View
229
Download
3
Category
Preview:
Citation preview
Research Methodology
97 Ph.D. Thesis
CHAPTER 3
RESEARCH METHODOLOGY
3.1. Introduction
In this chapter the theoretical framework and methodology adopted in the study has been
discussed. It outlines the various dimensions of the study and research objectives and the set of
methodologies adapted to accomplish those objectives. It explains in detail the pilot study
conducted for the identification of an appropriate online tool for the study after a comparative
analysis of three online tools. Netnography, which is a new qualitative, interpretive research
methodology, that uses internet optimized ethnographic research techniques to study the online
communities, has been applied, for the formulation of the research instrument. Further the
procedures followed for the collection of data and selection of the sample of online community
consumers and online community managers have been outlined. The tools and techniques
followed for analyzing the data for the study are also dealt in this section.
Using the concept of response modelling, four specific models have been developed during the
entire research study. These are-
(v) Consumer Trustworthiness Regression model using Netnography (CTR)
(vi) Co-creation model using INV based on Metcalf Law (C-INV)
(vii) Consumer Price Sensitivity model using K-means cluster analysis (CPS)
(viii) Business Online Community Credibility model (BOCC) using Linear
Programming
This chapter describes in detail the sample size, sampling techniques, tools of data collection and
tools of data analysis for the complete study.
3.1.1. Research Design
The research design is a framework for conducting the research. It involves the various steps
ranging from the definition of the information needed, specifying the measurement and scaling
procedures, constructing and pretesting the questionnaire, specifying the sampling process and
Chapter 3
98 Ph.D. Thesis
size to developing a plan for data analysis. The research design for my study is primarily
exploratory and descriptive in nature. It is exploratory because at the first stage it involved the
provision of insights into the research topic and comprehension of the problem situation. It led
me to formulate the research problem, develop the objectives of the study, isolate the key
parameters of the study and plan the future course of action. The descriptive research is a type of
conclusive research. It attempts to describe systematically a situation, problem, phenomenon,
service or programme; it also describes the characteristics of the respondents and the degree of
association or relationship between the variables being studied. It helps to make specific
predictions. These two research designs were apt for the present study.
3.2. Conducting pilot studies for selection of an online tool and a comparative analysis of
three tools for CRM and CEM.
3.2.1. Pilot Study
A pilot study was carried out as part of my exploratory research. This was conducted with a
focus group of 30 participants well versed with traversing the internet. This focus group
comprised a set of practitioners from the industry, who were already using these online web
spaces for consumer engagement and participation. They were asked to map the tools of the
collaborative web, namely, Blogs, Wikis and Online communities with regard to their ability to
create and deliver organisation, brand and product related value to the customer.
A focus group is an interview conducted by a trained moderator in a nonstructured and natural
manner with a small group of respondents. The group is homogeneous in terms of demographic
and socio economic characteristics and the respondents are pre-screened. The participants must
have had adequate experience with the object or issue being discussed. The moderator leads the
discussion. The value of this technique lies in the unexpected findings often obtained from a free
flowing group discussion.
3.2.1.1. Evaluation Grid 1: Selection of an online tool
The focus group was asked to rate the above stated tools across a set of 12 parameters (Table
3.1) by using a Rating scale (1-3) with 1 representing the best tool delivering consumer value
Research Methodology
99 Ph.D. Thesis
across a particular parameter. The Evaluation Grid 1 developed for the purpose is listed in
Annexure I.
A composite score was calculated for each respondent, for each tool. This was done by summing
up the scores for each participant across the entire parameters (Table 3.2). A mean consumer
perceived value score was calculated using SPSS 17.0 (Table 3.3).
Online communities were chosen because they create value for all their stakeholders including
the host members and any third parties such as advertisers. Increasing the perceived value of the
customers is the defining criteria in this respect. They can be used for value exploration in the
consumer cognitive space and enhancement of the relational equity of the firm. A community
can further serve as a value delivery mechanism to enhance the perceived value to the customers
by product promotions and enhancing customer cross selling and up selling, while also
stimulating greater content contribution and participation.
This study proceeds to outline individual features of online communities ranging from co-
presence, reciprocity and conviviality-these have well defined roles to play in building
relationships with customers and furthering the CRM goals of the organisation. Thus online
communities can become vital for organisations driving towards better profitability and
enhanced productivity by leveraging their internal and external customers.
Table 3.1: Construct of Consumer Perceived Value S.No. Parameter Source
1 Increasing customer perception of value in
organisation
(Zeithaml V., 1988)
2 Suitability for word of mouth marketing (Woerdl M. et al. 2008)
3 Enhancing customer participation (Alexander Ardichvili, Vaughn Page, Tim
Wentling, 2003)
4 Ease of navigation (Preece, J., 2001)
5 Ease of registration (Macaulay A. Linda ., 2007)
6 Ease of accessibility (Teo Hai H. et al. 2003)
7 Increasing customer knowledge of product (Sawhney M. et al. 2005)
8 Garnering traffic (Wayne G. Lutters, Mark S. Ackerman,
2003)
Chapter 3
100 Ph.D. Thesis
9 Facilitating communication ( Rovai P.A., 2007)
10 Congeniality of environment (Johnson M.C., 2001)
11 Customer centricity (Wagner C. and Majchrzak A., 2007)
12 Help build a community of customers (Wiertz C. and Ruyter K., 2007)
Table 3.2: Composite Respondent scores across all parameters for three Online Tools
Respondent Blog Online
community
Wiki
1 21 29 22
2 22 26 24
3 24 30 18
4 24 26 22
5 22 31 19
6 22 31 19
7 24 26 21
8 24 26 22
9 23 28 21
10 26 27 19
11 22 28 22
12 24 29 19
13 22 28 22
14 22 28 22
15 22 24 26
16 22 28 22
17 23 27 22
18 21 19 32
19 26 26 20
20 22 27 23
21 24 28 20
22 29 26 17
23 23 24 25
Research Methodology
101 Ph.D. Thesis
24 24 26 22
25 23 28 21
26 25 26 21
27 25 25 22
28 21 27 24
29 22 32 20
Table 3.3: Mean scores for Web 2.0 tools. Statistics
Blog Online
Community
Wiki
N Valid 29 29 29
Missing 0 0 0
Mean 23.2414 27.1034 21.6897
Online communities depicted the highest mean score amongst the three tools. They, hence,
appeared to be the tool of the collaborative web, best able to enhance a consumer’s perception of
value with regard to an organisation or brand, by virtue of ease of navigation and being the most
accessible to the consumer. The results of this study were further interpreted by a detailed study
of the features of online communities, through a literature review, to explore areas where they
could contribute to the CRM goals in organisations.
3.2.1.2. Evaluation Grid 2: Comparative analysis of three tools for CRM and CEM.
For the purpose, an evaluation grid was set up as a research instrument. Participants were asked
to rate the ability of the three online tools across each of the given parameters on a likert scale
(1-3).Thus the three touch points viz. Online Communities, Blogs, Wikis were rated with regard
to their ability to contribute to the value delivery and enhancement mechanism, which is the
backbone of Customer Experience Management. The Evaluation Grid 2 developed for the
purpose is listed in Annexure I.
Chapter 3
102 Ph.D. Thesis
A comparative study of Blogs, Online Communities and Wikis was undertaken across the above
discussed parameters. A focus group of consumers was identified by using the online-intercept
technique. The criteria for participation revolved around the internet usage of the participants
across 5 online consumer communities. These five online consumer communities were selected
through random stratified sampling conducted on an online directory of consumer communities.
Age and membership of the online community were used as stratification variables.
Stratified sampling is a probability sampling technique that uses a two-step process to partition
the population into subpopulations or strata. The strata should be mutually exclusive and
collectively exhaustive. Elements are selected from each stratum by a random procedure. The
variables used to partition the population into strata are referred to as stratification variables.
Figure 3.1: CEM using Collaborative Web tools
The above stated online tools are excellent customer feedback data streams for companies to
monitor perceptions and trends. Based on the need to create, deliver and enhance customer
perceived value, we proceeded to study the ability of three customer touch points viz. Blogs,
Wikis and Online communities with regard to their ability to contribute to the value delivery and
enhancement mechanism, which is the backbone of Customer Experience Management.
Research Methodology
103 Ph.D. Thesis
Figure 3.2: Customer Experience Management
3.2.2. Best practices in Customer Experience Management
3.2.2.1. Building Emotional Connection with Customers: Companies can improve
relationships with customers by participating in Online Communities. The literature review
shows in many ways that maintaining such relationships leads to loyalty. Customers who
participate in online discussions are motivated to repeat their purchases from the same company.
Customers will be loyal to these companies because they share the same ideology and views.
When companies take part in online forums, they are able to recognize the customers who want
some information, and who ask specific questions about their industry. Customers will
demonstrate loyalty to the company which has answered their specific questions. The
expectations of the customers have to be met consistently. Other uses include delighting
customers by meeting un-met needs, innovating new products, services, features and functions.
The result is that customers recommend the company and its products to their friends and family.
Customer Experience Management is effectively achieved when the customers step up to defend
the company and its product when they hear a negative comment about the brand. This can be
achieved only by connecting with the customer on an emotional level, showing you care and
appreciate them. Thus to achieve high degree of Customer Experience Management, the
Chapter 3
104 Ph.D. Thesis
companies should review their customer loyalty programs and strive to meet the basic needs and
un-met needs and make their customers feel valued.
Figure 3.3: Building Emotional connection with consumers
3.2.2.2. Listening to Voice of the customer and engineering influence: The bidirectional
communication can be used to practice listening to the customers. The bidirectional
communication is an important function of an online community. It further enhances customer
involvement and Customer Experience Management. The online web space is becoming a very
common place where people are connecting with other people in an emotional way. The
organisations have been listening to their customers for many years via surveys, via user tests for
interaction through their website. The online web space is an opportunity to engage and
communicate with customers, listening to them, showing them how the organisations are acting
on their feedback, and giving them feedback on what the organisation is doing. Customers want
to be listened to, they do not want to be passive receptors of the company’s sales pitch. The
online consumer communication tools can be blended with other voice of the customer sources
to create a holistic view of customer priorities.
Figure 3.4: Suitability for Voice of Customer
Research Methodology
105 Ph.D. Thesis
3.2.2.3. Building consumer participation: Online communities allow employees to share their
skills with other employees and consumers of a product; they are best in building a community
of customers. Continuous knowledge flow and capture due to steady inflow of quality data will
maximize the company’s intellectual resources. Intellectual resources will help organisations in
retaining old partners as well as capture new markets and clients. Consumer participation
enhances knowledge base for better knowledge acquisition from the explicit knowledge base.
The posts and reviews help in increasing the organisational brand equity as the consumer
perception for a product varies and depends a lot on the discussion on these communities. The
online communities have more discussions about the competitor’s products also and about new
features which increase the perceived value by stakeholders which can be utilized for growth of
the company.
Figure 3.5: Enhancing Customer Participation
3.2.2.4. Increasing customer knowledge and clear priorities for acting on customer
feedback: The organisations should establish a process internally where customer feedback from
online communities can be used as a basis for continual improvement of customer experience.
The importance of Customer Experience Management should be emphasized across all functions
of the organisation. A vital function of Customer Experience Management is to focus on
strategically significant customers. It is hence important to learn what these customers value and
what are the most important attributes driving customer loyalty and intent to purchase. A key
component of a branded customer experience is being differentiated in a way that is valuable to
targeted customers. This implies having a detailed understanding of the customer experience and
being intentional about designing it to deliver value at the key touch-points.
Chapter 3
106 Ph.D. Thesis
Figure 3.6: Increasing Customer Product knowledge
3.2.2.5. Increasing customer centricity: Greater customer centricity can be achieved by
nurturing a congenial environment, which aims at increasing customer value by increasing
customer satisfaction, increasing customer involvement and attempting to delight the customer.
Organisational objectives comprise nurturing brand advocates by connecting emotionally.
Figure 3.7: Customer centricity
3.2.2.6. Building a Community: This facilitates peer to peer consumer connect. There is a
concept of empathy and trust prevalent in Online Communities as it is said that greater
similarities amongst people forge better understanding. Furthermore when people discover they
have similar problems, requirements, opinions or experiences they may feel closer, more trusting
and be prepared to reveal even more. This has a “snowball effect” in that the more people
discover that they are similar to each other, the more they tend to like each other and the more
they will disclose about themselves. This is known as “self disclosure reciprocity” and it is
powerful online.
Figure 3.8: Help build a community of customers
Research Methodology
107 Ph.D. Thesis
Results were tabulated (Table 3.4). An individual score across each parameter was calculated for
each of the three tools. This was done by dividing the score across each parameter by the sum
total of the scores for each tool respectively.
Table 3.4: Individual Score across each parameter for three tools S.No. Parameter Blog Online
commun
ity
Wiki Blog-
Score
Online
community
-Score
Wiki-Score
1 Build Emotional
Connection with
consumers
59 54 51 0.087537
0.068878 0.09729
2 Suitability for Voice of
Customer
60 61 49 0.089021
0.077806 0.07815
3 Enhancing
customer
participation
65 70 39 0.096439
0.089286 0.0622
3.1 Ease of navigation 50 57 67 0.074184
0.072704 0.10686
3.2 Ease of registration 53 74 47 0.078635
0.094388 0.07496
3.3 Ease of accessibility 43 62 69 0.063798
0.079082 0.11005
3.4 Garnering traffic 55 71 48 0.081602
0.090561 0.07656
4 Increasing
customer
knowledge
of product
62 55 56 0.091988
0.070153 0.08931
5 Customer centricity 54 68 54 0.080119
0.086735 0.08612
5.1 Facilitating communication 55 67 52 0.081602
0.085459 0.08293
5.2 Congeniality of environment 63 63 48 0.093472
0.080357 0.07656
6 Help build a community of
customers
55 82 37 0.081602
0.104592 0.05901
Chapter 3
108 Ph.D. Thesis
Figure 3.9: Customer Experience Management using Online Communities
Online communities emerge as the best collaborative tools from the Customer Experience
Management perspective as they score highest with reference to building a community of
customers. They have greater accessibility which results in greater consumer participation,
hence building a group of loyal customers. People believe that they are breeding grounds for
new ideas and product improvements.
The outcomes of the above two pilot studies, hence formed the basis for my research thesis.
1. Online communities provide information that is updated frequently based on current
discussions and as it is online, it can also be consumed by people at diverse geographic
locations, thereby increasing knowledge share and exchange.
2. They provide better platform for communication which helps in increasing explicit
knowledge base of an organisation. They also help in knowledge transfer from their
database to the company’s database for analysis.
3. The posts made on the forums help an organisation in delivering good value to the
customer, based on their experiences and knowledge. In spite of blogs providing easy
search, people prefer these communities as a better substitute of word of mouth
marketing. Thus they provide continuous knowledge transmission back for better people
management and change.
Research Methodology
109 Ph.D. Thesis
4. Online Communities also provide a platform to both customers and employees to share
their observations, thus increasing socialization and interaction among the employees and
customers. Due to socialization, there is continuous knowledge reuse which helps
organisations for innovation on new products and developing an interaction friendly
environment for increasing product as well as brand equity.
5. Content delivery is better on wikis as they display the complete details of the products in
an organized manner. Wikis are the best alternative form of media as they provide
detailed knowledge of products, as compared to online communities or blogs, but blogs
provide the best data organisation as search is easy.
3.3. Proposed Conceptual Model
The conceptual model focuses on studying the ability of online communities, as a channel to
serve as an interface between organisation and consumer and aid the organisation in achieving its
CRM goals. The functions of the online communities which aid the process are listed in the
conceptual model below.
CRM and CEM using Business Online Communities
CRMCEMCLV
Co CreationEconomics of CRM
Operational, Analytical and
Collaborative CRMConsumer
Segmentation and Profiling
Customer Portfolio Analysis
Research Methods and Tools
Research InstrumentsRI1:-Applying Metcalf Law for INV & CNV-Identifying consumers with high relationship and profit potentialRI2: -Community Manager’s perspective and the OC
-Pilot Study-Netnography for identifying consumer trustworthiness
Online Community Dynamics
P¡ DOP¡ EA¡ OT¡ C¡ ML¡ AS¡ PAN¡ CP¡
Figure 3.10: Achieving CRM Goals by Using Online Communities
Chapter 3
110 Ph.D. Thesis
3.4. Research Objectives
This research thesis focuses on studying the ability of online communities, as a channel to serve
as an interface between organisation and consumer and aid the organisation in achieving its
CRM goals. This is accomplished through the following research objectives-
1. Develop a model to analyze the usage of online consumer communities for identifying the
components that build consumer trustworthiness for an organisation.
2. Create a framework for calculating the value of an online community based on its customers.
2. a. Identification of determinants of Individual Network Value (INV) and Community Network
Value (CNV) and creation of a framework to use INV as a basis for identifying consumer co-
creators.
3.Creation of a framework for selection of consumers demonstrating high future profit or
relationship potential and devise strategies to impact consumer price sensitivity for expensive,
medium and low cost products for organisations.
4. Creation of a model for identifying the credibility of a business online community from a
community manager’s perspective.
To accomplish Research Objective 1, Netnography was conducted on 40 online product
communities of Apple. Research objectives 2 and 3 were accomplished through self designed
Research Instruments (RI-1 and RI-2), (Annexure II) Research objective 4 was also
accomplished through self designed Research Instrument (RI-3), (Annexure II.).
3.5. Research Methodology for formulation of Research Instrument
I have applied the research technique of Netnography, which is very specific to the online
domain for formulation of two sets of research instruments. Experience is something singular
that happens to an individual and researchers cannot directly access (Caru, A. and Cova, B.,
2008). Therefore researchers only interpret what their subjects have expressed orally, in writing
or through their behaviour. Experience becomes more and more important to marketing,
Research Methodology
111 Ph.D. Thesis
however, the methodologies typically used to research experiences, such as interviews and focus
groups, have a number of drawbacks such as respondent inhibition (Elliott, R. and Jankel Elliot,
N., 2003). Verbatim comments instead, are important for understanding the private nature of the
experience to be studied.
3.5.1. Netnography
It is a new qualitative, interpretive research methodology that uses internet optimized
ethnographic research techniques to study the online communities. With the help of
Netnography, Online Community research can be done by either actively integrating the
members of the community or passively monitoring the community and integrating the gathered
information, knowledge and ideas into the new product development process, (Kozinets, Robert
V., 2002).
As a method, “Netnography” is faster, simpler, and less expensive than ethnography, and more
naturalistic and unobtrusive than focus groups or interviews. It provides information on the
symbolism, meanings, and consumption patterns of online consumer groups.
As a marketing research technique, “Netnography” uses the information publicly available in
online forums to identify and understand the needs and decision influences of relevant online
consumer groups. Compared to traditional and market oriented ethnography, “Netnography” is
far less-time consuming and elaborate. Another contrast with traditional and market-oriented
ethnography is that “Netnography” is capable of being conducted in a manner that is entirely
unobtrusive. Compared to focus groups and personal interviews, “Netnography” is far less
obtrusive, conducted using observations of consumers in a context that is not fabricated by the
marketing researcher. It also can provide information in a manner that is less costly and timelier
than focus groups and personal interviews. “Netnography” provides marketing researchers with a
window into naturally occurring behaviours, such as searches for information by, and communal
word-of-mouth discussions between, consumers. Because it is both naturalistic and unobtrusive,
an unprecedented unique combination not found in any other marketing research method-
“Netnography” allows continuing access to informants in a particular online situation. This
access may provide important opportunities for consumer-researcher and consumer-marketer
relationships.
Chapter 3
112 Ph.D. Thesis
Figure 3.11: Netnography as a Research Technique
3.5.2. Significance of Netnography
One of the main benefits of this methodology is the possibility to access unfiltered unbiased
information from very experienced and highly involved users, due to the huge amount of
conversations and the vivid online dialogue regarding consumer products marketing and
innovation. Managers are able to obtain deep insights into the everyday problems experienced by
consumers and their solutions to those problems. One of the main expectations of this new
technique of research methodology is to utilize a huge number of consumer statements for
qualitative analysis, to get unobtrusive and unbiased original consumer statements and to get
access to specialized user groups.
3.5.3. Procedure in Netnography
The following steps and procedures are included in a typical Netnography Research, (M. Bartl,
Steffen Huck, Stephan Ruppert., 2009).
3.5.3.1. Definition of Research field: It includes the definition of the field of innovation as well
as the systemization of topics, trends, markets and products which are of major interest. The
operating result of the first step is an extensive mind map that contains a classification and
Research Methodology
113 Ph.D. Thesis
structured set of topics which are used as starting point to define search strategies for the
identification of adequate online sources.
3.5.3.2. Identification and selection of Online Communities: The aim of the second step of
Netnography is to identify communities and internet sources where users exchange relevant
information on the defined research area. For this purpose, general online search engines, meta
search engines and specific online search engines that focus on blogs, groups, communities are
used. Having identified and sighted often, a couple of hundred relevant online sources for
Netnography, the researcher has now to select the communities which can be probed in for
further in-depth analysis. There exist a number of appropriate and well proven qualitative and
quantitative criteria which support the researcher in the selection procedure. Qualitative criteria
include e.g. “topic focus”, “data quality”, “language type”, “interaction type”, “profile editing”.
Quantitative criteria include criteria such as “number of messages”, “frequency of usage”,
“member activity”, “data quantity” or “interaction level”.
3.5.3.3. Community Observation and Data Collection: In this step, the selected online
communities are observed by the researcher who immerses in the community. This is
accomplished by extensive reading with focus on conversations which are recent, extensively
corresponded to, referenced and frequently viewed from the community members. While
before the emergence of the internet, it was necessary for the researcher to participate in the
considered group, nowadays Netnography enables observation and analysis of the consumer
communication without active participation. Hence the approach is a way to unobtrusively
study the nature and behaviour of online consumer groups. The analysis is conducted in the
natural context of the community and thus is free from the bias which may arise through the
involvement of the researcher or experimental research setting.
3.5.3.4. Data Analysis and Aggregation of consumer insights: The “thinking” about the
“noticed” and “collected” online consumer statements is part of the fourth step of
Netnography. In this step the aim is to look for patterns and relationships within and across the
collections of consumer statements and to make general discoveries about the subject matter of
Chapter 3
114 Ph.D. Thesis
research. Therefore, the researcher compares and contrasts the collected consumer records in
order to discover similarities and differences, build typologies, or find sequences.
3.5.3.5. Translation of Community insights into product and service solutions: The
Netnography process typically does not end with the generation of insights. A major challenge
is to transfer the obtained insights into innovative product and service solutions. The
implications of the results could be for product, brand, target group as well for the process of
communication e.g. source for product innovations and product modifications and
development of consumer oriented communication strategies.
3.5.4. Strengths of Netnography Revelatory depth of online communication
Ability to provide interesting and useful conclusions from small number of messages
Useful for contextualizing the data
Can be used as a standalone method for tracking the marketing related behaviours of
members
Is based primarily on observation of textual discourse
Utilizing carefully chosen message threads is akin to “purposive sampling” in market
research
3.5.5. Benefits of Netnography to Online Communities Greater consumer engagement and participation
Enhance their value perception
Co-creation and consumer evangelism
Relationship building, value creation and commitment
Refine marketing actions, while reducing the cost of routine sales
Identification of strategically significant community members
Research Methodology
115 Ph.D. Thesis
3.5.6. Limitations of Netnography
The limitations of “Netnography” draw from its more narrow focus on online communities, the
need for researcher interpretive skill, and the lack of informant identifiers present in the online
context that leads to difficulty generalizing results to groups outside the online community
sample. Marketing researchers wishing to generalize the findings of a “Netnography” of a
particular online group to other groups, must therefore apply careful evaluations of similarity and
employ multiple methods for triangulation.
3.5.7. Netnography Methods Adopted
The following Netnography methods were used for the formulation of the two research
instruments-
Gaining entree into the community or group I wanted to investigate
Gathering and analyzing data
Ensuring trustworthiness of data interpretation
Conducting ethical research
Member checking, or getting feedback from participants (Hammersley, M. and Atkinson, P.,
1995;, Lincoln, Y. and Guba., E 1985; Wolcott, H.F., 1994)
The entree involved identifying the online communities most relevant to my research as well as
learning as much as possible about the communities that are identified. The following features
were preferred -
A more focused, relevant segment, topic, or group with large number of questions
Higher traffic of postings
Larger numbers of discrete message posters
More detailed or descriptively rich data
Higher Interactivity between members
Netnography in the form of both participatory observation as well as non participatory
observation was conducted for the formulation of the constructs of the three research
instruments. For the formulation of the third research instrument for community managers, the
Chapter 3
116 Ph.D. Thesis
specific behaviour of managers and moderators was observed. For the participant observation, I
participated in the community and asked some specific questions to the members who depicted
very high status level in the community, high level of participation and provided maximum
number of correct answers which further depicted a higher consumer brand or product
knowledge.
Netnography in the form of participatory observation was performed for the business online
communities. Some examples of communities where I conducted participant observation are
Apple, Dell, HUL, Microsoft, CSC, Cisco and Shiksha.com.
Netnography in the form of non-participatory observation was performed for some other
business online communities wherein the customer reviews published on the internet contained
detailed information about their experiences with the products and services of the firm. The
reason for choosing a combination of participatory and non participatory observation is the
undesirable influence of the outsider to the group. The learning derived out of this was used to
study individual behaviour and subsequently formed the basis to develop specific constructs to
study consumer learning.
3.6. Consumer Trustworthiness Regression model using Netnography (CTR)
Research Objective 1: Develop a model to analyze the usage of online consumer communities for
identifying the components that build consumer trustworthiness for organisations.
In order to accomplish the first research objective the online marketing research technique of
Netnography has been applied and an analysis of 40 different online product communities of
Apple and one online community of Dell namely “Ideastorm” has been done.
3.6.1. Netnography of 40 online product communities of Apple
This study aimed to achieve the following:
To study Electronic Customer Relationship Management in Organisations (E-CRM)
IT Enabled Relationship Management between organisation and consumer
Increase consumer engagement, participation, and trustworthiness of participants
Research Methodology
117 Ph.D. Thesis
Value creation and consumer commitment
Before I outline the entire study in detail, the following variables which appear on the Apple’s
online product communities have been defined as they form the genesis for the further study.
These variables help us to measure the trustworthiness of participants.
1. No. of Points: It represents the status level of a member in the forum category or at the
main community level. These points are earned by replying to another members’ question
topic. This is considered as a measure of Reciprocity and correctness of replies made by
a forum member.
2. No. of Views: It represents the no. of times a participant post is viewed by other
members. This is indicative of trustworthiness of the participant.
3. No. of Days: It represents the total number of days spent by a consumer in the online
community of Apple from the date of registration till 1st Dec 2010. This is indicative of
longevity of community presence.
4. No. of Posts: It represents the volume of messages created by the community members.
This is considered as a measure of Participation.
A previous study (Alavi, S., Ahuja V., and Medury Y 2010) “Building Participation, Reciprocity
and Trust – A Netnography of an Online Community of Apple-Using regression analysis for
prediction” had already explored the implications of all the above variables. The number of posts
and number of days appeared to have no direct relationship with trust, hence the same were not
explored further, in this study.
3.6.1.1. Content Organisation
Content organisation on the site, is initiated to enable consumers to view relevant content in
order to induce greater consumer participation and for creating and maintaining value laden
relationships with current and potential customers. The typology of content that attracts greater
consumer interest and generates subsequent engagement by soliciting participation and
involvement through comments needs to be identified to enable organisations to post content in
accordance with consumer receptivity. The online community of Apple - Apple Discussions
Chapter 3
118 Ph.D. Thesis
(http://discussions.apple.com/category.jspa?categoryID=204) was used for our study and
secondary data for about 400 consumers was collected for the study and further regression
analysis was done.
The content in Apple Discussions is organized as follows:
1. Forum Categories—Categories represent a collection of topical forums as well as other
categories, and are used to organize forums. Most categories are generally defined by a product
name, such as "iPod," "iMac," or "Mac OS X v10.4 Tiger." The Apple Discussions Forum has
40 product categories.
2. Forums—Forums are the areas where individual discussions take place. The discussions are
displayed as a list of topics. For example, if a consumer is looking for conversations about
searching their Mac with Spotlight, they can click the Mac OS X v10.4 Tiger link on the
Discussions homepage, and then click the Spotlight link in the resulting Tiger page to visit the
Spotlight forum.
3. Topics—Topics refer to the actual topics of discussion, each of which consists of messages
displayed as a conversation.
4. Messages—Messages are the individual posts made by community members. If a consumer
clicks a topic to view a discussion, they will see messages posted by other members.
5. Replies—Replies are posts made in response to other messages and are organized in a flat or
threaded manner. For example, if someone posts a question in a topic, other members may post a
reply to that question.
The procedure has been carried out for 40 product communities of Apple.
3.6.1.2. Participation and Reciprocity in the Apple’s Online Community
When a community member posts a question as a topic starter, other members can post an
answer in reply. An answer can be just some hints or helpful information to help the poster solve
Research Methodology
119 Ph.D. Thesis
an issue. The originator of the topic can mark such a reply as a "Helpful" post. A small yellow
star appears next to that reply and the person who posted the reply is awarded 5 points. If a
community member posts a specific answer, that provides a solution to the original poster's
issue, the originator marks this reply as a "Solved" post. A large green star appears next to that
reply and the person who posted the reply will be awarded 10 points. Only the original topic
poster has the option to mark replies as either Helpful or Solved or to not mark a reply at all. The
originator can also end the discussion by marking the topic as "answered," which displays a
green star at the top of the topic page to let everyone know that the topic contains valid helpful
information. If a participant replies to another member's question topic, they are eligible to
receive points from that member, though this is at his or her sole discretion. The originator has
the option of marking a reply as either Helpful or Solved, which will add points to the
respondent's account. These points, in turn, increase a member's ranking (status level) in the
community over time. A member receives 10 points for each reply that a member marks as
"Solved" and 5 points for each reply that a member has marked as "Helpful." The reward system
helps to increase community participation. When a community member gives a reward to
another member for providing helpful advice or a solution to his or her question, the recipient's
points will help increase his or her status level within the community. Members can see their
status level by Forum, Category, or at the main Community level.
Table 3.5: Depiction of status level for Apple’s communities Status Level Point Range
5 50,000+
4 8,000 - 49,999
3 1,000 - 7,999
2 150 – 999
1 30 – 149
Under this model, the studies are conducted on a set of online communities of Apple and the
correlation and regression model is applied for forty product categories. The results of a regression
model are used to analyze factors contributing to the growth of trust in an online community and for
finding out the contribution of the independent variable, that is, number of points to the dependent
Chapter 3
120 Ph.D. Thesis
variable, that is, number of views. The information is subsequently applied for prediction by analyzing
how far the dependent variable depends on the independent variable.
The results and findings of this study of Netnography have been detailed in Section 4.1.1. of
Chapter 4 on Results and Findings.
3.6.2. Netnography of Dell Ideastorm
The online community of Dell named “Ideastorm” bypasses the middlemen and sells directly to
the customers. It aims to integrate the experiences of the customers with the process of co-
creation. It leverages user collaboration to act as a multiplier force for better collaborative CRM.
Dell wants to see what all products and services users want Dell to develop. IdeaStorm gives a
direct voice to the customers. In almost three years, Idea Storm has crossed the 10,000 idea mark
and implemented nearly 400 ideas. The customers can interact with each other as well as Dell.
Customers can voice their concerns as well as complaints. Though these complaints may not be
regarding Dell units that these customers possess (handled by the support section), it is the user’s
feedback about the product in general. Managers can judge needs and respond quickly. A posted
idea grants Dell a royalty free license to use and implement it without compensation to the
originator. Marketers already have an idea of product perception amongst the consumer base.
The managers have increased clarity regarding what customers want and such products are more
likely to have better sales.
The following variables formed the basis of the study-
1. No. of Posts: Users can post their experiences and ideas
2. No. of Comments: Users can comment on others ideas
3. No. of Votes: Users can promote and demote others
4. Voted up: No. of vote ups on user’s ideas
5. Voted down: No. of vote downs on user’s ideas
6. No. of ideas :The number of ideas submitted
The data for users in the online community of Dell is collected and compiled and further
regression analysis was done (Section 4.1.2).
Research Methodology
121 Ph.D. Thesis
3.7. Research Instruments
Three sets of self designed questionnaires have been drafted, based on Netnography conducted
and extensive literature review, in order to study and analyze the factors that play a pivotal role
for an online business community to be successful in achieving the Customer Relationship
Management and Customer Experience Management goals of an organisation, both from the
perspective of an online community consumer and a community manager. All research
instruments are detailed in Annexure II.
3.7.1. Questionnaire for Online Community Consumers (RI-1)
This research instrument (RI-1) was divided into two sections. The first section i.e. Section
gathered information about the demographics of the respondents which included their name, age
and gender. The second section i.e. Section B measures Consumer participation, Degree of
participation, Emotional Attachment, Online Trust, Usability of the features of an online
community, Commitment, Member Loyalty, Attitude towards switching and Period of
association. The individual items of each of the questions are highlighted in the table below
(Table 3.6).
Table 3.6: Individual items of RI-1
S.No. Constructs Items 1. Participation
(4 point likert scale: 1-Definitely not important ,4-Very important)
Ease of access Barriers to entry Responsiveness Ability to address participant needs Role assignment Member empowerment Content typology Membership duration Membership retainment Degree of flexibility Access to shared resources Shared goals
2. Degree of Participation (5 point likert scale: 1-Poor, 5-Excellent)
Frequency of Login Ability to navigate Ability to provide correct responses Reciprocity Interactivity Co-presence Increase in trust Volume of content contributed Computer and Internet proficiency
Chapter 3
122 Ph.D. Thesis
Abilility to navigate the internet
3. Emotional Attachment (5 point likert scale: 1-Never, 5-Always)
Emotional support Belongingness Familiarity Affinity Bonding and liking Aids recruitment
4. Online Trust (3 point likert scale: 1-Low, 3-High)
Resolve consumer problem Consumer Uncertainty Customer satisfaction Communication level Increase Reliability Increase Altruism Promote self disclosure
5. Usability of features of an online community (Rank Order Scaling )
Chat Wikis Q&A Debate Forum Blog File Manager
6. Commitment (5 point likert scale: 1-Poor, 5-Excellent)
Support members Explicit goals Focus on specific needs Promote same identity Motivate for contribution Interpersonal similarity Clustering members Named groups Promote frequent interaction Promote interdependent tasks Promote competition Reduce repelling forces Promote diversity Promote testimonials
7. Member Loyalty (5 point likert scale: 1-Strongly Disagree,5-Strongly Agree)
Recommend Product Suggesting others to join Resolve problems Proactive to help resolve problems Rebuy or repatronize Interaction with other members Growth in community membership Business and social goodwill Feedback Test a new product or idea
8. AttitudeTowards Switching (3 point likert scale: 1-Never, 3-Always)
Switch Brand Low switching costs Change brand Discount option Unique buying experience Emotional Attachment Perceive greater quality in product Reputation of the company
Research Methodology
123 Ph.D. Thesis
9. Period of association with the network
Less than a year 1-2 years 2-3 years 3-4 years More than 4 years
3.7.2. Questionnaire for Online Community Consumers (RI-2)
This research instrument (RI-2) measures Consumer Price Sensitivity (For expensive, medium
and low cost products). The individual items of the question are highlighted in the table below
(Table 3.7).
Table 3.7: Individual items of RI-2 S.No. Constructs Items 1. Consumer Price Sensitivity
(5 point likert scale: 1-Strongly Disagree,5-Strongly Agree)
Desist from purchase if price out of line Resist purchase if price is more Evaluate price of competing brand Factors impacting resistance to brand
switching Sensitivity to product price
3.7.3. Questionnaire for Online Community Managers (RI-3)
This research instrument (RI-3) measures Community Dynamics, Co-creation, Online
community as a collaboration enabler, CRM goals of an organisation, Return on Investment and
Customer Life Time Value. The individual items of each of the questions are highlighted in the
table below (Table 3.8).
Table 3.8: Individual items of RI-3 S.No. Constructs Items
1. Community Dynamics (5 point likert scale: 1-Strongly Disagree, 5-Strongly Agree)
Fixed Purpose Fixed Policy Sufficient Community Size Quality of Participation Latest technologies and tools Restriction on membership Easily Navigable Greater Interactivity Breaks cross cultural barriers Designed for growth and change Generation of profits Platform for consumer learning Flexible for change
Chapter 3
124 Ph.D. Thesis
Greater accessibility to members Communication (Linked to search
engines) Presence of knowledge manager Influence of Moderator Depicts Self Disclosure Reciprocity Encourages the level of empathy Encourages the level of trust Provides clarity of procedure Creates and maintains feedback loop Strategic Marketing tool for company Supports the role of a developer Strategic marketing tool for company
2. Co-creation (5 point likert scale: 1-Not at all effective,5-Very Effective)
Improving customer satisfaction Achieving high productivity levels Collecting innovative ideas for new products Capturing and storing of new ideas Application and discussion on new ideas Harvesting new skills and competencies Develop consumer experts Creating value for existing products
3. Collaboration Enabler (5 point likert scale : 1-Never ,5-Always)
Operational performance Lower costs and higher quality Improved customer service and satisfaction Higher value added relationships Faster responsiveness
4. CRM Goals (5 point likert scale : 1Poor ,5-Excellent)
Consumer Information System Increase Customer Satisfaction Referrals from satisfied customers Predicting Consumer Behaviour Identify leads to generate customer pipeline Increase profitability Improve business planning Assess size of target market Assess growth rate of target market Improved customer sales to target ratio Increase overall customer profitability Increase customer service quality Increase customer retention Caters to Mass Customization Provides customized communications Long term customers to enter into
collaborations Lower Price Sensitivity Less time duration for conversion of prospect
to consumer 5. Return on Investment
(Dichotomous Question)
True False
6. Customer Life Time Value (5 point likert scale : 1Poor 5-Excellent)
Identify profitable customers High cost to serve customers Incremental Revenue per customer Average amount per purchase Increase in retention rate
Research Methodology
125 Ph.D. Thesis
Identify lifetime value potential of customers Identify strategic customers Consumer segmentation Anticipated life time of the customer
relationships Reduce acquisition cost Decrease direct marketing costs Reduce cost of mailing Provide presale and post sale support Identify customers who value unique
functionality and features Increase average number of consumer
purchases Improved ordering and delivery relationships Raise switching costs Increase a consumer’s contribution margin Identify price sensitive consumers Reduce consumer price sensitivity Retention effort is aligned to customer life time
value Identify customer categories for cross selling
and up selling Development of service and product portfolios
3.7.4. Pretesting the Research Instruments (RI-1, RI-2 and RI-3)
Pretesting implies testing of the questionnaire on a small sample of respondents for the purpose of
improving the questionnaire by identifying and eliminating potential problems. Pretests were done by
conducting personal interviews. The respondents who were selected for pretesting of the
questionnaires were similar to those that were included in the actual survey. The respondents for the
first two questionnaires were online community consumers and for the third were community
managers. The sample size for pretesting was 30 respondents for RI-1, RI-2 and 10 respondents for
RI-3. The questionnaires were administered in an environment and context similar to that of the actual
survey. The debriefing procedure was used for pretesting. Debriefing occurred when the questionnaire
was completely filled up by the respondents when they were informed that the questionnaire filled up
by them was meant for a pretest. The objectives of the pretest were described to them and then they
were asked to a) describe the meaning of each question, b) explain their answers and c) state the
problems encountered by them at the time of answering the questions. After this, the questionnaires
were edited in the context of the identified problems.
Chapter 3
126 Ph.D. Thesis
3.7.5. Reliability and Validity of RI-1
In order to test the reliability and validity of RI-1, the same was administered to 100
respondents. The collected responses were tabulated. Reliability and validity tests were
conducted on the data to check the validity and usability of the instrument. Cronbach’s alpha,
KMO measure of adequacy and Bartlett’s test of sphericity were conducted.
Reliability refers to the extent to which a scale produces consistent results if repeated
measurements are made (Sinha P., 2000). Thus it is the degree to which selected constructs yield
similar results. It is assessed by determining the proportion of systematic variation in a scale by
determining the association between scores obtained from different administrations of the scale.
If the association is high, the scale yields consistent results and is therefore reliable. For my
research, I have used internal consistency method of reliability. The internal consistency
reliability of the research instruments was tested by using cronbach’s alpha. It is one of the most
frequently used methods to check the internal consistency of the survey items. A higher degree
of cronbach’s alpha coefficient demonstrates higher degree of inter item correlation among the
constructs. If the value of cronbach’s alpha is more than 0.7, then the instrument is considered to
be reliable.
Further, Kaiser-Meylen-Olkin test was done to measure the homogeneity of variables and
Bartlett’s test of sphericity was done to test for the correlation among the variables used. The
Bartlett’s test showed significant results for all the questions and thus the instrument was used
for further study. Table 3.9 summarizes the results. Thus the instrument was further circulated
for data collection.
Table 3.9: Tests of Reliability and Validity of RI-1 Questions No. of
Items Cronbach’s Alpha
Kaiser-Meyer-Olkin Measure of Sampling Adequacy
Bartlett's Test of Sphericity
Participation 12 0.897 0.916 Approx. Chi-Square
12088.012
Research Methodology
127 Ph.D. Thesis
df 190.00
Sig. .000
Degree of participation
10 0.803 0.855 Approx. Chi-Square
5052
df 210.00
Sig. .000
Emotional attachment
6 0.931 0.816 Approx. Chi-Square
4450
df 290.00
Sig. .000
Online Trust 7 0.858 0.789 Approx. Chi-Square
12354.031
df 180.00
Sig. .000
Usability of Online Tools
7 0.768 0.868 Approx. Chi-Square
11543.051
df 235.00
Sig. .000
Commitment 14 0.863 0.931 Approx. Chi-Square
3975.321
df 310.00
Sig. .000
Member Loyalty
10 0.936 0.882 Approx. Chi-Square
2495.720
df 156.00
Chapter 3
128 Ph.D. Thesis
Sig. .000
Attitude towards switching
8 0.832 0.796 Approx. Chi-Square
6058.012
df 190.00
Sig. .000
Period of association with the network
5 0.821 0.917 Approx. Chi-Square
5207.352
df 250.00
Sig. .000
3.7.6. Reliability and Validity of RI-2
In order to test the reliability and validity of RI-2, the same was administered to 100
respondents. The collected responses were tabulated. Reliability and validity tests were
conducted on the data to check the validity and usability of the instrument. Cronbach’s alpha,
KMO measure of adequacy and Bartlett’s test of sphericity were conducted. Table 3.10
summarizes the results. Thus the instrument was further circulated for data collection.
Table 3.10: Tests of Reliability and Validity of RI-2
Questions No. of Items
Cronbach’s Alpha
Kaiser-Meyer-Olkin Measure of Sampling Adequacy
Bartlett's Test of Sphericity
Consumer price sensitivity
12 0.901 0.891 Approx. Chi-Square
3495.620
df 165.00
Sig. .000
Research Methodology
129 Ph.D. Thesis
3.7.7. Reliability and Validity of RI-3
In order to test the reliability and validity of RI-3, the same was administered to 50 community
managers. The collected responses were tabulated. Reliability and validity tests were conducted
on the data to check the validity and usability of the instrument. Cronbach’s alpha, KMO
measure of adequacy and Bartlett’s test of sphericity were conducted. Table 3.11 summarizes
the results. Thus the instrument was further circulated for data collection.
Table 3.11: Tests of Reliability and Validity of RI-3 Questions No. of
Items Cronbach’s Alpha
Kaiser-Meyer-Olkin Measure of Sampling Adequacy
Bartlett's Test of Sphericity
Community Dynamics
25 0.946 0.926 Approx. Chi-Square
5039.522
df 252.00
Sig. .000
Co-creation 8 0.872 0.786 Approx. Chi-Square
4058.012
df 160.00
Sig. .000
Collaboration Enabler
5 0.913 0.861 Approx. Chi-Square
3450
df 180.00
Sig. .000
CRM goals 18 0.902 0.816 Approx. Chi-Square
2495.720
df 153.00
Chapter 3
130 Ph.D. Thesis
Sig. .000
Return On Investment
2 0.721 0.817 Approx. Chi-Square
5072.552
df 153.00
Sig. .000
Customer Life Time Value
23 0.926 0.891 Approx. Chi-Square
3039.522
df 162.00
Sig. .000
3.8. Co-creation model using INV based on Metcalf Law (C-INV))
This model aims at analyzing the determinants of Individual Network Value (INV), further goes
on to develop a framework for calculating Community Network Value and subsequently,
empirically identifies individuals who have high INV. The community value of a network is of
great importance to organisations, especially from the CRM, Marketing and Customer
Experience standpoint. The study segments consumers using hierarchical cluster analysis into
groups to identify consumer co-creators.
The study of INV addresses the following research questions:
1. Identification of determinants of Individual Network Value (INV) and Community Network
Value (CNV)
2. Weighting of determinants
3. Using Metcalf Law to create a framework for calculating Individual Network Value (INV)
4. In the purview of Metcalf’s law and Individual Network Value, create a framework for
calculating the value of an online community based on its customers
5. Creation of a framework to use INV as a basis for identifying consumer co-creators
3.8.1. Identification of Determinants of Individual Network Value
My previous research studies on E-CRM (Alavi, S, Ahuja V. and Medury Y., 2011), Online
Communities (Alavi, S, Ahuja V. and Medury Y., 2011) and Customer Experience Management
Research Methodology
131 Ph.D. Thesis
(Alavi, S, Ahuja V. and Medury Y., 2010) helped identify the following determinants of INV. To
calculate the value of an individual consumer in a network, we consider the following variables.
1. Participation in the community network (P¡)
2. Emotional Attachment (EA¡)
3. Online trust(OT¡)
4. Commitment(C¡)
5. Member Loyalty(ML¡)
6. Attitude towards switching (AS¡)
7. Period of association with the network(PAN¡)
3.8.2. Weighting of Determinants
An Evaluation Grid 3 mentioned in Annexure I, was circulated to a focus group of 30
respondents, who were asked to rate the seven determinants of INV on a scale of 1-5 (5-
Excellent, 1–Poor). Based on their rating the following weights were extracted for each of the
determinant-Participation (P¡)-0.25, Emotional Attachment (EA¡)-0.13, Online Trust (OT¡)-
0.16,Commitment (C¡)-0.12, Member Loyalty (ML¡)-0.13, Attitude towards switching (AS¡)-
0.10, Period of association with the Network (PAN¡)-0.11.
3.8.3. Framework for calculating Individual Network Value (INV)
Based on the above determinants, the weighting criterion and Metcalf law, we create the following formula for calculating Individual Network Value.
Individual network Value = (0.25* P¡ + 0.13* EA¡ + 0.16* OT¡ +0.12*C¡+0.13* ML¡+0.10*
AS¡+0.11* PAN¡)
3.8.4. Framework for calculating value of an Online Community Further in the purview of Metcalf’s law and Individual Network Value, we create a framework
for calculating the value of an online community based on its customers.
Community Network Value=∑INV1+∑INV2+∑INV3+…..∑INV100
Chapter 3
132 Ph.D. Thesis
This framework modifies the above formula and CNV can be calculated as follows-
n (CNV) = ∑ (0.25* P¡ + 0.13* EA¡ + 0.16* OT¡ + 0.12* C¡ +0.13* ML¡+0.10* AS¡ i=1 +0.11* PAN¡)
Under this model the values of INVi (Table 4.27, Annexure III) were subjected to Hierarchical
cluster analysis using SPSS 17.0 and five consumer clusters were identified (Table 4.28, Section
4.3). Based on Table 4.27, Annexure III, the profiles of the 200 consumers divided into five
clusters were identified and appropriate targeting strategies were formulated (Table 4.28,
Section 4.3).
3.9. Consumer Price Sensitivity model using k-means cluster analysis (CPS)
This model attempts to study the usage of online communities to study consumer price
sensitivity in the context of the type of product purchased, i.e. expensive, medium or low cost
products. It examines the impact of reference price effect, difficult comparison effect, price
quality effect and switching cost effect on consumer price sensitivity and proceeds to segment
consumers into groups which demonstrate similar characteristics. This model uses data collected
from the Apple Online Discussion Forums
(discussions.apple.com/category.jspa?categoryID=204). Apple has over 40 product communities
formulated for consumer engagement, interaction, feedback, building strong online consumer
trust and reciprocity. The consumer profiling was done on the data collected using K-means
clustering and cluster memberships were extracted and the most significant consumers across all
three product categories were identified. Further the detailed profiles for most significant
consumer clusters for each price category were created (Table 4.40, Section 4.5).
Organisations will benefit by identifying the strategically significant consumers in each category
and then target them appropriately rather than investing in a blanket promotion program. The
objective is to enable organisations to identify consumers demonstrating future profit or
relationship potential and devise strategies to impact price sensitivity by responding to price
search intentions, improving product perceptions, improving consumer experiences, informing
consumers about new schemes and improving product perceived value. Most valuable consumers
Research Methodology
133 Ph.D. Thesis
will be those who depict consistent behaviour across all three product typologies, as they will
tend to be the most loyal. Valuable consumers identified through a customer portfolio analysis
can be leveraged by hosting appropriate content in an online business community and
subsequently using customers engaged through business online communities as important
sources of competitive advantage. Due to these benefits, online business communities may
generate more profitable sales than transactional marketing methods.
3.10. Data Collection for RI-1
The data collection for the research instrument RI-1 has been done in two phases.
3.10.1. Phase I of primary data collection
The comprehensive list of 100 online consumer communities was used as the database for my
research. This list was sourced from the “Customer Community project, Building Professional
Peer Communities”, Leader Networks, 2009. The online questionnaire was drafted using
Qualtrics (www.qualtrics.com). The appropriate link was hosted in the respective community.
The data was primarily collected through a period of six months starting from July 2011 till
December 2011. As it was not possible to build one to one contact with all the online community
consumers, a sample size of 200 was chosen for the study. The sampling techniques are
discussed in detail.
Apart from this database, consumers from some other online business communities were also
tapped for data collection. The consumers of these communities were contacted through
professional networks like LinkedIn. Personal visits were made to locally available consumers in
order to gather the information through hard copy questionnaires. Thus data was collected
through hard copy questionnaires and also through the internet.
The sampling techniques involved “online intercept random sampling” as well as “snow ball
sampling”. In online intercept sampling, consumers to these communities were intercepted and
given an opportunity to participate in the survey. For a consumer to fill the questionnaire, the
criteria were (a) The consumer should have been a member of an online business/product
community for at least 1 year (b) The consumer should have a minimum frequency of
participation of at least once a month. The consumers who fulfilled the above criteria were
further selected on the basis of simple random sampling technique. Nevertheless, randomization
Chapter 3
134 Ph.D. Thesis
improved representativeness and discouraged multiple responses from the same respondent. In
order to increase the number of respondents in the survey and to improve the overall response
rates, up to two additional reminder invitations were typically mailed at two-to-four-days
intervals to those respondents who fulfilled the defined criteria but had not yet participated in
the survey.
I also used snow ball sampling technique, which is a non probability sampling technique for data
collection. Personal messages were mailed to certain distinguished consumers who depicted high
status level in these online communities to provide some referrals. The link of the questionnaire
was mailed to these referrals requesting them to fill the questionnaire. This process was carried
out in waves by obtaining referrals from referrals, thus leading to a snow balling effect. Even
though probability sampling is used to select the initial respondents, the final sample is a non
probability sample. It also results in relatively low sampling variance and costs.
In total, the questionnaire was sent to 300 online community consumers, of which 219
responded, thus making the response rate to be 73%. 19 questionnaires were discarded due to
incomplete information. 200 fully completed questionnaires were considered for the study. From
the 200 respondents, in total, 129 (64.5%) respondents were male and 71(35.5%) were female.36
(18%) of the respondents were under 25 years old, 51 (25.5%) were between 25 and 35, 72
(36%) were between 36 and 45 and 41 (20.5%) were older than 46.
3.10.2. Phase II of primary data collection (For Co-creation model using INV based on
Metcalf Law (C-INV))
The data collection for calculation of INV values using the framework created has been done
across communities of four companies, namely Apple (Apple I Pad), Cisco (Cisco Collaboration
Community), Dell (Ideastorm) and Microsoft (Microsoft Dynamics CRM) using “online
intercept random sampling Technique”. These companies are the only ones where information
with regard to the last date of consumer’s participation was available to facilitate online intercept
sampling technique. The participants to these communities were intercepted and given an
opportunity to participate in the survey. The criteria for participating in the survey was a status
level of 5 and a point range of 50,000 and above. The seven questions of RI-1 (participation,
emotional attachment, online trust, commitment, member loyalty, attitude towards switching and
Research Methodology
135 Ph.D. Thesis
period of association with the network) were filled again by the participants who fulfilled the
above mentioned criteria from these specific four communities, for conducting the study of INV
based on Metcalf law. The data was primarily collected through a period of six months starting
from July 2011 till December 2011. As it was not possible to build one to one contact with all the
online community participants, a sample size of 50 participants of one community was chosen
for the study.
The participants who fulfilled the above criteria were further selected on the basis of simple
random sampling technique. Nevertheless randomization improved representativeness and
discouraged multiple responses from the same respondent. In order to increase the number of
respondents in the survey and to improve the overall response rates, up to two additional
reminder invitations were typically mailed at two-to-four-days intervals to those respondents
who fulfill the defined criteria but have not yet participated in the survey.
Finally the data collection was accomplished from 50 participants for each of the four
communities. In total, 100 participants from each community were requested to complete the
survey. Some questionnaires from each community were incomplete, so they were discarded. 50
fully completed questionnaires from each community were considered for the study.
Thus, for each consumer, we were able to calculate the respective values for Participation (Pi),
Emotional Attachment (EAi), Online trust (OTi), Commitment (C¡), Member loyalty (ML¡) and
attitude towards switching (AS¡) which was the summation for each variable, across each of the
determinants, for 200 consumers on the basis of the framework mentioned in Section 3.8.3
(Table 4.27, Annexure III).
Framework for calculating Period of Association with the network (PAN¡) for each
consumer: The consumers were asked to identify the time span of consumer-product association
which would prevent them from switching to another product or brand. There were 5 categories
of Period of Association (Less than 1year, 1-2 years, 2-3 years, 3-4 years and more than 4
years).The number of consumers who opted for the respective categories were summed up
(∑CPAN¡). Further, the PAN¡ index for each consumer was calculated by using the formula
[(∑CPAN¡)/50] for each of the respective four communities.
Chapter 3
136 Ph.D. Thesis
3.11. Data Collection for RI-2
The data was collected across 200 consumers from the 40 online product communities of Apple
using online intercept sampling technique. In these communities of Apple, the information with
regard to the last date of consumer’s participation was available to facilitate online intercept
sampling technique. The participants were intercepted and given an opportunity to participate in
the survey.
The pre-requisite for a consumer to fill the questionnaire was:
(a) The consumer should have been a member of an online business/product community for at
least 1 year.
(b) The consumer should have a minimum frequency of participation of at least once a month.
3.12. Community Managers
Community managers serve as Community Advocates by representing the customer. This
includes listening, which results in monitoring, and being active in understanding what
customers are saying, in both the corporate community as well as external websites. They also
engage customers by responding to their requests and needs or just conversations, both in private
and in public. As a Brand Evangelist, the community manager promotes events, products and
upgrades to customers by using traditional marketing tactics and conversational discussions. The
community managers also gather community feedback for future product and services.
Perhaps the most strategic of all tenets, community managers are responsible for gathering the
requirements of the community in a responsible way and presenting it to product teams. This
may involve formal product requirement methods from surveys to focus groups, to facilitating
the relationships between product teams and customers. The opportunity to build better products
and services through this real-time live focus group are ripe. In many cases, customer
communities have been waiting for a chance to give feedback. Every network has an underlying
purpose and motivations for such network creation include: Mission, Business, Idea, Learning or
Personal. The Community Manager holds the collective vision to create and manage
relationships, manage collaborative processes, community avocation and brand
ambassadorship.
Research Methodology
137 Ph.D. Thesis
3.13. Data Collection for RI-3
The third research instrument RI-3 was administered to online community managers. The
community managers were asked to rate their respective online communities with respect to the
community dynamics, as also the ability of the communities to support the process of co-creation
within the community. Ability of the community to serve as a collaboration enabler, as also its
ability to cater to the process of Customer Relationship Management were measured. The
community managers were also asked to rate their communities with respect to the ability of the
community to provide a good ROI as also the aspects of Customer Lifetime Value.
In 2008, Jeremiah Owyang (Web strategist, Altimeteer Group, Research Analyst, Forrester 2008)
began maintaining a list of online community managers employed by large corporations (a
Fortune 5000 company or over 1,000 employees).
To fill Research Instrument RI-3, I have used this database of online community managers
(http://www.web-strategist.com/blog/2008/06/20/list-of-social-computing-strategists-and-
community-managers-for-large-corporations-2008/).
The contacts were established through LinkedIn/personal visits/and company websites and
communities. Community managers were contacted across Technology, Gaming, Consumer
Goods, Agriculture, Health, Education categories from the above mentioned database.
The Community Managers are listed in the following Table:
Chapter 3
138 Ph.D. Thesis
Table: 3.12: Database of Online Community Managers S.No. Sector Community Manager
1. Technology
Lionel Menchaca, Community Manager, Dell
Anton Chiang, Web Communities Manager,
Juniper Networks
Lacy Kemp, Social Media Communications
Specialist at Real Networks
Stephen Spector, Sr. Program Manager,
Xen.org Community, Citrix
Michael Sandoval, Global Communities
Manager, Texas Instruments
Vishal Ganeriwala, Sr. Manager of Citrix
Developer Network, Citrix
Amie Paxton, Channel Community Manager,
Dell
Angela LoSasso, Community & blogs
strategist, HP
Tom Diederich, Social Media/Web Community
Manager, Cadence Systems
Bill Pearson Bill, Manager, Intel Software
Network, Intel
Josh Hilliker, Community Manager of the vPro
Expert Center, Intel
Robyn Tippins, Community Manager, Yahoo!
Developer Network at Yahoo!
John Summers, Community Manager at
NetApp
Mario Sundar, Community Evangelist at
Tom Ablewhite, Community Manager,
Thomson Reuters
Craig Cmehil, Community Manager for the
SAP Developer Network
Lou Ordorica, Social Media Producer at Sun
Research Methodology
139 Ph.D. Thesis
Microsystems
John Earnhardt, Senior manager, media
relations and blogger in chief, Cisco Systems
Deirdre Walsh, Community Manager at
National Instruments
Rachel Luxemburg, Community Manager at
Adobe
Aaron Tersteeg, Software Developer
Community, Intel
Josh Bancroft, Software Developer
Community, Intel
Jeff Moriarty, Software Developer Community,
Intel
Cathy Ma, Yahoo Community Manager, Yahoo
Europe
Shashi Bellamkonda, Social Media Swami,
Network Solutions
Ian Kennedy, Product Guy, MyBlogLog,
Community Manager, Yahoo
David Kim, Manager, Online Marketing and
Communities at Symantec
Marilyn Pratt, Community Evangelist, SAP
Labs
Scott Jones, Community Manager and Content
Strategist, SDN at SAP Labs
Badshah Mukherji, Sr. Community Manager at
VMware
Jon Mountjoy, Community Manager & Editor-
In-Chief at Salesforce
Senior Director, OTN & Developer Programs
Oracle
Jake Kuramoto, Oracle Apps Labs, Oracle
USA
Chapter 3
140 Ph.D. Thesis
Kelly Feller, Web Marketing Manager leading
the IT Community site Open Port, Intel
Erica Kuhl, Sr. Producer & Community
Manager, Salesforce.com Community
Aaron Tersteeg, Community Manager (Multi-
core Development) Intel Software Network,
Intel
Jeff Moriarty, Community Manager (mobility)
for the Intel Software Network, Intel
Alison Bolen Editor, Sascom voices blog, SAS
Melissa Daniels, Community Manager for All-
Star group for Yahoo! Messenger, Yahoo!
Amy Barton, Strategic Programs Manager,
Intel Software Network, Intel
Holly Valdez, Community Manager, Cisco, the
WebEx Technology group
2. Electronics
Ray Haddow, Blogger Outreach, Nokia
Charlie Schick, Lead on Nokia corporate blog,
Nokia
3. Media, Gaming,
Entertainment
Kellie Parker, Online Community Manager at
Sega
Kristopher Shaw, Community Manager at MTV
Networks UK
EM Stock, Senior Community Manager at Sony
Online Entertainment
Katie Hamlin, Community Manager,
Fodors.com, Random House
Justin Korthof, Community Manager at
Microsoft
David Cushman, Digital Development Director,
Bauer Consumer Media UK
Laurent Courtines, Community Manager at
Games.com AOL
Research Methodology
141 Ph.D. Thesis
4. Research John Cass, Online Community
Manager, Forrester Research
5. Finance
Scott Moore, Senior Online Community
Manager at Schwab Learning
Jose Antonio Gallego, Community Manager at
BBVA (Spain)
Amy Worley, Director, Marketing Manager,
HR Block
Fran Sansalone, Community Manager for the
Open Calais Web Service, Thomson Reuters
6. Automotive
Karen Spiegler, Community Manager,
Edmunds.com, Inc.
Alicia Dorset, Blog editor, General Motors
7. Retail
Slaton Carter, Online Community
Development Manager, Whole Foods Market
Winnie Hsia, Online Community Moderator,
Whole Foods Market
8. Consumer Goods Jennifer Cisney, Chief Blogger, Kodak
9. Agriculture
Christopher Paton, Social Media Team Lead,
Monsanto
In addition to the above database this research instrument was also filled by some eminent
industry experts in this domain. The contacts were established through LinkedIn/ personal
visits/and company websites and communities. These were Jay Ehret: Marketing educator and
resource provider for entrepreneurs and small business owners, speaker, blogger, podcast host,
Gautam Ghosh: Platform Evangelist at Brave New Talent, Blogger, Speaker and Writer at
Gautam on Organisations 2.0, Ashok Krish: Head - Web 2.0 Innovation labs at Tata
Consultancy Services, Sampson Lee: President and Founder at Global CEM (Global Customer
Experience Management Organisation) , Andrew Calvert: Experienced professional delivering
measurable results in Leadership, Engagement, Sales and Customer Loyalty, Pradeep Gairola:
Chapter 3
142 Ph.D. Thesis
COO at Applect Learing Systems Pvt. Ltd, Ashok Lalla: Marketing 3.0 Evangelist and
Consultant, Founder and Author of 'The Future Of Digital For Brands', Sujith Nair: Vice
President - Engineering at Info Edge India Ltd., Community manager from Shiksha.com.
3.14. Business Online Community Credibility Model (BOCC) using Linear Programming
The Business Online Community Credibility model (BOCC) was developed to measure the
credibility of an online community by incorporating the concepts of community dynamics, co-
creation, collaboration enabler, CRM goals, return on investment and customer life time value.
The model uses a numeric weighting technique to calculate a business online community
credibility score, using an Evaluation Grid 4 in Annexure I. The model further uses the linear
programming technique to maximise the BOCC score and further to determine an optimal
solution for the variables contributing to the BOCC score. The results are discussed under
Section 4.6.6. in Chapter number 4 on Results and Findings.
3.15. Tools for Data Analysis
Statistical Package for Social Sciences (SPSS) version 17.0 was used for statistical analyses of
the collected and tabulated data. The following statistical techniques have been used for analyses
across all the three research instruments. A brief description of these techniques is as follows-
3.15.1. Correlation
It is the most widely used statistic method for summarizing the strength of association between
two metric (interval or ratio scaled) variables, say X and Y. It is an index used to determine
whether a linear, or straight–line relationship exists between X and Y. It indicates the degree to
which the variation in one variable, X, is related to the variation in another variable, Y. The
correlation coefficient r is also known as Pearson correlation coefficient as it was originally
proposed by Karl Pearson. It is also referred to as simple correlation, bivariate correlation or
merely the correlation coefficient. Often correlation analysis is used in conjunction with
regression analysis to measure how well the regression line explains the variation of the
dependent variable, Y. Correlation can also be used by itself, however to measure the degree of
association between two variables.
Research Methodology
143 Ph.D. Thesis
3.15.2. Regression
Regression analysis is a powerful and flexible statistical procedure for analyzing associative
relationships between a metric dependent variable and one or more independent variables. It can
be used in the following ways: 1. Determine whether the independent variables explain a
significant variation in the dependent variable. 2. Determine how much of the variation in the
dependent variable can be explained by the independent variables. 3. Determine the structure or
form of the relationship. 4. Predict the values of the dependent variable. 5. Control for other
independent variables when evaluating the contributions of a specific variable or set of variables.
Although the independent variables may explain the variation in the dependent variable, this
does not necessarily imply causation. The use of the terms dependent or criterion variables, and
independent or predictor variables, in regression analysis arises from the mathematical
relationship between the variables. These terms do not imply that the criterion variable is
dependent on the independent variables in a causal sense. Regression analysis is concerned with
the nature and degree of association between variables and does not imply or assume any
causality. The mathematical equation is derived in the form of a straight line by using the least –
squares procedure. When the regression is run on standardized data, the intercept assumes value
of 0, and the regression coefficients are called beta weights. The strength of association is
measured by the coefficient of determination, r 2.
Further it is the operation of learning a function that predicts the value of a real valued
dependent variable based on the values of other independent variables. It is a technique used for
the modelling and analysis of numerical data consisting of values of a dependent variable and of
one or more independent variables. The dependent variable in the regression equation is
modelled as a function of the independent variables, corresponding parameters (constants), and
an error term. The error term is treated as a random variable and represents unexplained variation
in the dependent variable. The standard error of estimate is used to assess the accuracy of
prediction and may be interpreted as a kind of average error made in predicting the dependent
variable from the regression equation. Parameters are estimated to give a "best fit" of the data.
Most commonly, the best fit is evaluated by using the least squares method, but other criteria
have also been used. Regression can be used for prediction (including forecasting of time-series
data), inference, hypothesis testing, and modelling of causal relationships.
Chapter 3
144 Ph.D. Thesis
3.15.3. Prediction
Prediction is the same as classification or estimation, except that the records are classified
according to some predicted future behaviour or estimated future value. In a prediction task, the
only way to check the accuracy of the classification is to wait and see. The primary reason for
treating prediction as a separate task from classification and estimation is that in predictive
modelling there are additional issues regarding the temporal relationship of the input variables or
predictors to the target variable.
3.15.4. Factor Analysis
Factor analysis, also called exploratory factor analysis (EFA), is a class of procedures used for
reducing and summarizing data. Each variable is expressed as a linear combination of the
underlying factors. Likewise, the factors themselves can be expressed as linear combinations of
the observed variables. The factors are extracted in such a way that the first factor accounts for
the highest variance in the data, the second the next highest, and so on. In formulating the factor
analysis problem, the variables to be included in the analysis should be specified based on past
research, theory, and the judgement of the researcher. These variables should be measured on an
interval or ratio scale Factor analysis is based on a matrix of correlation between the variables.
The appropriateness of the correlation matrix for factor analysis can be statistically tested. The
two basic approaches to factor analysis are principal components analysis and common factor
analysis. I have used the principal component analysis method. Here the total variance in the data
is considered. The diagonal of the correlation matrix consists of unities, and full variance is
brought into the factor matrix. Principal component analysis is recommended when the primary
concern is to determine the minimum number of factors that will account for maximum variance
in the data for use in subsequent multivariate analysis. The factors are called principal
components. In principal component analysis the initial or unrotated factor matrix indicates the
relationship between the factors and individual variables, it seldom results in factors that can be
interpreted, because the factors are correlated with many variables. Therefore rotation is used to
transform the factor matrix into a simpler one that is easier to interpret. The most commonly
used method of rotation is the varimax procedure, which results in orthogonal factors. The
rotated factor matrix forms the basis for interpreting the factors. Factor scores can be computed
for each respondent. Alternatively, surrogate variables may be selected by examining the factor
Research Methodology
145 Ph.D. Thesis
matrix and selecting for each factor a variable with the highest or near highest loading. The
differences between the observed correlations and the reproduced correlations, as estimated from
the factor matrix, can be examined to determine model fit. The principal components method of
extraction was used for data reduction. Components with Eigen values greater than 1 were
extracted. As the communalities were all high, the extracted components represented the
variables well.
3.15.5. Simplex Tableau Method of Linear Programming
The linear programming method is a technique for choosing the best alternative from a set of
feasible alternatives, in situations in which the objective function as well as the constraints can
be expressed as linear mathematical functions. The simplex tableau method of linear
programming provides an efficient technique which can be applied for solving LPPs of any
magnitude involving two or more decision variables. In this technique, the objective function is
used to control the development and evaluation of each feasible solution to the problem. The
simplex algorithm is an iterative procedure for finding, in a systematic manner, the optimal
solution to a linear programming problem. This method, according to its iterative search selects
this optimal solution from among the set of feasible solutions to the problem. The algorithm is
indeed very efficient because it considers only those feasible solutions which are provided by the
corner points, and that too not all of them. Thus by using this technique we have to consider a
minimum number of feasible solutions to obtain an optimal one. Also, this technique has the
merit to indicate whether a given solution is optimal or not. For applying simplex method to the
solution of an LPP, first of all an appropriately selected set of variables is introduced into the
problem. The iterative process begins by assigning values only to these variables and the primary
(decision) variables of the problem are all set equal to zero. This assumption is analogous to
starting the evaluation process in the graphic approach at the point of origin, where both x1 and
x2 are equal to zero. The algorithm then replaces one of the initial variables by another variable –
the variable which contributes most to the desired optimal value enters in, while the variable
creating the bottleneck to the optimal solution goes out. This improves the value of the objective
function. This procedure of substitution of variables is repeated until no further improvement in
the objective function value is possible. The algorithm terminates there indicating that the
optimal solution is reached, or that the given problem has no solution. For the purpose of this
Chapter 3
146 Ph.D. Thesis
research thesis, linear programming has been performed using an online tool for the simplex
tableau method (http://www.zweigmedia.com/RealWorld/simplex.html).
3.15.6. Numeric Weighting Technique
It is a statistical adjustment to the data in which each case or respondent in the data base is
assigned a weight to reflect its importance, relative to the other cases or respondents. The value
1.0 represents the unweighted case. The effect of weighting is to increase or decrease the number
of cases in the sample that possess certain characteristics. It is most widely used to make the
sample data more representative of a target population on specific characteristics. For example, it
may be used to give greater importance to cases or respondents with higher quality data. Yet
another use of weighting is to adjust the sample so that greater importance is attached to
respondents with certain characteristics.
3.15.7. Datamining for Customer Relationship Management
Datamining techniques are algorithms and methods used to carry out datamining tasks. They
differ from each other in type of data handled, assumptions about the data, scope and
interpretation of the output. Analysis of large quantities of data require approaches that are very
different from the traditional data analysis approaches and this has given birth to the field of
Knowledge Discovery in Databases (KDD), more popularly known as Datamining. Knowledge
Discovery in Databases is a non-trivial process of identifying valid, novel, potentially useful and
ultimately understandable patterns in data, (Fayyad, U., Gregory, Piatetsky S., Padhraic S.,
2007). Others look at data mining in terms of a set of tools and techniques that operate on and
extract implicit patterns from data. Knowledge Discovery and Data Mining (KDD) is an
interdisciplinary area focusing upon methodologies for extracting useful knowledge from data
for Business Intelligence. The ongoing rapid growth of online data due to the Internet and the
widespread use of databases has created an immense need for KDD methodologies. The
challenge of extracting knowledge from data draws upon research in a wide variety of fields to
draw upon tools that can synthesize and organize knowledge on any given topic of interest from
a corpus of documents. There is an increasing realization that effective customer relationship
management can be done only based on a true understanding of the needs and preferences of the
customers. Under these conditions, data mining tools can help uncover the hidden knowledge
Research Methodology
147 Ph.D. Thesis
and understand a customer better, while a systematic knowledge management effort can channel
the knowledge into effective marketing strategies. This makes the study of knowledge extraction
and management particularly valuable for marketing and customer relationship management,
(Shaw, M.J., Subramaniam, C., Tan, G.W. and Welge, M.E., 2001).
Datamining tools and Techniques operate on large databases and extract patterns that are implicit
in them, resulting in actionable information. Datamining can be directed or undirected. Directed
datamining attempts to explain or categorize some particular target field such as income or
response. Undirected datamining attempts to find patterns or similarities among groups of
records without the use of a particular target field or collection of predefined classes.
Customer understanding is the core of CRM. It is the basis for maximizing customer lifetime
value, which in turn encompasses customer segmentation and actions to maximize customer
conversion, retention, loyalty and profitability. Proper customer understanding and actionability
lead to increased customer lifetime value. Incorrect customer understanding can lead to
hazardous actions. Similarly, unfocused actions, such as unbounded attempts to access or retain
all customers, can lead to decrease of customer lifetime value (law of diminishing return).
Hence, emphasis should be put on correct customer understanding and concerted actions derived
from it.
In view of the above, corporates commenced accumulation of wide spectrum of consumer data
viz. transaction data, customer databases based on consumer behaviour and purchase transactions
and in this context, creation of data warehouses. But, due to lack of appropriate tools and
techniques to analyze these huge databases, a wealth of customer information and buying
patterns was permanently hidden and unutilized in such databases. But, memory is of little use
without intelligence. The central idea of datamining is that data from the past contains
information that will be useful in the future. It works because consumer behaviours captured in
corporate data are not random, but reflect the differing consumer needs and preferences.
Knowledge-based marketing, which uses appropriate datamining tools and knowledge
management framework, addresses this need and helps leverage knowledge hidden in databases.
Customer profiling is one of the major areas of the application of data mining for knowledge-
Chapter 3
148 Ph.D. Thesis
based marketing. This is of relevance because consumer behavioural data is a more valuable
source of information than consumer demographics data.
3.15.8. Clustering
Cluster Analysis, also called data segmentation, relates to grouping or segmenting a collection of
objects (also called observations, individuals, cases, or data rows) into subsets or “clusters”, such
that those within each cluster are more closely related to one another than objects assigned to
different clusters. Cluster analysis uses a number of techniques of sorting individuals into similar
groups, (J.Saunders, 1980). Hence, objects in a cluster are similar to each other. They are also
dissimilar to objects outside the cluster, particularly objects in other clusters, (Ahuja V., Medury,
Y., 2011). Cluster analysis is also called classification analysis or numerical taxonomy.
Clustering algorithms function such that intracluster similarity is the maximum and the inter-
cluster similarity is minimum. Clustering also has applications in the field of marketing
segmentation. What distinguishes clustering from classification is that clustering does not rely on
predefined classes. In classification, each record is assigned a predefined class on the basis of a
model developed through training on preclassified examples.
The following statistics and concepts are associated with cluster analysis
Agglomeration schedule: It gives information on the objects or cases being combined at
each stage of a hierarchical clustering process.
Cluster centroid: The cluster centroid is the mean values of the variables for all the
cases or objects in a particular cluster.
Cluster centers: They are the initial starting points in non hierarchical clustering.
Clusters are built around these centers or seeds.
Cluster membership: Cluster membership indicates the cluster to which each object r
case belongs.
Dendrogram: A dendrogram or tree graph, is a graphical device for displaying clustering
results.
Distances between cluster centers: These distances indicate how separated the
individual pairs of clusters are Clusters that are widely separated are distinct, and
therefore desirable.
Research Methodology
149 Ph.D. Thesis
There are two major methods of clustering-hierarchical clustering and k-means clustering.
3.15.8.1. Hierarchical Clustering
This is a statistical method for finding relatively homogeneous clusters of cases based on
measured characteristics. It starts with each case in a separate cluster and then combines the
clusters sequentially, reducing the number of clusters at each step until only one cluster is
left. When there are N cases, this involves N-1 clustering steps, or fusions. This hierarchical
clustering process can be represented as a tree, or dendrogram, where each step in the clustering
process is illustrated by a join of the tree.
3.15.8.2. K-means Clustering
The clustering algorithm is initiated by creating k-different clusters and subsequently the
distance measurement between each of the sample, within a given cluster, to their respective
cluster centroid is calculated, (Berry, Michael, J. A., Linoff, Gordon S., 2007). We use Euclidean
distance measure for our study. After obtaining initial cluster centers, the procedure, (i) Assigns
cases to clusters based on distance from the cluster centers and (ii) Updates the locations of
cluster centers based on the mean values of cases in each cluster. These steps are repeated until
any reassignment of cases would make the clusters more internally variable or externally similar.
The initial cluster centers are the variable values of the K well-spaced observations. The final
cluster centers are computed as the mean for each variable within each final cluster. The final
cluster centers reflect the characteristics of the typical case for each cluster.
3.15.9. Consumer Profiling
Consumer Profiling is one of the major areas of the application of data mining for knowledge-
based marketing.Consumer profiling involves creating consumer models, based on which a
marketer can decide on the right strategies and tactics to meet the needs of the customer.
Profiling is an innate tool used for consumer behaviour and preference prediction. Consumer
profiles can be formed based on their purchase data or any other behavioural data.
3.16. Research Models-Summary of Data Analysis Tools and Procedures
Using the CRM concept of response modelling, four specific models have been developed during
the entire research study. These are-
Chapter 3
150 Ph.D. Thesis
1. Consumer Trustworthiness Regression Model using Netnography (CTR) - Under this
model the studies are conducted on a set of online communities of Apple and correlation
and regression model is applied for forty product categories. The results of a regression
model are used to analyze factors contributing to the growth of trust in an online
community and for finding out the contribution of the independent variable, that is,
number of points to the dependent variable, that is, number of views. The information is
subsequently applied for prediction by analyzing how far the dependent variable depends
on the independent variable. The same techniques are applied on the data for users on the
online community of Dell.
2. Co-creation Model using INV based on Metcalf Law (C-INV) - Under this model the
values of INVi (Table 4.27, Annexure III) were subjected to Hierarchical cluster
analysis using SPSS 17.0 and five consumer clusters were identified. Based on Table
4.27, Annexure III the profiles of the 200 consumers divided into five clusters were
identified and appropriate targeting strategies were formulated (Table 4.28, Section 4.3).
3. Consumer Price Sensitivity Model using k-means cluster analysis (CPS) - This model
uses data collected from the Apple Online Discussion Forums
(discussions.apple.com/category.jspa?categoryID=204).Apple has over 40 product
communities formulated for consumer engagement, interaction, feedback, building strong
online consumer trust and reciprocity. The consumer profiling was done on the data
collected using K-means clustering and cluster memberships were extracted and the most
significant consumers across all three product categories were identified. Further, the
consumer segmentation was performed and a detailed profile for most significant
consumer clusters for each price category were created (Table 4.40, Section 4.5).
4. Business Online Community Credibility Model (BOCC) using Linear Programming-
The model uses a numeric weighting technique to calculate a business online community
credibility score. The model further uses the linear programming technique to maximise
the BOCC score and further to determine an optimal solution for the variables
contributing to the BOCC score (Table 4.50, Section 4.6.6).
Recommended