This paper develops an effective framework to determine a bidding strategy of an intermediary company for banner advertising. The intermediary company receives offers from clients and then bids for advertising spots for clientsu27 banner advertisings on websites in ad networks. In our framework, we first estimate the number of conversions and the probability of successful bids by using a multiplicative regression model. Then, based on these estimate values, we formulate an optimization problem for determining a bidding strategy, i.e., the frequency of bids and bid prices for each advertising spot, as a linear optimization problem. We solve the optimization problem to maximize the expected number of conversions, which bring benefits to the intermediary company, in consideration of the attributes of websites and banner advertisings. Our numerical experiments were conducted by using an actual data set which was provided by an intermediary company. One-month data was divided into the first half which was used as a training data set, and the latter half which was used as a testing data set. Through the regression analysis, we clarified the factors affecting the number of conversions and the probability of successful bids. Moreover, the simulation results showed that our bidding strategy can improve profits of the intermediary company.
展开▼
