Sample Data Explanation

This document explains WHAT data is seeded in sample-data-init.sql, WHY each data element exists, and WHEN it’s used during ad serving and campaign management.

Data Seeding Philosophy

The sample data represents a realistic, multi-tenant Vietnamese e-commerce advertising scenario. It includes:

Multiple advertisers operating simultaneously
Cross-channel campaigns (direct, Google, Shopee, Lazada affiliates)
Real-world pricing in VND (Vietnamese Dong)
Authentic A/B testing scenarios
Realistic user behaviors and event distributions
Complete targeting configurations

Purpose: Enable developers, QA, and stakeholders to test end-to-end ad serving workflows without requiring external data sources.

Data Layers & Meanings

Layer 1: Tenant Foundation

WHAT: Single tenant named “demo” with Vietnamese regional settings

WHY:

Multi-tenant isolation is core to ad server design
Tenant controls all data access and settings
All subsequent entities must belong to this tenant

WHEN USED:

Every API request filters by tenant_id = 1
Audit trails track tenant-level operations
Settings control currency (VND), country (VN), environment

{
  "tenant_key": "demo",
  "name": "LEO Ad Server Demo",
  "country": "VN",
  "currency": "VND",
  "environment": "demo"
}

Layer 2: Advertisers & Accounts

WHAT: 3 brands representing different business models

Coolmate (Direct Advertiser)

Business Model: Direct e-commerce brand (apparel)
WHY: Tests internal campaign management and direct placements
WHEN USED:
- When admin creates campaigns directly in the system
- Source account type: “local” (internal control)
- Example: Summer sale campaign (50M VND budget)

ABC Fashion (Affiliate Partner)

Business Model: Premium affiliate/brand partnership
WHY: Tests partner integrations and affiliate network handling
WHEN USED:
- When integrating with affiliate providers (Shopee, Lazada)
- Validates revenue sharing and performance tracking
- Example: Seasonal retargeting campaign via Shopee

TechStore Vietnam (Electronics)

Business Model: Multi-category electronics retailer
WHY: Tests cross-category campaign logic
WHEN USED:
- Demonstrates category-specific creative variants
- Electronics-specific pricing and product attributes
- Example: Tech awareness campaign (20M VND budget)

Source Accounts (Provider Integration):

Provider	Account Key	Type	WHEN USED
internal	coolmate_internal	Local	Direct Coolmate campaigns
google_ads	gam_demo_account	Ad Network	Google Ad Manager integration
shopee_affiliate	shopee_demo_account	Affiliate	Shopee marketplace syndication
lazada_affiliate	lazada_demo_account	Affiliate	Lazada marketplace syndication

WHY multiple accounts: Real advertisers use multiple channels simultaneously to maximize reach.

Layer 3: Placements (Publisher Inventory)

WHAT: 10 inventory slots representing different publisher positions and devices

WHY: Placements define where ads can appear. Different placements have:

Different performance characteristics
Different audience composition
Different daily capacity limits
Different size constraints

WHEN USED: During ad serving, placement determines:

Which ads are eligible to show
Which creative sizes are acceptable
Whether frequency caps are enforced
Performance attribution

Placement Inventory Map

Placement	Device	Type	Daily Cap	WHY
coolmate-banner-300x250	Desktop	Standard	5,000 impr	High-performing desktop placement for e-commerce
coolmate-banner-mobile	Mobile	Responsive	8,000 impr	Mobile banner (reduced height for thumb access)
coolmate-product-carousel	All	Responsive	3,000 impr	Carousel feed (limited to reduce user fatigue)
coolmate-native-article	All	Native	2,000 impr	Native ads in article content (low volume, high intent)
coolmate-search-ads	Mobile	Search	6,000 impr	Mobile search (high intent, captures mobile shoppers)
google-top-banner	Desktop	Flexible	4,000 impr	Premium top banner on Google partner network
google-native	All	Responsive	2,500 impr	Google native ad format (contextual)
shopee-sidebar	Desktop	Standard	7,000 impr	Shopee marketplace sidebar (affiliate channel)
lazada-interstitial	Mobile	Responsive	3,500 impr	Lazada mobile interstitial (intrusive but high-impact)
generic-feed-carousel	All	Responsive	4,500 impr	Generic feed carousel (cross-platform)

Daily caps WHY:

Prevent over-saturation (user fatigue, declining CTR)
Manage publisher inventory fairly
Enforce publisher rate limits
Simulate realistic ad exchange dynamics

Layer 4: Campaigns

WHAT: 4-5 active campaigns with different objectives and budgets

WHY: Campaigns represent business intentions:

Drive product sales (Conversions objective, CPC/CPA buying model)
Build brand awareness (Awareness objective, CPM buying model)
Generate traffic (Traffic objective, CPM buying model)

WHEN USED:

Campaign budget controls total spend across all ads
Campaign objective determines optimization direction
Campaign status (active/paused) controls serving eligibility
Campaign dates enforce temporal boundaries

Campaign Details

Campaign 1: Coolmate Summer Sale 2026

Advertiser: Coolmate
Objective: Conversions (primary goal: purchase)
Buying Model: CPC (Cost-Per-Click) - pay only for clicks
Budget: 50M VND total, 500K daily
Status: Active
Duration: 2026-06-01 to 2026-08-31
WHY: Seasonal campaign targeting shopping season; CPC model reduces risk (pay only for engagement)
WHEN USED:
- At serving time: Filter ads by campaign status/dates
- At billing: Calculate CPC costs per click event
- At reporting: Aggregate metrics by campaign

Campaign 2: ABC Fashion Retargeting

Advertiser: ABC Fashion
Objective: Conversions (cart abandonment recovery)
Buying Model: CPA (Cost-Per-Action) - pay per purchase
Budget: 30M VND
WHY: Retargeting has high conversion rates; CPA aligns costs with actual conversions
WHEN USED:
- At serving: Only show ads to users with recent cart abandonment
- At billing: Calculate CPA costs per conversion event

Campaign 3: TechStore Awareness

Advertiser: TechStore Vietnam
Objective: Awareness (brand visibility)
Buying Model: CPM (Cost-Per-1000-Impressions) - pay per thousand views
Budget: 20M VND
WHY: Awareness campaigns prioritize reach over immediate conversion; CPM offers guaranteed impressions
WHEN USED:
- At serving: Maximize impressions (don’t restrict by conversion likelihood)
- At billing: Calculate CPM costs per 1,000 impressions

Layer 5: Creatives (Ad Content)

WHAT: 10+ creative assets representing different messaging and variants

WHY: Creatives are the actual ad content users see. Multiple creatives enable:

A/B testing different messages
Testing different formats (banner, native, carousel)
Platform-specific optimization (desktop vs. mobile)
Seasonal/contextual variations

WHEN USED:

At serving: Select which creative to show based on placement/user
At analytics: Track performance metrics per creative
At optimization: Identify winning variants and pause underperformers

Variant A (Control):

Headline: "Khuyến mãi Hè 2026"
Subheadline: "Giảm giá đến 40%"
Image: Standard summer sale image
Priority: 100 (highest)
WHY: Conservative messaging, established brand trust approach

Variant B (Test - Urgency):

Headline: "Chỉ còn lại 2 ngày! ⏰"
Subheadline: "Giảm giá đến 40% - Mua ngay!"
Image: High-energy summer sale with countdown timer
Priority: 95 (slightly lower)
WHY: Urgency messaging tests whether FOMO increases clicks/conversions

When used:

Both variants run simultaneously on same placement
Analytics compares CTR and conversion rate
Winning variant gets higher score_weight after 1-2 weeks
Losing variant is paused

Creative Platform Variants

Some creatives optimized for specific devices:

Creative Key	Device	Format	WHY
retargeting_dynamic_01_mobile	Mobile	320x100	Responsive height for mobile screens
retargeting_dynamic_01_desktop	Desktop	300x250	Standard desktop square
product_carousel_01_mobile	Mobile	100% width, 3 items	Mobile thumb-friendly carousel
product_carousel_01_desktop	Desktop	100% width, 4 items	Desktop landscape orientation allows more items

WHY platform variants: Same message may perform differently on mobile vs. desktop. Different devices have different ergonomics.

Layer 6: Creative Items (Carousel Products)

WHAT: 9 products in product carousel with realistic Vietnamese e-commerce attributes

WHY: Product carousels drive e-commerce conversions by:

Showcasing multiple items per ad (increasing relevance)
Displaying real-time pricing and discounts
Building social proof (ratings, reviews)
Enabling quick browsing without leaving publisher

WHEN USED:

At serving: Display carousel creative on placements supporting product carousel format
At click: User clicks product → navigates to product page
At conversion: Track which product was purchased (via product_id in conversion event)

Product Attributes & WHY

Product	Price	Original	Discount	Rating	Stock	WHY
Áo sơ mi nam Classic	399K	499K	-20%	4.8⭐	45	Premium shirt, high discount, best-seller (high review count)
Áo thun nam Essential	249K	275K	-9%	4.6⭐	112	Budget option, good stock, trending
Áo polo nam Premium	425K	475K	-10%	4.7⭐	67	High-end option, seasonal demand
Quần shorts nam chino	449K	549K	-18%	4.9⭐	89	Summer seasonal, large discount attracts buyers

Vietnamese pricing rationale:

199K-499K range matches real Coolmate apparel pricing
20% average discount reflects e-commerce norm
Stock levels (45-156 units) suggest in-stock assurance
Ratings (4.6-4.9) reflect authentic e-commerce high-volume sellers

WHEN USED AT DIFFERENT STAGES:

At Ad Creation: Admin configures which products appear in carousel
At Serving: Platform fetches current price/stock/rating for each product
At Click: User clicks product in carousel → navigates to product page
At Analytics: Platform tracks which products were viewed/clicked/purchased

Layer 7: Audience Segments

WHAT: 8 audience definitions representing different user segments and behaviors

WHY: Audiences enable targeting:

Retargeting (users who viewed products but didn’t buy)
Lookalike (users similar to buyers)
Exclusion (users already customers → don’t re-acquire)
Behavioral (users showing purchase intent)

WHEN USED:

At campaign setup: Admin selects which audiences to target
At serving: Ad server checks if user is in included/excluded audiences
At optimization: Analytics identifies which audiences have highest ROI

Audience Definitions

High-Value (LTV > 3M VND)

Definition: Customers with lifetime value exceeding 3 million VND
Member Count: ~35,000 users
Lookback: All-time (permanent segment)
WHY: Target existing high-spenders with premium offers
WHEN USED: Premium campaigns, exclusive deals

Cart Abandoners (7-day)

Definition: Users who added items to cart but didn't purchase in last 7 days
Member Count: ~52,000 users
Lookback: 7 days (dynamic, updates hourly)
WHY: Cart abandoners have high conversion intent; strategic timing increases recovery
WHEN USED: ABC Fashion retargeting campaign, time-sensitive offers

Recent Product Viewers (30-day)

Definition: Users who viewed product pages in last 30 days
Member Count: ~125,000 users
Lookback: 30 days
WHY: Strong purchase intent signal; users actively shopping
WHEN USED: Retargeting campaigns, product-specific ads

Mobile Users VN

Definition: Users with mobile device in Vietnam (7-day active)
Member Count: ~285,000 users
Lookback: 7 days (active in last week)
WHY: Mobile users respond to mobile-optimized creatives
WHEN USED: Mobile placement campaigns, mobile-specific formats

New Users (30-day)

Definition: Users with first visit in last 30 days
Member Count: ~18,000 users
WHY: New users need brand education; use different creative approach
WHEN USED: Brand awareness campaigns, onboarding messages

Seasonal Summer Shoppers (60-day)

Definition: Users showing summer seasonal interest + mobile preference
Member Count: ~42,000 users
WHEN USED: Seasonal campaign (June-August), time-limited offers

Repeat Purchasers

Definition: Users with 3+ purchases (lifetime)
Member Count: ~78,000 users
WHY: Loyal customers; test premium products and exclusive offers

Corporate/Bulk Buyers

Definition: Users identified as B2B bulk purchasers
Member Count: ~8,500 users
WHY: Bulk buyers have different purchase patterns; use B2B messaging

Layer 8: Targeting Rules

WHAT: 8+ conditional rules that determine which users see which ads

WHY: Targeting rules maximize relevance and ROI:

Geographic targeting (show Vietnam ads only in Vietnam)
Temporal targeting (show urgency messages during peak hours)
Device targeting (optimize for mobile vs. desktop)
Behavioral targeting (target cart abandoners differently than new users)
Intent targeting (detect shopping mode via context)

WHEN USED: At serving time, before showing an ad:

1. Check if user in required audiences (INCLUDE)
2. Check if user NOT in excluded audiences (EXCLUDE)
3. Evaluate all targeting rules → at least one must match
4. If all rules pass → ad is eligible to show

Example Rule: Coolmate Retargeting

Priority: 100 (highest priority)
Countries: ['VN'] (only Vietnam)
Device Types: ['mobile', 'desktop'] (excludes tablet)
Languages: ['vi'] (Vietnamese language)
Context Keywords: ['fashion', 'menswear', 'style'] (relevant categories)

Custom Predicates:
  - retargeting: true (must be retargeting audience)
  - recentProductViewDays: 30 (viewed product in last 30 days)
  - minTimeOnSite: 30 seconds (spent >30s on site, showing interest)
  - dayOfWeek: Mon-Fri (exclude weekends, different behavior)

WHEN USED: High-priority cart recovery
WHY: Retargeting cart abandoners is highest-intent segment

Example Rule: Mobile Urgency (Peak Hours)

Priority: 95
Device Types: ['mobile'] (only mobile)
Time of Day: ['08:00-12:00', '18:00-23:00'] (morning commute + evening browsing)

Custom Predicates:
  - cartValue: >100,000 VND (higher-value orders only)
  - recentCartAbandonment: true (abandoned cart)
  - daysSinceLastPurchase: 7-30 (not recent buyer)

WHEN USED: Mobile-specific urgency messaging during peak hours
WHY: Mobile users at work/evening more likely to respond to urgency

Example Rule: Contextual (Article Context)

Priority: 80
Context: ['article_page'] (showing on article/news page)
Article Categories: ['fashion', 'lifestyle', 'style'] (relevant content)
Device Types: ['all'] (works on any device)

WHEN USED: Native ads in article content
WHY: Article context matching → higher relevance → better CTR

Layer 9: Ads (Delivery Configuration)

WHAT: 10 ad configurations linking campaign + creative + placement

WHY: Ads tie together:

Campaign (the business intent: “sell summer collection”)
Creative (the content: “Khuyến mãi Hè” banner)
Placement (where it shows: “Homepage banner”)

This configuration is the minimal serving unit.

WHEN USED:

At serving: Platform fetches eligible ads for a placement
At ranking: Ads sorted by score_weight (100 = top priority)
At frequency capping: Limit how many times one ad shows per user
At attribution: Track performance metrics per ad

Ad Configuration Example

Ad Key: "ad-coolmate-banner-01"
Campaign: "coolmate-summer-2026" (50M VND budget, conversions objective)
Creative: "summer-sale-banner" (v1 - control variant)
Placement: "coolmate-banner-300x250" (5K daily impressions)
Status: "active"
Score Weight: 100.0 (highest priority)
Frequency Cap: 5x per user (show max 5 times per user)

WHEN USED:
  1. User visits placement "coolmate-banner-300x250"
  2. Platform checks ad eligibility
  3. "ad-coolmate-banner-01" passes all filters
  4. "summer-sale-banner" v1 creative is shown
  5. Impression event logged
  6. If user clicks → click event logged
  7. If user purchases → conversion event logged
  8. Metrics aggregated to campaign for budget tracking

Ad Serving Index (Placement-Ad Mapping)

For performance, “placement_ad” table pre-computes which ads can serve on which placements:

placement_id=1, ad_id=1, rank_score=100.0 (highest rank)
placement_id=1, ad_id=2, rank_score=95.0
placement_id=1, ad_id=5, rank_score=80.0
-- ... sorted by score for fast candidate selection

WHY: Instead of scanning all 10 ads every request, platform pre-filters eligible ads per placement.

Layer 10: User Serving Profiles

WHAT: 10 synthetic user profiles with attributes for targeting

WHY: User profiles enable:

Segment-based targeting (show ads only to high-value users)
Device-based personalization (desktop vs. mobile creative)
Interest-based matching (fashion users → fashion ads)
LTV-based prioritization (spend more on high-value users)

WHEN USED: At serving time:

1. User visits placement
2. Platform fetches user profile (user_serving_key)
3. Platform checks if user matches targeting rules
   - Is user in "high-value" audience? ✓
   - Is user on mobile? ✓
   - Is user in Vietnam? ✓
4. If all checks pass → show ad

User Segment Examples

User Profile 1: High-Value Retargeter

LTV: 5M VND (high spender)
Purchase Count: 12 (loyal)
Primary Device: mobile
Segment: retargeting (cart abandonment)
Interests: fashion, sales, menswear
WHY: Show premium products + urgency messaging

User Profile 2: New Visitor

LTV: 0 VND (no purchases)
Purchase Count: 0
Primary Device: desktop
Segment: first_visit (just arrived)
Interests: (unknown)
WHY: Show brand intro + entry-level products

User Profile 3: Seasonal Summer Shopper

LTV: 3.2M VND
Purchase Count: 5 (occasional)
Primary Device: mobile
Seasonal Interest: summer, menswear
Purchase Pattern: Q2-Q3 focus
WHY: Show summer-specific products during peak season

Layer 11: Sample Events (Realistic Funnel)

WHAT: 50+ events representing user interactions (impressions, clicks, conversions)

WHY: Events are the raw data for:

Performance analytics (CTR, conversion rate)
Budget tracking (cost accumulation)
Attribution modeling (multi-touch paths)
Campaign optimization (identify winning ads)

WHEN USED:

At runtime: Events generated as users interact with ads
At analytics: Events aggregated into reports and dashboards
At billing: Events used to calculate costs (CPC, CPA, CPM)

Event Distribution Rationale

Timeline:  T-6h → T-1h  (impressions spread over 6 hours)
Impressions: 72 total
  ├─ Desktop: 35 impressions (mix of banners, natives)
  └─ Mobile: 37 impressions (mobile-optimized placements)

Clicks: 12 total (~2.5% CTR = realistic for e-commerce)
  ├─ From retargeting users: 8 clicks (high intent)
  └─ From new users: 4 clicks (lower intent)

Conversions: 9 total (~75% of clicks convert = realistic)
  ├─ Direct clicks: 9 conversions
  └─ Revenue: 249K + 199K + etc. (product-dependent)

VTC (View-Through): 1 conversion
  └─ User saw ad but didn't click, converted later
  └─ Demonstrates view-based attribution value

Event Type Explanations

Impression Event:

{
  "event_type": "impression",
  "user_id": "demo-user-001",
  "ad_id": 1,
  "placement_id": 1,
  "device": "mobile",
  "country": "VN",
  "timestamp": "2026-08-15 14:32:15",
  "duration_seconds": 3.2,
  "WHY": Track when ad was shown to user
}

Click Event:

{
  "event_type": "click",
  "user_id": "demo-user-001",
  "ad_id": 1,
  "placement_id": 1,
  "destination_url": "https://coolmate.com/product/1",
  "timestamp": "2026-08-15 14:33:45",
  "WHY": Track engagement - user clicked ad
}

Conversion Event (Purchase):

{
  "event_type": "conversion",
  "event_subtype": "purchase",
  "user_id": "demo-user-001",
  "ad_id": 1,
  "campaign_id": 1,
  "product_id": "p2",
  "revenue": 249000,
  "currency": "VND",
  "timestamp": "2026-08-15 14:48:30",
  "time_to_conversion_seconds": 945,
  "WHY": Track business outcome - user purchased
}

VTC Event (View-Through Conversion):

{
  "event_type": "conversion",
  "attribution_type": "view_through",
  "user_id": "demo-user-002",
  "ad_id": 2,
  "revenue": 199000,
  "timestamp": "2026-08-15 16:20:00",
  "hours_since_impression": 2.5,
  "WHY": Track delayed conversions from views alone (no click)
}

Data Flow Timeline

Campaign Lifecycle with Sample Data

T0 (Campaign Creation):
  Admin creates campaign "coolmate-summer-2026"
  → INSERT leo_ads.campaign (budget_amount=50M, status='draft')
  
T1 (Content Creation):
  Admin creates creatives and carousel products
  → INSERT leo_ads.creative (headline, body, images)
  → INSERT leo_ads.creative_item (products, prices, stock)
  
T2 (Targeting Configuration):
  Admin defines audiences and rules
  → INSERT leo_ads.audience (high_value_ltv, cart_abandoners)
  → INSERT leo_ads.targeting_rule (countries, devices, custom_predicates)
  
T3 (Ad Creation):
  Admin creates ad linking campaign + creative + placement
  → INSERT leo_ads.ad (campaign_id, creative_id, placement_id)
  → Platform computes placement_ad index
  
T4 (Activation):
  Admin activates campaign
  → UPDATE leo_ads.campaign SET status='active'
  
T5 (Serving):
  User visits placement → Platform selects ad → Shows creative
  → INSERT leo_ads.ad_event (type='impression')
  
T6 (Engagement):
  User clicks ad → Browser redirects to destination URL
  → INSERT leo_ads.ad_event (type='click')
  → User views product page
  
T7 (Conversion):
  User purchases product
  → INSERT leo_ads.ad_event (type='conversion', revenue=249K)
  → UPDATE leo_ads.campaign SET spent_amount = spent_amount + cost
  
T8 (Analytics):
  Campaign runs for 2 months
  Platform aggregates events → Reports show:
  → 72 impressions, 12 clicks (16.7% CTR), 9 conversions (75% conv rate)
  → ROI calculated: 9 × 249K revenue vs. actual CPC cost

Why Each Element Matters

Element	Why It Matters	Real-World Use
Tenant	Isolation & security	SaaS multi-customer data protection
Advertiser	Business entity	Track campaigns by brand
Campaign	Business objective	Budget allocation, performance goal
Creative	Message testing	A/B test messaging, formats
Placement	Ad location	Optimize by publisher/position/device
Ad	Serving configuration	Minimal unit of serving decision
Audience	User targeting	Reach right person, right time
Targeting Rule	Eligibility conditions	Context-based + behavioral filtering
Event	Business outcome	Attribution, optimization, billing
User Profile	Personalization	Segment-based creative selection

Verification Queries

After seed script runs, verify data integrity:

Check Campaign Budget Tracking Setup

SELECT 
  c.campaign_key, 
  c.budget_amount, 
  c.daily_budget_amount,
  c.status,
  COUNT(a.ad_id) as ad_count
FROM leo_ads.campaign c
LEFT JOIN leo_ads.ad a ON a.campaign_id = c.campaign_id
GROUP BY c.campaign_id, c.campaign_key;

Check A/B Test Variant Setup

SELECT 
  creative_key,
  version_no,
  priority,
  status
FROM leo_ads.creative
ORDER BY creative_key, version_no;

Check Audience Targeting Links

SELECT 
  aud.audience_key,
  COUNT(aa.ad_id) as ads_targeted,
  SUM(CASE WHEN aa.relation_type='include' THEN 1 ELSE 0 END) as includes,
  SUM(CASE WHEN aa.relation_type='exclude' THEN 1 ELSE 0 END) as excludes
FROM leo_ads.audience aud
LEFT JOIN leo_ads.ad_audience aa ON aa.audience_id = aud.audience_id
GROUP BY aud.audience_id, aud.audience_key;

Check Event Distribution (Funnel)

SELECT 
  event_type,
  COUNT(*) as event_count,
  ROUND(100.0 * COUNT(*) / (SELECT COUNT(*) FROM leo_ads.ad_event), 1) as pct
FROM leo_ads.ad_event
GROUP BY event_type
ORDER BY event_count DESC;

Check Placement-Ad Index (Serving Candidates)

SELECT 
  p.placement_key,
  COUNT(pa.ad_id) as eligible_ads,
  STRING_AGG(a.ad_key, ', ' ORDER BY pa.rank_score DESC) as ad_candidates
FROM leo_ads.placement p
LEFT JOIN leo_ads.placement_ad pa ON pa.placement_id = p.placement_id
LEFT JOIN leo_ads.ad a ON a.ad_id = pa.ad_id
WHERE p.tenant_id = 1
GROUP BY p.placement_id, p.placement_key;

Context for Development & Testing

For QA Testing:

Full campaign lifecycle end-to-end
A/B testing scenario validation
Targeting rule evaluation
Event tracking accuracy
Frequency capping enforcement

For Dashboards & Analytics:

Sample data populates reports
Realistic metrics (CTR ~2.5%, conversion ~75% of clicks)
Vietnamese pricing for localization testing
Multi-channel campaign comparison

For Performance Testing:

50+ events exercise event processing
10 placements × 10 ads test index performance
10 users × 8 audiences test targeting logic
Materialized view queries demonstrate optimization

For Integration Testing:

Multiple provider accounts (local, Google, Shopee, Lazada)
Source asset mapping validates provider sync
Event payload structure matches real providers

Documentation

For schema details, see: README.md (comprehensive schema documentation) For API usage, see: README.md (ad server API endpoints) For full database context, see: database-schema.sql

LEO Customer 360 — Documentation

Explorer

Sample Data Explanation

Sample Data Explanation

Data Seeding Philosophy

Data Layers & Meanings

Layer 1: Tenant Foundation

Layer 2: Advertisers & Accounts

Coolmate (Direct Advertiser)

ABC Fashion (Affiliate Partner)

TechStore Vietnam (Electronics)

Layer 3: Placements (Publisher Inventory)

Placement Inventory Map

Layer 4: Campaigns

Campaign Details

Layer 5: Creatives (Ad Content)

Creative A/B Test Example: Coolmate Summer Banner

Creative Platform Variants

Layer 6: Creative Items (Carousel Products)

Product Attributes & WHY

Layer 7: Audience Segments

Audience Definitions

Layer 8: Targeting Rules

Example Rule: Coolmate Retargeting

Example Rule: Mobile Urgency (Peak Hours)

Example Rule: Contextual (Article Context)

Layer 9: Ads (Delivery Configuration)

Ad Configuration Example

Ad Serving Index (Placement-Ad Mapping)

Layer 10: User Serving Profiles

User Segment Examples

Layer 11: Sample Events (Realistic Funnel)

Event Distribution Rationale

Event Type Explanations

Data Flow Timeline

Campaign Lifecycle with Sample Data

Why Each Element Matters

Verification Queries

Check Campaign Budget Tracking Setup

Check A/B Test Variant Setup

Check Audience Targeting Links

Check Event Distribution (Funnel)

Check Placement-Ad Index (Serving Candidates)

Context for Development & Testing

Documentation

Graph View

Table of Contents

Backlinks