Content uploaded by Masamichi Takahashi
Author content
All content in this area was uploaded by Masamichi Takahashi on Oct 15, 2019
Content may be subject to copyright.
CENTER FOR
COLLECTIVE
INTELLIGENCE
Working Paper Series
How can people and computers be connected so
that-collectively-they act more intelligently than
any individuals, groups, or computers have ever
done before?
THE SHIFT FROM CENTRALIZED TO PEER-TO-PEER
COMMUNICATION IN AN ONLINE COMMUNITY: PARTICIPANTS AS
A USEFUL ASPECT OF GENRE ANALYSIS
Masamich Takahashi
JoAnne Yates
George Herman
Atsushi Ito
Keiichi Nemoto
CCI WORKING PAPER 2008-NUMBER 001
MIT SLOAN SCHOOL OF MANAGEMENT WORKING PAPER 4677-08
© 2008 Takahashi, Yates, Herman, Ito, and Nemoto
Sloan School of Management
3 Cambridge Center, NE20-336
Cambridge, MA 02139-4307
U.S.A.
http://cci.mit.edu/
MIT Sloan School of Management
MIT Sloan School Working Paper 4677-08
1/1/2008
The Shift from Centralized to Peer-to-Peer Communication in an Online Community: Participants as a Useful Aspect of
Genre Analysis
© 2008 Masamichi Takahashi, JoAnne Yates, George Herman, Atsushi Ito, Keiichi Nemoto
Masamichi Takahashi, JoAnne Yates, George Herman, Atsushi Ito, Keiichi Nemoto
All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without
explicit permission, provided that full credit including © notice is given to the source.
This paper also can be downloaded without charge from the
Social Science Research Network Electronic Paper Collection:
http://ssrn.com/abstract=1084674
The Shift from Centralized to Peer-to-Peer Communication in an Online Community:
Participants as a Useful Aspect of Genre Analysis
Masamichi Takahashi
Fuji Xerox Co., Ltd.
430 Sakai,Nakai-machi
Ashigarakami-gun, Kanagawa 259-0157, Japan
Masamichi.Takahashi@fujixerox.co.jp
Formerly Visiting Researcher
Center for Collective Intelligence
MIT Sloan School of Management
JoAnne Yates
Center for Collective Intelligence
MIT Sloan School of Management
MIT E52-475
50 Memorial Drive
Cambridge, MA 02142-1347
jyates@mit.edu
George Herman
MIT Sloan School of Management
MIT NE20-336
50 Memorial Drive
Cambridge, MA 02142-1347
gherman@mit.edu
Formerly Research Scientist
Center for Collective Intelligence
Atsushi Ito
Fuji Xerox Co., Ltd.
430 Sakai,Nakai-machi
Ashigarakami-gun, Kanagawa 259-0157, Japan
Ito.Atsushi@fujixerox.co.jp
Keiichi Nemoto
Fuji Xerox Co., Ltd.
430 Sakai,Nakai-machi
Ashigarakami-gun, Kanagawa 259-0157, Japan
Keiichi.Nemoto@fujixerox.co.jp
Abstract
In this paper we analyzed an online community based on a mailing list that was created
as an internal marketing tool for launching a new network service. We focused on the change in
communication over time among dispersed Sales representatives and the employees in a
centralized Service Department. We conducted a genre analysis based on content (what),
purpose (why), timing (when), form (how) and participants (who communicates to whom) (Yates
and Orlikowski, 2002). Analyzing the participants in a genre and how those participants
changed over time highlighted a shift from centralized to dispersed, peer-to-peer communication
in this community. We highlight implications both for genre analysis and for organizational
practice.
Introduction
In this paper we look at how a relatively informal communication channel based on a
mailing list created an online community in a company and how communication in that
community evolved over time. As online communities have proliferated in recent decades, they
have drawn considerable attention among researchers and practitioners (Herring, 1999; Murray,
2000; Rheingold, 1993; Lewis and Knowles, 1997). In this paper we use the term ‘online
community’ as any virtual social space where people come together to get and give information
or support, to learn, or to find company, although we recognize that there is controversy
regarding the term and that there is no accepted definition (Preece, 2001). Much research in this
area has focused on publicly accessible online communities of interest that cross organizational
2
boundaries (e.g., Baym, 2000), since they are most readily available for analysis. Yet online
communities also exist within organizational settings. Some companies have already
incorporated online communities into their daily work. For example, a product development
team in a Japanese firm adopted a Usenet-based system, adapting it over time to support their
development effort (Orlikowski et al., 1995). Based on ethnographic research about the work of
customer service engineers, researchers at Xerox Corporation built the Eureka online system to
support and improve knowledge sharing over time among repair technicians (Bobrow &Whalen,
2002). Even though most companies have informal, IT-supported networks that cross
organizational boundaries in their work setting, relatively little research has been conducted on
the role of online communities within a company (Bobrow & Whalen, 2002; Füller et al, 2004;
Orlikowski et al, 1995; Quan-Haase & Wellman, 2005). Füller et al. (2004) indicated the
importance of community-based innovation and suggested a method for using the existing
innovative potential of online communities by integrating its members virtually into new product
development. Quan-Haase and Wellman (2005) applied social network analysis to investigate a
computer-mediated community in a software company and made visible the actual lines of
communication within departments, between departments, and outside of the organization in
order to understand how a collaborative community is maintained online and offline. As they
pointed out, more studies are needed that examine online communities in the actual business
context, rather than analytically isolating them. Little attention has focused on the changing role
of an online community in an organization over time.
In this paper, we examine use of an online community in the launching of a new business
in a company that already had an incumbent business. Such a company faces a challenge if it
wants to roll out a new type of product or service using its existing sales channels (Westerman,
3
McFarlan, & Iansiti, 2006). More communication is needed to resolve hitherto unexplored
issues. A list-based online community, we found, played a useful role in supporting the launch of
a new business.
In this paper, we use genre analysis (Yates and Orlikowski, 1992; Orlikowski and Yates,
1994; Yates and Orlikowski, 2002) to understand the changing communication in this
community over time. We focus particularly on participants, that is, who communicates to
whom (who/m) to illuminate subtle variations and shifts in the qualitative nature of the discourse
and how participants’ particular knowledge and powers shape the genres. Communication and
organization researchers have looked at the relationship among different players in
communications, including through social network analysis, to identify the structure of
communications and to consider the meaning and performance of a communication network or
structure (Cross and Cummings, 2004; Quan-Haase and Wellman, 2005). Examination of the
role of participants is quite rare in genre research, however.
In the following section, we’ll review relevant research about genre analysis of
organizational communication and justify our approach and focus. We will then explain the site
and methodology for our study. The heart of the paper presents the results of our analysis: the
genres and genre systems identified and their changes over time, including the intriguing shift
from centralized to peer-to-peer communication. In the subsequent section, we explore
implications of our results for genre researchers and for practitioners.
Genre research
Scholars have used genre as a lens for analysis and design in a variety of organizational
communication (Orlikowski and Yates, 1994; Yates and Orlikowski, 1992; Yates, Orlikowski,
4
and Fonstad, 2001; Yates and Orlikowski, 2002) and information systems research (Crowston
and Williams, 1997; Ihlström and Henfridsson, 2005; Karjalainen et al., 2001). Yates and
Orlikowski (1992) introduced the notion of genres, or socially recognized types of
communication into the organizational literature a decade and a half ago. They showed that
assuming communication is embedded in social processes, rather than the result of isolated
rational actions, was useful in the study of organizational communication. They characterized
genres based on socially accepted purpose, content, and form (with purpose the leading
characteristic). They also found the notion of genre systems, series of interconnected genres
comprising a social activity (Bazerman, 1994), as an especially useful lens for studying
interaction because it focuses on how people use communicative actions to coordinate their
activity over time and space, as in the peer reviewing of papers for a journal or conference. In
addition, they proposed the notion of a group’s genre repertoire, the set of genres used by that
particular group, and argued that it can reveal a rich and varied array of communicative practices
that characterize the group and its work (Orlikowski and Yates, 1994).
In a more recent paper (Yates and Orlikowski, 2002), they elaborated the framework for
studying genre and genre systems beyond the traditional dimensions of purpose (why), content
(what), and form (how), noting that genres and genre systems are organizing structures that also
provide a community with expectations about participants involved as both initiators and
recipients of communication (who/m), timing (when), and place of communicative interaction
(where). To our knowledge, this elaborated framework has not yet been used systematically in
any empirical studies. Yet the added dimensions illuminate additional expectations that
accompany genres and genre systems. In particular, the participants or who/m dimension may
add to our understanding of genres of organizational communication, for example, by
5
highlighting who has the authority or privilege of initiating particular genres or genre systems
and who can only play a receiving role in such genres. Only a few studies using genre analysis
(Yates, Orlikowski, and Fonstad, 2001; Yoshioka, Yates, and Orlikowski, 2002) have considered
the roles of the different players at all, and then typically only in a limited way.
We adopt this elaborated framework in studying change in genres over time in the online
community. The temporal dimension, when, is captured in the analysis over time. The spatial
dimension is all on a particular mailing list. Thus we focus particularly on the who/m dimension,
or participants in various genres and genres systems, which allows us to highlight a key
development in the communicative practices of employees in different occupations and
organizations of the company—a move from centralized to peer-to-peer communication.
Research site: Online community related to the introduction of a new business
The case we explored was at a Japanese manufacturing company (the Company) that
already had a well entrenched business based on hardware sales. In late 2002 a service
development department (the Service Dept.) in the Company launched a new business (which we
call 'the Network Service’ or ‘the Service’), a secure internet connection service including access
to hardware, software and consulting. The Service Dept. tried to sell this new service through
Sales representatives at the Company’s existing, geographically dispersed sales subsidiaries
(Sales). Because the Network Service was a novel business for the Company, most people in
Sales did not initially have the skills to sell it to their customers. Their customers also did not
expect the Company to offer such a network service. Thus, in addition to conducting general
marketing and advertising, the Service Dept. needed to conduct internal promotion and provide
an educational program to enable Sales representatives at the subsidiaries to sell the Network
6
Service. This program helped Sales to understand the Service better, improve their skills, and
improve processes and practices between Sales and the Service Dept.
Even before the Service Dept. announced the Network Service to their customers, they
established a Community Mailing List (CList) as part of the educational program aimed at Sales.
CList was initiated to cultivate communication between Sales and the Service Dept., and,
secondarily, among Sales representatives. While employees in the Service Dept. used CList to
announce formal organizational information to Sales, such as the release of a new function in the
Service or the establishment of a new organizational procedure around contracts or accounting,
CList was generally a fairly informal communication channel. The list participants discussed and
shared information related to the Network Service without any restrictions. They discussed how
to sell the Service, how to solve technical problems related to it, and how to improve their back
office procedures both before and after selling it. They shared and discussed information about
the emergence of competitors and trends in the networking industry. The Service Department
sometimes used it to discuss problems or ideas about the Network Service itself directly with
Sales.
Data and Methodology
The messages posted to CList over four years constitute our primary data for this paper.
We supplemented the primary message data with interviews, documents, and a questionnaire. In
this section we first describe the CList data and how it was coded, then turn to the supplementary
data.
CList data, coding scheme, and genre analysis
Table 1 shows the profile of the CList, including the time period, the number of
participants, the cumulative number of posters, the number of messages, the number of replied
7
messages (messages that received at least one reply). The number of participants is based on the
latest list of the members of the CList at the time the study was begun.1 The number of posters is
based on actual counts from the message logs. Although the employees in the Service Dept.
made up only 3% of the participants, they posted 23% of messages, reflecting the fact that the
Service Dept. regularly posted official announcements to the otherwise primarily informal CList.
Figure 1 shows activities in the CList during the time period studied.
[insert Table 1 and Figure 1 here]
The messages in CList for this period were then coded according to a scheme developed
to capture aspects of genre. We developed a manual coding scheme for content (what), purpose
(why), and form (how), as described below. Data on participants (who/m) and time (when) were
coded automatically from header information. Members of the Service Dept. were in one
location and members of Sales were geographically dispersed in multiple sales subsidiaries.
Location information other than that captured by department was judged not to be relevant in
this study.
Based on prior research in the field (Orlikowski and Yates, 1994), we established a
preliminary manual coding scheme for content, purpose, and form. Coders could identify one or
two categories for content or purpose, and as many as appropriate for form. We trained three
coders in this scheme and did a small trial sample. Based on this sample, we revised the scheme
slightly and conducted further training. After two more trials, we attained a level of agreement
that allowed us to stabilize the coding scheme. We tested this scheme for inter-rater reliability
using a sample of over 10% of the messages (256 messages). As Table 2 shows, all of the
1 A few individuals who had been members of the CList and even posted on it earlier in the period had moved to other departments and were not
members of the CList at the end of the period.
8
categories had agreements over 0.6, and all but one over 0.7.2 We then manually coded all 2269
messages.
[insert Table 2 here]
To give some sense of the resulting data, the coding results about content (what) and
purpose (why) are shown in Figures 2 and 3. The vertical axis in Figure 2 and Figure 3 shows the
ratio of the number of messages including each type of content and purpose to the total number
of messages in each fiscal year. We used fiscal years starting in April and ending in March, as
that is the time period used in the Company for planning and execution.
[insert Figures 2 and 3 here]
In performing a genre analysis on the coded data, we started by looking at the five
content areas (what) that we had discovered in the messages as we developed the coding
scheme—technical matters, sales, formal information, competitor information, and industry
trends—and identifying genres and genres systems related to each. Following previous literature
in this area (Orlikowski and Yates, 1994; Yates and Orlikowski, 2002), we took purpose (why)
and content (what) as the leading characteristics of a genre, with form (how) and participants
(who/m) as secondary characteristics that might change over time (when) within essentially the
same genre.3 Within each of the five content areas, we explored which purposes were used most
frequently, counting up the number of purpose/content combinations used in each fiscal year and
calculating the ratio of the messages with any particular combination to the total number of
messages posted in each fiscal year. We focused on combinations that accounted for at least 5%
2 We depended on Landis and Koch’s (1977) benchmarks for assessing the relative strength of agreement, which are as follows: Poor (< 0), Slight
(.0 - .20), Fair (.21-.40), Moderate (.41 - .60), Substantial (.61 - 80), and Almost Perfect (.81 - 1.0). By this standard all our categories had
either Substantial or Almost Perfect agreement.
3 None of the hand-coded form categories ended up being important to the story, though one automatically coded form feature (response to
previous message) did reinforce the why coding.
9
of the total messages except for some combinations including competitor and trend information.
While these two content areas had less than 5% of the messages in total, we had learned from the
questionnaire and interviews that they were considered particularly important by list members.
Consequently, for these, we analyzed the combinations with a relatively high percentage
compared with others in the same coding category.
Interviews, questionnaire, and documents related to the Service
The supplementary data came from several sources. We conducted semi-structured
interviews ranging from 1.5 to 3 hours with 10 employees in the Service Dept. and 20 Sales
representatives in order to get an overview of the Network Service business and the usage of the
CList and other communication channels. We collected relevant internal documents in the
Service Dept. and Sales in order to understand past events, strategy, and sales outcomes. We also
conducted a web-based questionnaire, described in more detail elsewhere [reference suppressed],
to investigate how members of the Service Dept. and Sales used seven communication channels
(three mailing lists, the largest and most inclusive of which was CList, two call centers,
conferences, and newsletters intended to be posted on a wall) in launching and selling the
Network Service. The targets of the questionnaire were the 1451 Sales representatives4 enrolled
in the CList, and we achieved a response rate of 36.2% (525 respondents). The self-reported
usage of the seven communication channels between the Service Dept. and Sales for each topic
in each year is primarily used elsewhere, but we refer to this data occasionally in order to
motivate our approach to, or interpret the results of, the genre analysis.
4 The questionnaire was sent to all sales representatives, but not to sales staff (e.g., secretarial assistants) who were included on the CList but who
never posted to it or really used it themselves. Eliminating staff reduced the 1564 members of the CList to 1451 members receiving the
questionnaire.
10
Results
Our genre analysis helped us to identify genres or genre systems for each content
category: in technical matters, technical queries and responses; in sales, sales queries, responses,
and announcements; in formal information, official announcements; in competitor information,
competitor queries and responses; and in industry trends, trend announcements. In the following
sections, we present each genre or genre system, then analyze its change over time.
Technical queries and responses
Sales representatives frequently posted technical queries (purpose: query; content:
technical) about configurations, compatibility, and capabilities of the Network Service to the
CList. For example, one Sales representative posted the following:
I’d like to know the compatibility of [a specific communication protocol] with the Service.
This is one of the conditions that our customer asked me about. But, I don’t know what it
is at all. (12/10/2002)5
Such a query could be answered in a technical response (purpose: respond; content: technical) by
someone in the Service Dept., or by another Sales representative (who/m). In this case, the
query was answered by a top technical manager in the Service Dept., who said, in part, the
following:
In conclusion, there is no limitation of using this protocol with the Service. For your
information, I will write down how the Service handles it in terms of internal technical
mechanism. (12/10/2002)
Together, the query and response form a genre system of technical queries and responses. This
particular instance of the genre system can be classified along the participants (who/m)
dimension as SR-SD (an interaction between a Sales representative and the Service Dept.). In
5 All messages cited are translated from the original Japanese, and proprietary details are disguised.
11
such interactions about technical information, the Service Dept. typically gave a theoretical
answer to the query, based on knowledge of the internal technical specifications of the Network
Service that they developed, even though members of the Service Dept. had never tested the
protocol in the customer environment.
This technical query/response genre system was also enacted solely among Sales
representatives (classified on the participants dimension as SR-SR), and with an important
qualitative difference. For example, a Sales representative made the following query, clearly
aimed at his peers in Sales:
One of our current customers using the Service is interested in making a contract for
using an optional function of the Service, which can be used with a mobile phone. Does
anyone have the same case where your customer used this function with the following
mobile phone service provided by company Z? (8/23/05)
He received a very helpful response from another Sales representative, who said:
My co-worker actually tested this mobile phone with our service when the mobile phone
service was started. We found that it is completely compatible without any trouble
(8/23/05).
In this variant of the genre system, in contrast to the SR-SD version of the genre system, other
Sales representatives’ answers typically gave real examples, in which their customer actually
used the Service in the same or a similar environment. Both types of interactions were, of course,
useful to Sales representatives who wanted to propose the Network Service to their customers,
but they are qualitatively different and could be used by the Sales representative in somewhat
different ways. Thus identifying the two genre variants along the participants (who/m)
dimension adds to our understanding of the interactions.
The SR-SR variant of the genre systems also turned out to have some unanticipated
positive benefits for the Service Dept. and ultimately for customers because the exchanges
12
sometimes helped identify a technical problem within the Service. For example, in several
instances when a Sales representative responding to a technical query indicated a problem under
particular settings or in particular customer environments, the Service Dept. initially assumed
that the problem was specific to the customer environment and unrelated to any problem with the
Network Service itself. However, after a few other Sales representatives also responded to the
query, pointing out the same problem under the same or similar conditions, the Service Dept.
started to investigate and eventually identified and solved the underlying technical problem.
Even though the Service Dept. provided a specific formal channel for gathering technical
problems from Sales (one of the Help Lines), that channel did not allow for sharing the problem
among different Sales representatives. The sharing of technical information among Sales
participants in CList in some cases encouraged the emergence of new information useful to the
Service Dept., as well as to those involved in the exchange. Indeed, the top technical manager
often decided to go to a customer site based on the customer's needs as revealed in the CList. A
similar positive consequence occurred around customers’ needs for additional functionality in
the Service. When Sales representatives used technical queries and responses to discuss the
needs of their customers among themselves, Service Dept. members of CList could consider
implementing functions responding to those needs.
Looking at technical queries and responses over time (when) and taking into account the
participant (who/m) variants (SD-SR and SR-SR) provides additional information. As we can
see in Figure 2, the number of messages including technical information decreased from F2002
to F2004, then increased somewhat (though not to their original level) in F2005, when a new set
of technical features were introduced. Figure 4 shows that the two genres of technical queries
13
and technical responses followed a similar pattern (except for a slight increase in responses in
F2003) if we ignore participants.
[insert Figures 4 and 5 here]
When we take participants (who/m) into account, however, an interesting new pattern emerges.
Focusing on responses to technical queries initiated by Sales representatives only, Figure 5
shows that the technical responses could come either from the Sales Dept. (SD) or from other
sales representatives (SR). Figure 5 breaks down the technical responses to queries initiated by
Sales representatives by their sources (SD or SR), showing an interesting shift over time. In the
first two years, responses to Sales representatives’ queries about technical information were
more often answered by the Service Dept. than by other sales representatives, but in F2004/2005,
Sales representatives were more likely to respond than members of the Service Dept.
Over time, then, the genre system of technical queries and responses shifted from being
used predominantly for interactions between Sales and the Service Dept. (SR-SD) to being used
predominantly for interactions among Sales representatives (SR-SR). Sales didn’t understand
the nature and purpose of the Service early on (especially in F2002) since it was technically
different from the established products and services of the company. It is not surprising that
employees in the Service Dept were most able to respond to queries then. By F2004, however,
responses to sales representatives’ technical queries more often came from other Sales
representatives than from the Service Dept. That is, the genre system of technical queries and
responses shifted from a centralized pattern in which the Service Dept. responded to queries
from Sales representatives to a peer-to-peer pattern in which dispersed Sales representatives
responded to technical queries from their Sales peers.
14
Based on reading the contents of the messages, we observed that Sales representatives
increasingly shared results of experiments about the compatibility of other companies’ services
with their own Network Service. This type of information could not easily be provided by the
Service Dept. because there were so many combinations of the compatibilities that they could
not test all of the possible combinations in advance. Sales, however, had thousands of
implementations from which to extract favorable or unfavorable combinations in a ‘real world’
environment. For example, the Company’s Network Service could be linked to compatible
internet services provided by hotel chains through a specific function of the Service. Doing so
required knowing the compatibility of internet services in each hotel chain. A motivated sales
force started to share this availability in the CList, creating one example of the peer-to-peer
genre system of technical queries and responses.
Sales queries and responses
A similar genre system of queries and responses occurred around sales information
(what). Queries about sales were generated by sales representatives but could be responded to
by either the Service Dept. or by another Sales representative. For example, the following query
was generated by a Sales representative:
Our customer in a bank is interested in the Service. Does anyone have an experience of
proposing the Service to a bank? (7/22/03)
Although the query seem aimed more to other Sales representatives than to the Service Dept., a
member of the Service Dept. responded to it, as follows:
That’s a very interesting case for the Service. I’d like to directly help to give a proposal to
your customer. Could you let me know the details? (8/26/03)
This instance of the genre system can be classified along the participants dimension as an SR-SD
interaction—that is, an interaction between the centralized Service Dept. and the dispersed Sales
15
representatives. In such interactions, the Service Dept. was eager to help, but had no direct
experience in selling the Service to this customer segment.
In other cases the response came not from the Service Dept. but from another Sales
representative, for an SR-SR variant of the genre system. For example, a Sales representative
posed the following sales query to CList:
Does anyone have a case of proposing the Service to medical institution? They are
interested in multiple connections using the Service in order to share medical records
among the distributed offices. (11/30/05)
In this instance, another Sales representative responded, beginning with the following statement:
I've provided the Service to a medical institution. They are very satisfied with the system
we developed. (11/30/05)
Here, the Sales representative is able to respond out of direct experience, not simply out of
interest in the customer segment. In this variant of the genre system, other Sales representatives’
answers often provided directly applicable experience of selling to similar customers.
Again, both SR-SR and SR-SD interactions were useful to Sales representatives
attempting to sell the Network Service to their customers. But responses from the two different
sources differed in nature and could be used by the Sales representatives in different ways. The
experience-sharing that occurred in the SR-SR variant offered particular advantages to Sales
representatives, allowing them to learn directly relevant information from each other. The SR-
SD variant, on the other hand, was more likely to provide general guidance than specific
guidance based on real experience.
Turning to a longitudinal analysis of the genre system of sales queries and responses over
time and classifying variants on the who/m dimension (SD-SR and SR-SR) again provides
additional information. We have seen in Figure 2 that messages with sales-related content
16
increased from F2002 to F2004, and evened off in F2005. Figure 6 shows that Sales queries and
responses followed a similar pattern of growth from F2002 to F2004, followed by relative
stability in F2005. But if we focus on Sales representatives’ queries and break down the
responses by type of participant (see Figure 7), we see an interesting new pattern. Service Dept.
responders outnumbered Sales responders in F2002. In F2003, both types of responses increased
in F2002, but responses by other Sales representatives slightly outnumbered those from the
Service Dept. In F2004 and F2005, responses from the Service Dept. rose slightly then dropped,
while responses from other Sales representatives kept increasing, considerably surpassing
responses from the Service Dept. in both years. In other words, queries from Sales were
increasingly responded to by others in Sales—the same change from centralized to peer-to-peer
communication we saw in technical queries and responses.
[insert figures 6 and 7 here]
There was little experience with sales information around the new Service in F2002.
Interviews and documentation reveal (and the questionnaire confirmed) that employees in the
Service Dept organized a face-to-face conference to promote the Network Service to sales
representatives and sometimes even accompanied them to a customer office to educate them in
sales features. Beginning in the second year, Sales came to believe in the potential of the Service
and started to gradually increase sales while learning the relevant skills or knowledge related to
the Service itself. While both questions about sales information posted by Sales and the number
of licenses for the Service sold rapidly increased in F2003, the responses to the questions were
still mainly posted by the Service Dept, as Figure 7 shows. Based on interviews, we know that
the know-how about how to sell the Service was still being transferred from the Service Dept to
Sales during this year, with the Service Dept going to the sales subsidiaries to educate the sales
17
force in sales and marketing techniques and knowledge. In F2004-F2005, however, Sales
representatives were responding to their peers elsewhere in Sales more often than the Service
Dept. was.
During that period, the Service Dept increased the number of internal marketing
conferences for Sales from twice to four times per year and also launched a formal help center to
sell the Service, but they decreased the number of visits to sales subsidiaries. These kinds of
changes in the formal organizational structure were intended to increase the skills related to
selling the Service as well as the number of Sales representatives selling it. Nevertheless, these
formal channels did not capture much specific knowledge being accumulated by Sales
representatives. When a Sales representative called the help center about whether or not a
similar sales case was in its database, the staff of the center often suggested that the
representative check in CList messages for F2004-2005.
In addition, sales representatives used CList to discuss how they could sell the Service to
customers that they had never approached. For example, a motivated sale representative posted a
plan to conduct an experiment confirming the compatibility of the Service with a mission critical
server provided by another company. He asked the participants of the CList about additional
experiments they were interested in. Another Sales representative asked about details of this
experiment in order to refer to it when he proposed the same combination to his customer. By
conducting this experiment and sharing its results, the Sales representative thus created a new
target for selling the Service, a target that became a popular sales pattern especially for middle-
sized customers who had such a mission critical server.
This kind of information could not be provided by the Service Dept. because it lacked a
formal communication channel to gather information about their customer’s environments, even
18
though they had one to gather past sales records as best practices. Thus Sales representative
often were better equipped to create ideas of how to sell the Service under a specific
environment than was the Service Dept.
Formal announcements
The Service Dept, as the organization in charge of the Network Service, provided formal
information such as news releases announcing a new function, an organizational change, an
implementation of the formal system that could support back office procedures between the
Service Dept. and Sales, and events or sales promotions. Such announcements sometimes
contained related requests, as well. A genre rather than a genre system, formal announcements
by definition were one-way communications from the Service Dept. to Sales. That is, only the
Service Dept. was empowered to make official announcements. For example, one such
announcement was as follows:
As we’ve already announced the new function on this list, we have a plan to update the
information about it on the intranet as follows:[…]. (3/7/2003)
Another announcement of a planned service stoppage offered the opportunity to respond if they
foresaw problems:
We have a plan to stop the Service just for 10 minutes in order to update it. If you have
any requirement and concern about the date/time for stopping the Service for your
customer for some reason, could you let us know? (12/0/04)
Typically, however, announcements received no responses over CList.
As Figure 8 shows, these formal announcements peaked in F2004, a peak probably
related to the release of many new, optional services. Because this genre is composed of one-way
19
announcements from the centralized Service Dept., analysis by participants does not reveal any
additional information or changes over time.
[insert Figure 8 here]
Competitor queries, responses, and announcements
Even though the total number of messages including competitor information is small (83
messages), one of the important roles of the CList, based on the results of the questionnaire
described earlier, was to exchange competitor information.6 In particular, the genre system of
competitor queries and responses, though small, was important to CList users. Competitor
queries from Sales representatives could ask advice on how to deal with particular competitors.
For example, one sales representative asked:
Our customer asked me about the difference between our Network Service and a
service provided by another company. If you were me, how would you explain in order to
clarify the difference? (6/14/04)
Another Sales representative responded, in part, as follows:
I will provide you a fundamental comparison between them. I believe that it [the
comparison] will help you to persuade your customer to choose our service. […]
(6/15/04)
Responses could also come from the Service Dept. In another example, a Sales
representative asked:
I heard from our customer that another company proposed the same kind of service to
our customer. If you have any good ideas of how to compete with them [the competitor]
in order to get the deal, could you let me know? (7/28/04)
In this case, the Service Dept. responded as follows:
6 44% of all questionnaire respondents had used the CList to get competitor information in each year.
20
Thank you for your information about a competitor. For your information, I will let you
know the information that we’ve gathered from various sources about the difference [of
this competitor service] from our service in terms of cost and function. I think that it’s
very important for all of us to share how to compete with this competitor from various
viewpoints. We would be happy to promote these conversations more and more in the
CList. (11/4/05)
In this response the Service Dept. both promises to pass along competitor information it has
obtained and encourages the Sales representatives to continue posting to CList, sharing their
knowledge about competing with specific rival services. This was a topic about which Sales
representatives had information the Service Dept. did not necessarily have, so the Service Dept.
encouraged them to share that information with each other on CList, thus also informing the
Service Dept.
[insert figures 9 and 10]
As Figure 9 shows, the genres of queries and responses about competitor information
(which combine to make a genre system) both spiked in 2004, about the time that one of the
critical competitors from their incumbent business emerged into the Service market. Figure 10,
focusing only on Sales representatives’ queries on competitor information and responses to those
queries, shows that the Service Dept. played very little role in responding (though the difference
between Figures 9 and 10 shows that Service Dept. members occasionally initiated their own
queries about competitors, which they or Sales representatives responded to) . From the start,
peers were most able to respond to Sales representatives’ questions in this area. The peer-to-
peer pattern peeked in F2004. Competitive knowledge thus seemed best shared among Sales
peers.
Trend announcements and discussion
A final genre that was small in total numbers but identified by questionnaire respondents
as important to CList members was the trend announcement. Messages instantiating this genre
21
announced some change in the market or in government regulation and speculated about its
probable affect on future sales of the Network Service. Such messages included, for example,
the following announcement from the Service Dept. talking about recent computer viruses and
how the Network Service could help combat such viruses:
I will inform you about the essential problems caused by computer virus incidents all
over the world. …Our Service can provide the most robust solution for this kind of attack.
(1/26/03)
Such announcements could also come from Sales representatives, as in the following:
I will share updated information about the IP phone provided by the most famous
internet service provider, which I’ve already posted in a different mailing list in our sales
subsidiary company in order to discuss the possibility. Based on this information, we will
be able to sell the Service with the IP phone more intensively, which might be a new
pattern for selling the Service. (9/26/2003)
Messages on trends sometimes initiated multiple follow-on messages, from both Sales
representatives and from the Service Department, discussing the potential impact on selling the
Service (though too few to analyze as a separate genre).
[insert Figure 11 here]
As Figure 11 shows, the genre of announcements about trends more frequently came
from the Service Dept (SD) than from Sales representatives (SR), throughout the period. For
example, a knowledgeable product manager in the Service Dept periodically made the latest
trend news understandable even for non-technical Salespeople and posted it to the CList. This
genre sometimes included ’breaking news’ that Sales couldn't get through mass communication
media such as a general industrial newspaper. In F2003 and F2005, one knowledgeable Sales
representative also provided his thoughts about trends affecting sales of the Service, which can
be seen in the peaks in SR trend announcements in those two years (see Figure 11).
Even though trend announcements posted in CList are few compared with other genres,
22
the questionnaire indicated that this type of information was obtained primarily through CList,
rather than from the marketing conferences or other more formal channels. Because this
information sometimes included individual interpretation, the Service Dept. may have
considered it not entirely official, and therefore more appropriately shared through the informal
CList than through other, more formal channels.
Implications for research and practice
This analysis of communication on CList reveals the importance of participants and
participation patterns in genre analysis more broadly. It also highlights the important role that
peer-to-peer communication such as what developed over CList can play in certain types of
business situations.
Participants (who/m) as an important aspect of genre analysis
During the four-year period studied, the company successfully launched a new product
that differed fundamentally from its normal products. The informal online community clearly
played an important role in allowing the geographically and organizationally dispersed Sales
force to learn about the new Network Service and how to sell it, both from the Service Dept. and
from each other. Our analysis, which used the broader definition of genre presented in Yates and
Orlikowski (2002), highlighted particularly the importance of analyzing participants, or who/m,
as well as the traditional purpose (why), content (what), and form (how) aspects of genre. While
purpose and content are still primary indicators of a genre, information regarding who
communicates to whom in a particular genre or genre system over time provides important
information, as well.
[insert Table 3 here]
23
The value of this aspect of genre is especially salient in Response to Query for both sales
and technical query-and-response genre systems over time. Table 3 summarizes the six aspects
of genre (why, what, how, when, where, and who/m) for the responses to both technical and
sales queries for two time periods—F2002/2003, and F2004/2005. Focusing on the first five
aspects, we see no fundamental change over time. The one category with a significant change is
‘who/m,’ which highlights the shift from a centralized pattern in which the Service Dept.
responded to Sales representatives’ queries in F2002/2003, to a peer-to-peer pattern in which
Sales representatives responded to each other in F2004/2005. This change in who answers a
query, whether about sales, technical, or competitor matters, is important to capture, as the
dynamic is much different for the participants in the genre, as is the qualitative nature of the
messages. In F2002/, the participation pattern was ‘hub and spoke’ (or perhaps we should say
‘spoke and hub’), with a central Service Dept answering questions from a geographically
dispersed Sales force. When the sales force started answering their peers’ questions they changed
both the breadth of responders (now drawing from a larger group) and the number of responses
based on real sales encounters and real technical implementations, rather than theoretical
solutions. Moreover, the interviews suggest that the level of ‘trust’ among peers might be higher.
CList moved from being a predominantly one-way channel to being a community of Sales
representatives as well as of Service Dept. members,.
A similar shift in participants occurred in the Competitor queries and responses, though it
happened even earlier, in F2003. Competitor information could be pieced together from the
Sales representatives’ shared observations of competitor behavior before it could be researched
and announced by the Service Dept.—indeed, interchanges among Sales representatives in this
24
genre system undoubtedly provided new and useful information to the Service Dept., as well, as
it did in the exchanges about technical compatibilities with other systems.
The who/m aspect is less salient in the other genres identified, but still potentially useful.
The formal announcements on the CList, by definition, were directed from the centralized
Service Dept. to the Sales representatives. Nevertheless, including participants in the genre
analysis highlights the fact that only members of the Service Dept. are empowered to issue
formal announcements. The participant information on trend announcements shows that both the
Service Dept. and Sales representatives felt empowered to make informal announcements about
trends. Even though the Service Dept. issued more trend announcements in all four years, Sales
representatives consistently contributed them, as well.
Thus by including participants as an aspect of genre, researchers can highlight who
participates in what role in genres and genre systems. This information says something about the
power structure of the community using the genres. This dimension combined with analysis
over time can highlight changes in genre participants that are linked to qualitative changes in
instances of the genre (e.g., the use of real examples rather than drawing on abstract technical
specifications) that might otherwise be overlooked. The linking of changes in participation with
subtle shifts in content (or purpose, form, location, or timing) made possible by genre analysis
adds value beyond what network analysis alone can provide. Changes in participation are often
particularly significant when a genre migrates from paper to electronic form, since additional
electronic copies are essentially costless. Thus adding the dimension of participants to genre
analysis is especially useful when using this framework to study shifts to electronic
communication.
25
Finally, when analyzing a group’s genre repertoire over time in order to understand the
group’s changes (e.g., Orlikowski and Yates, 1994), researchers may benefit from paying
attention to changes in genre participants. Such changes should signal shifts in power and
participation over time, as well as direct closer attention to subtle shifts in content and purpose
over time.
Peer-to-peer is more useful than centralized communication in some business situations
From the point of view of practitioners, rather than researchers, our study highlights the
shifting role of participants in this online community over time, and the value of that shift to the
Company. In F2004, whether intentionally or not, the CList shifted from being a place for
communication primarily between the Service Dept and Sales to being a place for
communication primarily among Sales representatives themselves.7 The utility of this list for
peer-to-peer communication emerged for a number of reasons. Over the period studied, other,
more formal communication channels, such as call centers and marketing conferences, were
established to take on some of the centralized communication needs. However, even if a query
was answered through a formal channel, we found that some Sales representatives would also
confirm the answer on the CList. This desire for confirmation suggests a higher degree of trust
for peers in similar positions. Also, as shown when Sales representatives shared examples of
implementations, peer-to-peer communication offered a broader set of participants with ‘real
world’ examples that were better shared in a distributed, peer-to-peer fashion. Most aspects of
7 When we conducted interviews with employees in the Service Dept in 2006, they typically regarded the CList as a place for Sales
representatives to communicate with each other.
26
the genres were the same, but the changes in participants apparently increased the perceived
utility of the channel.
These findings suggest that firms should not try to formalize and regularize all
communication channels as soon as possible. In cases such as this, where peers have
information that the centralized department does not have, peer-to-peer communication plays an
important role and should be encouraged. We can even identify a positive correlation between
the number of sales queries and responses from sales representatives and the number of licenses
sold, as Figure 12 shows. While we can’t prove causality, there is an intuitive correspondence
between getting satisfactory responses to sales questions among Sales themselves and more
effective selling.
[insert Figure 12 here]
Finally CList, increasingly dominated by peer-to-peer exchanges, was treated as a much less
formal channel than other channels (e.g., the Help line) added later. Genres focusing on certain
types of content were apparently felt by many list members (both Sales representatives and
Service Dept. personnel) to be best shared over a less formal channel. The questionnaire data
confirmed that CList members considered CList the channel from which they acquired the best
competitor and trend information, in spite of the existence of other, more formal (and typically
more centralized) channels. This finding supports the notion that peer-to-peer communication
can be more useful than centralized communication in some business situations. Although
centralized communication may have been necessary and valuable in the early days of launching
a new business, for example, over time, as sales representatives acquired more experience with
selling the new service, we might expect more networked, peer-to-peer communication to
become increasingly important. For another example, a company may not want to endorse
27
opinions about trends or competitors in a formal channel but may be willing to have employees
share this information informally. Thus, this research suggests to practitioners that they may
want to resist the impulse to formalize and centralize all communication, recognizing the value
of peer-to-peer communication in many situations.
Conclusion
In this paper we analyzed the communication within a mailing list that was launched as
an internal marketing tool and evolved into an online community fostering peer-to-peer
communication. We focused on the change in communication over time between Sales
representatives and the employees in a Service Dept. We conducted a genre analysis and found
that although the frequently studied aspects of genre and genre systems (purpose, content, and
form) capture much information, adding the participants dimension provides additional insights.
Changes in participants over time affect how genres are used and perceived. Thus researchers
using genre analysis should benefit from including the additional dimensions outlined by Yates
and Orlikowski (2002), and particularly the participant (who/m) dimension.
For practitioners, understanding participants and how they are using genres in online
communities can help identify shifts in communication patterns. Focusing on changes in
participants may also highlight emergent changes that can be leveraged into improved processes
or results. The change in participants (who/m) in genres can highlight a parallel change in the
formality of communication. In future work, we plan to focus on this shift in formality and its
interaction with research on organizational practices and processes.
28
Acknowledgement
We appreciate the excellent participation of employees at the case study company in the
questionnaires and interviews.
References
Baym, N. K. (2000). Tune in, log on: Soaps, fandom, and online community. Thousand Oaks:
CA: Sage Publications.
Bazerman, C. Systems of Genres and the Enactment of Social Intentions, Genre and the New
Rhetoric, A. Freedman A. and Medway, P. Eds., London, Taylor & Francis Ltd., 1994, 79-
101.
Bobrow, D.G. and Whalen, J. Community Knowledge Sharing in Practice: The Eureka Story,
REFLECTIONS, 4, 2, 2002.
Cross, R., and Cummings J. Tie and Network Correlates of Individual Performance in
Knowledge-Intensive Work, Academy of Management Journal, 47, 2, 2004, 928–937.
Crowston, K., and Williams, M. Reproduced and Emergent Genres of Communication on the
World-Wide Web, In IEEE Proceedings of the 30th HICSS. (Jan. 3-6, 1997, Maui Island,
Hawaii).
Füller, J., Bartl, M., Ernst, H., and Mühlbacher, H. Community Based Innovation A Method to
Utilize the Innovative Potential of Online Communities, In IEEE Proceedings of the 37th
HICSS. (Jan. 5-8, 2004 , Big Island, Hawaii).
Ihlström, C., and Henfridsson, O. Online Newspapers in Scandinavia: A Longitudinal Study of
Genre Change and Interdependency, IT & People, 18, 2, 2005, 172-192.
29
Karjalainen, A., Päivärinta, T., Tyrväinen, P., and Rajala, J. Genre Based Metadata for
Enterprise Document Management, In IEEE Proceedings of the 34th HICSS. (Jan. 3-6, 2001 ,
Big Island, Hawaii).
Landis, J.R. and Koch, G.G. “The Measurement of Observer Agreement for Categorical Data,”
Biometrics, 33, 1977: 159-174.
Orlikowski, W.J., and Yates, J. Genre Repertoire: The Structuring of Communicative Practices
in Organizations, Administrative Science Quarterly, 39, 4, Dec., 1994, 541-574.
Orlikowski, W.J., Yates, J., Okamura, K., and Fujimoto, M. Shaping Electronic Communication:
The Metastructuring of Technology in the Context of Use, Organization Science, 6, 1995,
423-444.
Orlikowski, W.J., and Hofman, J.D. An Improvisational Model for Change Management- The
Case of Groupware Technologies, Sloan Management Review, 38, 2, Winter 1997, 11–21.
Preece, J. Sociability and usability: Twenty years of chatting online, Behavior and Information
Technology Journal, 20, 5, 2001, 347-356.
Quan-Haase, A., and Wellman, B. Hyperconnected Net Work: Computer Mediated Community
in a High-Tech Organization., Collaborative Community in Business and Society. Heckscher,
C. and Adler, P. Eds., New York, Oxford University Press, 2005.
von Hippel, E., Innovation by User Communities: Learning from Open-Source Software, MIT
Sloan Management Review, 42, 4, Summer 2001, 82–86.
Westerman, G., McFarlan, F.W. and Iansiti, M., Organization Design and Effectiveness the
Innovation Life Cycle, Organization Science, 17, 2, 2006, 230–238.
30
Yates, J., and Orlikowski, W.J. Genres of Organizational Communication: A Structurational
Approach to Studying Communication and Media, Academy of Management Review, 17, 2,
1992, 299-326.
Yates, J., Orlikowski, W.J., and Fonstad, N., Sloan 2001: a virtual odyssey, Our Virtual World:
The Transformation of Work, Play, and Life via Technology, Zigurs, I., and Chidambaram, L.
Eds, Idea Group Publishing, Hershey, PA, 2001.
Yates, J., and Orlikowski, W.J., Genre systems: structuring interactions through communicative
norms, Journal of Business Communication, 39, 1, 2002, 13-35.
Yoshioka, T., Yates, J., and Orlikowski, W. Community-based interpretive schemes: exploring
the use of cyber meetings within a global organization, In IEEE Proceedings of the 35th
HICSS. (Jan. 7-10, 2002 , Big Island, Hawaii).
31
Tables *
Table 1. Basic profile of CList
Time period Jul. 2002 – Mar. 2006
# of participants 1612 (SR: 1564, SD: 48)
Cumulative # of posters 303 (SR: 232:, SD: 71)
Total # of messages 2269 (SR: 1120, SD: 1149)
Total # of reply message 825 (SR: 480, SD: 345)
SR: Sales representatives
SD: Service Dept. employees
Table 2. The coding scheme and Cohen’s Kappa
Coding
category Types in category Definition of types Reliability
(Cohen’s k) N %
c1. TECH technical matter / trouble / setting of
Service 0.83 649 28.6%
c2. SALES how to sell Service/ sales material /
past and similar sales case 0.76 475 20.9%
c3. FORMAL formal information of Service and
Service Dept 0.75 949 41.8%
c4. COMP competitor information 0.91 83 3.7%
What (content)
c5. TREND trend or general information in a
related industry 0.84 113 5.0%
p1. QUERY ask a question 0.81 408 14.1%
p2. RESPOND answer to the question 0.73 596 20.6%
p3. PROPOSE/REQUEST ask someone to do
something(together) 0.67 405 14.0%
p4. INFORM/ANNOUNCE let the participants know something 0.73 1060 36.6%
p5. THANK appreciate something 0.78 214 7.4%
p6. APOLOGIZE apologize for something 0.87 94 3.2%
Why (purpose)
p7. COMMIT commit to do something 0.73 122 4.2%
f1. EMBEDDED with an embedded message 0.82 673 29.7%
f2. GRAPHICAL contain a graphical element 1.00 42 1.9%
f3. INFORMALITY use an informal and colloquial
expression 0.76 431 19.0%
f4. NICKNAME use a nickname 0.74 27 1.2%
f5. OPEN SALUTATION started from open salutation 0.71 1382 60.9%
f6. LINK include the link to other information 0.85 715 31.5%
f7. LISTING/ITEMIZING use listing/itemizing 0.76 858 37.8%
How(form)
f8. CITED cited information on other source 0.84 236 10.4%
32
Table 3. Change in predominant who/m pattern in Response to Technical and Sales
Queries over time
Year F2002/2003 F2004/2005
Why Respond Respond
What Technical and Sales information Technical and Sales information
How Email mailing list Email mailing list
Where Geographically dispersed Geographically dispersed
When As queries are posted As queries are posted
Who/m Service Dept to Sales Sales to Sales
33
Figures *
0
20
40
60
80
100
120
140
Jun- 02
Jul-02
Aug- 02
Sep- 02
Oct - 02
Nov - 02
Dec - 0 2
Jan- 03
Feb- 03
Mar- 03
Apr- 03
May- 03
Jun- 03
Jul-03
Aug- 03
Sep- 03
Oct - 0 3
Nov - 03
Dec - 0 3
Jan- 04
Feb- 04
Mar- 04
Apr- 04
May- 04
Jun- 04
Jul-04
Aug- 04
Sep- 04
Oct - 0 4
Nov - 04
Dec - 0 4
Jan- 05
Feb- 05
Mar- 05
Apr- 05
May- 05
Jun- 05
Jul-05
Aug- 05
Sep- 05
Oct - 0 5
Nov - 05
Dec - 0 5
Jan- 06
Feb- 06
Mar- 06
# of messages, posters, and replies
month
# of messages(CList) # of poster s in each mon(CList ) # of replies(CList )
Figure 1. Activities in the CList (Monthly)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
F2002 F2003 F2004 F2005
# of messages including each information / # of total
mes sages post e d in e ach f isc al y ear
Fiscal ye ar
c1.TECH c2.SALES c3.FORMAL c4.COMP c5.TREN
D
Figure 2. Content (what) in each fiscal year
34
0.00
0.10
0.20
0.30
0.40
0.50
0.60
F2002 F2003 F2004 F2005
# of messages including each purpose / # of total
messages posted in each fisc al year
Fiscal ye ar
p1.Q UERY p2. RESPO ND p3.PROPOSE/ REQUEST p4.ANNO UNCE/ INFORM
p5.T HAN
K
p6.APOLOGIZE p7.CO MMIT
Figure 3. Purpose (why) in each fiscal year
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
14.00%
16.00%
18.00%
F2002 F2003 F2004 F2005
TECH‐QUER Y(TOTAL ) TECH+RES PO ND(TOTA L)
Figure 4. Queries and responses about technical information
35
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
8.00%
9.00%
10.00%
F2002 F2003 F2004 F2005
TECH+RES POND (SD‐SR) TEC H+RESP OND( SR‐SR)
Figure 5. Responses to Sales representatives’ queries about technical information
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
F2002 F2003 F2004 F2005
SALES+QUERY(TOTAL) SALES+RESPOND(TOTAL)
Figure 6. Queries and responses about sales information
36
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
F2002 F2003 F2004 F2005
SALES+RESPOND(SD‐SR) SALES+RESPOND(SR‐SR)
Figure 7. Responses to Sales representatives’ queries about sales information
0.00%
5.00%
10.00%
15.00%
20.00%
25.00%
30.00%
35.00%
40.00%
F2002 F2003 F2004 F2005
FORMA L +ANNO UNC E(S D)
Figure 8. Announcements of formal information
37
0.00%
0.50%
1.00%
1.50%
2.00%
2.50%
3.00%
F2002 F2003 F2004 F2005
COMP+QUERY(TOTAL) COMP+RESPOND(TOTAL)
Figure 9. Queries and responses about competitor information
0.00%
0.50%
1.00%
1.50%
2.00%
F2002 F2003 F2004 F2005
COMP+RESPOND(SD‐SR) COMP+RESPOND(SR‐SR)
Figure 10. Responses to Sales representatives’ queries about competitor information
38
0.00%
0.50%
1.00%
1.50%
2.00%
2.50%
3.00%
F200
2
F2003 F2004 F2005
TREND+ANNOUNCE(SD) TREND+ANNOUNC E(S R)
Figure 11. Announcements of trend information
y = - 0.0114 + 0.0701x
(
r = .94143,
p
=0.059*
)
y = 0.011 + 0.0466 x
(
r =.99461,
p
=0.0054* * *
)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0 0.2 0.4 0.6 0.8 1 1.2
Tot al # of l ic enses in each
y
ear
(
normalized
)
Ratio of # of messages includi ng quer y or response
s
about sales information to tot al # of messages in
each year
SR: SALES+ QUERY
SR
-
SR: SALES
+
RESPOND
Figure 12. Total number of Service licenses related to Sales queries and responses
(r = Pearson's product-moment correlation, p = p-values)
39
