Combining behaviors and demographics to segment online audiences

Experiments with a youtube channel

Bernard Jansen, Soon Gyo Jung, Joni Salminen, Jisun An, Haewoon Kwak

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

Abstract

Social media channels with audiences in the millions are increasingly common. Efforts at segmenting audiences for populations of these sizes can result in hundreds of audience segments, as the compositions of the overall audiences tend to be complex. Although understanding audience segments is important for strategic planning, tactical decision making, and content creation, it is unrealistic for human decision makers to effectively utilize hundreds of audience segments in these tasks. In this research, we present efforts at simplifying the segmentation of audience populations to increase their practical utility. Using millions of interactions with hundreds of thousands of viewers with an organization’s online content collection, we first isolate the maximum number of audience segments, based on behavioral profiling, and then demonstrate a computational approach of using non-negative matrix factorization to reduce this number to 42 segments that are both impactful and representative segments of the overall population. Initial results are promising, and we present avenues for future research leveraging our approach.

Original languageEnglish
Title of host publicationInternet Science - 5th International Conference, INSCI 2018, Proceedings
EditorsSvetlana S. Bodrunova
PublisherSpringer Verlag
Pages141-153
Number of pages13
ISBN (Print)9783030014360
DOIs
Publication statusPublished - 1 Jan 2018
Event5th International Conference on Internet Science, INSCI 2018 - St. Petersburg, Russian Federation
Duration: 24 Oct 201826 Oct 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11193 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other5th International Conference on Internet Science, INSCI 2018
CountryRussian Federation
CitySt. Petersburg
Period24/10/1826/10/18

Fingerprint

Strategic planning
Factorization
Decision making
Chemical analysis
Experiment
Experiments
Strategic Planning
Non-negative Matrix Factorization
Social Media
Profiling
Segmentation
Decision Making
Tend
Interaction
Demonstrate

Keywords

  • Audience analytics
  • Audience segmentation
  • User profiling

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Jansen, B., Jung, S. G., Salminen, J., An, J., & Kwak, H. (2018). Combining behaviors and demographics to segment online audiences: Experiments with a youtube channel. In S. S. Bodrunova (Ed.), Internet Science - 5th International Conference, INSCI 2018, Proceedings (pp. 141-153). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11193 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-01437-7_12

Combining behaviors and demographics to segment online audiences : Experiments with a youtube channel. / Jansen, Bernard; Jung, Soon Gyo; Salminen, Joni; An, Jisun; Kwak, Haewoon.

Internet Science - 5th International Conference, INSCI 2018, Proceedings. ed. / Svetlana S. Bodrunova. Springer Verlag, 2018. p. 141-153 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11193 LNCS).

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

Jansen, B, Jung, SG, Salminen, J, An, J & Kwak, H 2018, Combining behaviors and demographics to segment online audiences: Experiments with a youtube channel. in SS Bodrunova (ed.), Internet Science - 5th International Conference, INSCI 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11193 LNCS, Springer Verlag, pp. 141-153, 5th International Conference on Internet Science, INSCI 2018, St. Petersburg, Russian Federation, 24/10/18. https://doi.org/10.1007/978-3-030-01437-7_12
Jansen B, Jung SG, Salminen J, An J, Kwak H. Combining behaviors and demographics to segment online audiences: Experiments with a youtube channel. In Bodrunova SS, editor, Internet Science - 5th International Conference, INSCI 2018, Proceedings. Springer Verlag. 2018. p. 141-153. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-01437-7_12
Jansen, Bernard ; Jung, Soon Gyo ; Salminen, Joni ; An, Jisun ; Kwak, Haewoon. / Combining behaviors and demographics to segment online audiences : Experiments with a youtube channel. Internet Science - 5th International Conference, INSCI 2018, Proceedings. editor / Svetlana S. Bodrunova. Springer Verlag, 2018. pp. 141-153 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d117d30d30c04dfc82233bf893e451b6,
title = "Combining behaviors and demographics to segment online audiences: Experiments with a youtube channel",
abstract = "Social media channels with audiences in the millions are increasingly common. Efforts at segmenting audiences for populations of these sizes can result in hundreds of audience segments, as the compositions of the overall audiences tend to be complex. Although understanding audience segments is important for strategic planning, tactical decision making, and content creation, it is unrealistic for human decision makers to effectively utilize hundreds of audience segments in these tasks. In this research, we present efforts at simplifying the segmentation of audience populations to increase their practical utility. Using millions of interactions with hundreds of thousands of viewers with an organization’s online content collection, we first isolate the maximum number of audience segments, based on behavioral profiling, and then demonstrate a computational approach of using non-negative matrix factorization to reduce this number to 42 segments that are both impactful and representative segments of the overall population. Initial results are promising, and we present avenues for future research leveraging our approach.",
keywords = "Audience analytics, Audience segmentation, User profiling",
author = "Bernard Jansen and Jung, {Soon Gyo} and Joni Salminen and Jisun An and Haewoon Kwak",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-030-01437-7_12",
language = "English",
isbn = "9783030014360",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "141--153",
editor = "Bodrunova, {Svetlana S.}",
booktitle = "Internet Science - 5th International Conference, INSCI 2018, Proceedings",

}

TY - GEN

T1 - Combining behaviors and demographics to segment online audiences

T2 - Experiments with a youtube channel

AU - Jansen, Bernard

AU - Jung, Soon Gyo

AU - Salminen, Joni

AU - An, Jisun

AU - Kwak, Haewoon

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Social media channels with audiences in the millions are increasingly common. Efforts at segmenting audiences for populations of these sizes can result in hundreds of audience segments, as the compositions of the overall audiences tend to be complex. Although understanding audience segments is important for strategic planning, tactical decision making, and content creation, it is unrealistic for human decision makers to effectively utilize hundreds of audience segments in these tasks. In this research, we present efforts at simplifying the segmentation of audience populations to increase their practical utility. Using millions of interactions with hundreds of thousands of viewers with an organization’s online content collection, we first isolate the maximum number of audience segments, based on behavioral profiling, and then demonstrate a computational approach of using non-negative matrix factorization to reduce this number to 42 segments that are both impactful and representative segments of the overall population. Initial results are promising, and we present avenues for future research leveraging our approach.

AB - Social media channels with audiences in the millions are increasingly common. Efforts at segmenting audiences for populations of these sizes can result in hundreds of audience segments, as the compositions of the overall audiences tend to be complex. Although understanding audience segments is important for strategic planning, tactical decision making, and content creation, it is unrealistic for human decision makers to effectively utilize hundreds of audience segments in these tasks. In this research, we present efforts at simplifying the segmentation of audience populations to increase their practical utility. Using millions of interactions with hundreds of thousands of viewers with an organization’s online content collection, we first isolate the maximum number of audience segments, based on behavioral profiling, and then demonstrate a computational approach of using non-negative matrix factorization to reduce this number to 42 segments that are both impactful and representative segments of the overall population. Initial results are promising, and we present avenues for future research leveraging our approach.

KW - Audience analytics

KW - Audience segmentation

KW - User profiling

UR - http://www.scopus.com/inward/record.url?scp=85055827574&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055827574&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-01437-7_12

DO - 10.1007/978-3-030-01437-7_12

M3 - Conference contribution

SN - 9783030014360

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

SP - 141

EP - 153

BT - Internet Science - 5th International Conference, INSCI 2018, Proceedings

A2 - Bodrunova, Svetlana S.

PB - Springer Verlag

ER -