Discovery of repost patterns by topic analysis in enterprise social networking
| Date | 20 March 2017 |
| DOI | https://doi.org/10.1108/AJIM-08-2016-0128 |
| Published date | 20 March 2017 |
| Pages | 158-173 |
| Author | Jianhong Luo,Xuwei Pan,Xiyong Zhu |
| Subject Matter | Library & information science,Information behaviour & retrieval,Information & knowledge management,Information management & governance,Information management |
Discovery of repost patterns by
topic analysis in enterprise
social networking
Jianhong Luo, Xuwei Pan and Xiyong Zhu
Department of Management Science and Engineering,
Zhejiang Sci-Tech University, Hangzhou, China
Abstract
Purpose –An increasing number of users are inspired by enterprises to repost social media messages, which
greatly contributes to the dissemination of such messages in an online social network. The purpose of this
paper is to discover the repost patterns of users regarding enterprise social media messages to help
enterprises improve information management abilities for social media.
Design/methodology/approach –This paper proposes a novel method to discover the repost patterns of
users in enterprise social networking (ESN) at the macro-level through topic analysis. Specifically, it proposes
the message-diversity metric to measure the latent topic diversity degree of the social media messages.
Through this technique, the paper analyzes the message-diversity characteristics of the enterprise social
media messages and then explores the repost patterns of users.
Findings –The experimental results show that a high repost rate is more prominent for the messages with
diverse latent topics, where message-diversity is as high as 0.5.
Practical implications –The findings have great potential in several management areas, such as
employing social media marketing, predicting popular messages, helping enterprises strengthen their online
presence, and gathering more potential customers.
Originality/value –This study explores how the repost patterns of users in ESN can be determined through
general macro-level behavior of users instead of their micro-level processes. The patterns can also lead to a
deeper understanding of which contents can drive people to diffuse information. This study gives an
important insight into the information behavior of social media users for enterprise management researchers.
Keywords Content analysis, Enterprise social networking, Message-diversity, Repost pattern,
Retweet pattern, Topic analysis
Paper type Research paper
1. Introduction
Online social media has become a part of our daily routines. People can easily share and
spread messages to their friends or strangers through the internet and their smartphones
regardless of time and place. Social media have various forms, such as Sina Weibo,
Twitter and Facebook, where users can exchange messages. Enterprise social networking
(ESN) is an organization’s use of online social media among people who share business
interests or activities.
Many enterprises have begun to utilize social media to communicate with their
consumers and then formed their ESN. The accumulated data on their ESN contain
important information for analyzing their consumers. Therefore, smart social media
marketing (SMM) strategies must be developed to help businesses attract more customers at
a low cost. In this way, the business value of the ESN can be improved. SMM have vast
space for further development and will become a new research trend in the future
(Gan and Wang, 2015). Medium and small enterprises play an important role in the
Aslib Journal of Information
Management
Vol. 69 No. 2, 2017
pp. 158-173
© Emerald PublishingLimited
2050-3806
DOI 10.1108/AJIM-08-2016-0128
Received 12 August 2016
Revised 20 November 2016
22 February 2017
27 February 2017
Accepted 28 February 2017
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/2050-3806.htm
This work was supported by the National Nature Science Foundation of China (Grant No. 71501172)
and Zhejiang Provincial Natural Science Foundation of China (Grant No. LQ14G020016). This research
was also supported by the National Nature Science Foundation of China (Grant No. 71471165) and 2014
university visiting scholars professional development programs of Education Department of Zhejiang
Province of China (Grant No. FX2014017).
158
AJIM
69,2
economy. Given that all enterprises attempt to investigate new methods of growth, they can
expand their efforts, attract customers, and ultimately achieve growth by utilizing SMM
methods. However, medium and small enterprises are slow in terms of embracing
SMM channels as outlets to promote their products and services. The task of balancing and
managing a business, as well as attempting to reach potential customers through new social
media channels, may be difficult for these organizations. In sum, these businesses have a
limited understanding of the social media behavior of customers.
Social media are important tools for information propagation. In a social media
platform, messages are mainly spread through repost/retweet behaviors. When a user
posts a message, such a message can be pushed to his/her followers, who in turn decide
whether to repost the same message. If reposted, the message will be pushed to another set
of followers. In this way, the message will keep spreading throughout the social network.
Repost number, an important indicator of repost behavior, denotes how many times a
message has been reposted.
Some studies (Ma et al., 2013; Yang et al., 2010) have investigated the information
dissemination and retweet behaviors in microblogs. Most of these studies (Sasa Petrovic,
2011) have treated such problems as classification problems, such as a two-class
classification problem or a multi-class classification problem. Two-class classification
problem studies have investigated whether a message will be retweeted, while multi-class
classification problem studies have created retweet prediction models by extracting the
appropriate features and choosing suitable classifiers. Most of the existing methods for
studying repost/retweet messages check the interest of an individual toward the content or
his/her influence on the social network, but ignore the repost behavior of the “crowd”toward
the contents of the message. In other words, most of the existing methods have focused
solely on repost/retweet behaviors of the individuals at the micro-level and lacked of
considering repost behavior of the crowds at the macro-level.
The repost process on Sina Weibo/Twitter shows that the reposted messages mainly come
from two sources, namely, messages from the direct and indirect followers of a user. Those
users who repost the social media messages of an enterprise are considered latent customers.
Therefore, more users must be encouraged to repost a message to improve the dissemination
process and for such message to reach more latent customers in the social network. In addition,
appropriate messages must be posted to attract more followers and to increase the repost
numbers from the direct and indirect followers of an enterprise. However, what kinds of
content can attract users to repost? In other words, can we find the repost patterns of the
“crowds”users regarding the messages in an enterprise social network? If so, what are these
patterns? To answer this question, the messages that are reposted by users must be analyzed.
In particular, the content of these messages must be characterized in a way that the repost
patterns of users can be naturally measured. The repost patterns of users, especially in ESN,
have not yet been thoroughly investigated in the literature.
Despite the growing body of resear ch on social media and repost behavi or,
understanding the responses of users to the reposts in an enterprise social network
remains a challenge. Identifying the strategies for disseminating messages online has high
practical significance in SMM. Thus, the current paper aims to provide insights into the
repost behavior of the “crowds”users for business SMM to help improve the information
management ability of enterprise social media.
In order to contribute to these research objectives, we offer a perspective based on
contents analysis to mine social media messages and get insights directly from the users’
early reaction for the messages. We consider the special case in which Sina Weibo is used as
a channel by enterprises to post messages for SMM. In this special case the users’early
repost reaction is occurred according to their interest in the content of the message, and to
repost the message.
159
Enterprise
social
networking
To continue reading
Request your trial