Budget optimization for online campaigns with positive carryover effects

Nikolay Archak, Vahab Mirrokni, Shanmugavelayutham Muthukrishnan

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    While it is relatively easy to start an online advertising campaign, proper allocation of the marketing budget is far from trivial. A major challenge faced by the marketers attempting to optimize their campaigns is in the sheer number of variables involved, the many individual decisions they make in fixing or changing these variables, and the nontrivial short and long-term interplay among these variables and decisions. In this paper, we study interactions among individual advertising decisions using a Markov model of user behavior. We formulate the budget allocation task of an advertiser as a constrained optimal control problem for a Markov Decision Process (MDP). Using the theory of constrained MDPs, a simple LP algorithm yields the optimal solution. Our main result is that, under a reasonable assumption that online advertising has positive carryover effects on the propensity and the form of user interactions with the same advertiser in the future, there is a simple greedy algorithm for the budget allocation with the worst-case running time cubic in the number of model states (potential advertising keywords) and an efficient parallel implementation in a distributed computing framework like MapReduce. Using real-world anonymized datasets from sponsored search advertising campaigns of several advertisers, we evaluate performance of the proposed budget allocation algorithm, and show that the greedy algorithm performs well compared to the optimal LP solution on these datasets and that both show consistent 5-10% improvement in the expected revenue against the optimal baseline algorithm ignoring carryover effects.

    Original languageEnglish (US)
    Title of host publicationInternet and Network Economics - 8th International Workshop, WINE 2012, Proceedings
    Pages86-99
    Number of pages14
    DOIs
    StatePublished - Dec 26 2012
    Event8th International Workshop on Internet and Network Economics, WINE 2012 - Liverpool, United Kingdom
    Duration: Dec 10 2012Dec 12 2012

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume7695 LNCS
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349

    Conference

    Conference8th International Workshop on Internet and Network Economics, WINE 2012
    CountryUnited Kingdom
    CityLiverpool
    Period12/10/1212/12/12

    Fingerprint

    Carry-over Effects
    Marketing
    Optimization
    Greedy Algorithm
    Task Allocation
    Constrained Control
    MapReduce
    User Behavior
    Markov Decision Process
    User Interaction
    Parallel Implementation
    Distributed Computing
    Efficient Implementation
    Markov Model
    Distributed computer systems
    Optimal Control Problem
    Baseline
    Trivial
    Optimal Solution
    Optimise

    ASJC Scopus subject areas

    • Theoretical Computer Science
    • Computer Science(all)

    Cite this

    Archak, N., Mirrokni, V., & Muthukrishnan, S. (2012). Budget optimization for online campaigns with positive carryover effects. In Internet and Network Economics - 8th International Workshop, WINE 2012, Proceedings (pp. 86-99). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7695 LNCS). https://doi.org/10.1007/978-3-642-35311-6_7

    Budget optimization for online campaigns with positive carryover effects. / Archak, Nikolay; Mirrokni, Vahab; Muthukrishnan, Shanmugavelayutham.

    Internet and Network Economics - 8th International Workshop, WINE 2012, Proceedings. 2012. p. 86-99 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7695 LNCS).

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Archak, N, Mirrokni, V & Muthukrishnan, S 2012, Budget optimization for online campaigns with positive carryover effects. in Internet and Network Economics - 8th International Workshop, WINE 2012, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7695 LNCS, pp. 86-99, 8th International Workshop on Internet and Network Economics, WINE 2012, Liverpool, United Kingdom, 12/10/12. https://doi.org/10.1007/978-3-642-35311-6_7
    Archak N, Mirrokni V, Muthukrishnan S. Budget optimization for online campaigns with positive carryover effects. In Internet and Network Economics - 8th International Workshop, WINE 2012, Proceedings. 2012. p. 86-99. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-35311-6_7
    Archak, Nikolay ; Mirrokni, Vahab ; Muthukrishnan, Shanmugavelayutham. / Budget optimization for online campaigns with positive carryover effects. Internet and Network Economics - 8th International Workshop, WINE 2012, Proceedings. 2012. pp. 86-99 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
    @inproceedings{49b9e148f0c741c3988ec5810b1bdfb9,
    title = "Budget optimization for online campaigns with positive carryover effects",
    abstract = "While it is relatively easy to start an online advertising campaign, proper allocation of the marketing budget is far from trivial. A major challenge faced by the marketers attempting to optimize their campaigns is in the sheer number of variables involved, the many individual decisions they make in fixing or changing these variables, and the nontrivial short and long-term interplay among these variables and decisions. In this paper, we study interactions among individual advertising decisions using a Markov model of user behavior. We formulate the budget allocation task of an advertiser as a constrained optimal control problem for a Markov Decision Process (MDP). Using the theory of constrained MDPs, a simple LP algorithm yields the optimal solution. Our main result is that, under a reasonable assumption that online advertising has positive carryover effects on the propensity and the form of user interactions with the same advertiser in the future, there is a simple greedy algorithm for the budget allocation with the worst-case running time cubic in the number of model states (potential advertising keywords) and an efficient parallel implementation in a distributed computing framework like MapReduce. Using real-world anonymized datasets from sponsored search advertising campaigns of several advertisers, we evaluate performance of the proposed budget allocation algorithm, and show that the greedy algorithm performs well compared to the optimal LP solution on these datasets and that both show consistent 5-10{\%} improvement in the expected revenue against the optimal baseline algorithm ignoring carryover effects.",
    author = "Nikolay Archak and Vahab Mirrokni and Shanmugavelayutham Muthukrishnan",
    year = "2012",
    month = "12",
    day = "26",
    doi = "10.1007/978-3-642-35311-6_7",
    language = "English (US)",
    isbn = "9783642353109",
    series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
    pages = "86--99",
    booktitle = "Internet and Network Economics - 8th International Workshop, WINE 2012, Proceedings",

    }

    TY - GEN

    T1 - Budget optimization for online campaigns with positive carryover effects

    AU - Archak, Nikolay

    AU - Mirrokni, Vahab

    AU - Muthukrishnan, Shanmugavelayutham

    PY - 2012/12/26

    Y1 - 2012/12/26

    N2 - While it is relatively easy to start an online advertising campaign, proper allocation of the marketing budget is far from trivial. A major challenge faced by the marketers attempting to optimize their campaigns is in the sheer number of variables involved, the many individual decisions they make in fixing or changing these variables, and the nontrivial short and long-term interplay among these variables and decisions. In this paper, we study interactions among individual advertising decisions using a Markov model of user behavior. We formulate the budget allocation task of an advertiser as a constrained optimal control problem for a Markov Decision Process (MDP). Using the theory of constrained MDPs, a simple LP algorithm yields the optimal solution. Our main result is that, under a reasonable assumption that online advertising has positive carryover effects on the propensity and the form of user interactions with the same advertiser in the future, there is a simple greedy algorithm for the budget allocation with the worst-case running time cubic in the number of model states (potential advertising keywords) and an efficient parallel implementation in a distributed computing framework like MapReduce. Using real-world anonymized datasets from sponsored search advertising campaigns of several advertisers, we evaluate performance of the proposed budget allocation algorithm, and show that the greedy algorithm performs well compared to the optimal LP solution on these datasets and that both show consistent 5-10% improvement in the expected revenue against the optimal baseline algorithm ignoring carryover effects.

    AB - While it is relatively easy to start an online advertising campaign, proper allocation of the marketing budget is far from trivial. A major challenge faced by the marketers attempting to optimize their campaigns is in the sheer number of variables involved, the many individual decisions they make in fixing or changing these variables, and the nontrivial short and long-term interplay among these variables and decisions. In this paper, we study interactions among individual advertising decisions using a Markov model of user behavior. We formulate the budget allocation task of an advertiser as a constrained optimal control problem for a Markov Decision Process (MDP). Using the theory of constrained MDPs, a simple LP algorithm yields the optimal solution. Our main result is that, under a reasonable assumption that online advertising has positive carryover effects on the propensity and the form of user interactions with the same advertiser in the future, there is a simple greedy algorithm for the budget allocation with the worst-case running time cubic in the number of model states (potential advertising keywords) and an efficient parallel implementation in a distributed computing framework like MapReduce. Using real-world anonymized datasets from sponsored search advertising campaigns of several advertisers, we evaluate performance of the proposed budget allocation algorithm, and show that the greedy algorithm performs well compared to the optimal LP solution on these datasets and that both show consistent 5-10% improvement in the expected revenue against the optimal baseline algorithm ignoring carryover effects.

    UR - http://www.scopus.com/inward/record.url?scp=84871396553&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84871396553&partnerID=8YFLogxK

    U2 - 10.1007/978-3-642-35311-6_7

    DO - 10.1007/978-3-642-35311-6_7

    M3 - Conference contribution

    SN - 9783642353109

    T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

    SP - 86

    EP - 99

    BT - Internet and Network Economics - 8th International Workshop, WINE 2012, Proceedings

    ER -