Offer Optimization, a Dual Approach
summary
A practical near-optimal bidding strategy for demand-side platforms (DSPs) in real-time bidding (RTB) auctions that scales to thousands of campaigns and billions of auctions daily. The method formulates two stylized but realistic constrained global optimization problems, relaxes the combinatorial problem to a convex program, and uses Lagrangian duality: compute real-time first- and second-price bids from shadow prices, while updating shadow prices daily via a subgradient descent algorithm that converges logarithmically and is robust to market noise and shifts.
AI-written summary · about
chapters (9)
- 0:00Offer optimization intro
- 1:13Auction problems in ads
- 8:35Ad contract types
- 11:250-1 knapsack reduction
- 16:20Optimal campaign allocation
- 21:58Convex relaxation & Lagrange multipliers
- 25:30Dual function & conjugate
- 32:57Algorithm implementation
- 36:58Q&A
related talks
40mOnline Ad Allocation: Robust Optimization for Repeated Auctions
44mOptimizing Marketing Spend with Auto-Bidding Algorithms
01h 07mMulti-Dimensional and Non-Linear Mechanism Design and Approximation I
19mMarket Algorithms for Autobidding
26mWin More Effectively in Real-Time Bidding
18mProduct Management for Optimization
29mAI and ML for Computational Advertising
21mQuestions you Should be Asking Your DSP