Attribution Models in Retail Media
Understand Ring DAS's three-tier attribution system: Direct, Halo attribution for comprehensive retail media measurement.
Attribution Models in Retail Media
Overview
Attribution is the process of assigning credit to advertising touchpoints that lead to customer conversions. In retail media, accurate attribution is critical for measuring campaign ROI and optimizing advertising spend.
Ring DAS provides a three-tier attribution system that captures different levels of advertising impact:
- Direct Platform Attribution - Online click and impression-based conversions
- Halo Attribution - Brand and category-level cross-product impact
Why Two Tiers? Each tier captures different aspects of advertising effectiveness. Direct attribution shows immediate product impact, halo attribution reveals brand-level influence.
graph TB
subgraph "Tier 1: Direct Platform Attribution"
A[User sees/clicks<br/>Product A ad] --> B[User purchases<br/>Product A]
style A fill:#e1f5ff
style B fill:#c8e6c9
end
subgraph "Tier 2: Halo Attribution"
C[User sees/clicks<br/>Brand X Product A] --> D[User purchases<br/>Brand X Product B]
style C fill:#fff3e0
style D fill:#ffecb3
end
Tier 1: Direct Platform Attribution
What is Direct Platform Attribution?
Direct platform attribution tracks conversions where a user interacts with an ad and purchases the exact product advertised. This is the most straightforward attribution model and the foundation of retail media measurement.
Attribution Methods
1. Click-Based Attribution
Conversions are attributed to ads that the user clicked before purchasing.
Requirements:
- User must click the sponsored product ad
- Purchase must occur within the lookback window
- Purchased product GTIN must match advertised product
Example:
1. User searches for "chocolate"
2. User clicks sponsored product: "Cadbury Dairy Milk 200g" (GTIN: 7733636)
3. User adds product to cart
4. User completes purchase within 14 days
→ Conversion attributed to campaign
Ring DAS Implementation:
- Click tracking via Activity Pixel
- Purchase event tracking with product GTINs
- Automatic attribution matching in Decision Engine
2. Impression-Based Attribution (View-Through)
Conversions are attributed to ads that the user saw (but did not necessarily click) before purchasing.
Requirements:
- User must see the sponsored product ad (impression tracked)
- Purchase must occur within the lookback window
- Purchased product GTIN must match advertised product
Example:
1. User browses "Snacks" category
2. User sees sponsored product: "Cadbury Dairy Milk 200g" (impression tracked)
3. User doesn't click, continues browsing
4. User returns later and purchases the product directly
→ View-through conversion attributed to campaign
Ring DAS Implementation:
- Impression tracking via ad serving
- View-through window configuration (typically shorter than click window)
- Attribution priority: Click attribution > View-through attribution
Configuration Parameters
Lookback Windows
Define how far back to look for attribution touchpoints.
| Touchpoint Type | Default Window | Recommended Range | Ring DAS Config |
|---|---|---|---|
| Click-based | 14 days | 7-30 days | Configurable per deal |
| Impression-based | 7 days | 1-14 days | Configurable per deal |
Configuration Example:
attribution:
direct:
click_lookback_days: 14
impression_lookback_days: 7
priority: "click" # Click attribution takes precedenceProduct Matching
Exact GTIN Match (Default):
- Purchased product GTIN must exactly match advertised product
- Most accurate but may undercount conversions
Example:
- Advertised: Cadbury Dairy Milk 200g (GTIN: 7733636)
- Purchased: Cadbury Dairy Milk 200g (GTIN: 7733636) ✅ Attributed
- Purchased: Cadbury Dairy Milk 600ml (GTIN: 7733628) ❌ Not attributed (see Halo)
Tier 2: Halo Attribution
What is Halo Attribution?
Halo attribution captures conversions where a user interacts with an ad for one product but purchases a different product from the same brand or category. This measures the broader brand awareness and cross-product influence of advertising.
Prerequisite: Halo attribution requires the
brandfield to be populated in your product catalog for accurate attribution.
Attribution Types
1. Brand-Level Halo (Click Attribution)
User clicks an ad for Brand X Product A and purchases Brand X Product B.
Example:
1. User clicks sponsored product: "Cadbury Dairy Milk 200g"
2. User browses and purchases: "Cadbury Twirl 43g"
→ Halo conversion attributed to original Cadbury campaign
Requirements:
- Both products share the same brand name
- Purchase occurs within lookback window
- Click interaction on original ad
2. Brand-Level Halo (Impression/View-Through)
User sees an ad for Brand X Product A and purchases Brand X Product B.
Example:
1. User sees (but doesn't click): "Cadbury Dairy Milk 200g"
2. User later searches and purchases: "Cadbury Oreo 120g"
→ Halo view-through conversion attributed to campaign
Requirements:
- Both products share the same brand name
- Purchase occurs within lookback window
- Impression recorded for original ad
3. Category-Level Halo (Optional)
User interacts with a product in Category A and purchases another product in Category A.
Example:
1. User clicks sponsored product: "Cadbury Dairy Milk" (Category: Chocolate)
2. User purchases: "Lindt Excellence" (Category: Chocolate)
→ Category-level halo conversion (if enabled)
Brand Field Requirements
For halo attribution to work, ensure product catalogs include accurate brand information.
Product Catalog Example:
{
"products": [
{
"gtin": "7733636",
"name": "Cadbury Dairy Milk 200g",
"brand": "Cadbury", // ← Required for halo
"category": "Snacks/Chocolate",
"price": "3.99"
},
{
"gtin": "7733628",
"name": "Cadbury Twirl 43g",
"brand": "Cadbury", // ← Same brand enables halo
"category": "Snacks/Chocolate",
"price": "1.99"
}
]
}Halo Attribution Priority
When multiple attribution models apply, Ring DAS uses this priority:
- Direct Click Attribution (highest priority)
- Halo Click Attribution
- Direct Impression Attribution
- Halo Impression Attribution (lowest priority)
Example Scenario:
User clicks Ad 1 (Product A) → Direct click attribution
User sees Ad 2 (Product B, same brand)
User purchases Product A
→ Attributed to Ad 1 (direct click), not Ad 2 (halo impression)
Updated 7 days ago
