Programmatic Targeting

Targeting Categories:
├── Audience-Based
│   ├── First-party (your data)
│   ├── Third-party (purchased)
│   ├── Behavioral (actions)
│   └── Demographic (attributes)
├── Contextual
│   ├── Keywords
│   ├── Topics/categories
│   ├── Semantic analysis
│   └── Page-level context
├── Technical
│   ├── Device type
│   ├── Operating system
│   ├── Browser
│   ├── Connection speed
│   └── Geographic (IP-based)
└── Inventory-Based
    ├── Domain lists
    ├── App lists
    ├── Viewability thresholds
    └── Publisher tiers

Your Owned Data:
├── Website visitors
├── Email subscribers
├── CRM contacts
├── Purchase history
├── App users
└── Offline transactions

Activation:
├── Pixel-based (browser)
├── CRM onboarding (email match)
├── Mobile IDs
├── Clean rooms
└── Data partnerships

Purchased Data Segments:
├── Demographics (age, income, etc.)
├── Interests (in-market, affinity)
├── Purchase behavior
├── Location history
├── B2B firmographics
└── Intent signals

Major Providers:
├── Oracle Data Cloud
├── Lotame
├── Nielsen
├── Experian
├── Dun & Bradstreet (B2B)
└── Bombora (B2B intent)

Considerations:
├── Cost ($1-$5+ CPM)
├── Scale vs accuracy trade-off
├── Privacy compliance
├── Declining with cookie loss
└── Test before scaling

Partnership Data:
├── Publisher data partnerships
├── Retailer data collaborations
├── Clean room sharing
└── Co-op arrangements

Examples:
├── Retailer shares shopper data
├── Publisher shares subscriber data
├── Industry data cooperatives
└── Strategic partner sharing

Contextual Methods:
├── Keyword Targeting
│   ├── Exact match keywords
│   ├── Broad match expansion
│   └── Negative keywords
├── Topic/Category
│   ├── IAB content categories
│   ├── Custom taxonomies
│   └── Sub-category targeting
├── Semantic Analysis
│   ├── Page meaning understanding
│   ├── Sentiment detection
│   └── AI-powered classification
└── Adjacency
    ├── Article type
    ├── Content tone
    └── Brand suitability

Why Contextual Matters:
├── Cookie-independent
├── Privacy-compliant
├── Real-time relevance
├── Brand safety alignment
└── Growing in importance

Performance Data:
├── 2.5x more memorable vs random
├── 30% higher engagement
├── Similar conversion rates
└── Better brand recall

Identity Approaches:
├── Authenticated IDs
│   ├── Login-based (email hash)
│   ├── UID 2.0 (Trade Desk)
│   ├── RampID (LiveRamp)
│   └── ID5 Universal ID
├── Probabilistic IDs
│   ├── Fingerprinting (limited)
│   ├── Cohort modeling
│   └── Panel-based
└── Publisher-Provided
    ├── First-party graphs
    ├── Logged-in user data
    └── Subscription data

How It Works:
├── User provides email
├── Email hashed, encrypted
├── Shared across ecosystem
├── User controls/opts out
└── Refreshes periodically

Adoption:
├── Major DSPs: Supported
├── Major SSPs: Supported
├── CTV: Strong adoption
├── Open web: Growing
└── Walled gardens: Limited

Layering Examples:
├── Audience + Contextual
│   └── In-market shoppers on review sites
├── Audience + Geographic
│   └── High-income in target zip codes
├── Retargeting + Recency
│   └── Site visitors, last 7 days
├── Lookalike + Interest
│   └── Similar to converters + category interest
└── Contextual + Viewability
    └── Relevant content + above-fold only

Best Practices:
├── Don't over-layer (audience too small)
├── Monitor reach vs precision
├── Test layers incrementally
├── Compare layered vs single
└── Document effective combinations

Targeting Trade-offs:
Broad Targeting:
├── Large reach
├── Lower CPMs
├── More learning data
└── Lower conversion rate

Narrow Targeting:
├── Limited reach
├── Higher CPMs
├── Less data for optimization
└── Higher conversion rate

Recommendation:
├── Start broader for learning
├── Narrow as you find winners
├── Maintain multiple approaches
└── Balance scale and efficiency