Spend Efficiency Analyzer

You are running a spend efficiency audit across all campaigns. The goal: find where money is being wasted and rank opportunities by potential savings.

Before you start, read the ads-playbook skill to load target CPA, wasted spend thresholds, and business context.

---

Step 1: Pull Campaign-Level Performance

Run this GAQL query to get all enabled campaigns:

SELECT
  campaign.name,
  campaign.status,
  campaign.advertising_channel_type,
  metrics.cost_micros,
  metrics.conversions,
  metrics.cost_per_conversion,
  metrics.clicks,
  metrics.impressions,
  metrics.ctr,
  metrics.average_cpc,
  metrics.search_impression_share
FROM campaign
WHERE segments.date DURING LAST_30_DAYS
  AND campaign.status = 'ENABLED'
ORDER BY metrics.cost_micros DESC

Cost conversion: Divide cost_micros, cost_per_conversion, and average_cpc by 1,000,000.

---

Step 2: Pull Ad Group-Level Performance

For each campaign, pull ad group data to find zero-conversion ad groups:

SELECT
  campaign.name,
  ad_group.name,
  ad_group.status,
  metrics.cost_micros,
  metrics.conversions,
  metrics.clicks,
  metrics.impressions,
  metrics.ctr,
  metrics.cost_per_conversion
FROM ad_group
WHERE segments.date DURING LAST_30_DAYS
  AND campaign.status = 'ENABLED'
  AND ad_group.status = 'ENABLED'
  AND metrics.cost_micros > 0
ORDER BY metrics.cost_micros DESC

---

Step 3: Pull Keyword Match Type Data

Check for broad match bleeding:

SELECT
  campaign.name,
  ad_group.name,
  ad_group_criterion.keyword.text,
  ad_group_criterion.keyword.match_type,
  metrics.cost_micros,
  metrics.conversions,
  metrics.clicks,
  metrics.impressions,
  metrics.cost_per_conversion,
  ad_group_criterion.quality_info.quality_score
FROM keyword_view
WHERE segments.date DURING LAST_30_DAYS
  AND campaign.status = 'ENABLED'
  AND ad_group.status = 'ENABLED'
  AND metrics.cost_micros > 0
ORDER BY metrics.cost_micros DESC

---

Step 4: Flag Issues

Analyze the data and flag these categories:

A. High CPA Campaigns

Campaigns where CPA exceeds the target from the ads-playbook.

For each flagged campaign:

• Current CPA vs target CPA
• How much was spent above the target
• Potential savings if CPA hit target

B. Zero-Conversion Ad Groups with Significant Spend

Ad groups that spent above the playbook's zero-conversion kill threshold without a single conversion.

For each:

• Ad group name, campaign name
• Total spend (all wasted)
• Number of clicks (to gauge if there's enough data)

C. Broad Match Bleeding

Keywords on broad match that are consuming disproportionate budget with poor results. Flag if:

• Broad match keyword has CPA > 2x the target
• Broad match keyword has CTR < half the campaign average
• Broad match keyword has zero conversions and significant spend

For each:

• Keyword text, match type
• Spend, conversions, CPA
• Recommendation: tighten to phrase/exact or add negatives

D. Low Quality Score Keywords

Keywords with Quality Score below 5 that are still accumulating spend. Low QS = paying more per click for worse positions.

For each:

• Keyword text, Quality Score
• Spend, CPA
• Recommendation: improve ad relevance, improve landing page, or pause

---

Step 5: Rank by Potential Savings

Create a ranked list of all flagged items sorted by estimated savings (highest first).

Savings estimation:

• Zero-conversion ad groups: savings = total spend (it's all wasted)
• High CPA campaigns: savings = spend - (conversions * target CPA)
• Broad match bleeding: savings = spend on flagged broad terms with zero conversions
• Low QS keywords: savings = estimated CPC reduction * clicks (rough: assume QS improvement saves ~20% of CPC)

---

Step 6: Present the Report

Summary

• Total spend analyzed: $X,XXX
• Total potential savings identified: $X,XXX (X% of spend)
• Number of issues found: X

Prioritized Recommendations

Priority	Type	Item	Spend	Savings	Action
1	Zero-conv AG	"Brand Awareness" in Campaign X	$450	$450	Pause ad group
2	High CPA	Campaign Y (CPA: $120 vs $45 target)	$2,400	$1,200	Review targeting
3	Broad match	"meeting software" (broad)	$380	$380	Tighten to phrase
...	...	...	...	...	...

For each recommendation, include a specific action the user can take.

---

Step 7: Offer to Execute

For actionable items (pausing ad groups, changing match types), ask:

"I can pause the zero-conversion ad groups and tighten the broad match keywords. Want me to do any of these? I'll confirm each change before executing."

Wait for explicit approval on each change.

---

Edge Cases

• No MCP connected: Ask the user to paste campaign performance data (from Google Ads export) and analyze that.
• New account with limited data: Note that 30 days may not be enough for statistical significance. Flag items but caveat the recommendations.
• Single campaign: Still run all checks at ad group and keyword level.
• Meta Ads: Pull campaign and ad set performance. Flag high-CPA ad sets, zero-conversion ad sets, and audience overlap issues.
• Apple Search Ads: Pull campaign and keyword performance. Flag high-CPI keywords and discovery match bleeding (Apple's equivalent of broad match).