Maximizing revenue for publishers using header bidding and ad exchange auctions

Jason Rhuggenaath; Reza Refaei Afshar; Alp Akcay; Yingqian Zhang; Uzay Kaymak; Fatih Çolak; Muratcan Tanyerli

doi:10.1016/j.orl.2021.01.008

Maximizing revenue for publishers using header bidding and ad exchange auctions

Jason Rhuggenaath (Corresponding author), Reza Refaei Afshar, Alp Akcay, Yingqian Zhang, Uzay Kaymak, Fatih Çolak, Muratcan Tanyerli

Research output: Contribution to journal › Article › Academic › peer-review

1 Citation (Scopus)

Abstract

We study how web publishers should set their floor prices in order to maximize expected revenues when they have access to two selling mechanisms, namely an ad exchange and header bidding, in order to sell impressions on the real-time bidding market. We consider the publisher’s problem under incomplete information, propose bandit-type algorithms, and show that their regret – the performance loss compared to the optimal algorithm – is sub-linear in the time horizon. Experiments illustrate the effectiveness of our algorithms.

Original language	English
Pages (from-to)	250-256
Number of pages	7
Journal	Operations Research Letters
Volume	49
Issue number	2
DOIs	https://doi.org/10.1016/j.orl.2021.01.008
Publication status	Published - Mar 2021

Keywords

Auctions
Floor price
Header bidding
Machine learning
Multi-armed bandits
Pricing

Access to Document

10.1016/j.orl.2021.01.008

Cite this

@article{7eb0293639e14323be63a0884f2bf6f7,

title = "Maximizing revenue for publishers using header bidding and ad exchange auctions",

abstract = "We study how web publishers should set their floor prices in order to maximize expected revenues when they have access to two selling mechanisms, namely an ad exchange and header bidding, in order to sell impressions on the real-time bidding market. We consider the publisher{\textquoteright}s problem under incomplete information, propose bandit-type algorithms, and show that their regret – the performance loss compared to the optimal algorithm – is sub-linear in the time horizon. Experiments illustrate the effectiveness of our algorithms.",

keywords = "Auctions, Floor price, Header bidding, Machine learning, Multi-armed bandits, Pricing",

author = "Jason Rhuggenaath and Afshar, {Reza Refaei} and Alp Akcay and Yingqian Zhang and Uzay Kaymak and Fatih {\c C}olak and Muratcan Tanyerli",

year = "2021",

month = mar,

doi = "10.1016/j.orl.2021.01.008",

language = "English",

volume = "49",

pages = "250--256",

journal = "Operations Research Letters",

issn = "0167-6377",

publisher = "Elsevier",

number = "2",

}

Maximizing revenue for publishers using header bidding and ad exchange auctions. / Rhuggenaath, Jason (Corresponding author); Afshar, Reza Refaei ; Akcay, Alp ; Zhang, Yingqian ; Kaymak, Uzay; Çolak, Fatih; Tanyerli, Muratcan.

In: Operations Research Letters, Vol. 49, No. 2, 03.2021, p. 250-256.

Research output: Contribution to journal › Article › Academic › peer-review

TY - JOUR

T1 - Maximizing revenue for publishers using header bidding and ad exchange auctions

AU - Rhuggenaath, Jason

AU - Afshar, Reza Refaei

AU - Akcay, Alp

AU - Zhang, Yingqian

AU - Kaymak, Uzay

AU - Çolak, Fatih

AU - Tanyerli, Muratcan

PY - 2021/3

Y1 - 2021/3

N2 - We study how web publishers should set their floor prices in order to maximize expected revenues when they have access to two selling mechanisms, namely an ad exchange and header bidding, in order to sell impressions on the real-time bidding market. We consider the publisher’s problem under incomplete information, propose bandit-type algorithms, and show that their regret – the performance loss compared to the optimal algorithm – is sub-linear in the time horizon. Experiments illustrate the effectiveness of our algorithms.

AB - We study how web publishers should set their floor prices in order to maximize expected revenues when they have access to two selling mechanisms, namely an ad exchange and header bidding, in order to sell impressions on the real-time bidding market. We consider the publisher’s problem under incomplete information, propose bandit-type algorithms, and show that their regret – the performance loss compared to the optimal algorithm – is sub-linear in the time horizon. Experiments illustrate the effectiveness of our algorithms.

KW - Auctions

KW - Floor price

KW - Header bidding

KW - Machine learning

KW - Multi-armed bandits

KW - Pricing

UR - http://www.scopus.com/inward/record.url?scp=85100202202&partnerID=8YFLogxK

U2 - 10.1016/j.orl.2021.01.008

DO - 10.1016/j.orl.2021.01.008

M3 - Article

VL - 49

SP - 250

EP - 256

JO - Operations Research Letters

JF - Operations Research Letters

SN - 0167-6377

IS - 2

ER -

Maximizing revenue for publishers using header bidding and ad exchange auctions

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this