Online System Budget Recommendation via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advertisers specifying budgets for online advertising campaigns without contextual information may result in reduced revenue for online systems and deter advertisers from advertising, as they may not achieve desired impressions or conversion events, leading to discouraged participation.
Innovation Solution
The online system presents a range of candidate budgets to advertisers, including an estimated number of conversion events, and identifies a default budget that maximizes the likelihood of advertiser participation and revenue, using machine learned models trained on historical data to determine optimal budget options based on advertiser and advertisement characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If advertisers specify budgets for advertising campaigns without contextual information, then the advertising process is simple and quick, but revenue to the online system is reduced and advertiser participation is deterred
Solution Approach 1:
The online system performs preliminary analysis of historical advertising data and contextual information before the advertiser specifies a budget. The system pre-calculates optimal budget ranges and projected outcomes, presenting this information to the advertiser in advance. This allows the advertiser to make an informed budget decision without requiring complex real-time analysis, thus maintaining ease of operation while maximizing revenue potential.
Solution Approach 2:
The system implements a feedback mechanism where historical performance data, contextual factors, and projected outcomes are fed back to the advertiser during the budget specification process. This feedback loop enables the advertiser to adjust their budget based on predicted performance metrics, ensuring both user-friendly operation and optimized revenue generation for the online system.
2Loss of time
If advertisers specify budgets without contextual information, then the advertising setup is fast, but desired conversion events are not achieved and advertisers are discouraged from participating
Solution Approach 1:
The online system performs preliminary analysis of historical advertising data and contextual information before the advertiser specifies a budget. The system pre-calculates optimal budget ranges and projected outcomes, presenting this information to the advertiser in advance. This allows the advertiser to make an informed budget decision without requiring complex real-time analysis, thus maintaining ease of operation while maximizing revenue potential.
Solution Approach 2:
The system implements a feedback mechanism where historical performance data, contextual factors, and projected outcomes are fed back to the advertiser during the budget specification process. This feedback loop enables the advertiser to adjust their budget based on predicted performance metrics, ensuring both user-friendly operation and optimized revenue generation for the online system.
3Adaptability or versatility
If the online system allows advertisers to specify arbitrary budgets, then advertiser freedom is maximized, but revenue optimization is reduced
Solution Approach 1:
The system dynamically adjusts budget parameters based on contextual analysis. Instead of allowing completely arbitrary budget specification, the system modifies the budget parameters by presenting recommended ranges and adjusting bids based on contextual factors such as historical performance, market conditions, and campaign goals. This maintains advertiser flexibility while optimizing revenue potential through data-driven parameter adjustments.
Data Source
AI summary
An online system provides identifies multiple candidate budgets to an advertiser requesting presentation of advertisements via the online system. The advertiser may select a candidate budget used for presenting advertisements via the online system. A default budget may be identified from the candidate budgets to increase the likelihood of the advertiser selecting the default budget. The candidate budgets and the default budget are determined by the online system to maximize the likelihood that an advertiser elects to present advertisements via the online system or to maximize revenue received the online system for presenting advertisements from the advertiser. Various factors are specific to the advertiser, the object being advertised, and other contextual information may be used to determine the candidate budgets.