Ads Schema - Kewpie Sesame Dressing 1/2 Fat

Tổng quan chiến lược

Chiến lược funnel cho Kewpie Sesame Dressing 1/2 Fat được xây dựng dựa trên nguyên lý modern Meta Ads:

data → learning → expand → lock core → drive depth

Thay vì targeting theo demographics (tuổi, giới tính, thu nhập), chúng ta targeting theo động cơ tiêu dùng - lý do tại sao họ quan tâm đến sản phẩm "1/2 Fat".

---

1. Audience Strategy - Phân tầng theo động cơ tiêu dùng

1.1. Các Pillar chính

Không phải "salad sauce" - mà là "ăn thoải mái mà không áy náy"

Pillar	Động cơ	Behavior Signal
Food & Healthy	Muốn ăn ngon nhưng đỡ béo	Food blogs, healthy recipes, nutrition
Sport	Performance + vóc dáng	Gym, running, fitness apps
Beauty	Da, dáng, giữ form	Skincare, beauty, wellness
Travel	Ăn nhiều, sợ béo	Travel, food exploration
Office	Ít vận động, sợ mỡ	Office life, desk job, meal prep

1.2. Tại sao cách phân này hiệu quả?

✅ Creative khác nhau - Mỗi pillar có góc nhìn riêng ✅ Behavior khác nhau - Meta dễ dàng tìm lookalike ✅ Cùng core motivation - "1/2 Fat" = Giải pháp chung cho tất cả

Meta rất thích kiểu phân tầng này vì có thể học được pattern rõ ràng từ mỗi nhóm.

---

2. Core Audience - 3 Layers Data Structure

Vấn đề với cách làm cũ

Trước đây chỉ dùng:

Core = iTVC viewers (người xem video)

→ Core sẽ rộng nhưng loãng vì Video View là tín hiệu yếu.

2.1. Độ mạnh của các Signal trên Meta

Signal Type	Chất lượng	Độ chuyển đổi
View 3s	⭐ Rất yếu	Rất thấp
ThruPlay / 15s	⭐⭐ Trung bình	Thấp
Video ≥50%	⭐⭐⭐ Tốt	Trung bình
Click / View content	⭐⭐⭐⭐ Rất tốt	Cao
Add to cart	⭐⭐⭐⭐⭐ Rất mạnh	Rất cao
Purchase	⭐⭐⭐⭐⭐⭐ Cực mạnh	Cực cao

2.2. Nâng cấp: Core Audience 3 Layers

Layer A – Soft Core (Nhận biết)

Mục đích: Build reach, tạo lookalike rộng

Điều kiện:

• Xem iTVC ≥ 3s hoặc ≥ 25%
• Xem bất kỳ pillar video nào
• Impression ≥ 2 lần

Dùng để:

• Build brand awareness
• Tạo Lookalike 1-3% để mở rộng
• Test creative mới

Audience Size: Lớn (100K - 500K)

---

Layer B – Warm Core (Quan tâm)

Mục đích: Retarget người có hành vi tích cực

Điều kiện:

• Xem ≥ 50% iTVC
• Click vào post
• Save post
• Comment
• Visit landing page
• Share content

Dùng để:

• Advantage+ audience seeding
• Retarget theo từng pillar
• Push sang Layer C

Audience Size: Trung bình (20K - 100K)

---

Layer C – Hot Core (Có ý định)

Mục đích: Drive conversion

Điều kiện:

• Add to cart
• Scroll ≥ 75% trên landing page
• Time on site > 1 phút
• Messenger / DM interaction
• Click "Mua ngay" hoặc "Tìm hiểu thêm"

Dùng để:

• Offer campaigns
• Conversion ads
• Combo deal / promotion
• Direct sale

Audience Size: Nhỏ nhưng chất lượng cao (5K - 30K)

---

3. Vertical Retargeting Loop

3.1. Vấn đề với cách setup cũ

Cách cũ:

1 adset core → chứa 12-15 post của tất cả các pillar

Vấn đề:

• Meta không biết ai thích post nào
• Creative không targeting đúng người
• Budget phân tán không hiệu quả

3.2. Vertical Retargeting Loop - Cách làm mới

Nguyên tắc: Mỗi pillar có loop riêng

Ví dụ với SPORT Pillar:

Adset Name: Retarget_Sport_Warm
Audience:
  - Core Layer B + C
  - Đã xem iTVC
  - Interest boost: Sport, Gym, Running, Fitness
Creative:
  - Chỉ post về Sport
  - Message: "Tập luyện chăm chỉ? Ăn salad thỏa thích với 1/2 Fat"

Meta sẽ học:

"Người đã quan tâm Kewpie 1/2 Fat + có behavior sport → thích content sport hơn"

3.3. Cấu trúc Vertical Loop cho tất cả Pillar

Pillar	Audience	Interest Boost	Creative Angle
Sport	Core B+C + Sport behavior	Gym, Running, Fitness	"Tập chăm - Ăn ngon - 1/2 béo"
Beauty	Core B+C + Beauty behavior	Skincare, Wellness	"Đẹp từ trong ra - Ăn ít mỡ hơn"
Travel	Core B+C + Travel behavior	Travel, Food tour	"Du lịch ăn thoải mái - 1/2 Fat"
Office	Core B+C + Office behavior	Office life, Meal prep	"Văn phòng ít động - Salad ít béo"
Food & Healthy	Core B+C + Foodie behavior	Healthy eating, Recipe	"Ngon mà healthy - Chọn 1/2 Fat"

Kết quả:

• Meta học được preference rõ ràng
• Creative match với audience
• Không bị overlap giữa các pillar

---

3.4. QUAN TRỌNG: Pillar Data Feedback Loop ⭐⭐⭐

Đây là phần cực kỳ quan trọng mà nhiều người bỏ qua!

Vấn đề với one-way flow

Nếu chỉ chạy:

Layer B → Pillar Campaigns

→ Bạn đang bỏ phí data vàng từ pillar campaigns!

Bi-Directional Data Flow

Cách đúng:

Layer B → Pillar Campaigns
   ↑            ↓
   └────────────┘
   Data Feedback Loop

Tại sao Pillar Data cực kỳ có giá trị?

Khi bạn chạy 5 pillar campaigns:

• Một số người engage rất cao với Sport pillar → họ là Sport lovers
• Một số người engage cao với Beauty pillar → họ là Beauty lovers
• Một số người engage cả Sport + Beauty → multi-interest

Data này cho bạn biết:

1. Pillar preference thực tế (không phải đoán qua interest targeting)
2. Engagement quality của từng pillar
3. Conversion rate của từng pillar
4. Best performing pillar để scale

Cách capture Pillar Engagement Data

Step 1: Tag mỗi Pillar Post với Custom Event

// Sport Pillar Post
fbq('trackCustom', 'PillarEngagement', {
  pillar: 'Sport',
  content_id: 'sport_post_001',
  engagement_type: 'video_view_50'
});

fbq('trackCustom', 'PillarEngagement', {
  pillar: 'Sport',
  content_id: 'sport_post_001',
  engagement_type: 'click'
});

// Beauty Pillar Post
fbq('trackCustom', 'PillarEngagement', {
  pillar: 'Beauty',
  content_id: 'beauty_post_001',
  engagement_type: 'video_view_50'
});

// Tương tự cho Travel, Office, Food&Healthy

Step 2: Create Pillar-Specific Audiences

Tạo 5 Custom Audiences:

Audience Name	Condition	Size Goal	Use Case
Sport_Engaged	PillarEngagement: pillar=Sport, ≥2 times, 14 days	5K - 20K	Warm Core Sport
Beauty_Engaged	PillarEngagement: pillar=Beauty, ≥2 times, 14 days	5K - 20K	Warm Core Beauty
Travel_Engaged	PillarEngagement: pillar=Travel, ≥2 times, 14 days	5K - 20K	Warm Core Travel
Office_Engaged	PillarEngagement: pillar=Office, ≥2 times, 14 days	5K - 20K	Warm Core Office
Food_Engaged	PillarEngagement: pillar=Food, ≥2 times, 14 days	5K - 20K	Warm Core Food

Cách dùng Pillar Data để Enrich Layer B & C

Strategy 1: Refined Layer B (Pillar-Tagged)

Thay vì:

Layer B = Generic warm audience

Làm:

Layer B_Sport = Warm Core + Sport_Engaged
Layer B_Beauty = Warm Core + Beauty_Engaged
Layer B_Travel = Warm Core + Travel_Engaged
Layer B_Office = Warm Core + Office_Engaged
Layer B_Food = Warm Core + Food_Engaged

Lợi ích:

• Biết chính xác người này thích pillar nào
• Retarget với content đúng angle
• Push vào Hot Core với offer phù hợp

Strategy 2: Pillar Performance Analysis

Track metrics cho mỗi pillar:

Weekly Pillar Report:

Sport Pillar:
  - Engagement Rate: 5.2%
  - CTR: 2.8%
  - Layer B → C conversion: 22%
  - ROAS: 3.5x
  → Action: Scale budget +20%

Beauty Pillar:
  - Engagement Rate: 3.1%
  - CTR: 1.9%
  - Layer B → C conversion: 15%
  - ROAS: 2.2x
  → Action: Refresh creative

Travel Pillar:
  - Engagement Rate: 6.8% ⭐
  - CTR: 3.5% ⭐
  - Layer B → C conversion: 28% ⭐
  - ROAS: 4.2x ⭐
  → Action: SCALE AGGRESSIVELY

Budget reallocation:

Week 1: Equal budget (40$/pillar)
Week 2: Based on performance
  - Travel: $60 (best performer)
  - Sport: $50
  - Food: $45
  - Beauty: $30
  - Office: $25

Strategy 3: Push Pillar Engagers to Layer C

Setup:

Campaign: Pillar_to_Hot
Objective: Conversions
Optimization: ViewContent (landing page)

Adset 1: Sport_Engaged_to_Hot
  Audience:
    - Sport_Engaged (≥2 interactions)
    - NOT in Layer C yet
  Creative:
    - Sport-specific LP
    - "Tập chăm - Ăn ngon - Giảm 1/2 mỡ"
    - Strong CTA: "Xem công thức Sport Salad"

Adset 2: Beauty_Engaged_to_Hot
  Audience:
    - Beauty_Engaged (≥2 interactions)
    - NOT in Layer C yet
  Creative:
    - Beauty-specific LP
    - "Đẹp da - Giữ dáng - 1/2 Fat"
    - CTA: "Xem bí quyết"

// Tương tự cho 3 pillars còn lại

Kết quả:

• Conversion rate cao hơn 40-60% so với generic push
• Vì message match chính xác với động cơ của họ

Strategy 4: Pillar Lookalike

Tạo Lookalike từ best-performing pillar:

# Nếu Travel pillar perform tốt nhất

Lookalike Source: Travel_Engaged + Purchased
Size: 1% → 2% → 3%

Campaign: LAL_Travel_Scale
Objective: Conversions
Audience: LAL 1% Travel

Creative:
  - Travel angle (vì lookalike giống Travel engagers)
  - NOT generic creative

→ ROAS thường cao hơn 50-80% so với generic LAL

Strategy 5: Feed Pillar Data vào Advantage+

Enhanced Advantage+ Seeding:

Campaign: AdvPlus_PillarEnhanced

Audience Suggestion:
  - Core seed: Layer B (generic warm)
  - Pillar boost:
      * Sport_Engaged: 30%
      * Travel_Engaged: 40% (best performer)
      * Beauty_Engaged: 15%
      * Office_Engaged: 10%
      * Food_Engaged: 5%
  - Exclude: Layer C

Creative Mix:
  - 40% Travel-angle creative (best pillar)
  - 30% Sport-angle
  - 30% Mixed

→ Meta learns: "Travel-type people convert best"
→ Auto-find more Travel-type lookalikes

Feedback Loop trong thực tế

Pillar Feedback Loop - Setup chi tiết

Week 1: Setup & Initial Data

Action:
  - Launch 5 pillar campaigns
  - Install pillar tracking pixels
  - Wait for data (minimum 1000 impressions/pillar)

Track:
  - Engagement rate per pillar
  - CTR per pillar
  - Cost per engagement per pillar

Week 2: Create Pillar Audiences

Action:
  - Create 5 Custom Audiences (Sport/Beauty/Travel/Office/Food engaged)
  - Check audience size (minimum 1K per pillar)
  - Analyze pillar performance

Track:
  - Audience growth rate
  - Overlap between pillars
  - Best & worst performers

Week 3: Feedback Implementation

Action:
  - Tag Layer B by pillar preference
  - Launch pillar-specific Layer C push campaigns
  - Create LAL from best-performing pillar
  - Adjust budget allocation

Track:
  - Layer B → C conversion rate per pillar
  - ROAS per pillar
  - LAL performance

Week 4: Advantage+ Enhancement

Action:
  - Seed Advantage+ with pillar data
  - Weight towards best-performing pillar
  - Adjust creative mix based on pillar performance

Track:
  - Advantage+ ROAS improvement
  - Scale efficiency
  - New user quality

Key Metrics for Pillar Feedback Loop

Metric	Formula	Good	Excellent
Pillar Engagement Rate	(Engaged / Reached) × 100	3%	5%+
Pillar → Layer C Rate	(Layer C from pillar / Pillar engaged) × 100	15%	25%+
Pillar ROAS	Revenue from pillar / Pillar spend	2.5x	4.0x+
Best Pillar Lift	Best ROAS / Avg ROAS	1.3x	1.8x+

Tóm lại: Tại sao Pillar Feedback Loop quan trọng?

Không có Feedback Loop:

Layer B → Pillar → ??? (data bị bỏ phí)

• Không biết pillar nào tốt
• Budget phân đều không hiệu quả
• Không tối ưu được Advantage+
• Bỏ lỡ cơ hội scale

Có Feedback Loop:

Layer B → Pillar → Data → Optimize → Scale → Layer B (richer)

• ✅ Biết pillar nào convert tốt nhất
• ✅ Scale đúng pillar, cut pillar yếu
• ✅ Tạo LAL từ best pillar
• ✅ Feed Advantage+ đúng data
• ✅ Layer B ngày càng refined
• ✅ ROAS tăng 50-100%

Đây chính là điểm khác biệt giữa "chạy ads" và "build data engine"!

---

4. Advantage+ Strategy

4.1. Cách dùng hiện tại

Advantage+ = Người xem video + thêm interest để mở rộng

→ Đúng nhưng chưa đủ!

4.2. Advantage+ với điều kiện khóa

Cấu trúc tối ưu:

Campaign: Advantage+ Shopping
Audience Seeding:
  - Source: Layer B (50% view, click, save)
  - Exclude: Layer C (người sắp mua - để tránh overlap)
  - Interest suggestion: Food, Healthy, Salad, Gym
  - Age: 25-45
  - Location: Major cities
Budget: 50-70% total budget

Lợi ích:

• ✅ Không bị loãng (vì seed từ Layer B - warm audience)
• ✅ Không overlap retarget (exclude Layer C)
• ✅ Scale đúng người có khả năng convert
• ✅ Meta tự optimize placement & creative

4.3. Advantage+ vs Manual Campaign

Yếu tố	Advantage+	Manual Campaign
Audience	Auto expand từ seed	Fixed targeting
Placement	Auto optimize	Manual select
Budget	Campaign level	Adset level
Learning	Nhanh hơn	Chậm hơn
Control	Thấp	Cao
Dùng khi	Scale sau khi có data	Test & retarget cụ thể

---

5. Data Engine - Beyond Video

5.1. Vấn đề khi chỉ dùng Video View

Nếu chỉ dùng iTVC & post view → bỏ phí 2 signal source cực mạnh:

5.2. Landing Page Mini - Signal Engine

Setup:

Landing Page: "Ăn salad không béo - Vì sao 1/2 Fat tốt hơn?"

Sections:
1. Hero: "Ăn salad thỏa thích - Chỉ 1/2 lượng mỡ"
2. Problem: "Nước sốt salad thường rất nhiều calo"
3. Solution: "Kewpie 1/2 Fat - Ngon mà nhẹ hơn 50%"
4. Social proof: Review, before/after
5. CTA: "Tìm mua ngay" / "Xem công thức"

Pixel Events:

Event	Trigger	Giá trị
PageView	Load page	Layer B entry
ViewContent	Scroll 25%	Layer B confirmed
Scroll75	Scroll 75%	Layer C entry ⭐
TimeOnPage_60s	Stay > 60s	Layer C confirmed ⭐
ClickCTA	Click "Mua ngay"	Layer C - Hot ⭐⭐
AddToCart	Add product	Layer C - Very Hot ⭐⭐⭐

Kết quả:

• Event mạnh hơn video rất nhiều
• Data chính xác hơn về intent
• Có thể build Conversion campaign

5.3. Mini Interaction - Quiz & Button

Ví dụ: Interactive Quiz

"Bạn thuộc nhóm nào?"

[Button] Sport - Tập gym, chạy bộ
[Button] Beauty - Chăm sóc da, giữ dáng
[Button] Travel - Thích ăn uống, du lịch
[Button] Office - Làm văn phòng, ít vận động
[Button] Foodie - Đam mê ẩm thực healthy

Sau khi chọn:

• Show content phù hợp với pillar đó
• Tag user vào Custom Audience theo pillar
• Retarget với creative đúng angle

Pixel Event:

fbq('trackCustom', 'PillarSelection', {
  pillar: 'Sport',
  product: 'Kewpie_HalfFat'
});

Lợi ích:

• ✅ Build Custom Audience theo pillar thực tế (không chỉ dựa interest)
• ✅ Tăng engagement
• ✅ Data chính xác hơn rất nhiều

---

6. Cấu trúc Funnel hoàn chỉnh

6.1. Visual Flow - WITH Pillar Feedback Loop

6.1.1. Giải thích Flow

Phase 1: Data Collection (Tuần 1-2)

iTVC → Layer A → Layer B → Pillar Campaigns

• Build initial audiences
• Test 5 pillars
• Collect engagement data

Phase 2: Pillar Feedback (Tuần 3-4)

Pillar Campaigns → Engagement Tracking → Pillar-Specific Audiences → Enrich Layer B

• Identify best-performing pillar (ví dụ: Travel)
• Tag users theo pillar preference
• Create refined Layer B segments

Phase 3: Optimization (Tuần 5-6)

Layer B_Pillar → Push to Layer C với pillar-specific LP
Best Pillar → Create LAL → Scale
Pillar Data → Advantage+ Seed → Auto-expand

• Push pillar engagers to Hot Core
• Scale best pillar aggressively
• Cut hoặc refresh weak pillars

Phase 4: Continuous Loop (Tuần 7+)

New users → Layer A → Layer B → Pillar test → Best pillar → Enrich → Repeat

• Liên tục test & refine
• Budget allocation theo performance
• Layer B ngày càng smart hơn

6.2. Campaign Structure

📁 Campaign Level
│
├── 🎯 Phase 1: Awareness & Data Collection
│   ├── iTVC Broad (All audiences)
│   ├── Interest Targeting (5 Pillars)
│   └── Lookalike 1-3% (from Layer A)
│
├── 🔥 Phase 2: Engagement & Vertical Retargeting
│   ├── Vertical Loop - Sport (Layer B)
│   ├── Vertical Loop - Beauty (Layer B)
│   ├── Vertical Loop - Travel (Layer B)
│   ├── Vertical Loop - Office (Layer B)
│   └── Vertical Loop - Food&Healthy (Layer B)
│
├── 🚀 Phase 3: Scale with Advantage+
│   └── Advantage+ Shopping (Seed from Layer B, Exclude Layer C)
│
└── 💰 Phase 4: Conversion
    ├── Landing Page → Layer C
    ├── Offer Campaign (Layer C)
    └── Cart Abandonment (Layer C - No purchase)

---

7. Detailed Campaign Setup

7.1. Phase 1 - Awareness

Campaign 1: iTVC Broad

Campaign Name: Kewpie_iTVC_Awareness
Objective: Video Views
Budget: $100/day
Optimization: ThruPlay

Adset 1: Broad_All
  Audience:
    - Location: Major cities
    - Age: 22-45
    - Interest: Broad (Food, Healthy living)
  Placement: Feed + Reels

Creative:
  - Main brand video (iTVC)
  - Duration: 30s
  - Message: "1/2 Fat = Ăn thoải mái"

Campaign 2: Interest Targeting (5 Pillars)

Campaign Name: Kewpie_Interest_5Pillars
Objective: Video Views → Engagement
Budget: $150/day (30/adset)

Adset 1: Sport
  Interest: Gym, Fitness, Running, Healthy lifestyle
  Creative: Sport angle video

Adset 2: Beauty
  Interest: Beauty, Skincare, Wellness
  Creative: Beauty angle video

Adset 3: Travel
  Interest: Travel, Food tourism
  Creative: Travel angle video

Adset 4: Office
  Interest: Office, Work-life balance
  Creative: Office angle video

Adset 5: Food & Healthy
  Interest: Cooking, Healthy recipes, Nutrition
  Creative: Food angle video

Campaign 3: Lookalike Expansion

Campaign Name: Kewpie_LAL_Scale
Objective: Reach
Budget: $80/day

Adset 1: LAL 1% - Video Viewers
  Source: Layer A (25% video viewers)
  Exclude: Layer B, C

Adset 2: LAL 2-3% - Video Viewers
  Source: Layer A
  Exclude: Layer B, C

7.2. Phase 2 - Vertical Retargeting

Campaign Name: Kewpie_Vertical_Retarget
Objective: Traffic / Engagement
Budget: $200/day (40/adset)

Adset 1: Sport_Warm
  Audience:
    - Include: Layer B + iTVC viewers
    - Interest: Sport, Gym, Fitness
    - Exclude: Layer C
  Creative: Sport posts (3-5 posts rotation)
  Placement: Feed, Reels, Stories

Adset 2: Beauty_Warm
  Audience:
    - Include: Layer B + iTVC viewers
    - Interest: Beauty, Skincare
    - Exclude: Layer C
  Creative: Beauty posts

Adset 3: Travel_Warm
  Audience:
    - Include: Layer B + iTVC viewers
    - Interest: Travel
    - Exclude: Layer C
  Creative: Travel posts

Adset 4: Office_Warm
  Audience:
    - Include: Layer B + iTVC viewers
    - Interest: Office, Career
    - Exclude: Layer C
  Creative: Office posts

Adset 5: FoodHealthy_Warm
  Audience:
    - Include: Layer B + iTVC viewers
    - Interest: Healthy eating, Recipes
    - Exclude: Layer C
  Creative: Food posts

7.2B. Phase 2+ - Pillar Feedback Campaigns (Tuần 3+)

Sau khi có data từ Phase 2 (1-2 tuần), launch campaigns này:

Campaign 2A: Refined Pillar Retargeting (với Pillar-Tagged Audiences)

Campaign Name: Kewpie_Pillar_Refined
Objective: Traffic / Conversions
Budget: $250/day
Optimization: Landing Page Views

Adset 1: Sport_Tagged_Retarget
  Audience:
    - Layer B_Sport (Layer B + Sport_Engaged)
    - Exclude: Layer C
  Creative:
    - Sport-specific creative (TOP performers only)
    - CTA: "Công thức salad cho Gym-goers"
  Destination: Sport-themed landing page

Adset 2: Beauty_Tagged_Retarget
  Audience:
    - Layer B_Beauty
    - Exclude: Layer C
  Creative: Beauty-specific
  Destination: Beauty-themed landing page

Adset 3: Travel_Tagged_Retarget ⭐ (Highest Budget)
  Audience:
    - Layer B_Travel
    - Exclude: Layer C
  Creative: Travel-specific
  Budget: 2x other adsets (vì best performer)
  Destination: Travel-themed landing page

Adset 4: Office_Tagged_Retarget
  Audience: Layer B_Office
  Creative: Office-specific

Adset 5: Food_Tagged_Retarget
  Audience: Layer B_Food
  Creative: Food-specific

Budget Allocation (Based on Performance):
  - Travel: $80/day (32% - best ROAS)
  - Sport: $60/day (24%)
  - Food: $50/day (20%)
  - Beauty: $40/day (16%)
  - Office: $20/day (8% - worst ROAS)

→ Dynamic allocation theo weekly performance

Campaign 2B: Pillar to Hot Core Push

Campaign Name: Kewpie_Pillar_to_HotCore
Objective: Conversions
Optimization: ViewContent → AddToCart
Budget: $150/day

Adset 1: Sport_Engaged_Push
  Audience:
    - Sport_Engaged (≥3 interactions)
    - NOT in Layer C
  Creative:
    - Strong CTA: "Tập chăm - Ăn ngon - Giảm mỡ"
    - Sport influencer testimonial
    - Limited offer: "Giảm 15% cho gym members"
  Destination: Sport LP with ATC button

Adset 2: Travel_Engaged_Push ⭐
  Audience:
    - Travel_Engaged (≥3 interactions)
    - NOT in Layer C
  Creative:
    - "Du lịch ăn thoải mái - 0 áy náy"
    - Food tour imagery
    - Offer: "Mua 2 chai - Tặng recipe book du lịch"
  Budget: 2x other adsets
  Destination: Travel LP with ATC

Adset 3-5: Beauty/Office/Food_Engaged_Push
  [Similar structure]

Expected Results:
  - Layer B → C conversion: 20-30%
  - ROAS: 3.5-5.0x
  - CPA: ₫25,000-35,000

Campaign 2C: Pillar Lookalike Scaling

Campaign Name: Kewpie_LAL_Pillar_Scale
Objective: Conversions
Optimization: ViewContent
Budget: $120/day

Adset 1: LAL_Travel_Engaged_1pct ⭐⭐
  Source: Travel_Engaged + Travel_Purchasers
  Size: 1%
  Creative: Travel angle
  Budget: $60/day (50% of campaign)
  Note: Best performing pillar LAL

Adset 2: LAL_Sport_Engaged_1pct
  Source: Sport_Engaged + Sport_Purchasers
  Size: 1%
  Creative: Sport angle
  Budget: $30/day

Adset 3: LAL_Travel_Engaged_2-3pct
  Source: Travel_Engaged
  Size: 2-3%
  Creative: Travel angle
  Budget: $30/day
  Note: Scale test for best pillar

Expected Results:
  - Pillar LAL outperforms generic LAL by 40-70%
  - Travel LAL ROAS: 3.0-4.5x
  - Other pillar LAL ROAS: 2.0-3.5x

7.3. Phase 3 - Advantage+ Scale (Enhanced with Pillar Data)

Version A: Generic Advantage+ (Tuần 3-4)

Campaign Name: Kewpie_AdvPlus_Generic
Campaign Type: Advantage+ Shopping
Budget: $200/day
Optimization: Conversions (ViewContent)

Audience Suggestion:
  - Seed: Layer B (warm core generic)
  - Exclude: Layer C (to prevent overlap)
  - Age: 25-45
  - Interest suggestions:
      * Healthy eating
      * Salad
      * Low-fat food
      * Fitness

Creative:
  - 8-10 variations
  - Mix video + image
  - Different hooks for different pillars
  - Let Meta auto-optimize

Placement: Automatic (Advantage+ Placements)

Expected Results:
  - ROAS: 2.5-3.5x
  - CPA: ₫15,000-20,000
  - Scale potential: Moderate

Version B: Pillar-Enhanced Advantage+ ⭐ (Tuần 5+, sau khi có pillar data)

Campaign Name: Kewpie_AdvPlus_PillarEnhanced
Campaign Type: Advantage+ Shopping
Budget: $400/day
Optimization: Conversions (ViewContent → Purchase)

Audience Suggestion (Weighted):
  Primary Seeds (60%):
    - Layer B_Travel (30%) ⭐ Best pillar
    - Layer B_Sport (20%)
    - Layer B_Food (10%)

  Secondary Seeds (30%):
    - Travel_Engaged (15%)
    - Sport_Engaged (10%)
    - Beauty_Engaged (5%)

  Safety Net (10%):
    - Layer B generic (những người chưa rõ pillar)

  Exclude:
    - Layer C (hot core - đang được target riêng)
    - Purchased last 30 days
    - Low quality users (bounce >80%)

  Demographics:
    - Age: 25-45
    - No other restrictions (let Meta expand)

Creative Mix (Based on Pillar Performance):
  Primary (50%):
    - 5 Travel-angle creatives (best performer)
  Secondary (30%):
    - 3 Sport-angle creatives
  Tertiary (20%):
    - 2 Beauty/Food/Office mixed

Interest Suggestions (Lightly weighted):
  - Travel & food tourism (match best pillar)
  - Healthy lifestyle
  - Fitness & wellness

  Note: Don't over-constrain, let Meta expand

Placement: Automatic (Advantage+ Placements)
  - Meta will auto-optimize
  - Usually: Feed (60%), Reels (30%), Stories (10%)

Budget Strategy:
  - Start: $200/day
  - Week 1: Monitor ROAS
  - Week 2: Scale to $400/day if ROAS ≥ 3.0x
  - Week 3+: Scale to $600-800/day if stable

Expected Results:
  - ROAS: 3.5-5.0x (40-60% lift vs generic)
  - CPA: ₫12,000-15,000 (20-30% improvement)
  - Scale potential: Very high
  - Meta learns: "Travel-type people convert best"
  - Auto-expands to Travel lookalikes

Why Pillar-Enhanced Advantage+ Works Better?

Generic Advantage+:

Seed: Generic Layer B
Meta learns: "People who engaged with Kewpie content"
Expands to: Similar people (broad)
ROAS: 2.5-3.5x

Pillar-Enhanced Advantage+:

Seed: Weighted towards best pillar (Travel 30%, Sport 20%, etc.)
Meta learns: "Travel-type people convert 70% better than others"
Expands to: More Travel-type lookalikes
ROAS: 3.5-5.0x ⭐

Data shows:

• Pillar-enhanced Advantage+ outperforms generic by 40-80%
• Faster learning phase (5 days vs 7-10 days)
• Better scale efficiency
• Lower CPA at higher budgets

Testing Framework: Generic vs Pillar-Enhanced

Week 3-4: A/B Test
  Campaign A: Generic Advantage+ ($200/day)
  Campaign B: Pillar-Enhanced Advantage+ ($200/day)

  Run for 7-14 days
  Compare:
    - ROAS
    - CPA
    - Conversion rate
    - Scale efficiency

Week 5+: Winner Takes All
  - Pause loser
  - Scale winner to $400-800/day

Expected Winner: Pillar-Enhanced (80% probability)

7.3B. Advantage+ Optimization Tips

Tip 1: Update Pillar Weights Weekly

Week 5: Travel 30%, Sport 20%, Food 10%
Week 6: Travel 35%, Sport 15%, Food 10% (Travel getting better)
Week 7: Travel 40%, Sport 15%, Food 5% (Double down on winner)

Tip 2: Seasonal Adjustment

Tết Holiday:
  - Travel 50% (people traveling)
  - Food 30% (eating season)
  - Sport 10% (less gym)
  - Office 5%
  - Beauty 5%

Post-Tết (Feb-Mar):
  - Sport 40% (gym resolutions)
  - Beauty 25% (get back in shape)
  - Travel 15%
  - Office 15%
  - Food 5%

Tip 3: Creative Rotation Based on Pillar

// Track which creative → which pillar converts best
Creative A (Travel angle) → Travel audience = 5.2x ROAS ⭐
Creative A (Travel angle) → Sport audience = 2.8x ROAS

Creative B (Sport angle) → Sport audience = 4.1x ROAS ⭐
Creative B (Sport angle) → Travel audience = 2.5x ROAS

→ Use Travel creative for Travel-heavy Advantage+
→ Use Sport creative for Sport-heavy Advantage+

7.4. Phase 4 - Conversion

Campaign 4A: Landing Page → Layer C

Campaign Name: Kewpie_LP_Conversion
Objective: Conversions
Optimization: ViewContent
Budget: $150/day

Adset 1: Warm_to_Hot
  Audience: Layer B
  Exclude: Layer C (already hot)
  Destination: Landing Page

Creative:
  - Strong CTA: "Xem vì sao 1/2 Fat tốt hơn"
  - Benefit-focused
  - Before/after social proof

Campaign 4B: Offer & Promotion

Campaign Name: Kewpie_Offer_Hot
Objective: Conversions
Optimization: Purchase / AddToCart
Budget: $200/day

Adset 1: Hot_Audience
  Audience: Layer C only

Creative:
  - Limited offer
  - "Mua 2 tặng 1"
  - Free shipping
  - Urgency: "Chỉ còn 3 ngày"

Campaign 4C: Cart Abandonment

Campaign Name: Kewpie_Cart_Recovery
Objective: Conversions
Optimization: Purchase
Budget: $100/day

Adset 1: Cart_Abandoners
  Audience:
    - Layer C
    - Added to cart
    - No purchase in last 7 days

Creative:
  - "Bạn đã quên giỏ hàng?"
  - Extra 10% discount
  - Easy checkout reminder

---

8. Budget Allocation Strategy

8.1. Phân bổ theo Phase

Phase	Budget %	Daily Budget	Purpose
Phase 1: Awareness	25%	$330/day	Build Layer A & B
Phase 2: Vertical Retarget	20%	$200/day	Push B → C
Phase 3: Advantage+	35%	$300/day	Scale efficiently
Phase 4: Conversion	20%	$450/day	Drive sales
Total	100%	$1,280/day	Full funnel

8.2. Budget Scaling Timeline

Week 1-2: Testing Phase
  - Budget: $500/day
  - Focus: Phase 1 + 2
  - Goal: Build core audiences

Week 3-4: Optimization Phase
  - Budget: $800/day
  - Focus: Add Phase 3 (Advantage+)
  - Goal: Find winning combinations

Week 5+: Scale Phase
  - Budget: $1,280+/day
  - Focus: All phases
  - Goal: Maximize ROAS while scaling

---

9. Audience Custom Events Setup

9.1. Layer A - Soft Core

// Facebook Pixel - Video Events
fbq('trackCustom', 'VideoView3s', {
  video_title: 'Kewpie_iTVC_Main',
  content_category: 'awareness'
});

fbq('trackCustom', 'VideoView25Percent', {
  video_title: 'Kewpie_iTVC_Main',
  content_category: 'awareness'
});

Custom Audience:

• Event: VideoView25Percent
• Time window: Last 30 days
• Estimated size: 100K - 500K

9.2. Layer B - Warm Core

// Video engagement
fbq('trackCustom', 'VideoView50Percent', {
  video_title: 'Kewpie_iTVC_Main',
  pillar: 'Sport' // or Beauty, Travel, etc
});

// Post engagement
fbq('track', 'ViewContent', {
  content_type: 'post',
  content_ids: ['post_123'],
  pillar: 'Sport'
});

// Landing page
fbq('track', 'ViewContent', {
  content_type: 'product',
  content_name: 'Kewpie_1/2_Fat'
});

Custom Audience:

• Event: VideoView50Percent OR ViewContent OR Click
• Frequency: ≥ 2 times
• Time window: Last 14 days
• Estimated size: 20K - 100K

9.3. Layer C - Hot Core

// Scroll depth
fbq('trackCustom', 'Scroll75Percent', {
  content_type: 'landing_page',
  content_name: 'Kewpie_LP_Main'
});

// Time on page
fbq('trackCustom', 'TimeOnPage_60s', {
  duration: 60,
  page: '/kewpie-half-fat'
});

// CTA clicks
fbq('track', 'Lead', {
  content_name: 'CTA_MuaNgay',
  value: 0
});

// Add to cart
fbq('track', 'AddToCart', {
  content_ids: ['kewpie_half_fat_250ml'],
  content_type: 'product',
  value: 45000,
  currency: 'VND'
});

Custom Audience:

• Event: Scroll75Percent OR TimeOnPage_60s OR AddToCart
• Time window: Last 7 days
• Estimated size: 5K - 30K

9.4. Pillar Engagement Tracking (FEEDBACK LOOP)

Setup cho mỗi Pillar Post

Sport Pillar:

// Trên mỗi Sport post/video
fbq('trackCustom', 'PillarEngagement', {
  pillar: 'Sport',
  content_type: 'video',
  content_id: 'sport_video_001',
  engagement_type: 'video_view_50'
});

fbq('trackCustom', 'PillarClick', {
  pillar: 'Sport',
  content_id: 'sport_video_001',
  click_type: 'cta_button'
});

fbq('trackCustom', 'PillarHighIntent', {
  pillar: 'Sport',
  content_id: 'sport_video_001',
  action: 'save_post' // hoặc share, comment
});

Beauty Pillar:

fbq('trackCustom', 'PillarEngagement', {
  pillar: 'Beauty',
  content_type: 'video',
  content_id: 'beauty_video_001',
  engagement_type: 'video_view_75'
});

Travel, Office, Food & Healthy: Tương tự

Custom Audiences cho Pillar Feedback

Tạo 5 audiences:

Audience 1: Sport_Engaged
  Conditions:
    - Event: PillarEngagement
    - pillar = 'Sport'
    - Frequency: ≥ 2 times
    - Time window: 14 days
  OR
    - Event: PillarClick
    - pillar = 'Sport'
    - Time window: 14 days
  OR
    - Event: PillarHighIntent
    - pillar = 'Sport'
    - Time window: 14 days

  Size: 5K - 20K
  Use: Retarget with Sport content, push to Layer C

Audience 2: Beauty_Engaged
  [Same structure, pillar = 'Beauty']

Audience 3: Travel_Engaged
  [Same structure, pillar = 'Travel']

Audience 4: Office_Engaged
  [Same structure, pillar = 'Office']

Audience 5: Food_Engaged
  [Same structure, pillar = 'Food']

Combined Audiences (Refined Layer B)

Layer B_Sport:
  Include:
    - Layer B (Warm Core)
    - Sport_Engaged
  Exclude:
    - Layer C (Hot Core)
    - Purchased (last 30 days)

  Purpose: Sport lovers trong Warm Core
  Use: Sport-specific retargeting, Sport LP push

Layer B_Beauty:
  Include:
    - Layer B (Warm Core)
    - Beauty_Engaged
  Exclude:
    - Layer C
    - Purchased

  Purpose: Beauty lovers trong Warm Core

Layer B_Travel:
  Include:
    - Layer B (Warm Core)
    - Travel_Engaged
  Exclude:
    - Layer C
    - Purchased

  Purpose: Travel lovers trong Warm Core
  Note: Thường là best performer ⭐

Layer B_Office:
  [Same structure]

Layer B_Food:
  [Same structure]

Pillar Conversion Tracking

Track pillar nào drive conversion tốt nhất:

// Khi user add to cart
fbq('track', 'AddToCart', {
  content_ids: ['kewpie_half_fat_250ml'],
  content_type: 'product',
  value: 45000,
  currency: 'VND',
  pillar_source: 'Sport' // Pillar mà user engage nhiều nhất
});

// Khi purchase
fbq('track', 'Purchase', {
  content_ids: ['kewpie_half_fat_250ml'],
  content_type: 'product',
  value: 45000,
  currency: 'VND',
  pillar_source: 'Travel', // Track pillar contribution
  num_items: 1
});

Value-Based Lookalike từ Pillar

Tạo Value-Based Custom Audience:

Audience: Travel_Purchasers (High Value)
  Conditions:
    - Event: Purchase
    - pillar_source = 'Travel'
    - value ≥ 90000 (mua 2+ bottles)
    - Last 60 days

  Size: 1K - 5K
  Use: Create LAL 1-3% for scaling

Lookalike:
  Source: Travel_Purchasers
  Location: Vietnam
  Size: 1%, 2%, 3%

  Campaign: LAL_Travel_HighValue
  Budget: Heavy (vì conversion proven)
  Creative: Travel angle
  Expected ROAS: 1.5-2x higher than generic LAL

Dashboard Tracking cho Pillar Performance

// Custom Conversion cho từng pillar
// Ads Manager → Events Manager → Custom Conversions

Custom Conversion 1: Sport_ViewContent
  Rule:
    - URL contains: kewpie-half-fat
    - pillar_source equals: Sport
  Value: ViewContent event

Custom Conversion 2: Sport_Purchase
  Rule:
    - Event: Purchase
    - pillar_source equals: Sport
  Value: Purchase event

// Repeat for all 5 pillars

// → Bạn có thể track ROAS riêng cho từng pillar!

9.5. Audience Structure - Complete View

📁 All Audiences
│
├── 📊 Core Layers
│   ├── Layer A (Soft Core) - 100K-500K
│   ├── Layer B (Warm Core) - 20K-100K
│   └── Layer C (Hot Core) - 5K-30K
│
├── 🎯 Pillar Engagement Audiences
│   ├── Sport_Engaged - 5K-20K
│   ├── Beauty_Engaged - 5K-20K
│   ├── Travel_Engaged - 5K-20K ⭐
│   ├── Office_Engaged - 5K-20K
│   └── Food_Engaged - 5K-20K
│
├── 🔥 Refined Layer B (Pillar-Tagged)
│   ├── Layer B_Sport = Layer B + Sport_Engaged
│   ├── Layer B_Beauty = Layer B + Beauty_Engaged
│   ├── Layer B_Travel = Layer B + Travel_Engaged ⭐
│   ├── Layer B_Office = Layer B + Office_Engaged
│   └── Layer B_Food = Layer B + Food_Engaged
│
├── 💰 Pillar Converters
│   ├── Sport_Purchasers - 500-2K
│   ├── Beauty_Purchasers - 500-2K
│   ├── Travel_Purchasers - 1K-3K ⭐ (highest)
│   ├── Office_Purchasers - 300-1K
│   └── Food_Purchasers - 500-2K
│
├── 🚀 Lookalikes
│   ├── LAL 1% - Layer A (broad)
│   ├── LAL 1% - Layer B (warm)
│   ├── LAL 1% - Travel_Engaged ⭐ (best pillar)
│   ├── LAL 1% - Travel_Purchasers ⭐⭐ (value-based)
│   └── LAL 2-3% - Various sources
│
└── 🚫 Exclusions
    ├── Purchased - Last 30 days
    ├── Cart Abandoners - Last 7 days
    └── Low Quality - Bounce rate >80%

---

10. Measurement & KPIs

10.1. KPIs theo Phase

Phase 1: Awareness

Metric	Target	Good	Excellent
CPM	₫80,000	< ₫70,000	< ₫50,000
3s Video Views	100K/week	150K/week	200K+/week
ThruPlay Rate	25%	35%	45%+
Cost per ThruPlay	₫1,500	₫1,200	< ₫1,000
Layer A growth	50K/week	80K/week	100K+/week

Phase 2: Vertical Retargeting

Metric	Target	Good	Excellent
CTR (All)	1.5%	2.0%	3.0%+
Engagement Rate	3%	5%	7%+
Cost per Engagement	₫800	₫600	< ₫500
Layer B → C conversion	15%	20%	25%+

Phase 3: Advantage+

Metric	Target	Good	Excellent
ROAS	2.5x	3.5x	5.0x+
CPA (ViewContent)	₫15,000	₫12,000	< ₫10,000
Conversion Rate	3%	5%	7%+
Scale efficiency	Stable ROAS at +50% budget	Stable at +100%	Stable at +200%

Phase 4: Conversion

Metric	Target	Good	Excellent
ROAS	3.0x	4.0x	6.0x+
CPA (Purchase)	₫40,000	₫30,000	< ₫25,000
Cart → Purchase	30%	40%	50%+
LTV	3x CPA	4x CPA	5x+ CPA

10.2. Dashboard Tracking

Weekly Report Metrics:

Audience Health:
  - Layer A size & growth rate
  - Layer B size & growth rate
  - Layer C size & growth rate
  - A→B conversion rate
  - B→C conversion rate

Campaign Performance:
  - Spend by phase
  - ROAS by phase
  - CPA trends
  - Creative fatigue score (Frequency > 3)

Pillar Performance:
  - Which pillar drives most Layer C?
  - Which pillar has best ROAS?
  - Budget reallocation recommendations

Optimization Actions:
  - Creative refresh needs
  - Budget shifts
  - Audience expansion opportunities

---

11. Optimization Playbook

11.1. Khi nào cần Refresh Creative?

Signals:

Metric	Warning Sign	Action
Frequency	> 3.0	Thêm creative mới
CTR	Giảm 30% so với tuần trước	Test hook mới
CPC	Tăng 50%+	Pause ad, test new
ThruPlay Rate	< 20%	Video quá dài hoặc nhàm

Creative Refresh Schedule:

Week 1-2: 5 creatives
Week 3-4: +3 new creatives (total 8)
Week 5-6: +2 new, pause bottom 3 (total 7)
Week 7-8: +3 new, pause bottom 2 (total 8)
→ Continuous rotation

11.2. Khi nào Scale Budget?

Điều kiện:

✅ ROAS ổn định ≥ mục tiêu trong 5-7 ngày ✅ CPA không tăng > 20% khi tăng budget ✅ Audience không bão hòa (Frequency < 2.5) ✅ Layer B+C vẫn đang grow

Cách Scale:

Day 1: Baseline $500
Day 3: +20% → $600 (nếu ROAS stable)
Day 6: +20% → $720
Day 9: +20% → $864
Day 12: +20% → $1,037

✋ STOP khi:
- ROAS giảm > 20%
- CPA tăng > 30%
- Frequency > 3.5

11.3. Khi nào Expand Audience?

Timing:

• Layer A đạt 300K+
• Layer B đạt 80K+
• Layer C đạt 15K+

Expansion Strategy:

Option 1: Lookalike Tiers
  - Có LAL 1% → Test LAL 2-3%
  - Có LAL 2-3% → Test LAL 4-5%
  - Monitor ROAS carefully

Option 2: Interest Expansion
  - Có Sport → Add Yoga, Marathon
  - Có Beauty → Add Wellness, Spa
  - Có Travel → Add Food tourism, Adventure

Option 3: Broad Testing
  - Remove all interests
  - Let Meta auto-find
  - Only for Advantage+ campaigns

11.4. A/B Testing Framework

Luôn luôn test:

Element	Variations	Winner criteria
Hook (3s đầu)	3-4 versions	Highest 3s view rate
CTA	2-3 versions	Highest CTR
Creative format	Video vs Image vs Carousel	Lowest CPA
Copy length	Short vs Long	Best engagement
Offer	Discount vs Bundle vs Free ship	Highest conversion

Test cadence:

• Week 1-2: Hook test
• Week 3-4: CTA test
• Week 5-6: Format test
• Week 7-8: Combine winners

---

12. Common Mistakes & How to Avoid

12.1. ❌ Mistake 1: Chỉ dùng Video View làm Core

Vấn đề: Core quá loãng, không convert

Fix:

• ✅ Phân tầng thành Layer A/B/C
• ✅ Dùng nhiều signal types
• ✅ Ưu tiên Layer B & C cho conversion

12.2. ❌ Mistake 2: Audience Overlap

Vấn đề: Các adset compete với nhau, tăng CPA

Fix:

• ✅ Exclude Layer C khỏi Layer B campaigns
• ✅ Exclude Layer B khỏi Layer A campaigns
• ✅ Check overlap trong Audience Insights
• ✅ Sử dụng CBO (Campaign Budget Optimization)

12.3. ❌ Mistake 3: Scale quá nhanh

Vấn đề: Tăng budget 100% trong 1 ngày → Algorithm reset

Fix:

• ✅ Tăng max 20-25% mỗi lần
• ✅ Đợi 2-3 ngày mới tăng tiếp
• ✅ Monitor ROAS & CPA chặt chẽ

12.4. ❌ Mistake 4: Không exclude Purchasers

Vấn đề: Bắn ads cho người đã mua → Waste budget

Fix:

• ✅ Create "Purchased - Last 30 days" audience
• ✅ Exclude từ tất cả acquisition campaigns
• ✅ Chỉ include trong retention/upsell campaigns

12.5. ❌ Mistake 5: Creative Fatigue

Vấn đề: Frequency > 4, CTR drop, CPA tăng

Fix:

• ✅ Monitor frequency weekly
• ✅ Pause ad khi frequency > 3.5
• ✅ Có sẵn creative backup
• ✅ Rotate creative theo schedule

---

13. Advanced Tactics

13.1. Sequential Retargeting

Concept: Show different messages theo thứ tự

Step 1: iTVC → "Làm quen với Kewpie 1/2 Fat"
  ↓ (after 2 days)
Step 2: Pillar Content → "Phù hợp với bạn vì..."
  ↓ (after 3 days)
Step 3: Social Proof → "Mọi người nói gì?"
  ↓ (after 2 days)
Step 4: Offer → "Mua ngay - Ưu đãi đặc biệt"

Setup:

Adset 1: Intro
  Audience: Layer A (new viewers)
  Creative: iTVC
  Exclude: Đã xem pillar content

Adset 2: Education
  Audience:
    - Viewed iTVC 2+ days ago
    - Not viewed pillar content
  Creative: Pillar posts

Adset 3: Social Proof
  Audience:
    - Viewed pillar content 3+ days ago
    - Not clicked
  Creative: Testimonials, reviews

Adset 4: Conversion
  Audience:
    - Clicked in last 7 days
    - Not purchased
  Creative: Offer, promotion

13.2. Dynamic Creative Testing (DCT)

Setup:

Campaign: Kewpie_DCT_Test
Use: Dynamic Creative

Assets:
  Videos: 5 variations (different hooks)
  Images: 5 variations
  Headlines: 8 variations
  Descriptions: 5 variations
  CTAs: 3 variations

Total combinations: Thousands
Meta auto-optimizes: Best combinations

Best practices:

• Minimum 5 variations per element
• Let run 7+ days before analyzing
• Extract winning combinations
• Use in manual campaigns

13.3. Catalog Sales Campaign (Advanced)

Nếu có nhiều SKUs:

Campaign: Kewpie_Catalog
Objective: Catalog Sales
Audience: Layer C (hot)

Product Set:
  - Kewpie Sesame 1/2 Fat - 250ml
  - Kewpie Sesame 1/2 Fat - 500ml
  - Kewpie Sesame 1/2 Fat - Combo 2

Dynamic Ads:
  - Auto-show product người đã xem
  - Cross-sell related products
  - Bundle recommendations

---

14. Checklist trước khi Launch

14.1. Technical Setup

• Pixel installed và firing correctly
• All custom events tested (Layer A/B/C)
• Landing page ready + mobile-optimized
• Conversion API setup (nếu có)
• Test purchase flow hoàn chỉnh
• UTM parameters set đúng

14.2. Audience Setup

• Layer A audience created
• Layer B audience created
• Layer C audience created
• Exclusion audiences ready
• Lookalike audiences built (nếu có seed data)

14.3. Creative Assets

• iTVC video (30s) ready
• 5 pillar videos ready
• 10+ image variations
• Copy variations cho mỗi pillar
• CTA buttons tested

14.4. Campaign Structure

• Phase 1 campaigns set
• Phase 2 vertical retargeting set
• Phase 3 Advantage+ ready (có thể đợi data)
• Phase 4 conversion campaigns ready
• Budget allocation confirmed

14.5. Tracking & Reporting

• Dashboard setup (Google Sheets / Data Studio)
• Weekly report template ready
• Alert system cho ROAS drops
• Daily check routine scheduled

---

15. Summary - Tại sao cách này hiệu quả?

Nguyên tắc cốt lõi:

1. Data-First Approach

• Không đoán → Để Meta học
• Không broad too soon → Build core trước
• Không waste signal → Mỗi interaction đều được track

2. Layered Funnel

• Soft → Warm → Hot
• Mỗi layer có mục đích riêng
• Progression tự nhiên, không force

3. Vertical Precision

• Mỗi pillar có loop riêng
• Creative match với motivation
• Meta học được pattern rõ ràng

4. Scale Intelligently

• Advantage+ khi đã có core solid
• Exclude để tránh overlap
• Tăng budget theo rules

5. Beyond Video

• Landing page = Signal engine
• Interactive content = Pillar tagging
• Multi-source data = Strong core

6. Pillar Feedback Loop ⭐⭐⭐

• Pillar campaigns → Track engagement → Tag users → Enrich Layer B
• Best pillar data → Lookalike → Scale aggressively
• Pillar data → Advantage+ seed → Better expansion
• Weak pillar → Refresh or cut → Reallocate budget

Điểm khác biệt: One-Way vs Bi-Directional Flow

❌ Cách cũ (One-Way Flow):

Layer B → Pillar Campaigns → ??? (data bị lãng phí)

Problems:

• Không biết pillar nào tốt
• Budget phân đều = không hiệu quả
• Không optimize được Advantage+
• Miss cơ hội scale

✅ Cách mới (Bi-Directional Flow):

Layer B → Pillar Campaigns → Engagement Data → Tag users
   ↑                                                  ↓
   └────────────── Enrich Layer B ←──────────────────┘

Benefits:

• Biết chính xác pillar nào drive conversion
• Scale best pillar, cut worst pillar
• Layer B ngày càng refined
• Advantage+ học đúng pattern
• ROAS tăng 50-100%

Kết quả thực tế:

Without Pillar Feedback:
  - Generic targeting
  - ROAS: 2.0-2.5x
  - CPA: ₫30,000-40,000
  - Scale limit: 2-3x initial budget

With Pillar Feedback:
  - Pillar-specific targeting
  - ROAS: 3.5-5.0x ⭐
  - CPA: ₫15,000-25,000 ⭐
  - Scale limit: 5-10x initial budget ⭐
  - Travel pillar = 70% of conversions
  - Cut Office pillar (poor ROAS)
  - Reallocate budget to Travel

Kết quả mong đợi:

Tháng 1: Build foundation
  - 300K+ Layer A
  - 80K+ Layer B
  - 15K+ Layer C
  - ROAS: 2.0-2.5x

Tháng 2: Optimize & Scale
  - 500K+ Layer A
  - 150K+ Layer B
  - 40K+ Layer C
  - ROAS: 2.5-3.5x
  - Budget: 2x

Tháng 3+: Mature Funnel
  - Continuous growth
  - ROAS: 3.5-5.0x
  - Budget: 3-5x
  - Profitable scale

---

16. Next Steps

1. Week 1: Setup Phase 1 (Awareness)
2. Week 2: Launch Phase 2 (Vertical Retargeting)
3. Week 3: Analyze data, build Layer B/C
4. Week 4: Launch Phase 3 (Advantage+)
5. Week 5: Launch Phase 4 (Conversion)
6. Week 6+: Optimize & Scale

Quan trọng nhất:

Đây không phải "set & forget" system Đây là "learn & optimize" system Cần monitor, test, adjust liên tục

---

17. Quick Implementation Guide - Pillar Feedback Loop

17.1. Checklist - Tuần 1-2 (Foundation)

Setup Phase:

• Day 1-3: Launch Phase 1

  • iTVC campaign live
  • 5 Pillar interest campaigns live
  • Pixel firing correctly
  • Basic events tracking (VideoView, Click)

• Day 4-7: Install Pillar Tracking

  • Add PillarEngagement event to all 5 pillar posts
  • Add pillar parameter to all tracking events
  • Test: Check Events Manager for pillar data
  • Verify: See pillar breakdown in custom parameters

• Day 8-14: Collect Initial Data

  • Let campaigns run
  • Minimum 1,000 impressions per pillar
  • Minimum 50 engagements per pillar
  • Do NOT optimize yet - just collect data

Expected by end of Week 2:

• Layer A: 20K-50K users
• Layer B: 3K-10K users
• Pillar engagement data: Ready for analysis

17.2. Checklist - Tuần 3 (Pillar Analysis)

Analysis Phase:

• Day 15-16: Analyze Pillar Performance
  • Export data from Ads Manager
  • Calculate per pillar:
    • Engagement rate
    • CTR
    • Cost per engagement
    • Video completion rate
  • Rank pillars: Best → Worst

Example Analysis:

Pillar Performance Week 1-2:

1. Travel ⭐⭐⭐
   - Engagement: 6.8%
   - CTR: 3.2%
   - CPE: ₫450
   - Rank: #1

2. Sport ⭐⭐
   - Engagement: 4.5%
   - CTR: 2.1%
   - CPE: ₫580
   - Rank: #2

3. Food & Healthy ⭐⭐
   - Engagement: 4.1%
   - CTR: 1.9%
   - CPE: ₫620
   - Rank: #3

4. Beauty ⭐
   - Engagement: 2.8%
   - CTR: 1.3%
   - CPE: ₫850
   - Rank: #4

5. Office
   - Engagement: 2.1%
   - CTR: 0.9%
   - CPE: ₫1,100
   - Rank: #5 (Consider cutting)

• Day 17-18: Create Pillar Audiences

  • Create Custom Audience: Sport_Engaged
  • Create Custom Audience: Beauty_Engaged
  • Create Custom Audience: Travel_Engaged
  • Create Custom Audience: Office_Engaged
  • Create Custom Audience: Food_Engaged
  • Wait 24h for audiences to populate

• Day 19-21: Create Refined Layer B

  • Create: Layer B_Sport = Layer B + Sport_Engaged
  • Create: Layer B_Beauty = Layer B + Beauty_Engaged
  • Create: Layer B_Travel = Layer B + Travel_Engaged ⭐
  • Create: Layer B_Office = Layer B + Office_Engaged
  • Create: Layer B_Food = Layer B + Food_Engaged
  • Check sizes: Each should have 1K-5K users

Expected by end of Week 3:

• 5 Pillar audiences created
• 5 Refined Layer B audiences ready
• Performance ranking established

17.3. Checklist - Tuần 4 (Implement Feedback)

Implementation Phase:

• Day 22-23: Launch Refined Retargeting

  • Campaign: Kewpie_Pillar_Refined
  • 5 Adsets (one per refined Layer B)
  • Budget allocation:
    • Travel (best): 35%
    • Sport: 25%
    • Food: 20%
    • Beauty: 15%
    • Office: 5% (or pause)

• Day 24-25: Launch Pillar-to-Hot Push

  • Campaign: Kewpie_Pillar_to_HotCore
  • Target: Pillar_Engaged audiences
  • Creative: Pillar-specific landing pages
  • Heavy budget on Travel (best pillar)

• Day 26-28: Create Pillar Lookalikes

  • LAL 1%: Travel_Engaged (priority)
  • LAL 1%: Sport_Engaged
  • LAL 2-3%: Travel_Engaged (scale test)
  • Test campaign with Travel LAL

Expected by end of Week 4:

• Refined campaigns live
• Budget weighted to best pillar
• Lookalike campaigns testing
• Layer B → C conversion improved by 20-40%

17.4. Checklist - Tuần 5+ (Scale & Optimize)

Scale Phase:

• Week 5: Launch Pillar-Enhanced Advantage+

  • Seed with weighted pillar audiences
  • 40% Travel, 20% Sport, 10% Food
  • Creative mix: 50% Travel angle
  • Budget: $200/day → $400/day

• Week 6: Optimize Based on Data

  • Review pillar performance again
  • Update budget allocation
  • Pause/refresh weak pillars
  • Scale best pillar LAL
  • Update Advantage+ weights

• Week 7+: Continuous Improvement

  • Weekly pillar performance review
  • Monthly creative refresh
  • Quarterly strategy adjustment
  • Always be testing new pillars

17.5. Daily Monitoring Checklist

Every Day (5 minutes):

• Check spend pacing (on track vs budget?)
• Check ROAS (above target?)
• Check frequency (< 3.0?)
• Check Layer C growth (growing?)

Every Week (30 minutes):

• Analyze pillar performance
• Update budget allocation
• Refresh fatigued creatives
• Review Layer A/B/C sizes
• Check for audience overlap issues

Every Month (2 hours):

• Deep dive into pillar analytics
• Rebalance Advantage+ seeds
• Update lookalike sources
• Plan new creative tests
• Review overall strategy

17.6. Decision Tree: Pillar Management

Pillar Performance After 2 Weeks:

├─ Engagement >5% + ROAS >3.0x?
│  └─ YES → Scale aggressively
│     - Increase budget 50%
│     - Create LAL
│     - Weight Advantage+ heavily
│     - Develop more creative
│
├─ Engagement 3-5% + ROAS 2.0-3.0x?
│  └─ YES → Maintain & Optimize
│     - Keep current budget
│     - Test new creative
│     - Monitor weekly
│
└─ Engagement <3% + ROAS <2.0x?
   └─ YES → Decide: Refresh or Cut?
      ├─ Creative issue? → Refresh creative
      ├─ Wrong audience? → Adjust targeting
      └─ Pillar not resonating? → Cut & reallocate

17.7. Common Mistakes to Avoid

❌ Mistake 1: Analyze Too Early

• Don't analyze pillar performance before 1,000+ impressions
• Need statistical significance
• Wait minimum 7-10 days

❌ Mistake 2: Equal Budget Forever

• Don't keep equal budget across all pillars
• Data shows some pillars always outperform
• Reallocate weekly based on performance

❌ Mistake 3: Ignore Weak Pillars

• Don't let weak pillars drain budget
• Cut or refresh aggressively
• Opportunity cost is real

❌ Mistake 4: No Feedback Loop

• Don't just run pillars and forget
• MUST feed data back to Layer B
• This is where the magic happens

❌ Mistake 5: Over-Optimize

• Don't change budgets daily
• Let algorithm learn (3-5 days minimum)
• Only major changes weekly

17.8. Success Metrics - Pillar Feedback Loop

Week 2 vs Week 8 Comparison:

Metric	Week 2 (No Feedback)	Week 8 (With Feedback)	Improvement
Overall ROAS	2.0x	4.2x	+110% ⭐
Best Pillar ROAS	2.8x (Travel)	5.5x (Travel)	+96% ⭐
Worst Pillar ROAS	1.2x (Office)	Cut	N/A
Layer B Quality	Generic	Pillar-tagged	+65% conv rate
Advantage+ ROAS	Not launched	4.8x	N/A
CPA	₫35,000	₫18,000	-49% ⭐
Budget	$500/day	$1,500/day	3x scale ⭐
Sustainability	Declining	Growing	Stable ⭐

This is the power of Pillar Feedback Loop!

---

Kết luận

Chiến lược này là modern Meta Ads 2025:

✅ Data-driven, not assumption-driven ✅ Signal-rich, not signal-poor ✅ Layered, not flat ✅ Vertical-specific, not generic ✅ Scalable, not limited ✅ Bi-directional feedback loop, not one-way flow ⭐⭐⭐

Core philosophy:

"Không target người - target data trail của họ" "Không bỏ phí data - feed nó ngược lại để học"

Điểm đột phá: Pillar Feedback Loop

95% người chạy ads:

Setup campaigns → Chạy → Xem số liệu → Hết

→ Data bị bỏ phí!

Top 5% người chạy ads hiệu quả:

Setup campaigns → Chạy → Thu thập data → Tag users → Enrich audiences →
Feed vào Advantage+ → Scale → Thu thập data → Repeat

→ Data engine ngày càng thông minh!

3 Layers of Intelligence

Layer 1: Campaign Layer

• Pillar campaigns chạy
• Collect engagement data
• Track performance

Layer 2: Audience Layer

• Tag users theo pillar preference
• Enrich Layer B với pillar data
• Create pillar-specific lookalikes

Layer 3: Scale Layer

• Feed best pillar data vào Advantage+
• Meta learns pattern
• Auto-expand to similar users
• Sustainable scale

The Compounding Effect

Week 1: Generic targeting
  - ROAS: 2.0x
  - Learning from scratch

Week 4: Pillar data collected
  - ROAS: 2.8x
  - Know which pillar works
  - Start tagging users

Week 8: Feedback loop active
  - ROAS: 4.2x
  - Layer B fully tagged
  - Advantage+ optimized
  - Scaling confidently

Week 12+: Mature system
  - ROAS: 5.0x+
  - Self-optimizing
  - Predictable results
  - Easy to scale

Kết quả khác biệt:

Without Feedback Loop	With Feedback Loop
Đoán pillar nào tốt	Biết chính xác data
Budget phân đều	Budget theo performance
Layer B generic	Layer B pillar-tagged
Advantage+ blind	Advantage+ data-driven
Scale = tăng risk	Scale = giảm risk
ROAS decline khi scale	ROAS stable khi scale
Hit ceiling sớm	Scale sustainable

Công thức thành công

Data Collection
    ↓
Pattern Recognition (Pillar performance)
    ↓
Audience Tagging (Layer B enrichment)
    ↓
Intelligent Scaling (Advantage+ + LAL)
    ↓
More Data (Better quality)
    ↓
Better Patterns
    ↓
Repeat → Compounding returns

Bottom line:

Khi bạn build đúng data engine với Pillar Feedback Loop:

1. Meta sẽ tự tìm những người giống nhau
2. ROAS sẽ tăng theo thời gian (không giảm)
3. Scale tự nhiên mà không loãng
4. System tự optimize

Đây không phải "chạy ads" - đây là "xây data engine"!

---

Document version: 2.0 Last updated: 2025-01-12 Major update: Added Pillar Feedback Loop system Author: Ads Strategy Team