For Agents · MCP

Your task.
The right skill.

The tunnel uses the same cards as the catalogue. Browse only as deep as needed — or load a broad bundle immediately.

No local installationNo M11 loginRead-only

Broad task

Load 5–10 and go

SEO, Sales, Agents or another broad area → one bundle call → work.

MCP endpoint: https://skills.m11.ch/mcp

Read-only access to published skills. Default 8, maximum 10 skills / 120,000 characters.

Task Packs

Marketing Foundation

Compact two-skill starter: clarify positioning and choose a lead magnet. Use Marketing Launch for the broader eight-skill go-to-market workflow.

Marketing Launch

Build an evidence-led marketing plan from ICP and competition through positioning, campaigns, growth and measurement.

Agent Audit Essentials

Diagnose architecture and context, plan agent-team responsibilities, then organize project context and session handoffs. Memory and cost-runtime reviews remain outside this pack.

Conversion and Activation Review

Review the journey from landing page and lead capture through registration, first value and transparent upgrades.

Campaign Content and Brand Review

Plan a campaign, draft its channel content and review the work against actual brand guidance.

Content and Launch Planning

Prioritize an editorial roadmap and plan how to launch and distribute it across suitable channels.

Customer, Market and Community Research

Understand customer needs, compare competitors and plan a community around real member value.

Lead Magnet to Lifecycle Nurture

Choose a relevant lead magnet, then draft a permission-based nurture journey with entry, suppression and exit rules.

API Contract and Observability

Define the API contract, then plan how to observe its latency, failures and retries. Guidance and checklist; no production changes.

Data Analysis Foundation

Profile a dataset, choose and interpret statistical methods, then validate calculations and conclusions before sharing.

Choose an area

01 · Category

What the MCP returns at this step

88 · Marketing · Ads Analysis

Google Ads Deep Analysis

Organize an eight-dimension GoMarble Google Ads audit; use Google Ads Account Diagnostics for the Google-authored issue workflow.

googleanzeigentiefenanalysegoogle-ads-depth-of-analysis

Google Ads Deep Analysis Framework · Original SKILL.md

---
name: google-ads-depth-of-analysis
description: "Use for any Google Ads audit, performance review, optimization request, strategy question, or account restructuring. Covers 8 analytical dimensions: hierarchical drill-down, breakdowns, temporal trends, attribution rigor."
---
# Google Ads - Depth of Analysis

> **When to apply**: Audits, performance reviews, optimization requests, strategy questions, account restructuring.
> **Skip for**: Simple data fetches, campaign listings, metric lookups.

## 1. Hierarchical Drill-Down

Never make recommendations based solely on campaign-level aggregates. Before concluding on any entity, drill deeper.

**Required flow:**
- Asked about a campaign → check ad group performance within it
- Asked about an ad group → check keyword + search term performance within it
- Asked about PMax → check asset group performance, search term categories, URL expansion config

**How:**
```
google_ads_run_gaql: SELECT campaign, ad_group, metrics FROM ad_group WHERE campaign.id = X
google_ads_run_gaql: SELECT search_term_view, metrics FROM search_term_view WHERE campaign.id = X
```

For PMax: search term data is limited — check search term category insights, asset group performance differences, and correlate PMax delivery changes with Brand Search impression share.

Identify which specific child entities drive the aggregate. Quantify: "Ad Group X accounts for 65% of campaign spend but CPA is 2x the campaign average."

## 2. Objective-Contextual Evaluation

Evaluate each campaign against its bid strategy target, not just absolute performance.

**Check first:**
- Bid strategy type: tCPA, tROAS, Maximize Conversions, Manual CPC, etc.
- Bid strategy target value (if set)
- Campaign type: Search, PMax, Shopping, Display, YouTube

**Apply correct evaluation:**
| Bid Strategy | Evaluate against | Red flag |
|-------------|-----------------|----------|
| tCPA | Actual CPA vs target CPA | CPA 30%+ above target |
| tROAS | Actual ROAS vs target ROAS | ROAS 30%+ below target |
| Maximize Conversions (no target) | CPA trend direction | Rising CPA with no constraint |
| Manual CPC | CPC vs Quality Score implied CPC | Paying premium due to low QS |

**Flag unconstrained strategies:** If tROAS target is 0.5x on a campaign achieving 3x — the algorithm isn't actually constrained. If tCPA is set 3x above actual CPA — same problem. Call this out.

**PMax maturity gate:** PMax must meet ONE of: age ≥ 30 days (ideally 60) OR total conversions ≥ 50. If neither met, flag as "insufficient data — premature to optimize." Exception: if campaign is clearly burning money (ROAS < 0.3x target after spend ≥ 3x AOV), recommend pausing.

## 3. Breakdown & Segmentation Analysis

Before concluding, check at least 2 of these breakdown dimensions:

**Required breakdowns:**
```
google_ads_run_gaql: SELECT segments.device, metrics FROM campaign WHERE campaign.id = X
google_ads_run_gaql: SELECT geographic_view, metrics FROM geographic_view WHERE campaign.id = X
google_ads_run_gaql: SELECT segments.day_of_week, metrics FROM campaign WHERE campaign.id = X
google_ads_run_gaql: SELECT campaign, metrics.search_impression_share WHERE segments.ad_network_type = 'SEARCH' vs 'SEARCH_PARTNERS'
```

**What to look for:**
- Device performance divergence — mobile vs desktop CPA gap
- Search Partners conversion rate vs Search network — quantify the gap (e.g., "Search Partners converts at 35% the rate of Search")
- Geographic variations — regions with high spend and poor conversion rates
- Day-of-week patterns — budget waste on low-converting days
- Cross-reference bid adjustments against actual segment performance — are adjustments aligned with data?

**Statistical significance:** Do not make segment-level recommendations from <50 clicks or <5 conversions. Call out low-confidence segments.

## 4. Temporal & Trend Analysis

Never treat the date range as a single number. Always compare time windows.

**Required comparisons:**
```
google_ads_run_gaql: SELECT segments.date, metrics FROM campaign WHERE segments.date BETWEEN 'YYYY-MM-DD' AND 'YYYY-MM-DD'
```

Analyze at minimum:
- Last 7 days vs previous 7 days (week-over-week)
- Separate day-of-week seasonality from structural trends
- Identify direction for each key metric: improving, declining, stable

**Distinguish:** "Metric is bad" vs "metric is getting worse" — the direction matters more than the absolute value.

**Connect the chain:** Impression share declining → CPC increasing → this is competitive pressure, not quality degradation. Correlate timing: "CPC started rising on [date], coinciding with impression share loss — suggests new competitor entered the auction."

## 5. Cross-Entity Dependency Recognition

Do not analyze campaigns in isolation. Map relationships.

**Check for:**
- Brand Search ↔ PMax overlap: Are both active on brand terms? Check search term reports for brand term leakage in PMax.
- Non-brand Search → Remarketing: Does non-brand search activity feed remarketing lists?
- Shopping ↔ Search overlap: Same products advertised in both? Check for cannibalization.

**How to detect PMax brand cannibalization:**
```
google_ads_run_gaql: SELECT search_term_view.search_term, metrics FROM search_term_view WHERE campaign.advertising_channel_type = 'PERFORMANCE_MAX'
```
Check for brand terms appearing in PMax search terms. Correlate PMax spend/conversion trends with Brand Search impression share trends over the same period.

**Flag dependencies:** "PMax and Brand Search are both converting on brand terms. Total reported conversions across both campaigns likely exceed actual unique conversions. Before scaling PMax based on its ROAS, consider that some conversions are cannibalized from Brand Search."

## 6. Platform Mechanic Awareness

Explain how Google Ads mechanics affect the specific situation. Do not just mention them — explain the compound interaction.

**Common compound effects:**

| Signal | Mechanic | Consequence |
|--------|----------|-------------|
| Broad match + rising CPC | Query expansion | Algorithm expanding to less relevant queries → QS drops → CPC rises → tCPA struggles |
| Low QS + high CPC | Ad relevance | Paying premium for same position — fix ad relevance, not just bid |
| tCPA + low conversion volume | Smart Bidding data | Insufficient signal for Smart Bidding if below minimum conversion thresholds (see guardrails) — consider Manual CPC |
| PMax + declining Brand Search IS | Brand cannibalization | PMax claiming brand conversions → inflated PMax ROAS, deflated Brand Search |
| Maximize Conversions + no target | Unconstrained spend | Algorithm will spend full budget at any CPA — add a tCPA target |

**Quality Score diagnosis:** When CPC is high, decompose QS into its 3 components (expected CTR, ad relevance, landing page experience). Identify which component is dragging and recommend specific fixes for that component.

## 7. Proactive Signal Detection

After answering the user's question, always surface 2-3 additional concerns or opportunities they did not ask about.

**Scan for:**
- Quality Score trending downward in any high-spend ad group
- Impression share declining toward a delivery constraint (budget-lost or rank-lost)
- Search terms report showing irrelevant queries consuming budget
- Any campaign with conversion volume below Smart Bidding minimum thresholds
- CPC inflation trend that will erode profitability if unchecked
- PMax impression share eating into Brand Search impression share

**Format:** "Outside your question, I noticed [signal] in [entity] — this suggests [predicted consequence] within [timeframe]."

## 8. Attribution & Measurement Rigor

Do not take reported conversion numbers at face value.

**Always check:**
- Attribution model in use (last click, data-driven, etc.)
- Conversion lag: are recent days' conversions still incomplete?
- What share of conversions are "estimated" (modeled) vs observed?

**Flag these situations:**
- Display/YouTube campaigns with high view-through conversions — separate view-through from click-through when assessing ROAS
- PMax + Brand Search both active on brand terms — total reported conversions likely exceed actual unique conversions
- Long attribution windows (90-day) on retargeting — inflates attributed conversions
- Conversion count includes micro-conversions (page views, scroll depth) mixed with macro-conversions (purchases, leads) — separate them

**Frame as caveat:** "Before scaling this campaign based on its reported 5x ROAS, note that [specific attribution concern] means the true incremental ROAS is likely lower."

When to use

Organize an eight-dimension GoMarble Google Ads audit; use Google Ads Account Diagnostics for the Google-authored issue workflow.

What you get

The complete original GoMarble guidance, with source attribution and declared limits.

How it works

Clarify scope, use available evidence, and apply the relevant source workflow within actual user authorization.

Requirements

Account goals, hierarchy, campaign types, full report coverage and attribution definitions.

Delivery and review notes

This framework does not certify attribution or causation. Account restructuring requires separate authorized decisions. GoMarble connectors, referenced sibling skills and runtime dependencies are not bundled or installed by M11. Brief obvious-danger screening only; no functional test or comprehensive safety certification. Loading this text authorizes no external calls, account mutations or ad spend.

Starting prompt

Use Google Ads Deep Analysis for [TASK]. Clarify Account goals, hierarchy, campaign types, full report coverage and attribution definitions. Separate observed evidence from heuristics and recommendations. Check actual connector availability and approval semantics before any external action.

Not for

Claiming installed GoMarble tools, proven campaign outcomes or authorization to launch ads merely from loading this source.

What M11 added

German routing and explicit scope. Original attribution: GoMarble. This framework does not certify attribution or causation. Account restructuring requires separate authorized decisions. GoMarble connectors, referenced sibling skills and runtime dependencies are not bundled or installed by M11. Brief obvious-danger screening only; no functional test or comprehensive safety certification. Loading this text authorizes no external calls, account mutations or ad spend.

Original authorship remains with gomarble-ai/marketing-agent · Original source ↗

Original license & copyright

MIT License

Copyright (c) 2026 GoMarble

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Source SHA-256: f4e847c8f842774737823ae52bff3d67ec80f80e51601bdd8abb88ef8d6d029f
Snapshot checked: 2026-09-28T16:19:52.867Z

Find a skill.Get to work.

All skills

Multi-Agent Orchestration

AI Context Window Audit

AI Agent Architecture Audit

AI Skill Discovery

Lead Magnet Strategy

Ideal Customer Profile

Go-to-Market Strategy

Marketing Campaign Ideas

Product Growth Loops

Competitor Analysis

North Star Metric

Product Positioning

Product Vision

Go-to-Market Motion Selection

Customer Feedback and JTBD Analysis

PESTLE Market Environment Analysis

Customer Journey Mapping

Ansoff Growth Options

Data Analysis Validation

Dataset Profiling

Statistical Analysis Guidance

Programmatic SEO Planning

Landing Page and Form Conversion Review

Paywall and Upgrade Planning

Signup and Registration Review

User Onboarding and Activation

Popup and Modal Planning

Email Sequence Copy and Flow

Marketing Campaign Planning

Marketing Content Drafting

Brand Voice and Content Review

Marketing Performance Reporting

Sales Company Research

Company and Contact Enrichment

Website Information Architecture

Content Strategy and Editorial Roadmap

Product Launch Planning

Customer Research and Voice of Customer

Community Growth and Member Experience

Competitor Research Profiles

Roadmap and Release Communication

Lifecycle Email Sequences

API Contracts and Interface Design

Observability and Instrumentation Planning

Agent Context and Session Handoff

Product Discovery Sprint

Deal Quality Scoring

Enrichment Waterfall Design

Customer Retention Playbook

Sales Coaching Practice

Revenue Cohort Analysis

Intent Signal Scoring

Customer Identity Matching

Segment Activation Planning

Sales Call Review

Retention Dashboard Design

Sales Coaching Competencies

Offer Delivery Design

Offer Value Communication

Service Productization

Conversion Hypothesis Prioritization

Braze Engagement Guidance

Brand Voice Guide Design

Content Pipeline Orchestration

Website Brand Profile

Buyer Objection Clarification

Klaviyo Integration Guidance

Offer Bonus Planning

Klaviyo Marketing Review

Cross-Platform Advertising Review

Expertise Business Model

LinkedIn Advertising Guidance

TikTok Advertising Guidance

Evidence Research Brief

Cinematic Video Prompts

Visual Prompt Model Adaptation

Visual Prompt Diagnostics

Short Film Development

Find a skill.
Get to work.