moreminimore-service-system/skills/mm-analytics-orchestrator/SKILL.md

---
name: mm-analytics-orchestrator
description: >
  Pull SEO + Ad performance data for a client, analyze trends, and produce
  prioritized strategy recommendations. Saves report to vault and presents
  in chat. Use when: "วิเคราะห์ผล", "analytics report", "ดูสถิติ",
  "แนะนำกลยุทธ์", "strategy recommendations", "seo report", "ad performance",
  "content ideas from data", "เขียนอะไรดี", "analytics".
---

# mm-analytics-orchestrator

Pulls analytics data, analyzes trends, and produces prioritized strategy
recommendations for a client.

## When This Must Trigger

- "วิเคราะห์ผล", "analytics report", "ดูสถิติ"
- "แนะนำกลยุทธ์", "strategy recommendations"
- "seo report", "ad performance"
- "content ideas from data", "เขียนอะไรดี"
- "analytics"

## Process

### Step 1: Determine client

If not specified, ask:

```
วิเคราะห์ผลของลูกค้าไหน?

1. moreminimore
2. [other clients...]
```

### Step 2: Pull data

Run analytics scripts from `~/Gitea/moreminimore-service-system/scripts/`:

```bash
# GSC data
python3 mm_gsc.py --client <client-id> --days 30 --json

# Google Ads data (if available)
python3 mm_google_ads.py --client <client-id> stats --days 30 --json

# Meta Ads data (if available)
python3 mm_meta_ads.py --client <client-id> stats --days 30 --json
```

If a script fails (no data, API error), skip that source and note it.

### Step 3: Analyze

#### SEO Analysis

**a. Quick Wins (Position 4-10)**
- Queries on page 1 but below top 3
- High impression, low CTR → title/description optimization
- Action: ปรับ meta title/description

**b. Content Gaps (Position 11-30)**
- Queries with impressions but no dedicated content
- Action: เขียนบทความใหม่

**c. CTR Opportunities**
- High impressions, low CTR (< expected for position)
- Action: ปรับ title/description, add schema

**d. Declining Pages**
- Pages with >30% traffic decline
- Action: Refresh content, update data

**e. GEO Opportunities**
- Question-format queries, definition queries
- Action: Optimize for AI citation

#### Ad Analysis (if data available)

**a. High CPA Campaigns**
- CPA > 2x target
- Action: ปรับ targeting, ทดสอบ creative ใหม่

**b. Low QS Keywords**
- Quality Score < 5
- Action: ปรับ ad relevance, landing page

**c. Budget Optimization**
- High spend, low conversion
- Action: ย้าย budget

**d. Creative Fatigue**
- CTR declining over time
- Action: Refresh creative

#### Cross-Channel

**a. SEO + Ad Synergy**
- Organic ranking well for paid keyword
- Action: ลด paid spend

**b. Content → Ad Pipeline**
- Article with good traffic → create ad
- Action: สร้าง campaign จาก content

**c. Ad → Content Pipeline**
- Keyword with good ad conversion → write article
- Action: เขียนบทความ SEO

### Step 4: Prioritize recommendations

Rank by impact × effort:

| Priority | Criteria |
|----------|----------|
| 🔴 Priority 1 | High impact, low effort (quick wins) |
| 🟡 Priority 2 | High impact, medium effort (content opportunities) |
| 🟢 Priority 3 | Medium impact, optimization (ad tuning) |
| 🔵 Priority 4 | Cross-channel opportunities |

### Step 5: Present and ask

Present the report in chat (match user's language):

```
📋 Strategy Report: [Client Name]
📅 Period: [date range]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📊 SEO Performance
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
| Metric | Value | Trend |
|--------|-------|-------|
| Clicks | X | ↑/↓ X% |
| Impressions | X | ↑/↓ X% |
| CTR | X% | ↑/↓ |
| Position | X | ↑/↓ |

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🎯 Priority 1: Quick Wins
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

1. [Action]
   Evidence: [data]
   Expected: [outcome]

2. [Action]
   ...

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📝 Priority 2: Content Opportunities
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

1. เขียนบทความ: "[topic]"
   Keyword: [kw], [X] impressions, position [X]
   Search intent: [type]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📢 Priority 3: Ad Optimization
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
(ถ้ามีข้อมูล ads)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔄 Priority 4: Cross-Channel
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
```

Then ask:

```
อยากทำตามแนะนำไหม?

1. ✅ ทำทั้งหมด — สร้าง brief / optimize ตามลำดับ
2. 📝 เฉพาะ content — สร้าง brief จาก content gaps
3. 📢 เฉพาะ ads — optimize campaigns
4. ⏭️ ข้าม — แค่ดูรายงาน
```

If user chooses to act:
- Content gaps → load `mm-article-idea-extract` with the recommended topic
- Ad optimization → provide specific commands/changes
- SEO fixes → provide specific title/description rewrites

### Step 6: Save report

Save to vault:

```
~/vault/50_Resources/Strategy for Customers/<client-id>/YYYY-MM-DD-report.md
```

Create the folder if it doesn't exist.

Report format:

```markdown
---
type: strategy-report
client: <client-id>
period: <start-date> to <end-date>
generated: YYYY-MM-DD
---

# Strategy Report: [Client Name]
[full report content]
```

### Step 7: Confirm

```
✅ รายงาน saved: ~/vault/50_Resources/Strategy for Customers/<client-id>/YYYY-MM-DD-report.md
```

## Output

- Chat: full report with recommendations
- Vault: `~/vault/50_Resources/Strategy for Customers/<client-id>/YYYY-MM-DD-report.md`

## Notes

- Match user's language (Thai if they write Thai, English if English)
- If GSC data is empty (new site), note it and focus on ad data
- If no ad data, focus on SEO only
- Always ask before taking action (creating briefs, optimizing)
- Reports are cumulative — each report builds on previous data