Ad Inventory Estimation via Probability Distribution

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for estimating ad inventory in online video streaming systems are inaccurate due to the assumption of independent targeted attributes, leading to unanticipated surpluses or shortages and high uncertainty, especially when attributes exhibit interdependence.

Innovation Solution

A method that determines a probability distribution of ad impressions based on electronic data records and targeted attributes, using a computer to forecast available impressions and manage ad inventory by decrementing sold impressions, with the ability to translate targeted attributes into a bit mask and organize impressions into a state table for ad pods, and forecasting viewership using linear functions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If prior methods assume targeted attributes can be handled independently for estimating purposes, then the estimation process becomes simpler, but the accuracy of ad inventory estimates deteriorates when attributes exhibit interdependence

Engineering Contradiction:

Improveestimation process complexityVSAvoidad inventory estimation accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent segments the ad inventory estimation process into distinct components: (1) determining a probability distribution of samples from electronic data records, (2) populating a data structure with forecasted numbers of available impressions for each combination of targeted attributes, and (3) estimating inventory numbers based on the populated data structure. This segmentation allows the system to handle attribute interdependencies systematically while maintaining manageable process complexity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent changes the estimation approach from assuming independent attributes to using a probability distribution that accounts for attribute interdependencies. By transitioning from simple independent attribute handling to a probabilistic model that captures relationships between attributes, the system improves estimation accuracy without proportionally increasing complexity.

Inventive Principle:

Principle #35Parameter changes

2Adaptability or versatility

If sophisticated video content platforms use complex targeted attributes with interdependencies, then the system can provide more precise advertising targeting, but the accuracy of ad inventory estimation deteriorates due to attribute interdependence

Engineering Contradiction:

Improveadvertising targeting capabilityVSAvoidad inventory estimation accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent incorporates feedback mechanisms by using electronic data records that capture actual ad impression data, which is then used to determine probability distributions. This feedback loop allows the system to learn from actual performance data and adjust estimates accordingly, maintaining accuracy even when handling complex, interdependent targeted attributes.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent performs preliminary actions by pre-populating a data structure with forecasted numbers of available impressions for each combination of targeted attributes before actual advertising campaigns begin. This preliminary estimation and population of the data structure enables the system to account for attribute interdependencies in advance, providing accurate inventory estimates for sophisticated targeting scenarios.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If the system provides detailed estimates for each combination of targeted attributes, then the precision of inventory information improves, but the complexity of managing and processing the data increases

Engineering Contradiction:

Improveinventory information precisionVSAvoiddata management complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates a universal data structure that can handle any combination of targeted attributes through a standardized population process. The same data structure and estimation methodology work for simple attribute combinations as well as complex interdependent attributes, providing precise inventory information without requiring separate management systems for different complexity levels.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS9313533B2Ad inventory management system

Publication Date: 2016.04.12 ADEIA MEDIA HOLDINGS INC

AI summary

Estimating ad inventory in an online video streaming system accurately handles interdependencies among targeted attributes. The estimating includes determining a probability distribution of samples taken from electronic data records of ad impressions in an interactive online video streaming service, among a population comprising each different combination of targeted attributes recorded for each of the samples. In addition, the estimating may include populating an electronic data structure with data relating the each different combination of the targeted attributes to corresponding forecasted number of available impressions in a defined time period, based on the probability distribution and a forecasted total number of available impressions in the time period. Then, estimating the ad inventory is based on the electronic data structure and targeting attributes for an ad campaign, optionally including summing forecasted impressions for combinations of the targeted attributes that include all of the defined set of targeted attributes.