BUZZ NETWORK ON THE WEBLOG COMMUNITY
Shinya Takami, Katsumi Tanaka
Graduate School of Informatics, Kyoto University
Keywords: Buzz Marketing, Advertising Effectiveness, Weblog, Community.
Abstract: As we don't give priority to the mass media as trustful information, effective expansion of buzz (word-of-
mouth) will be expected as an important marketing approach. It seems that some buzzes on the Internet
spread by information on weblogs. Thus it is thought that a kind of buzz network is formed in weblog
communities. It is important theme in our research to analyze the structure of the buzz network on the
weblog communities and to find the hubs and the connectors on the network. We define a set of entries on a
weblog as a 'Buzz-Node' and each buzz-node has a specific characteristic. 'Buzz-Hubs' and 'Buzz-
Connectors' are the important weblogs that have the buzz-node connecting to many other buzz-nodes or that
connect two or more clusters. Now we are developing the analysis system of buzz network for weblog
communities called 'WeBz Analyzer' for finding the buzz-hubs and the buzz-connectors.
1 INTRODUCTION
We can easily touch vast and various information
today, not only through TV and newspaper but also
through the Internet. It is not easy for us to trust such
information now, because we are surrounded by too
much information. These days, we often check some
web sites and are sometimes influenced with their
information. The authors may not be famous even if
Yahoo! or Google lists such websites in the high
rank. We also refer to comments in the bulletin
board and the recommendations on Amazon.com.
It becomes more and more important in recent
years although word-of-mouth exists for a long time.
When some hot topics are expanded by word-of-
mouth, each of them is called 'Buzz'. Emanuel Rosen
says that three reasons enhance the importance of
buzz. First, we are exposed to information too much
to process. Second, we doubt the advertisements.
Third, we can share information easily with each
other on the Internet. It is popular for us to access
the buzzes especially on the Internet (Rosen, 2000).
2 BUZZ MARKETING
Our preferences are diversified and the scale of the
group sharing a specific interest tends to become
small. In addition, various advertisement-
distributing media increased enormously by the
appearance of the Internet. Thus the importance of
the media planning to reflect the intention of the
advertisers effectively has increased (Takami et al.,
2004). However, obtaining enough advertising
effectiveness becomes difficult if only using the
mass marketing in recent years.
As we don't give priority to the mass media as
trustful information, effective expansion of word-of-
mouth will be expected as an important marketing
approach. That is called 'Buzz Marketing'. In buzz
marketing, it is important to support the potential
opinion leaders who can expand the reputation of
some topics by word-of-mouth effectively (See
Figure 1). And finding them and estimating the
influence of advertising effectiveness by their
Figure 1: Marketing approaches.
136
Takami S. and Tanaka K. (2006).
BUZZ NETWORK ON THE WEBLOG COMMUNITY.
In Proceedings of WEBIST 2006 - Second International Conference on Web Information Systems and Technologies - Society, e-Business and
e-Government / e-Learning, pages 136-139
DOI: 10.5220/0001250301360139
Copyright
c
SciTePress
contribution are the most critical theme.
3 BUZZ NETWORK ON THE NET
3.1 Buzz Network
It is thought that people construct a kind of network
that consists of the people who are interested in a
specific topic and their friends. People who can
expand some topics by word-of-mouth are called
opinion leaders and they are regarded as the hubs of
the network (See Figure 2). In such a network, the
hubs can be evaluated on the following two aspects.
It is important for opinion leaders that they have
a lot of acquaintances. So the hubs are the nodes that
have many links with other nodes. People who are
similar to each other tend to form clusters. Then it
may not spread outside the cluster though the topic
spreads rapidly inside the cluster. The people who
have links with different clusters are also important
because some topics can spread by word-of-mouth
to the people who are interested in them via the
people who are not. The nodes that have a role of
connecting two or more different clusters are called
‘Connectors’.
When people who have some authority, we tend
to trust or refer their information. We often willing
to buy something which celebrities wear. Celebrities
can be hubs in the network and they are called social
hubs. Professionals who have the charisma in the
specific area look like celebrities. We also trust or
refer the information from the semi-professionals
even if they are not so famous but we can feel that
they are familiar. Some professionals also may be
hubs in the network and they are called expert hubs.
3.2 Buzz on the Internet
In recent years, we can talk about some topics and
share some reputations on the Internet. More and
more products and services become famous by
word-of-mouth on the Internet. Which means that
buzzes are born and expanded on the Internet. So the
Internet has the power to expand some topics by
word-of-mouth and buzz marketing is effective on
the Internet.
Advertising effectiveness by word-of-mouth has
not been focused on as a research area because it is
difficult to analyze the structure and estimate the
influence. The objective in many researches is to
find the hot topics on the Web and we can easily
find the buzzes on the Internet as the contribution of
such researches (Carter, 2005).
According to the report of Japan Advertisers
Association Web Advertising Bureau, over 70% of
writings are positive stance when a product is
referred in weblogs. They analyzed about one
million weblogs and found that there are a lot of
weblogs that have suitable contents for advertisers
but they have only a few readers. They say that it is
profitable mechanism to inform such detailed and
excellent weblogs for potential buyers (WAB,
2005).
Then we pay attention to weblogs as the potential
media that can generate buzzes and expand them by
word-of-mouth. A weblog consists of entries that are
linked with other weblogs or web sites. Most
weblogs have some categories and each category has
entries about a specific topic. The connections with
other weblogs are hyperlinks, trackbacks and
comments (See Figure 3).
4 BUZZ NETWORK AND
WEBLOG COMMUNITY
4.1 Weblog Community
Many opinions are exchanged through the weblogs,
and we are influenced with them. Generally, word-
of-mouth spreads by people. But it seems that word-
Figure 2: A structure of the Buzz Network.
Figure 3: A structure of the Weblog Network.
BUZZ NETWORK ON THE WEBLOG COMMUNITY
137
of-mouth on the Internet spreads by information on
weblogs. Thus some weblog groups can be regarded
as buzz networks.
There are some famous and popular weblogs
which authors are celebrities. They are linked with
enormous other weblogs and some topics on their
entries have strong influences on the Internet. It
seems that celebrity's weblogs can be regarded as
social hubs on buzz network. On the other hand,
there are various professional weblogs which
authors are specialists in each area. Such weblogs
also have some trustful influence on many people in
each specific area. They can be also regarded as
expert hubs on buzz network. Thus, the weblog has
the characteristic as the hubs on the buzz network
now.
Obtaining some advertising effectiveness by
supporting authors to take up a specific topic in their
weblogs is practiced at Gawker as a new advertising
strategy. Moreover, some researches consider the
relation between the author and the readers by
analyzing the comments on weblogs, and try to use
the contribution for marketing (Glance et al., 2005).
4.2 Buzz-Nodes and Buzz-Hubs
When buzz networks are formed in weblog
communities, it is necessary to define the node and
the link. Each entry of a weblog is used for criteria
in most researches on weblogs. But in our research,
we have to consider categories of weblogs because
entries in each category share a specific topic and
the characteristics of each category define the
clusters. Then we define a set of entries in a category
as a ‘Buzz-Node’.
A weblog has some connections with each other.
Hyperlinks are used to connect with other web sites
or entries on other weblogs. Trackbacks,
characteristic feature of weblog, show that an entry
is referred by an entry of another weblog. Several
entries have some comments from readers. Some of
the readers who write comments have their own
weblogs. So some comments on the weblogs can be
regarded as a link from the commentator's weblogs.
Thus we consider these three connections as links on
buzz network.
A weblog consists of some buzz-nodes and has
some links to other buzz-nodes. Each buzz-node has
a specific characteristic (keyword) that is regarded
as a characteristic of a cluster on the buzz network.
Then we can consider such network structure as the
buzz network. In a word, a weblog has some buzz-
nodes and each buzz-node has a specific keyword
and buzz-nodes are linked with each other (See
Figure 4).
The characteristics of the buzz-nodes are
evaluated by various methods. We can find the
important keywords on the entry by analyzing the
document. The name of category may be used for
the keyword. Several approaches to find such
keywords are proposed in many researches.
Some of the advertisements on the weblogs also
can be used for characterizing them because there
are advertisements automatically displayed suitable
for each web site's characteristics. Then keywords
can be extracted from such advertising contents or
meta data of them. It is important to refer the
advertisements actually displayed on weblogs
because advertisers need to consider the
compatibility with advertising topics on the weblogs.
By extracting the keywords from advertisements, we
can find some potential interests not only for the
genre of topics but also for a specific product or
service.
The hubs are important nodes on the buzz
network. It is thought that such nodes have not only
just many links but also many connections with
similar characteristic nodes. In the weblog
community, a Buzz-Hub is not one of the buzz-
nodes but an important weblog. So a buzz-hub may
consist of some buzz-nodes. It is important for buzz-
hubs to have the buzz-node connecting to many
similar buzz-nodes. The buzz-hubs that have two or
more such buzz-nodes are more important.
Furthermore, the connectors are also important in
buzz marketing for expanding some topics by word-
of-mouth even if the nodes are not hubs. On the buzz
network, such nodes have a role of connecting two
or more different clusters. It looks difficult to find
such connectors on the buzz network by people. In
the weblog community, we consider that a ‘Buzz-
Figure 4: Buzz Network on the Weblog Community.
WEBIST 2006 - SOCIETY, E-BUSINESS AND E-GOVERNMENT
138
Connector’ is the weblog that has two or more buzz-
nodes and each buzz-node belongs to a different
cluster. And the weblog is also buzz-connector that
has the buzz-node connecting two or more similar
clusters even if the characteristic of it is different
from such clusters.
4.3 WeBz Analyzer
Now we are developing the analysis system of buzz
network for weblog communities. We call buzz on
the Web ‘WeBz’ and we named the system ‘WeBz
Analyzer’. WeBz Analyzer can find the buzz-hubs
and buzz-connectors in weblog communities by
analyzing the network structure of weblogs.
The system calculates the number of links with
buzz-nodes that has a same or similar keyword, and
it finds the clusters and priority order of buzz-nodes
is shown. Similarly, it is done in the case of other
keywords. Next, the system analyzes the links with
the buzz-nodes that have different keywords. When
the buzz-node connects similar or different clusters,
Gaussian distribution defines the importance of the
connections for buzz-connectors. Then the system
calculates the total score in each buzz-node and it
shows the important buzz-hubs and buzz-connectors.
5 RELATED WORKS
With the spread of the weblog, hot topics in the
weblog community are often collected and analysed
in some researches (Bar-Ilan, 2004). Technorati
provides a real-time search engine that keeps track
of what is going on in the blogosphere. We can find
the ten words that got into the news most in 24 hours
on weblogs at kizasi.jp. In the meaning of finding
widespread buzzes, such services are interesting. But
it is important in our research to find the potential
web sites suitable for buzz marketing.
In the area of the next generation web search,
some researchers try to rank websites by regarding
two or more websites as a set (Yumoto and Tanaka,
2005). Yumoto and Tanaka propose the idea of
‘multiple viewpoint retrieval’ in web searches
because there are websites that have been described
from different viewpoints for a specific topic. We
also plan to use a feature-vector based approach for
characterising buzz-nodes.
6 CONCLUSION
Many researches try to find the hot topics on the
Web and we can easily find some buzzes on the
Internet as the contribution of them. But it is still
difficult to analyze how buzzes are expanded and to
estimate the advertising influences by word-of-
mouth.
With the spread of Weblog, it seems that buzzes
on the Internet spreads by the information on
weblogs. Then we consider that weblog
communities have a kind of buzz network. In our
research, it is critical to find the buzz-hubs and buzz-
connectors because those are the potential weblogs
with the possibility of expanding buzzes.
We introduced the analysis system of buzz
network for weblog communities called 'WeBz
Analyzer'. Although we are still developing it, we
hope that the system can simulate the advertising
effectiveness by word-of-mouth on the Internet.
REFERENCES
Takami, S., et al. 2004. Support system of advertising
distribution contract for Media-Mix strategy. In Proc.
of CEC04-EAST, IEEE International Conference on e-
Commerce Technology for Dynamic E-Business,
pp122-125.
Bar-Ilan, J. 2004. An outsider's view on topic-oriented
blogging. In Proc. of 13th international World Wide
Web conference on Alternate track papers & posters,
pp28-34.
Yumoto, T. and Tanaka, K. 2005. Finding Pertinent Page-
Pairs from Web Search Results. In Proc. of ICADL
2005, 8th International Conference on Asian Digital
Libraries, pp301-310.
Carter, S. 2005. The role of the author in topical blogs. In
CHI '05 extended abstracts on Human factors in
computing systems, pp1256-1259.
Glance, N. et al. 2005. Analyzing online discussion for
marketing intelligence. In Special interest tracks and
posters of the 14th international conference on World
Wide Web, pp1172-1173.
Japan Advertisers Association Web Advertising Bureau
(WAB), 2005. Press Release on Dec. 13. Retrieved
from http://www.wab.ne.jp/pdf/20051213.pdf
Rosen, E. 2000. The book, Anatomy of Buzz: How to
Create Word-of-Mouth Marketing.
BUZZ NETWORK ON THE WEBLOG COMMUNITY
139