Home Developer Guide Decision Engine

Decision Engine

Learn how Cogniad's decision engine selects and serves ads using priority systems, auctions, frequency capping, and AI-powered optimization.

Overview

The decision engine is the core system that determines which ad to serve for each ad request. It evaluates campaigns and banners based on multiple factors including priority, targeting rules, frequency capping, budget availability, and optimization algorithms.

How It Works

The decision engine follows this process:

  1. Receive Ad Request: Publisher requests an ad for a zone
  2. Filter by Priority: Evaluate campaigns in priority order (highest first)
  3. Check Targeting: Verify user matches campaign and banner targeting rules
  4. Check Frequency Capping: Ensure user hasn't exceeded impression limits
  5. Check Budget: Verify campaign has available budget
  6. Select Best Ad: Choose optimal banner based on optimization algorithms
  7. Serve Ad: Return selected ad to publisher

Priority System

Campaigns are evaluated in priority order. Higher priority campaigns are considered first. The priority system ensures guaranteed delivery for contract campaigns while allowing remnant campaigns to fill unsold inventory.

Learn more about the Priority System.

Auctions & Bidding

For eCPM and programmatic campaigns, the decision engine runs real-time auctions to select the best ad based on bid values and performance metrics.

See the Auctions & Bidding Guide for details.

Frequency Capping

Frequency capping limits how many times a specific ad or campaign can be shown to the same user within a given time period. This prevents ad fatigue and improves user experience.

Learn more about Frequency Capping.

Optimization

The decision engine uses AI-powered algorithms to optimize ad selection for maximum performance. Optimization considers factors like CTR, conversion rates, user behavior, and historical performance.

See the Optimization Guide for details.

Performance

The decision engine is optimized for speed and scale:

  • Sub-10ms Latency: Ad selection happens in milliseconds
  • High Throughput: Handles billions of requests per day
  • Smart Caching: Caches frequently accessed data for speed
  • Edge Computing: Processes requests close to users

Related Topics