Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Modified second-price auctions in real-time bidding

Stigland, Oskar LU (2019) NEKH01 20182
Department of Economics
Abstract (Swedish)
The surge in revenue from real-time auctions for online display advertisement has spurred great research interest into how to design optimal mechanisms for these auctions, with respect to both buyers and sellers. This thesis focuses on some recent proposals on how to modify the Vickrey auction, which has been dominating the world of real-time auctions for the past decade, in order to better accommodate for the dynamics of this specific auction environment while also increasing the revenue of the seller. Specifically, this thesis aims to evaluate the so-called boosted second-price auction on a dataset provided by Adform, which is a global advertisement technology company based in Copenhagen. In this specific case,
the implementation of the... (More)
The surge in revenue from real-time auctions for online display advertisement has spurred great research interest into how to design optimal mechanisms for these auctions, with respect to both buyers and sellers. This thesis focuses on some recent proposals on how to modify the Vickrey auction, which has been dominating the world of real-time auctions for the past decade, in order to better accommodate for the dynamics of this specific auction environment while also increasing the revenue of the seller. Specifically, this thesis aims to evaluate the so-called boosted second-price auction on a dataset provided by Adform, which is a global advertisement technology company based in Copenhagen. In this specific case,
the implementation of the boosted second-price auction does not seem to be justified. The changes in allocation as a result of implementing the mechanism are unstable and hard to constrain, while there doesn't seem to be a reason to assume that the Vickrey auction does not already achieve an ecient allocation. All of the code used is provided in the following GitHub repository: https://github.com/Ostigland/econ-rtb (Less)
Please use this url to cite or link to this publication:
author
Stigland, Oskar LU
supervisor
organization
course
NEKH01 20182
year
type
M2 - Bachelor Degree
subject
keywords
Real-time bidding, Vickrey auction, second-price auction, modified second-price auction, boosted second-price auction, adverse selection
language
English
id
8974220
date added to LUP
2019-04-09 08:00:25
date last changed
2019-04-09 08:00:25
@misc{8974220,
  abstract     = {{The surge in revenue from real-time auctions for online display advertisement has spurred great research interest into how to design optimal mechanisms for these auctions, with respect to both buyers and sellers. This thesis focuses on some recent proposals on how to modify the Vickrey auction, which has been dominating the world of real-time auctions for the past decade, in order to better accommodate for the dynamics of this specific auction environment while also increasing the revenue of the seller. Specifically, this thesis aims to evaluate the so-called boosted second-price auction on a dataset provided by Adform, which is a global advertisement technology company based in Copenhagen. In this specific case,
the implementation of the boosted second-price auction does not seem to be justified. The changes in allocation as a result of implementing the mechanism are unstable and hard to constrain, while there doesn't seem to be a reason to assume that the Vickrey auction does not already achieve an ecient allocation. All of the code used is provided in the following GitHub repository: https://github.com/Ostigland/econ-rtb}},
  author       = {{Stigland, Oskar}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Modified second-price auctions in real-time bidding}},
  year         = {{2019}},
}