Optimal budget allocation strategies for real time bidding in display advertising

We present a simulation model to determine optimal budget allocation strategies for real-time bidding (RTB) based display advertising. A common challenge across RTB exchanges is to optimize both budget spend and performance attainment. Our simulation model uses a stochastic dynamic programming approach based budget allocation to determine budget for each time instant. We report on results from a real-world pilot in which our approach delivered an average 18% performance gain.