Ad Inventory Curation Engine for Demand-Side Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Programmatic advertising systems fail to effectively align supply side inventory with demand side interest due to opacity in advertising space availability and interest communication, leading to inefficient inventory matching and reduced ad performance.
Innovation Solution
A curation engine that monitors demand side bidding activity and provides responsive inventory selection criteria to enhance market signaling, aligning supply side offerings with advertiser objectives through machine learning and real-time filtering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If supply side entities throttle offers of inventory to the market due to high volume and dynamic nature, then inventory management becomes manageable, but inventory matching efficiency deteriorates
Solution Approach 1:
The patent introduces a curation engine as an intermediary between supply side entities and demand side entities. This curation engine receives curated inventory data from supply side entities and uses machine learning models to predict which inventory is most likely to be bid on, then provides targeted recommendations to demand side entities. This intermediary layer resolves the contradiction by enabling efficient matching without requiring supply side entities to manually throttle offers to every possible buyer.
Solution Approach 2:
The system changes the parameters of inventory offering by transforming raw inventory data into predicted performance metrics using machine learning models. Instead of offering inventory based on basic attributes, the system offers inventory based on predicted bid probability, conversion likelihood, and other derived parameters. This parameter transformation enables supply side entities to manage inventory efficiently while maintaining high matching efficiency.
2Ease of operation
If demand side entities communicate interest only via specific price/parameter bids, then bidding process remains simple, but market signaling effectiveness deteriorates
Solution Approach 1:
The curation engine performs preliminary analysis of inventory and demand patterns before actual bidding occurs. By pre-processing inventory data and predicting which inventory is most likely to be of interest to which demand side entities, the system prepares targeted recommendations in advance. This preliminary action enriches the bidding process with informative signals without complicating the actual bidding mechanics, resolving the contradiction between simplicity and information effectiveness.
Solution Approach 2:
The system implements feedback loops where bidding outcomes and performance data are fed back into the machine learning models to continuously improve predictions. This feedback mechanism enhances market signaling effectiveness by learning from actual bidding behavior and refining inventory recommendations, while maintaining the simplicity of the bidding process itself for participants.
3Reliability
If conventional auction-type transactions are used, then market transparency is maintained, but inventory relevance to advertiser objectives deteriorates
Solution Approach 1:
The patent segments the inventory offering process into distinct layers: raw inventory data, curated inventory data with predicted performance metrics, and targeted recommendations to specific demand side entities. This segmentation allows the system to maintain transparent auction processes for final transactions while applying sophisticated relevance filtering and prediction in the preparatory layers, thus preserving market transparency while improving inventory relevance.
Data Source
AI summary
A curation engine monitors the performance of demand side bidding activity, relative to advertiser objectives, and provides responsive inventory selection criteria for use in generating more relevant supply side offerings to an ad exchange.