See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/319472120
Netnographic Analysis: Understanding Culture through Social Media Data
Chapter · January 2014
CITATIONS
12
READS
4,060
1 author:
Some of the authors of this publication are also working on these related projects:
Food porn View project
Ethnography View project
Robert V. Kozinets
University of Southern California
86 PUBLICATIONS   10,281 CITATIONS   
SEE PROFILE
All content following this page was uploaded by Robert V. Kozinets on 04 September 2017.
The user has requested enhancement of the downloaded file.
2
Netnographic Analysis:
Understanding Culture through
Social Media Data
R o b e r t V. K o z i n e t s , P i e r r e - Ya n n D o l b e c , a n d
A m a n d a E a r l e y
In the past two decades, participation in
online conversations has grown from a relatively
marginal activity of hackers, geeks,
and early cyberculture members to a mainstream
activity recognized and supported by
mainstream businesses and media. Starting
from tiny numbers of enthusiasts, over a
billion people now use social media to
communicate, create, and share information,
opinions, and insights. Online social spaces
have become increasingly recognized as
important fields for qualitative social scientific
investigation because of the richness and
openness of its multifarious cultural sites. At
the same time, online data present unique
challenges for researchers, as it is voluminous,
optionally anonymous, and often difficult
to categorize. This chapter introduces
readers to netnography, a technique for the
cultural analysis of social media and online
community data. The purpose of this chapter
is to discuss the distinctive cultural features
of online, or social media, qualitative data
and to overview, develop, and illustrate techniques
for their rigorous analysis as they
have been developed through the research
approach of netnography.
OVERVIEW
Defining Netnography
Netnography is an established approach to
qualitative research, whose name draws
together the terms “Internet” and “ethnography”
(see Kozinets, 2010). Netnography
shares many of the characteristics of ethnography
(see Gubrium and Holstein, Chapter 3,
this volume), from which it has been adapted,
in that it is a flexible approach that allows
scholars to explore and explain rich, diverse,
cultural worlds. Naturalistic in orientation, it
approaches cultural phenomena in their local
contexts, providing windows on naturally
occurring behaviors. The output of a netnography
can be descriptive as well as analytical
and the method tends to generate rich, thick
description through grounded interpretations
(see Willig, Chapter 10, this volume), thereby
18-Flick_Ch-18.indd 262 7/30/2013 8:21:31 PM
Cite as: Kozinets, Robert V., Pierre-Yann Dolbec, and
Amanda Earley (2014), Netnographic Analysis:
Understanding Culture through Social Media Data,” in Uwe
Flick, ed. Sage Handbook of Qualitative Data Analysis, Sage:
London, 262-275.
NETNOGRAPHIC ANALYSIS: UNDERSTANDING CULTURE THROUGH SOCIAL MEDIA DATA 263
providing a detailed representation of the
lived online experience of cultural members.
Netnography also emphasizes the role of
the “researcher-as-instrument” (Guba and
Lincoln, 1981), and the immersion of the
researcher into the computer-mediated context
of study.
Netnography is also different from ethnography
in many respects and thus requires a new
set of skills (Kozinets, 2010). These differences
emerge from the distinct nature of
computer-mediated communications. The specificities
of a rigorous and disciplined approach
to netnography are organized around the entrée
and data collection procedures, choices about
field sites, decisions about the types of data to
gather and analyze (see Marotzki et al., Chapter
31, this volume), the evaluation of the quality
of netnographic research, and ethics, which we
will review in the following paragraphs
(Kozinets, 2002).
As netnography is a naturalistic method, its
interpretations can be built from a combination
of elicited and, more often, non-elicited data.
These data emerge and are captured through
the researcher’s observation of and participation
with people as they socialize online in
regular environments and activities (Kozinets,
2010). Online cultural research is far less intrusive
than traditional ethnography, as the online
researchers can gather a vast amount of data
without making their presence visible to
culture members (Beaulieu, 2004; Kozinets,
2010). There are occasions where such nonparticipative
activity (aka “lurking”) is appropriate;
however, it tends not to be not appropriate
when the researcher is interested in the experience
of participation in an online field site.
The field of social media offers several
advantages rarely found in the traditional field.
First, researchers generally have access to vast
amounts of data, archived through forums and
search engines. These resources can provide an
unprecedented amount of information on cultural
members, values, and structures, allowing
researchers to better choose their field sites
and plan their entrée (Kozinets, 2010). Second,
ethnographers that employ interview methods
(see Roulston, Chapter 20, this volume) can
transcend geographic and time limitations by
using asynchronous communication technologies
such as email to conduct interviews.
Third, netnography can leverage the connective
power of the Internet and the search and
organizing capabilities of contemporary search
engines to offer accessibility and openness to a
vast variety of virtual voices (see Marotzki
et al., Chapter 31, this volume).
Characteristics of Netnographic
Field Data
The unique characteristics that distinguish
netnographic data from face-to-face cultural
data necessitate the ongoing development of
new adaptations of ethnographic procedures.
Some of the features we will briefly discuss in
this chapter are: (1) increased field site accessibility;
(2) increased communicative variety;
(3) communication connectedness across multiple
forms/fields; and (4) auto-archiving.
First, online fields offer dramatically
increased field site accessibility. The explosion
in online social worlds offers participants a
virtually (although not absolutely or uniformly)
borderless environment where geographically
dispersed members can meet and
communicate. Because these social fields are
accessible to anybody with an Internet connection,
it has been argued that the position of
“ethnographer as sole and privileged witness
[as found in ethnographies] may be more difficult
to uphold as a subject position and authorial
voice” in Internet-enabled social research
(Beaulieu, 2004). On the Web and elsewhere,
however, the difference between mere social
observer and social scientist is clear. As
Kozinets (2010: 113) elaborates, the difference
is analogous to that between journalist and
ethnographer – researchers adhere to respected,
legitimate, and rigorous methodological proscriptions
of their field, and in the development
of research answers and theories, as data
are analyzed and interpretive insights are
structured and incorporated “into a known and
respected body of codified knowledge.”
The communicative variety of netnographic
data refers to the many ways that researchers
18-Flick_Ch-18.indd 263 7/30/2013 8:21:31 PM
264
and community members relate across multiple
online and mobile platforms, as well as
through face-to-face interactions. Where once
there were face-to-face meetings and conversations,
supplemented by letters and phone
calls, researchers now must also consider
blogs, Twitter accounts, Facebook postings,
Linkedin groups and meetings, and many other
forms of social media meeting and communication.
Moreover, the online interactions are
themselves complex: they can happen both
privately and publicly, both asynchronously
and synchronously, over different time peri-
ods,andwithnumerouscontributors(Ruhleder,
2000), as well as from a number of different
site sources (e.g., corporate-owned forum or
grassroots blog pages) and in different formats
(e.g., textual, visual, audio).
Furthermore, there is added complexity in
that these forms of communication are now
often linked one to the other. Conversations are
happening on multiple sites, between multiple
community members, and recent technological
advances make it possible to easily post and
share content on multiple platforms. An entry
might be cross-posted on a blog and a microblog,
tagged in a geo-localized platform,
shared on personal and professional social
networks and social bookmarking services,
talked about on independent and commercialowned
web forums, then re-blogged and ridiculed
in an online video before ending up on
national news. This marks a sharp contrast to
both ethnographic work and earlier netnographic
studies (e.g., Correll, 1995).
In order for a netnographer to understand
the social circumstances of this activity, studying
one manifestation of a particular message
can often be insufficient. If a researcher
attempted to mine, scrape, or download a
descriptive posting on a web forum and
thought that this posting “told the whole
story,” that researcher would be sadly mistaken.
In order to gain a contextualized and
nuanced cultural understanding of the social
media phenomenon, the netnographer must be
attuned to this multiform communicative connectedness,
and be willing to follow multiple
links to multiple sites and postings in order to
gain a fuller and embedded understanding of
the overlapping social worlds enacted through
social media.
Finally, there is a constant and automatic
saving, sorting, classifying, and archiving of
all types of asynchronous – and much
synchronous – media (e.g., a feed from the
micro-bloggingserviceTwitter)ontheInternet,
similar to having access to recordings of every
public discourse, interaction, and social contact
in a given real-world community
(Kozinets, 2010).Archiving by search engines
and specialized sites such as The Wayback
Machine further this process. Again, this
makes the Internet a very different site for
ethnographic research compared to “real”life
social interactions, and requires levels of
procedural adaptation.
Challenges of Netnographic
Fieldwork
In addition to these technical considerations,
the structural characteristics of netnographic
data present a number of interesting theoretical
challenges. The first is ontological and
concerns the somewhat false distinction
between online and offline social “worlds.”
Because social worlds cut across complex
networks of face-to-face and technologically
mediated communications, the use of netnography,
as with the use of any single method or
focus, offers an incomplete view. Whereas
before, community members would discuss
face to face and over the phone, they are now
supplementing their exchanges with online
conversations on web forums and through
emails (Wellman et al., 2001). Information
search that happened offline is now radically
altered and amplified by the Internet
(Kayahara and Wellman, 2007). For example,
the experience of preparing for childbirth
might involve in-person conversations, doctor’s
visits, a baby shower, childbirthing
classes, and other in-person activities.
However, if netnography offers only a partial
view of many online–offline phenomena, the
reverse is also true. That is, in the current
environment – and increasingly in a rapidly
18-Flick_Ch-18.indd 264 7/30/2013 8:21:31 PM
NETNOGRAPHIC ANALYSIS: UNDERSTANDING CULTURE THROUGH SOCIAL MEDIA DATA 265
computerized and mobile Internet world –
many social activities cut across both online
and offline worlds (Garcia et al., 2009;
Kozinets, 2010; Miller and Slater, 2000). For
example, adjusting to a new college might
involve sharing information via email, reading
student and university blogs, getting information
from websites, participating in online communities,
and joining conversations on social
networks (DeAndrea et al., 2012). Because the
Internet is part of our everyday life, ethnographies
of aspects of contemporary society should
carefully consider the importance of studying
related online behaviors and the social worlds
of social media (Garcia et al., 2009; Miller and
Slater, 2000). “Pure” netnographies, that is,
netnography without an offline component,
should be reserved for phenomena which are
happening strictly in the online world, such as
self-presentation on personal websites (Schau
and Gilly, 2003) or online word of mouth
(Kozinets et al., 2010).
Netnographic data also raise a number of
epistemological and pragmatic questions.
Because the Internet provides access to so
much data, so easily, netnographers face the
inherent challenge of data overload.
Identification and classification of data can be
challenging as the optional anonymity that the
Web provides can lead to an absence of demographic
markers. This seemingly convenient,
easy, and anonymous datastream can also open
a Pandora’s box of ethical issues related to
privacy, consent, and appropriate representation.
Finally, although qualitative methods
like netnography are contextually embedded,
researchers often are drawn or directed to collect
corroborating evidence in order to generalize
findings or make more them more
transferable to diverse contexts (Kozinets,
2002). The next section begins the discussion
of netnographic procedures developed to
address these important challenges.
NETNOGRAPHIC PROCEDURES
Netnography is a relatively new method,
and analysis techniques thus far have been
developed in relation to analogous procedures
in ethnography. Here, we provide
explicit guidelines for adapting face-to-face
data collection and analysis techniques to the
new contingencies of computer-mediated
cultural communications. We will work from
five essential ethnographic considerations:
(1) preparing for data collections and cultural
entrée; (2) collecting and creating the data;
(3) performing ethical research; (4) conducting
an insightful and trustworthy analysis;
and (5) representing the data analysis in a
meaningful and appropriate manner. The section
that follows will provide an illustration
of how these procedures can be applied to
actual social media data.
Netnographic Sites and Entrée
There are many different sites of culture on the
Internet, and each one can be explored using
the netnographic approach. Kozinets’list of the
main types of netnographic field sites includes
bulletin boards, chat rooms, playspaces (where
videogame and other game play occurs), virtual
worlds, lists and web-rings (a largely defunct
form now replaced by blogrolls) (Kozinets,
2002), and blogs, wikis, audiovisual sites,
social content aggregators, and social networking
sites (Kozinets, 2010). It is crucial
for researchers to be attuned to the format of
different online field sites as it influences the
“types, forms and structures of online communication”
found within it (Kozinets, 2010:
87), and the possible roles that can be assumed
by users (e.g., being the reader of a blog, a follower
of a Twitter poster, a member of a forum,
or a friend on a social networking site).
Research questions addressed by netnographic
studies may relate to a phenomenon
that exists in both offline and online worlds;
that can only be experienced virtually; or that
concerns the very nature and structure of webmediated
communications. When choosing a
field site to study, researchers should favor
communities that (1) are more “research question
relevant”; (2) have a “higher traffic of
postings”; (3) have larger numbers of discrete
message posters; (4) have more detailed or
18-Flick_Ch-18.indd 265 7/30/2013 8:21:31 PM
266
descriptively rich data; and (5) have more
between-member interactions of the type
required by the research question (Kozinets,
2002: 63). We suggest identifying research
topics, or a set of research questions that will
help in pinpointing relevant online sites; studying
the sites and their participants to understand
the social dynamics at play; and finding
out if the community has been “tapped out”
(Kozinets, 2010: 79) by other researchers or
“turned off” by inconsiderate researchers in the
past. Along the way, never assume you know
more than the community about its own culture.
Most importantly, pay attention to the
kinds of social data that are available (textual,
visual, audiovisual, graphic, and so on) and get
prepared to collect and organize them.
Making a successful cultural entrée into
these field sites and with their members
requires understanding the data while collecting
them, and, even more importantly, understanding
and being sensitive to the needs and
functioning of the social media community.
Because netnography is attuned to the cultural
realities of living, breathing communities of
communicating people, netnographic research
requires an initial and deepening cultural
understanding of the community. As in any
social settings, online gathering “places” have
particular histories, social structures, codes of
etiquettes, particular ways of speaking, and
unique rituals and identities. An entrée can
make or break the interactions that will follow
with a community; a researcher can be rejected
if the researcher does not understand the customs
of the community he or she is talking to
(e.g., Kozinets, 2010: 77).
Given the abundance of “intelligence” data
available on the Web, researchers should conduct
background research to ease the entrée.
One example is “lurking” – a form of online
reconnaissance – to gain information on the
community before making the entrée, an opinion
shared by other researchers (e.g., Shoham,
2004). Lurking should only be the first step of
a netnographic study, though: Beaulieu (2004)
maintains that “lurking” netnographers can
miss parts of the phenomenon that are not publicly
visible and Kozinets (2010) emphasizes
the importance of engagement and participation
in cultural worlds.
Once background research is complete and
the various types of data and their approximate
contents are identified and accommodated,
active participation can move to the forefront.
Participation in a netnography allows researchers
to experience what it feels like to be a community
member. Here it may be helpful to think of
Walstrom’s (2004) term “participant experience,”
offered as a counterpoint to the traditional
but misleading ethnographic term “participant
observer.”
From the perspective of data collection, it is
important for this participation to be captured
in field notes, screen captures, recordings, and
other permanent records. When performed
well, ethnographic participation online enables
enhanced cultural understanding, the confirmation
of interpretation, and new opportunities to
recast the research enterprise as collaboration
rather than appropriation. Ideally, the
researcher’s participation should provide some
benefit to the community, for example, by
posting something that is useful and/or thought
provoking and sharing it with a comment.
Alternatively, a researcher could begin as a
participant in a particular site and become a
culture member before starting to study this
community. As the introspective approach
of “auto-ethnography” (Hayano, 1979)
morphs into “auto-netnography” (Kozinets
and Kedzior, 2009), pre-existing cultural membership
can smooth entrée into an unnoticeable
and much more personally complicated introspective
event (e.g., Cherny, 1999).
UNDERSTANDING NETNOGRAPHIC
DATA AND DATA COLLECTION
Three general types of data are available for
collection in netnography: archival data, elicited
data, and field note data. Archival data
comprise anything the researcher can gather
from the Web that is not a product of his or
her involvement to create or prompt the creation
of data. These types of data can constitute
a “cultural baseline,” serving as a portrait
18-Flick_Ch-18.indd 266 7/30/2013 8:21:31 PM
NETNOGRAPHIC ANALYSIS: UNDERSTANDING CULTURE THROUGH SOCIAL MEDIA DATA 267
of what the community was doing before the
researcher made his or her incursion into that
social media environment. From this point,
the researcher seeks to deepen his or her
knowledge of the cultural context (see
Kozinets, 2010: 104). These types of so-called
observational data can be difficult to find,
depending upon the specificity of the topic,
but once found are relatively straightforward
to access, and can be gathered at a very low
cost. Importantly, text (whether in text, image,
or video format) also has context. Beyond the
analysis of the text, researchers can also look
at the way that a webpage and postings are
formatted as a type of conversation between
various users. They might analyze the pictures
that people use to represent themselves,
the way that they describe themselves, the
various signs and signals that are used, and
the interactions between those.
Numerous ways exist to record online data.
The two most fundamental techniques are:
(1) to copy and paste the content of a forum
post, for example, into word processing software
files such as a Microsoft Word document; and
(2) to capture a screenshot of data using a program
such as Windows 7’s snipping tool or the
Apple Grab utility. It is also possible to record
in real time a researcher’s visit to a particular
website through programs such as Camtasia. It
may be tempting for researchers to look into
programs such as HTTrack Website Copier
that allow the researcher to archive the full
content of websites and forums, which seemingly
provides an interesting alternative to
copy–pasting text content, and/or screengrabbing
pages. However, the temptation to “mine”
large amounts of data can overshadow realtime
engagement with the cultural context.
Although greatly facilitating data collection,
automatic methods can thus can create a barrier
to understanding. Such tools must be
understood deeply and used judiciously in
coordination with actual, real-time, engaged,
confusing, all-too-human participation in the
social media community; its conversations; its
people; and its temporal unfolding. Archival
data can be present in the form of text, such as
messages exchanged on a forum; in visual
forms, such as the layout and logos of websites,
pictures of members of a social networking
sites, or images of avatars in online worlds;
by way of audio, such as songs exchanged on
forums; and through video, such as webcam
conversations and user-generated videos.
Elicited data refer to content that is co-created
by the researcher and members of the social
media community through processes of social
interaction. This includes the products of
online interviewing, whether by asynchronous
modes of communication (e.g., email, forums)
or (quasi-)synchronous ones such as chat and
video calling. Early literature on eliciting data
through online channels raised concerns about
how the medium itself can shape the interaction.
In the now-classic work from the Dark
Ages of netnographic inquiry, Markham (1998:
62–75) notes that textual online interviews
(e.g., in chat rooms) limit the information that
can be gathered from non-verbal cues.
Moreover, as typing is usually significantly
slower and more deliberate than text-based
interviews, it could limit participant spontaneity
(Catterall and Maclaran, 2002).At the same
time, netnographers can use these structural
characteristics to their advantage. Like conversations
in the traditional, face-to-face field, the
synchronous, real-time format of online chatting
can provide the researcher with quick
insight into a cultural phenomenon and can
even evolve into an informal interview. The
deliberate nature of email exchange lends itself
well to the goals of formal interviews, as it can
enhance reflexivity, foster a sense of intimacy,
and, if time is taken, deepen rapport (Kozinets,
2010; see McCracken, 1988, for methodological
recommendations regarding offline interview
methods). A high-bandwidth audiovisual
interview can be offered, if the interviewer and
interviewee possess a strong Internet connection,
to simulate a face-to-face interview. As it
restores access to participant body language,
video chat provides a significant advantage
over the well-established technique of telephone
interviewing.
The third and final type of data associated
with a netnographic research project are field
note data, which are generated directly by the
18-Flick_Ch-18.indd 267 7/30/2013 8:21:31 PM
AQ1
268
researcher for the purpose of research recording,
reflection, and analysis. These data are not
shared with the social media community,
although they may contain captures of data
such as texts, screenshots, moving images, and
so on. In order to analyze the important transitions
involved in cultural entrée and acculturation,
netnographic researchers should begin
recording reflective field notes as early in the
research process as possible, ideally as soon as
a project is initiated. These notes should continue
through the search for a specific site,
adding notes when experiencing the community,
contacting and interacting with culture
members, or even simply when thinking and
reflecting about the research project. Field
notes should document the journey of the netnographer
from an outsider to an inside cultural
member. Although netnographers have
many opportunities to automatically capture
online images and actions, these actions are by
no means a substitute for internal reflection,
the capture of in-the-moment impressions and
experiences, and the deep culture-bound introspective
analysis that marks all strong anthropological
and sociological ethnography.
With these three types of data in mind, along
with the diversity of communication media
they may involve, it is important to remember
that netnographic data collection is not simply
one thing, but now offers a range and continuum
of different offerings with different benefits,
drawbacks, and concomitant tradeoffs.
This holds for almost every element of conducting
netnographic research. The online
world has become enormously complex and
variegated, and netnographic research
approaches have followed suit.
ETHICS IN NETNOGRAPHY
Research ethics (see Mertens, Chapter 35,
this volume) are a complex and difficult
topic in netnographic exploration. As with all
social inquiry, ethical netnographic practice
is grounded in the principle of informed consent,
and consideration of potential benefits
and risks to individuals and communities.
Since social media blend the public and private
into a novel hybrid form, netnography
demands new thinking and methodological
innovation on issues of risk and privacy.
The online representation of identity is one
factor that affects how we think about consent
and the potential for harm, especially for elicited
data. Some researchers approach the issue
by requesting legal names and signed consent
forms, or by combining online and offline
methods.Although knowing a user’s “real-life”
name and having access to more information
about offline lives may enhance perceptions of
netnography’s credibility, this bears with it the
same responsibility for protecting identities
faced by ethnographers, interviewers, and all
other social researchers who obtain data from
human subjects.
For archived data, the researcher may
approach web content as published content.
Although information posted publicly on the
Web is technically published, and subject to
criticism and quotation with citation, netnographers
should consider the ethical issues inherent
in quoting directly from online sources
(Bruckman, 2006; Kozinets, 2002). Although
a web user is responsible for the consequences
of publicly posting information on the Internet
(either with an original name or a pseudonym),
re-publication or citation in an academic publication
may have unexpected consequences
for the individual and/or the community. For
example, it may be unsettling or injurious to
the individual if the quote appears with the
researcher’s critical reading every time the
individual’s online alias is entered into a major
search engine.
As such, when harm is likely, netnographers
follow the ethnographic tradition of pseudonymization,
protecting individuals’ names –
both legal and assumed. Many web users have
valued pseudonymous identities that they have
invested quite a deal of effort into creating and
protecting, and netnographers should always
be careful to treat these identities as if they
were legal identities, creating further pseudonyms
when human subjects need protection.
Nevertheless, using real names may be
appropriate in some cases. For example, an
18-Flick_Ch-18.indd 268 7/30/2013 8:21:31 PM
NETNOGRAPHIC ANALYSIS: UNDERSTANDING CULTURE THROUGH SOCIAL MEDIA DATA 269
art historian using netnography may need to
cite the blogs of famous art critics or artists.
From a legal perspective, three legitimate
privacy concerns may arise in the course of
such netnographic study. One is the use of
online information found in “semi-private”
web spaces. Are all members of a particular
online community bound by a Terms of Use
agreement that defines the forum as a de facto
private space? Do users have a reasonable
expectation of privacy? Does the group stipulate
other restrictions on how the content may
be used? Generally, groups that anyone can
join are still considered to be public, but the
netnographer should be careful with any Terms
of Use agreements that are “signed” when
signing up. The second privacy issue goes
hand in hand with earlier discussions of anonymity,
pseudonymity, and confidentiality.
Whether or not a netnographer uses a poster’s
real name or “real pseudonym” remains an
ethical question, not a legal one, as the information
is publicly published under that name.
In addition to privacy laws, copyright laws
may come into play when a researcher wants
to publish a picture, quote, or other material.
From an international level, this issue is so
complex and dynamic that questions are best
handled by experts (e.g., a university’s lawyers
advising about local copyright restrictions and
fair use guidelines) on a case-by-case basis.
NETNOGRAPHY IN ACTION: A BRIEF
ILLUSTRATION
In principle, the analysis of qualitative data
yielded by a netnography will proceed in a
fashion very similar to that of any other
comparable types of qualitative data. As we
mentioned earlier, netnography shares the
inductive and iterative aspects of ethnography.
Because of the nature of social media
and online communications, netnography
places the field site and research participants
within easy access of the researcher. Thus the
collection of data and their analysis are even
more likely than they are with other methods
to become blurred into a single ongoing
process. Add to this the ubiquity and variety
of search analytic engines and the convenience
of qualitative data management software
and coding programs (see Gibbs,
Chapter 19, this volume), and we have the
opportunity to automate many elements of
the netnographic data collection and analysis.
Although ethnography requires the
researcher to become a finely tuned instrument
(see Gubrium and Holstein, Chapter 3,
this volume), the use of such software risks
reducing the researcher to a mere buttonpusher
who finds the preprogrammed output
of various web-mining and content analysis
programs sufficient as either the outcome or
primary basis of an interpretation. To avoid
this, we advocate an “old-fashioned,” handson
approach to netnographic data analysis.
Although netnography uses some of the most
contemporary data available, and benefits
from the most recent technological developments,
this approach to data analysis grounds
the researchers in the basics of inductive (see
Reichertz, Chapter 9, this volume) and reflexive
analysis (see May and Perry, Chapter 8,
this volume).
We have four important prescriptions to
guide researchers in conducting netnographic
data analysis. The first core principle of quality
netnography is that of “ethnographic siting”
(see Gubrium and Holstein, Chapter 3, this
volume). This means that netnographers
should go site specific, initially at least, concentrating
on a small number of postings or a
very constrained data set in order to gain a
deep cultural sense of “what is going on” in
that particular social space. From there, the
analysis can broaden in scope as well as
deepen. The second tenet of rigorous netnographic
analysis is to undertaken cultural
analysis (see Winter, Chapter 17, this volume)
while engaged as a participant in a manner
appropriate to other cultural participants or
community members. This is the rule of “ethnographic
engaging.” The third maxim is that
of “ethnographic communicating,” where
communications are, at least initially, experienced,
processed, and understood exactly as
cultural members experience them. This means
18-Flick_Ch-18.indd 269 7/30/2013 8:21:31 PM
270
that they should be analyzed and viewed in
their natural “real-text” format rather than
experienced through the filters of some language
processor or compressive software
engine. The final recommended netnographic
convention is to allow “ethnographic timing”
to unfold, such that messages and posts are
experienced, read, interpreted, and analyzed in
real time, as they become available, rather than
all at once.
As with ethnography, the purpose of netnographic
data analysis is to organize the collected
products of participation and observation
into a rigorous, meaningful, and useful form of
research output, such as an article, a report, a
presentation, or even a book. The data will
likely include various downloaded textual,
graphical, photographic, audio, and audiovisual
files, screen captures, online interview transcripts,
and reflective field notes. In this section,
we present coding and hermeneutic
approaches to analyzing these materials. As
researchers who frequently work with textual
data, we find a combination of these two
approaches to be optimal, but other methods
may be more appropriate for other types of
netnographic data. For example, netnographers
working in image or video-heavy field sites
may want to reference the chapters of this
handbook that explicate semiotic and visual
studies approaches (see Banks, Chapter 27, and
Knoblauch et al., Chapter 30, this volume).
Coding is a qualitative data analysis method
commonly used by sociological researchers.
Here, “open coding” is arguably the most
popular approach (Corbin and Strauss, 2008;
see Thornberg and Charmaz, Chapter 11, this
volume). Open coding begins when the
researcher labels and categorizes data by
“emic,” field-level meanings, and then groups
these categories into other abstract categories.
The ultimate goal of open coding is to reach a
theoretically relevant understanding of the
phenomena of interest. Netnographers can
organize many different types of data with the
same codes. For example, a photograph, a
scanned drawing, a blog entry, the format of a
webpage, a few seconds of a YouTube video,
the color of text in a posting, and the name of
someone’s avatar could possibly all be coded
with the same term.
The “grounded theory” that emerges from
coding is tested as new data are collected and
analyzed; indeed, data may be collected specifically
for that purpose. Because netnography
allows such easy access to data, and
often offers up large amounts of data, such
cross-checking is facilitated in a rather
unprecedented manner. Comparisons (see
Palmberger and Gingrich, Chapter 7, this
volume) look for convergence and divergence
among the coded data and the categories.
Generalizations try to explain the occurrences
of the data and are used to construct new
theories.
Simultaneous with such inevitably micro but
effectively piecemeal approaches are more
macro and holistic approaches that seek a kind
of transcendent interpretation. As consumer
researcher Susan Spiggle (1994: 497) describes
them, these interpretations are like a type of
decoding that “occurs as a gestalt shift and
represents a synthetic, holistic, and illuminating
grasp of meaning.”Another way to conceptualize
this form of insightful reading of texts is the
notion of hermeneutics, where larger order
conceptual readings are garnered from readings
of the parts of the text in light of the text as a
whole. In a netnography, this “text” can encompass
particular sites; particular forms of social
media; the Web or Internet itself; interview data
about online interaction; the researcher’s ongoing
experience with online media; and a multitude
of different types of communications such
as email, blog posts, comments, ratings, photographs,
and videos.
In netnography, coding (see Thornberg and
Charmaz, Chapter 11, this volume) and hermeneutic
interpretation (see Wernet, Chapter 16,
this volume) may overlap in a variety of interesting
ways. Rarely do hermeneutic insights
simply burst into being like a light illuminating
a dark room. Because netnographers must
approach field sites as participants as well as
observers, their interpretations of online communications
and communities will emerge
gradually, as they build up the cultural codes to
make sense of virtual social spaces.
18-Flick_Ch-18.indd 270 7/30/2013 8:21:31 PM
NETNOGRAPHIC ANALYSIS: UNDERSTANDING CULTURE THROUGH SOCIAL MEDIA DATA 271
In order to demonstrate how coding and
hermeneutic analysis can be utilized in
netnographic data analysis, we offer a
short example that involves both coding
and hermeneutic interpretation. The 64-word
sample text was drawn from a discussion
about netnography on a LinkedIn group
dedicated to netnography. It was located
through a search of the term “netnography”
on a public, free (but ad-supported) semantic
search engine named “Social Mention”
(which can be found at socialmention.
com). The search engine looks for mentions
of the word netnography across a range of
domains on the Internet, including blogs,
micro-blogs, bookmarks, comments, news
stories, video and other forums and formats.
The software attempts to recognize
positive, neutral, and negative sentiments
in the mention of netnography, and it also
looks at the co-occurrence of the term and
other terms. As Figure 18.1 shows, Social
Mention classifies the term as having very
little strength, but positive sentiment and
reasonable passion.
Social Mention and other semantic search
engines are very useful to netnographers
because they can help to identify where to find
particular mentions of topics or terms. In this
case, the Social Mention output led us to a
Twitter tweet, which had an embedded link
that directed us to a discussion thread that was
useful to use as an example of netnographic
data analysis. The post, written by a consultant
and trainer from the UK, asked the following
question:
Does anybody do small scale, thick description
Netnography using naturally formed communities
on the net? I’m seeing a lot of stuff about
Netnography that is getting hard to distinguish
from social media listening. I want to know if
people do research as observers / participant
observers in naturally formed communities (not
supported by brands) and if so, how you deal with
the ethical issues?
Figure 18.1
Source: socialmention.com website search using search term “netnography”, conducted April 2013
18-Flick_Ch-18.indd 271 7/30/2013 8:21:31 PM
272
Because of the space constraints of this
chapter, we are prevented from entering into
extreme detail; however, we can code a range
of different topics from this posting. They
include: research terms (small scale, thick
description); method concerns (observers /
participant observers, ethical issues); and distinctions
for research and researchers (hard to
distinguish, naturally formed). From these we
can proceed to some higher order categories
or theories, noting that some researchers are
seeking to distinguish netnographies from
other techniques such as “social media listening”
and that the terms “naturally formed
communities” may also be powerful and
prevalent in setting up this distinction.
From these coding and higher order categorizations,
our hermeneutic interpretation
can build out a more general theory that
attends to commercialization in the sphere of
social media. It might see this posting as an
authenticating strategy that seeks to establish
the boundaries of authenticity and to link
them in a notional space with morality, ethics,
and particular research practices. The poster
views organically developed communities as
different in some sense from those which are
cultivated by companies or brands, and simultaneously
relates the use of “small scale, thick
description” ethnographic techniques to the
study of these communities as a way to distinguish
a particular kind of research and
researcher as more authentic than others.
In this way, coding and hermeneutics are
complementary techniques that can help to
illumine various aspects of even short postings
such as this one. Of course, such interpretations
often require an insider’s depth of knowledge,
linguistic terms, and understanding, and
such a deep understanding can only take place
through prolonged engagement and immersion
in a culture over a prolonged period of time.
There are a vast variety of different
approaches we can take to analyze netnographic
data. Another method that can be used
to glean insights for analysis is to use word
clouds. In Figure 18.2, we went to the
Wikipedia entry on “netnography,” which is
Figure 18. 2
Source: output from wordle.com input of wikipedia.com entry on “netnography”, conducted
April 2013
18-Flick_Ch-18.indd 272 7/30/2013 8:21:32 PM
NETNOGRAPHIC ANALYSIS: UNDERSTANDING CULTURE THROUGH SOCIAL MEDIA DATA 273
the top listing on the Google search engine
when we searched for that term. We selected
the entire article, and then pasted it into the
free word cloud generator “Wordle” (available
at wordle.com). The same method could of
course be used to generate word clouds for
entire postings, threads, or much larger sets of
words. We can then proceed to analyze the
output. The first author’s name shows up as the
largest word in the cloud, which perhaps suggests
how closely the method is still associated
with its pioneer and chief advocate (see also
Kozinets, 2012). Other words that are large
and significant (besides netnography) are communities,
social, research, and consumer. The
terms are still largely descriptive, which is to
be expected in an encyclopedia entry such as
Wikipedia, although the inclusion of the term
“communities” underscores the importance of
the role of online community studies, and the
use of consumer also indicates that the term is
still largely associated with business schools
and consumer research. It may be interesting to
note some important words which do not
appear prominently in the word cloud, such as
marketing, business, method, digital, and
anthropology. Obviously, these analyses are
brief, rather superficial, and illustrative only.
They invite considerably deeper exploration
than our space constraint allows.
MOVING FORWARD
We conclude this chapter with a brief discussion
of some of the challenges, strengths, and
new directions that accompany the analysis
of netnographic data. Many challenges have
previously been identified in past literature
on the topic, such as the anonymity of data;
the difficulty of maintaining research ethics
in a space that is both public and private; the
nature of cyberculture; and its technological
mediation of social communications (see,
e.g., Kozinets, 2002; 2010; Langer and
Beckman, 2005). We focus our closing
comments on several overlapping areas:
anonymity, legal issues, data overload, and
the emergence of new social media forms.
Public perceptions about anonymity on the
Internet have changed dramatically over the
past decade or so, and with them netnography
must also change. The Internet is now mainstream.
Documentation mechanisms have
emerged in many web forums, with the goal of
verifying relevant parts of a person’s identity,
character, or behavior. Perhaps most importantly,
increasingly, online interactions are no
longer anonymous. Social networks like
Facebook and LinkedIn strongly and successfully
encourage people to use their real or legal
names, often going so far as demanding it
within Terms of Use agreements. These identities
are further validated as users “friend” others
they know in real life, and activity in the
physical world becomes fodder for conversation
in the world of social networking. The
integration of those websites with newspaper
and magazine sites connects commentary to
one’s social network identity, furthering this
trend. Portable technologies and services like
Twitter allow advanced documentation of
what is going on in “real life.” There is a clear
tension between the forces favoring anonymous
interactions and the structures demanding
or encouraging its decline. In short, the
once more demarcated lines between offline
and online identities are disappearing.
However, the netnographer now must pay
much more close attention to the limitations he
or she has to conduct research on social networks
such as Facebook that are governed by
restrictive terms of use, and where social
media community data are, in some sense,
“owned” by a corporation and bought, sold,
and utilized mainly for its own private benefit.
Since social media present an overt risk of
data overload, netnographic procedures suggest
that converging concentration and centric
sampling and analysis be used to cope with
copious amounts of data. This is the netnographic
edge, in which the “researcher-asinstrument”
demonstrates what a trained
anthropologist can do that a sophisticated datamining
program cannot. To cope with overwhelming
quantities of data, netnographers
must clearly establish the boundaries of the
phenomenon being studied. These boundaries
18-Flick_Ch-18.indd 273 7/30/2013 8:21:32 PM
274
should link the analysis, in an iterative and
inductive manner, of the data collected with
the research focus. As in any qualitative
research project, the data collection should be
rigorous and follow established methodological
guidelines. Findings should be grounded
and emerge from the data. Using a small data
set does not equate to purposively sampling
data to confirm a researcher’s feeling of what
is going on in the field site. If the researcher
plans to analyze the data set by hand, Kozinets
(2010) recommends limiting the amount of
data to no more than 1000 double-spaced
pages of text. If the researcher is using qualitative
data analysis software, this amount can be
increased to 5000 pages, but the analyst should
proceed with caution and not lose sight of the
social site’s forest for the QDA’s categorical
trees (QDA: Qualitative Data Analysis – see
Gibbs, Chapter 19, this volume).
Our advice is to think carefully about what
kinds of data and how much data are necessary
for a given project. Consider whether the project
would be better served with more depth,
analyzing the comments and any linked content
for a certain time period. It may also be instructive
to think about the level of analysis and the
boundaries of the community. Are you interested
in comparing organizations or websites,
or is your community defined by a particular
group of users? How much data, and of what
kind, are required to learn about the community
of interest? Clear definitions of the community
and strategies for collecting the necessary data
should be established at the outset and reviewed
regularly throughout the project, to ensure that
the data are appropriate and not overwhelming
(Kozinets, 2010). Throughout, be sure to start
with the most promising areas of online content,
in the event that even these amounts of
data are overwhelming.
However, if you decide that your research
project requires you to sample widely or
broadly, and that large amounts of social media
data are necessary, there are many useful tools
to help address some of the organizational
issuesassociatedwithdataoverload.Depending
upon the project, software for automatically
downloading web content may be helpful. For
others, however, this may produce data that
require a considerable quantity of manual
cleaning. Regardless of the archiving method,
keep in mind that online data often contain
large quantities of noise, spam, or otherwise
unusable material. Website searches often yield
documents that are not relevant. Online community
discussion boards frequently feature
“off-topic” discussion that may or may not be
essential to your analysis. However, automated
QDA software (often abbreviated to CAQDAS
or Computer-Assisted or Aided Qualitative
DataAnalysis Systems or Software; see Lewins
and Silver, 2007) such as ATLAS.ti or NVivo
(see Gibbs, Chapter 19, this volume) offer
assistance in coding, searching, classifying, and
organizing large sets of qualitative data.
There are a range of different ways to judge
the quality of a netnography as a tool for revealing
insights from social media and online interactions.
Kozinets (2010) follows the history of
ethnography and its corresponding quality
judgments and evaluations and suggests a pliable
set of criteria that can be adapted to
achieve many rhetorical research purposes.
Some of these criteria contradict one another,
indicating the need for a custom-made evaluative
solution to be devised by the researcher, in
the spirit of ethnographic research. The arrows
in this evaluative quiver for judging the quality
of a netnographic analysis are that the analysis
is internally coherent; that it follows accepted
procedures; that it recognizes relevant literature
and approaches; that it follows from and links
to the data; that the ideas provide new understandings;
that a sensitizing connection is
gained; that a believable sense of culture is
presented; that the analysis is open to alternative
interpretations; that the text inspires social
action; and that the analysis accounts for the
interaction of online and offline social interactions
(see Kozinets, 2010, for more detail).
These 10 criteria offer netnographic researchers
a pragmatic “toolkit” orientation for the evaluation
of their data analysis.
In summary, the chapter offers various sets
of advice that may lead the researcher to highquality
social media data analysis, and provides
some best practices advice for conducting
18-Flick_Ch-18.indd 274 7/30/2013 8:21:32 PM
NETNOGRAPHIC ANALYSIS: UNDERSTANDING CULTURE THROUGH SOCIAL MEDIA DATA 275
netnographic studies and netnographic data
analysis. At its core, netnographic data analysis
is about maintaining the cultural quality of the
social media phenomenon through the careful
consideration of the researcher’s own role and
social intelligence throughout the process of
online social scientific research.
FURTHER READING
Garcia, Angela Cora, Standlee, Alecea J., Bechkoff,
Jennifer, and Cui, Yan (2009) “Ethnographic
approaches to the Internet and computer-mediated
communication,” Journal of Contemporary
Ethnography, 38 (1): 52–84.
Kozinets, Robert V. (2002) “The field behind the screen:
Using netnography for marketing research in online
communities,” Journal of Marketing Research, 39:
61–72.
Kozinets, Robert V. (2010) Netnography: Doing
Ethnographic Research Online. Thousand Oaks, CA:
Sage.
REFERENCES
Beaulieu, Anne (2004) “Mediating ethnography:
Objectivity and the making of Ethnographies of the
Internet,” Social Epistemology, 18 (2–3): 139–63.
Bruckman, Amy (2006) “Teaching students to study
online communities ethically,” Journal of Information
Ethics, 15 (2): 82–98.
Catterall, Miriam and Maclaran, Pauline (2002)
“Researching consumers in virtual worlds: A cyberspace
Odyssey,” Journal of Consumer Behavior, 1
(3): 228–38.
Cherny, Lynn (1999) Conversation and Community:
Chat in a Virtual World. Chicago: University of
Chicago Press.
Corbin, Juliet M. and Strauss, Anselm L. (2008) Basics
of Qualitative Research: Techniques and Procedures
for Developing Grounded Theory. Thousand Oaks,
CA: Sage.
Correll, Shelly (1995) “The ethnography of an elec-
tronicbar:Thelesbiancafé,”JournalofContemporary
Ethnography, 24 (3): 270–98.
DeAndrea, David C., Ellison, Nicole B., LaRose, Robert,
Steinfield, Charles, and Fiore, Andrew (2012)
“Serious social media: On the use of social media for
improving students’ adjustment to college,” The
Internet and Higher Education, 15 (1), 15–23.
Garcia, Angela Cora, Standlee, Alecea J., Bechkoff,
Jennifer, and Cui, Yan (2009) “Ethnographic
approaches to the Internet and computer-mediated
communication,” Journal of Contemporary
Ethnography, 38 (1): 52–84.
Guba, Egon G. and Lincoln, Yvonna S. (1981) Effective
Evaluation: Improving the usefulness of evaluation
results through responsive and naturalistic
approaches. San Francisco: Jossey-Bass.
Hayano, David M. (1979) “Auto-ethnography:
Paradigms, problems, and prospects,” Human
Organization, 38 (1): 99–104.
Kayahara, Jennifer and Wellman, Barry (2007) “Searching
for culture – high and low,” Journal of Computer
Mediated Communication, 12 (4): 825–45.
Kozinets, Robert V. (2002) “The field behind the screen:
Using netnography for marketing research in online
communities,” Journal of Marketing Research, 39:
61–72.
Kozinets, Robert V. (2010) Netnography. Thousand
Oaks, CA: Sage.
Kozinets, Robert V. (2012), “Marketing netnography:
prom/ot(ulgat)ing a new research method,”
Methodological Innovations Online (MIO), 1 (7):
37–45, online at: www.pbs.plym.ac.uk/mi/index.
html (accessed May 8, 2013).
Kozinets, Robert V. and Handelman, Jay M. (2004)
“Adversaries of consumption: Consumer movements,
activism, and ideology,” Journal of Consumer
Research, 31: 691–704.
Kozinets, Robert V. and Kedzior, Richard (2009) “I,
avatar: Auto-netnographic research in virtual
worlds,” in Michael Solomon and Natalie Wood
(eds.), Virtual Social Identity and Social Behavior.
Armonk, NY: M.E. Sharpe. pp. 3–19.
Kozinets, Robert V., de Valck, Kristine, Wojnicki, Andrea
C., and Wilner, Sarah J. S. (2010) “Networked narratives:
Understanding word-of-mouth marketing in
online communities,” Journal of Marketing, 74: 71–89.
Langer, Roy and Beckman, Suzanne C., (2005)
“Sensitive research topics: Netnography revisited,”
Qualitative Market Research: An International
Journal, 8 (2): 189–203.
Lewins, Ann and Christina, Silver (2007) Using Software
in Qualitative Research: A Step-by-Step Guide. Los
Angeles: Sage.
McCracken, Grant (1988) The Long Interview. Thousand
Oaks, CA: Sage.
Miller, Dame I and Slater, Don (2000) The Internet: An
Ethnographic Approach. New York: Berg.
Muñiz, Albert M. Jr. and O’Guinn, Thomas C. (2001)
“Brand community,” Journal of Consumer Research,
27: 412–32.
18-Flick_Ch-18.indd 275 7/30/2013 8:21:32 PM
AQ2
AQ3
276
Ruhleder, Karen (2000) “The virtual ethnographer:
Fieldwork in distributed electronic environments,”
Field Methods, 12 (1): 3–17.
Schau, Hope Jensen and Gilly, Mary C. (2003) “We are
what we post,” Journal of Consumer Research, 30
(3): 385–404.
Shoham, Aviv (2004) “Flow experiences and image
making: An online chat-room ethnography,”
Psychology and Marketing, 21 (10): 855–82.
Spiggle, Susan (1994) “Analysis and interpretation of
qualitative data in consumer research,” Journal of
Consumer Research, 21: 491–503.
Walstrom, Mary K. (2000) “The eating disorder is not
YOU: Applying Bakhtin’s theories in analyzing narrative
co-construction in an Internet support group,” in
Norman Denzin (ed.), Studies in Symbolic Interaction.
Greenwich, CT: JAI Press. pp. 241–60.
Walstrom, Mary K. (2004) “Ethics and engagement
in communication scholarship: Analyzing public,
online support groups as researcher/participantexperiencer,”
in Elizabeth A. Buchanan (ed.),
Virtual Research Ethics: Issues and controversies.
Hershey, PA: Information Science Publishing.
pp. 174–202.
Wellman, Barry, Quan Haase, Anabel, Witte, James,
and Hampton, Keith (2001) “Does Internet increase,
decrease or supplement social capital?”, American
Behavioral Scientist, 45 (3): 436–55.
18-Flick_Ch-18.indd 276 7/30/2013 8:21:32 PMView publication statsView publication stats
AQ4