
  25-50%: Severe — consider switching to phrase match
  > 50%: Critical — switch to exact match or pause keyword

Step 4: Remediation decision matrix

Workflow 3: Negative Keyword List Management

Campaign-level vs Shared lists:

Recommended list structure:

Shared List: "Universal Negatives"
  → Apply to ALL campaigns
  → Contains: jobs, careers, salary, free, cheap, DIY, tutorial, wiki

Shared List: "Competitor Brands"
  → Apply to non-competitor campaigns
  → Contains: exact match competitor brand names

Shared List: "Geographic Exclusions"
  → Apply to geo-targeted campaigns
  → Contains: irrelevant cities, countries, regions

Campaign-Level:
  → Campaign "Running Shoes" excludes: "basketball", "hiking", "dress shoes"
  → Campaign "Running Apparel" excludes: "shoes", "sneakers"

Managing list size:

• Google Ads limit: 5,000 negative keywords per list
• Google Ads limit: 20 shared lists per account
• Campaign-level limit: 10,000 negative keywords per campaign
• Practical recommendation: Keep each list under 1,000 terms

Workflow 4: Cross-Campaign Conflict Detection

What is a negative keyword conflict? A negative keyword in Campaign A blocks a search query that Campaign B is actively bidding on. This causes lost traffic.

Step 1: Export all keywords and negatives

-- Active keywords
SELECT
  campaign.name AS campaign,
  ad_group.name AS ad_group,
  ad_group_criterion.keyword.text AS keyword,
  ad_group_criterion.keyword.match_type AS match_type,
  'POSITIVE' AS polarity
FROM keyword_view
WHERE ad_group_criterion.status = 'ENABLED'

-- Negative keywords
SELECT
  campaign.name AS campaign,
  '' AS ad_group,
  campaign_criterion.keyword.text AS keyword,
  campaign_criterion.keyword.match_type AS match_type,
  'NEGATIVE' AS polarity
FROM campaign_criterion
WHERE campaign_criterion.type = 'KEYWORD'
  AND campaign_criterion.negative = TRUE

Step 2: Cross-reference for conflicts

For each negative keyword, check if it would block a positive keyword:

Step 3: Resolution

1. Remove the conflicting negative
2. Add the negative as campaign-level instead of shared
3. Use exact match negatives to be more precise
4. Move the positive keyword to the campaign with the negative to consolidate

Workflow 5: Automated Negative Mining Cadence

Weekly routine:

Monthly routine:

• Full 30-day search term analysis
• Broad match bleed rate calculation
• Shared list audit (remove overly aggressive negatives)
• Cross-campaign conflict resolution

Negative Keyword Match Types

How Negative Match Types Work (Different from Positive)

Critical difference from positive match types:

• Negative broad match does NOT include close variants, synonyms, or related searches
• Negative keywords don't match plurals, misspellings, or similar terms
• You must add each variation explicitly

Example:

Negative broad match: running shoes

• ❌ Blocks: "running shoes", "shoes running", "red running shoes"
• ✅ Still shows: "running", "shoes", "run shoe", "running sneakers"

Negative phrase match: "running shoes"

• ❌ Blocks: "buy running shoes", "running shoes sale"
• ✅ Still shows: "shoes for running", "running shoe"

Negative exact match: [running shoes]

• ❌ Blocks: "running shoes" only
• ✅ Still shows: "buy running shoes", "running shoes sale"

Industry Negative Keyword Templates

E-Commerce (General)

Broad negatives: free, cheap, used, refurbished, wholesale, bulk, diy, 
  repair, fix, manual, instructions, jobs, careers, salary, hiring,
  review, reviews, reddit, forum, comparison, vs, alternative

Phrase negatives: "how to make", "how to build", "how to fix",
  "near me" (if online only), "open box", "second hand"

Exact negatives: [ebay], [amazon], [walmart], [alibaba], [craigslist]

SaaS / B2B Software

Broad negatives: free, open source, template, tutorial, course,
  certification, training, salary, jobs, intern, internship,
  download, crack, pirated, github, stackoverflow

Phrase negatives: "what is", "how to use", "definition of",
  "vs", "alternative to", "open source", "free trial expired"

Exact negatives: [competitor names as exact match]

Local Services (Plumbing, HVAC, Legal)

Broad negatives: diy, tutorial, youtube, jobs, careers, salary,
  hiring, school, certification, course, training, license,
  requirements, exam, practice test

Phrase negatives: "how to fix", "how to install", "how to repair",
  "near me" (for excluded cities), "DIY guide"

Exact negatives: [cities you don't serve], [states you don't serve]

Healthcare / Medical

Broad negatives: jobs, careers, salary, nursing, school, degree,
  certification, free, symptoms, wiki, wikipedia, webmd, mayo clinic,
  home remedy, natural cure, reddit

Phrase negatives: "how to become", "medical school", "nursing program",
  "is it normal", "home treatment"

Exact negatives: [competitor names], [irrelevant specialties]

Financial Services

Broad negatives: free, jobs, careers, salary, scam, fraud, complaint,
  reddit, review, lawsuit, class action, calculator, formula, excel,
  template, worksheet

Phrase negatives: "how to calculate", "what is the formula",
  "free calculator", "excel template", "is it a scam"

Exact negatives: [competitor names], [irrelevant product types]

Real Estate

Broad negatives: jobs, careers, salary, license, exam, course,
  school, training, zillow, trulia, realtor.com, free, craigslist,
  for rent (if selling), for sale (if renting)

Phrase negatives: "how to become", "real estate school",
  "real estate license", "average salary"

Exact negatives: [cities not served], [competitor agent names]

Reference Data

Negative Keyword Impact on Performance

Typical improvements after systematic negative keyword mining:

Budget Waste Benchmarks

Examples

Example 1: SaaS Company Search Term Audit

Scenario: B2B SaaS spending $30K/month on Google Ads. CPA target: $80. Actual CPA: $135.

Analysis of 30-day search term report (2,400 unique terms):

Negatives added: 145 terms across 6 categories Projected savings: $8,200/month (27% of spend) Projected new CPA: $80 (hitting target)

Example 2: E-commerce Broad Match Bleeding Fix

Scenario: Online shoe store, keyword "running shoes" on broad match.

Bleed analysis:

• Total search term impressions: 45,000
• Relevant impressions: 28,000 (62%)
• Irrelevant impressions: 17,000 (38%) ← Severe bleed

Top bleed categories:

1. "running shoes drawing" (art related) — 4,200 impressions, 0 conversions
2. "running shoes for flat feet" (medical) — 3,100 impressions, 2 conversions (acceptable)
3. "free running shoes" — 2,800 impressions, 0 conversions
4. "running shoes factory outlet near me" — 2,400 impressions, 0 conversions (online only)
5. "running shoes repair" — 1,900 impressions, 0 conversions

Actions:

1. Added negatives: drawing, free, repair, factory outlet, near me
2. Kept "flat feet" (relevant variation)
3. Moved keyword from broad to phrase match
4. Bleed rate dropped from 38% to 11% in 2 weeks

Example 3: Cross-Campaign Conflict Resolution

Scenario: 3 campaigns competing with conflicting negatives.

Conflict found:

• Campaign "Brand" has negative phrase: "reviews"
• Campaign "Non-Brand" bids on keyword: "product name reviews"
• Result: "product name reviews" blocked everywhere (Brand neg + Non-Brand doesn't target it)

Resolution:

1. Changed Brand campaign negative from phrase to exact: [reviews]
2. Now "product name reviews" routes to Non-Brand campaign correctly
3. Impressions for review-related queries increased 340%
4. Added 12 high-intent review queries to Non-Brand campaign

About this skill

Part of the Synter free skills collection — open-source agent skills for advertising, PPC, and marketing automation.

Want this skill (and 40+ more) running on autopilot against your live ad accounts? Try Synter — AI Agent Media Buyers that connect to Google, Meta, LinkedIn, TikTok, Reddit, Amazon, and 7+ more platforms.

• 🌐 syntermedia.ai
• 📚 All free skills
• 💬 Built by @syntermedia — questions? Open an issue.