Shaded Bid Value Determination via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing auction systems for content delivery, such as advertisements, face challenges in efficiently determining bid values that balance winning auctions with minimizing costs, particularly in first-price auctions where the winning bid equals the auction price.
Innovation Solution
The system employs a machine learning model trained on auction data to determine shaded bid values, which are reduced bid values submitted to auctions, based on feature parameters associated with client devices and content requests, optimizing bid surplus and auction winnings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a high bid value is submitted to win the auction, then the auction win probability increases, but the cost increases
Solution Approach 1:
The system changes the bid parameter by introducing a shaded bid value that is lower than the true bid value. The machine learning model adjusts the bid parameter dynamically based on features such as client device characteristics, content type, and auction history to optimize the balance between winning probability and cost reduction.
Solution Approach 2:
The system implements feedback by using historical auction data and actual bid outcomes to train the machine learning model. The model continuously learns from past auction results to improve its predictions of win probability and optimize future bid shading decisions.
2Loss of energy
If a low bid value is submitted to reduce cost, then the bid cost decreases, but the auction win probability decreases
Solution Approach 1:
The system replaces the simple mechanical bid submission with an intelligent machine learning-based bid optimization system. The ML model processes multiple features and historical data to determine the optimal shaded bid value that achieves the desired win probability at lower cost.
Solution Approach 2:
The system dynamically adjusts the bid parameter by applying a shading factor to the true bid value. The machine learning model calculates the optimal shaded bid value based on predicted win probability, ensuring sufficient bid reduction while maintaining acceptable win rates.
3Loss of energy
If bid shading is applied to reduce bid cost, then the bid surplus increases, but the bid value may fall below the minimum bid to win
Solution Approach 1:
The system makes the bid shading dynamic by adjusting the shaded bid value based on the specific auction context, including the minimum bid requirement. The machine learning model adapts the shading factor in real-time to ensure the shaded bid remains above the minimum bid threshold while maximizing bid surplus.
Solution Approach 2:
The system uses feedback from auction results and bid outcomes to continuously improve the machine learning model's ability to predict optimal shading. The model learns from instances where bidding close to minimum thresholds to adjust future bid shading decisions.
Data Source
AI summary
Shaded bid values may be determined and/or submitted to one or more auction modules for participation in auctions. Auction information including at least one of impression indications associated with the auctions, sets of features associated with the auctions, the shaded bid values associated with the auctions, etc. may be stored in a database. A machine learning model may be trained using the auction information to generate a first machine learning model with feature parameters associated with features. A bid request, indicative of a second set of features, may be received. The first machine learning model may be used to determine win probabilities and/or expected bid surpluses associated with multiple shaded bid values based upon one or more feature parameters, of the feature parameters, associated with the second set of features. A shaded bid value for submission may be determined based upon the win probabilities and/or the expected bid surpluses.