Skip to content
Back to skills

Gsc Reporting

ASecurity

Report SEO via Search Console, GA4, index, CTR, experiments. Use when proving what moved, winner/loser analysis, forecasts, stakeholder digests. For audits, see seo-mastery. For attribution, see analytics-attribution.

  • 5 stars
  • 0 votes
  • 0 copies
  • 4 views
  • Added September 19, 2026
content-marketinggoperformance

Works with

  • cli
  • mcp

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned September 19, 2026

npx -y skills add DoddleHQ/DoddleOS --skill gsc-reporting --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Gsc Reporting?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Gsc Reporting
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/doddlehq-gsc-reporting/badge)](https://www.skillsdirectory.com/skills/doddlehq-gsc-reporting)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: gsc-reporting
id: doddle.marketing.gsc-reporting
version: 1.0.0
blueprint: ./blueprint.yaml
description: Report SEO via Search Console, GA4, index, CTR, experiments. Use when proving what moved, winner/loser analysis, forecasts, stakeholder digests. For audits, see seo-mastery. For attribution, see analytics-attribution.
---

# GSC Reporting

GSC + GA4 proof layer: performance, deltas, cannibalization evidence, forecasts, stakeholder outputs. Covers infographic block 9 (177-200).

## Language & Quality Standards

**CRITICAL**: Respond in same language user uses.

**Standards**: Token efficiency, sacrifice grammar for concision, list unresolved questions at end.

---

## When to Use This Skill

Apply when:
- Proving clicks, impressions, CTR, position moves
- Winner/loser, period deltas, incident impact
- Index coverage, cannibalization evidence
- Weekly / monthly stakeholder reporting, forecasts

## Initial Assessment

1. **Access**
   - GSC / GA4 connected? Date range?
   - Brand vs non-brand split needed?
2. **Scope**
   - Pages, queries, countries, devices?
   - Experiment or incident to explain?

---

## Inputs Schema

| Input | Type | Required | Description |
|-------|------|----------|-------------|
| site_or_property | string | yes | GSC property or domain; ask if missing |
| date_range | string | no | Comparison period; default last 28d vs prior |
| segment | string | no | Brand/non-brand, device, country filter |

---

## Outputs Schema

| Output | Type | Description |
|--------|------|-------------|
| report | markdown | Performance, deltas, evidence, forecast |
| action_list | json | Next actions from findings |

---

## Core Framework

### Step 1: Pull
GSC performance, pages, queries, CTR, coverage, GA4 assists.

### Step 2: Prove
Deltas, winners/losers, cannibalization evidence, incident review.

### Step 3: Report
Digest, recap, backlog, forecast baseline. No fabrication.

---

## Detailed Guidance

### Performance Pull

**Checklist:**
- [ ] GSC review (177), page table (178), query trends (179)
- [ ] CTR list (180), impression share (181)
- [ ] Landing delta (182), device split (188), country cut (189)
- [ ] Branded view (187), non-brand growth (186)
- [ ] Index summary (190), dashboard pack (185)

### Evidence & Deltas

**Checklist:**
- [ ] Winner/loser (183), annotation log (184)
- [ ] Cannibalization evidence (191) — prove overlap
- [ ] Experiment readout (193), incident review (197)
- [ ] Conversion assist (198), revenue tracker (199)

### Reporting Out

**Checklist:**
- [ ] Opportunity backlog (194) — findings → tasks
- [ ] Weekly digest (195), monthly recap (196)
- [ ] Action sheet (200) — next actions plainly
- [ ] Forecast baseline (192) — likely trend, mark assumptions

---

## Common Mistakes

### Strategy Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| Report vanity impressions | No decisions | Tie to clicks + revenue |
| No annotations | Unexplained drops | Log changes/releases |
| Mix brand + non-brand | Hides SEO | Split views |

### Execution Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| No date compare | No movement proof | Prior-period delta always |
| Ignore CTR | Missed quick wins | CTR opportunity list |
| Skip index check | False alarms | Coverage first |

### Analysis Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| Fabricate missing data | Violates reliability | State NOT AVAILABLE |
| Over-forecast | False precision | Baseline + range |
| No backlog | Report dies | Every finding → task |

---

## Metrics to Track

| Metric | Definition | Target | Tool |
|--------|------------|--------|------|
| Clicks / impressions | GSC totals | Growing | GSC |
| Avg CTR / position | Click rate, rank | CTR >3%, pos improving | GSC |
| Winners / losers | Top movers | Net positive | GSC delta |
| Index coverage | Indexed vs submitted | >90% | GSC |
| Assists / revenue pages | Search → conversion | Growing | GA4 |

---

## Decision Tree

No GSC access → manual template + NOT AVAILABLE tags. Drop detected → incident review + annotations. Overlap suspected → cannibalization evidence. Stakeholder ask → digest/recap, not raw dump.

---

## Quick Assessment Checklist

1. [ ] Property + date range?
2. [ ] Brand split needed?
3. [ ] Device / country cuts?
4. [ ] Incident / experiment context?
5. [ ] Tool access or manual?

---

## Expected Output Format

### Performance
[Clicks, CTR, pages, queries]

### Movers + Evidence
[Winners/losers, cannibalization, incidents]

### Next Actions
[Backlog, forecast, digest]

---

## Common Failure Modes

| Failure | Symptom | Fix |
|---------|---------|-----|
| No tool access | Empty tables | NOT AVAILABLE, manual fallback |
| Sampling noise | Tiny deltas overread | Threshold movers |
| Missing annotations | Mystery drops | Rebuild change log |

---

## MCP Tool Integration

| Tool ID | When to Use | Data to Pull | Required |
|---------|-------------|--------------|----------|
| doddle.tool.v1.gsc.getSearchAnalytics | All reporting | Queries, pages, CTR, position | no |
| doddle.tool.v1.ga4.getReport | Assists, revenue | Sessions, conversions | no |
| doddle.tool.v1.dataforseo.serpGoogle | SERP context | Features, volatility | no |

Fallback: if tool unavailable, state data as NOT AVAILABLE per `data-reliability-rules.md`. Never fabricate.

---

## Agent Collaboration

| Agent | When to Collaborate | What They Provide |
|-------|--------------------|--------------------|
| seo-specialist | Evidence review | Cannibalization, index triage |
| researcher | Trend context | Market shifts |
| planner | Backlog grooming | Task prioritization |

---

## Related Skills

- **seo-mastery**: Audits feeding reports
- **analytics-attribution**: Cross-channel attribution
- **content-strategy**: Refresh queue from decay

---

## Questions to Ask

1. GSC property + date range?
2. Brand / non-brand split?
3. Weekly digest or monthly recap?

Files in this skill

  • SKILL.md5.9 KB
  • blueprint.yaml1.5 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.