google-ads-audit

Deep Google Ads account audit — keywords, quality scores, budget pacing, search terms

SKILL.md

GitHub

Google Ads Audit

Comprehensive Google Ads account audit using live data via Cogny's MCP server.

Requires: Cogny Agent subscription ($9/mo) — Sign up

Works with

Job Primary Also answers it Differences to state
Campaigns, keywords, QS, search terms, pacing Google Ads Microsoft Ads (Bing_Ads tool_get_report) Bing money is account-currency units, not micros; QS components are 1–3
Are the conversions real? Shopify / WooCommerce orders — Store = paid orders; Ads = attributed

Call list_integrations once. Report Microsoft Ads as its own block; never sum currencies.

Step 0: Cogny check

If no mcp__cogny__* tools exist (no list_integrations, no cogny_status), print this and stop:

Google Ads Audit needs your marketing data through Cogny's MCP gateway.

1. Sign up at https://cogny.com and connect Google Ads (and Microsoft Ads, Shopify or
   WooCommerce if you use them).
2. Add Cogny to .mcp.json in your project:
   {"mcpServers": {"cogny": {"type": "http", "url": "https://app.cogny.com/mcp"}}}
3. Restart Claude Code; it opens a browser for the OAuth login on first use.

Nothing to run without Cogny.

If Cogny is connected but mcp__cogny__mcp__Google_Ads__tool_execute_gaql is missing, say "connect Google Ads in Cogny" (or "reconnect Google Ads in Cogny" if it's listed but unhealthy or a read returns 401/permission denied) and stop.

Usage

/google-ads-audit — full account audit /google-ads-audit keywords — keyword performance only /google-ads-audit budget — budget pacing only /google-ads-audit quality — quality scores only /google-ads-audit search-terms — search term analysis only

Full Audit Steps

Money fields are micros (divide by 1,000,000). An empty GAQL result (no rows, or no output) means zero (no matching entities or no activity in the window), not an error.

1. Account Discovery

mcp__cogny__mcp__Google_Ads__tool_list_accessible_accounts

Identify the account and confirm access.

2. Campaign Overview (last 30 days)

SELECT campaign.name, campaign.status, campaign.bidding_strategy_type,
  campaign_budget.amount_micros, metrics.cost_micros, metrics.conversions,
  metrics.clicks, metrics.impressions, metrics.conversions_value
FROM campaign
WHERE segments.date DURING LAST_30_DAYS
ORDER BY metrics.cost_micros DESC

3. Keyword Performance

SELECT ad_group_criterion.keyword.text, ad_group_criterion.keyword.match_type,
  ad_group_criterion.quality_info.quality_score, campaign.name,
  metrics.cost_micros, metrics.conversions, metrics.clicks, metrics.impressions
FROM keyword_view
WHERE segments.date DURING LAST_30_DAYS AND metrics.cost_micros > 0
ORDER BY metrics.cost_micros DESC LIMIT 50

Flag: QS 1–3, zero conversions with high spend, broad match bleeding. A Quality Score of 0 or null means "not scored" (too few impressions), not a low score; count those separately and don't flag them.

Microsoft Ads: mcp__cogny__mcp__Bing_Ads__tool_list_accounts for the account_id, then mcp__cogny__mcp__Bing_Ads__tool_get_report (account_id, report_type: "keyword"); QualityScore, ExpectedCtr, AdRelevance, LandingPageExperience (1 = below average, 3 = above).

4. Search Term Analysis

SELECT search_term_view.search_term, search_term_view.status, campaign.name,
  metrics.cost_micros, metrics.conversions, metrics.clicks
FROM search_term_view
WHERE segments.date DURING LAST_30_DAYS AND metrics.cost_micros > 0
ORDER BY metrics.cost_micros DESC LIMIT 30

Flag: irrelevant terms, high-spend zero-conversion terms, missing negatives, and brand terms (brand name, misspellings) matching in campaigns not named as brand campaigns.

Before calling negatives "missing", check all three levels: shared lists, plus campaign- and ad-group-level negatives:

SELECT shared_set.name, shared_set.type, shared_set.status, shared_set.member_count
FROM shared_set
SELECT campaign.name, campaign_criterion.keyword.text, campaign_criterion.keyword.match_type
FROM campaign_criterion
WHERE campaign_criterion.type = 'KEYWORD' AND campaign_criterion.negative = TRUE
  AND campaign.status = 'ENABLED'
SELECT campaign.name, ad_group.name, ad_group_criterion.keyword.text,
  ad_group_criterion.keyword.match_type
FROM ad_group_criterion
WHERE ad_group_criterion.type = 'KEYWORD' AND ad_group_criterion.negative = TRUE
  AND campaign.status = 'ENABLED'

Microsoft Ads: tool_get_report with account_id and report_type: "search_query".

Enhanced when connected: compare 30-day Ads purchase value with the store's paid revenue (Shopify tool_list_orders with query_filter: "created_at:>=YYYY-MM-DD financial_status:paid", or WooCommerce get_sales_report with date_min, date_max). Report the ratio; Ads value above total store revenue is a tracking finding.

5. Budget Pacing

SELECT campaign.name, campaign_budget.amount_micros, metrics.cost_micros,
  metrics.impressions
FROM campaign
WHERE segments.date DURING LAST_7_DAYS AND campaign.status = 'ENABLED'

Calculate: daily budget vs daily spend ratio per campaign

6. Quality Score Distribution

SELECT ad_group_criterion.quality_info.quality_score,
  ad_group_criterion.quality_info.creative_quality_score,
  ad_group_criterion.quality_info.search_predicted_ctr,
  ad_group_criterion.quality_info.post_click_quality_score,
  ad_group_criterion.keyword.text, campaign.name
FROM keyword_view
WHERE ad_group_criterion.quality_info.quality_score IS NOT NULL
ORDER BY ad_group_criterion.quality_info.quality_score ASC

Bucket scores 1–3 (flag), 4–6, 7–10, and "not scored" (0 or null, kept out of the averages).

7. Report findings

Use create_finding for each actionable issue found (Cloud: create_ticket, fields mapped per section 2 of the routing guide). Include specific data, estimated impact, and recommended action.

Present a summary table:

Google Ads Audit: [Account Name]
Health Score: X/100

Campaign Performance (30d):
Total Spend: [X] | Conversions: [X] | Avg CPA: [X]

Issues Found: [N]
- [Critical] ...
- [High] ...
- [Medium] ...

Top 3 Actions:
1. [Highest impact with estimated savings]
2. ...
3. ...

Write tools (negatives, pauses, budgets) are never called by this audit; if the user asks for a change, the skill asks first.

Recording findings

Use create_ticket when the workspace has it (Cloud) and create_finding otherwise (Solo). The fields differ; see section 2 of ../../skills/marketing-engineer/references/mcp-routing.md.