Overview
In the era of the end of 3rd party cookies (Apple ATT + Google cookie deprecation), the ad efficiency gap between brands with 1st party data and those without is widening. This skill measures and diagnoses that gap.
One-line responsibility: Diagnose Meta/Google pixel installation status + 1st Party data + Lookalike seed quality → automatically generate improvement priorities.
The 3 most common mistakes in pixel design
Mistake 1: Not excluding purchasers
Continuing to show "Haven't you bought yet?" ads to people who already bought.
Diagnosis: In Meta Ads Manager → Ad Set → Audience exclusions, check whether Purchase event 30-day trigger has been added.
- ❌ Not configured → ad spend leakage (ads shown to purchasers too)
- ✅ Configured → shown only to new purchase prospects
Mistake 2: Event parameters not configured
ViewContent·AddToCart에 Product ID and price parametersmust be attached for dynamic ads (DPA) to work.
Diagnostic code:
// ❌ 잘못된 예
fbq('track', 'ViewContent');
// ✅ 올바른 예
fbq('track', 'ViewContent', {
content_ids: ['SKU123'],
content_type: 'product',
value: 19800,
currency: 'KRW'
});
Mistake 3: Conversion API not installed (CAPI)
With the pixel alone, conversion tracking for iOS users is incomplete. Server-side Conversion API (CAPI) must be installed as well.
After iOS 14.5+:
- pixel-only: tracks only 50-70% of actual conversions
- pixel + CAPI: can track 90%+
3 types of 1st Party data
| Data | Utilization intensity | How to use |
|---|---|---|
| Email/phone number | Strongest | Collect via newsletter/purchase/events → upload to Meta/Google → Custom Audience + Lookalike seed |
| Pixel event data | Strong | PageView → ViewContent → AddToCart → InitiateCheckout → Purchase 5 stages |
| Purchase/CRM data | Very strong | Predict repurchase cycles, VIP segments, reactivate dormant customers |
Lookalike — seed quality is everything
The quality limit of Lookalike = the quality limit of the seed data.
| Seed data | Lookalike quality |
|---|---|
| All visitors | Low (includes people with no purchase intent) |
| Add-to-cart users | 중 |
| All purchasers | High |
| VIP purchasers (top 20%) | Best ⭐ |
Practical recommendations
1. VIP 구매자(상위 20%) 데이터 추출 (LTV·구매 횟수·평균 객단가 상위)
2. 메타에 커스텀 오디언스 업로드
3. Lookalike 생성 (유사도 1~3%부터 시작 → 포화되면 확장)
4. 기존 고객 반드시 제외
5. 월 1회 이상 씨앗 데이터 최신 상태 업데이트
Workflow — diagnostic checklist
[Step 1] 사이트 URL + 광고 채널 사용 여부 입력
↓
[Step 2] A. 메타 픽셀 / B. 구글 픽셀 / C. 1st Party / D. Lookalike 4영역 점검
↓
[Step 3] 영역별 점수 산출 (100점 만점)
↓
[Step 4] Phase 1 (즉시) / Phase 2 (단계적) / Phase 3 (최적화) 우선순위 자동 분류
↓
[Step 5] 진단 보고서 출력 + (선택) moai-marketer:marketing-campaign-planner·moai-seller:commerce-marketplace-coupang-ads 체이닝
A. Meta Pixel
- pixel installation (check with Facebook Pixel Helper)
- PageView event operation
- ViewContent event + content_ids/value/currency parameters
- AddToCart event + parameters
- Purchase event + parameters
- CAPI (server-side Conversions API) setup
- Add Purchase 30 days to ad set audience exclusions
- Aggregated Event Measurement (AEM) 8-event setup
B. Google Pixel (GA4 + Ads)
- GA4 setup
- Enable Enhanced Conversions
- Define conversion events (purchase, sign-up, cart)
- Google Ads integration
- Register negative keywords ('free', 'DIY', 'how to')
C. 1st Party Data
- Email/phone number collection channels (newsletter, purchase, events)
- Upload Meta Custom Audiences (1+ per month)
- CRM system (Customer Lifecycle tracking)
- Automatic extraction of top 20% VIP purchasers possible
D. Lookalike
- Use VIP seed data
- Start with 1-3% similarity
- Exclude existing customers
- Update seed at least once per month
Trigger keywords
Pixel, Meta Pixel, Google Pixel, Facebook Pixel, CAPI, Conversions API, Lookalike, similar target, 1st Party Data, VIP purchasers, email list, Custom Audience, iOS conversion loss, AEM, Aggregated Event
Input/Output
Input slots
| Item | Required | Example |
|---|---|---|
| Site URL | Required | https://example.com |
| Whether Meta Ads is used | Required | Yes/No |
| Whether Google Ads is used | Required | Yes/No |
| Has email list | Recommended | 5,000개 |
| Currently used pixels/tools | Recommended | Meta Pixel, GA4, CAPI, etc. |
Output (diagnostic report)
## 픽셀·1st Party 데이터 진단 결과
### 📊 점수
- 메타 픽셀: 60/100
- 구글 GA4: 70/100
- 1st Party 데이터: 30/100 ⚠️
- Lookalike: 미구축 ⚠️
- **종합 점수: 53/100 (C)**
### 🔴 즉시 수정 (Phase 1, 1주)
1. CAPI 미설치 → iOS 전환 30% 손실 추정
2. VIP 씨앗 Lookalike 미구축 → Cold 광고비 효율 손실
3. 이메일 리스트 메타 업로드 미시행
### 🟡 단계적 개선 (Phase 2, 1개월)
4. ViewContent 이벤트 파라미터 보강
5. Enhanced Conversions 활성화
### 🟢 최적화 (Phase 3, 분기)
6. 코호트 분석 + LTV/CAC 측정 자동화
Chaining
marketing-pixel-audit (본 스킬, 픽셀·1st Party 인프라 진단)
↓ Phase 1 즉시 수정 후
moai-marketer:marketing-campaign-planner (인프라 기반 캠페인 기획 + 9 인지 편향 적용)
↓ 메타·구글 광고 운영 시
moai-marketer:content-sns-content (채널별 심리 상태 매트릭스 + CAPI 검증)
↓ 쿠팡 광고는 별도로
moai-seller:commerce-marketplace-coupang-ads (메타·구글 픽셀 아닌 쿠팡 전용 자동규칙)
The narrative text of the diagnostic report is post-processed with
moai-coworker:ai-slop-reviewer → moai-writer:korean-humanizebefore publishing. Score/checklist tables are not subject to this.
Pair separation
| Pair skills | Difference |
|---|---|
moai-marketer:marketing-performance-report |
Comprehensive GA4/Meta/Naver ROAS analysis (this skill verifies pixel/data infrastructure) |
moai-marketer:marketing-campaign-planner |
Campaign planning/operation (this skill diagnoses data infrastructure) |
moai-seller:commerce-marketplace-coupang-ads |
Coupang Ads (this skill covers Meta/Google pixels) |