On February 25, 2014, ISET hosted Dr. Sergei Izmalkov of the Moscow New Economic School. Dr. Izmalkov, who works with the most popular Russian search engine, Yandex, presented a paper he co-authored with Dmitry Arkhanglesky, Dilyara Khakimova, and Sergey Lyashenko, titled: “On the evaluation of click-through rates and conversion rates of different positions in sponsored search auctions”.

At the beginning of his presentation, Dr. Izmalkov introduced the problem of advertisement allocation. Using the examples of the most popular search engines, he explained how advertisements are generally distributed and how the payment for those is defined. Payment is proportional to the number of clicks a link gets, through the search results that the advertisement will appear on are decided by auction.

Auctions are used extensively in online commerce for allocating advertisements, namely generalized second-price auctions because of their efficiency. Firms make one bid for the advertisement and there are three different positions. The first three bidders get the right to allocate advertisements and pay the price of the next highest bid multiplied by some coefficient determined by the click-through rates (CTR) for each position. The CTR for each position is the probability of clicking on an ad placed in that position Knowing the CTRs is crucial for running these auctions, e.g. for determining the optimal bid for each participant and for predicting the individual CTRs necessary for the optimal placement of ads. It is important to note that estimations of the CTRs based only on actual data are likely to be biased. That is why Yandex runs an experiment for estimating the CTRs.

After the presentation of the experiment and its results, the audience asked a number of questions. ISET would like to thank Dr. Izmalkov for providing an engaging presentation about a very interesting topic.