Design Ads Data Modeling

Netflix's "Design Ads Data Modeling" interview question focuses on creating a logical schema for a scalable ads platform. It emphasizes database design for global traffic with low-latency reporting needs.[3]

Problem Statement

Design the data model (logical schema) for an ads platform that supports advertisers creating campaigns with line items and creatives, along with targeting capabilities like geo, device, demographics, and interests. Assume large-scale global traffic and reporting available with low delay (minutes).[3]

Key Entities

Typical entities include:

• Campaigns: Top-level containers with budget, duration, and goals.
• Line Items: Sub-units under campaigns defining bid strategies and targeting.
• Creatives: Ad assets (images, videos) linked to line items.
• Targeting Rules: Geo, device, user segments for ad delivery.[3]

Constraints

• Handle high-volume global impressions/clicks.
• Support real-time reporting (e.g., via OLAP stores like Druid) within minutes.
• Scale for Netflix's ad-supported tier with frequency capping and vendor integration.[1][2]

No full input/output examples or detailed constraints were found beyond scale and latency requirements. Related questions mention schemas for ad impressions/clicks.[2][3]