Content uploaded by Sanda Erdelez
Author content
All content in this area was uploaded by Sanda Erdelez on Feb 15, 2019
Content may be subject to copyright.
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic
Library Websites
Anindita Paul
Indian Institute of Management Kozhikode, IIMK Campus PO, Kozhikode, Kerala – 673570, India
E-mail: aninditapaul@gmail.com
Sanda Erdelez
University of Missouri, USA
School of Information Science and Learning Technologies, 303 Townsend Hall, University of
Missouri, Columbia, MO- 65211
E-mail: erdelezs@missouri.edu
Abstract
Academic libraries are often required to support decision-making related to data that are both
credible and readily obtained. Library researchers and practitioners have recognized the need for
tools, such as Web Analytics (WA) for online researching. However, few studies have analysed
the library management’s perception towards the WA use. Understanding the management’s
perspective towards WA can further help to understand the optimum usage of WA and the various
challenges and obstacles faced while using it for the library purposes. Six members of the Website
Usability Committee of an academic library were engaged in an interactive group interview about
WA. Analysis revealed the library management’s various views about WA and the challenges faced
while interpreting the WA reports. Four themes emerged out of the interview regarding the use of
WA—decision about services, decision about web-design, library’s interest in user behavior, and
challenges of WA use for the library. These themes are then discussed to shed light on current and
future prospects of using WA for improving academic library website. Factors affecting web analytics
implementation are also discussed.
Keywords: Academic library website, Library decision–making, Web analytics, Google analytics, Web
site evaluation, Web metrics
Anindita Paul and Sanda Erdelez
116
World Digital Libraries 6(2): 115–132
1. Introduction
Use of Web analytics (WA) by academic libraries
has been widely advocated in various literatures
for quite some time. Covey (2002) concluded
in her study that there is a need for a product
that manages, analyzes, and graphically presents
library data. WA enables capturing of user
actions on the websites. It tracks users’ behaviour
on a website and provides precise information
about various usage metrics, including website
visits, sources of these visits, time on site, page
hits per visit, network properties, and user
computer configurations. There has been stress
for quite some time on academic libraries to
build more user-friendly systems (Hiller and
Self, 2004; Kim, 2006a; Kim, 2006b). With the
use of WA packages, library websites can capture
the users’ behaviour on their website through
various metrics and reports that will help them in
decision-making.
The extent of WA use in academic libraries
has not been properly documented in the
literature. Lakos (2007) found that in spite
of libraries being aware that data would help
them in decision-making, they stuck to their
traditional mode of operations based on
intuition rather than techniques that provide
data through “evidence, analytics, and results”.
Similarly, in a survey on methods of assessment
of web services, Manuel, Dearnley, and Walton
(2010) found that some academic libraries are
inclined towards the use of analytics tools, but
many others have yet to understand their utility.
WA has been extensively used in the
commercial sector (Jacoby and Luqi, 2007; Sen,
Dacin, and Pattichis, 2006; Srinivasan, Amir,
Deshpande et al, 2004). Akkus, Chen, Hardt et
al. (2012) has also addressed one of the most
talked about facet of WA, privacy, and proposed
a solution design. Bose (2009) has stressed on
the opportunities and challenges offered by
the advanced analytics for business intelligence
that integrates data mining with text and web
mining. Academic libraries have also started to
recognize the usefulness of WA. Various studies
have revealed that many academic libraries
are nowadays recommending the use of WA
(Manuel, Dearnley, and Walton, 2010; Fang,
2007; Fang and Crawford, 2008; Herold, 2010;
Finch, 2010; Betty, 2009; Loftus, 2012; Farney,
2011; Black, 2009; Cooper and May, 2009). These
studies have acknowledged the need to tap into
the immense potential that WA tools offer for
libraries. Nevertheless, many libraries across the
world are yet to utilize WA for their purpose.
A case study has been discussed in this paper
where the WA implementation of an academic
library in a mid-western American university
is studied in order to understand the extent
of utilization of WA in the library and the
factors affecting its use. This paper discusses
the adoption of WA and its use for academic
library decision-making. It looks at the existing
WA reports within the library through the eyes
of the librarian (instead of the tech-savvy IT
executive) and understands their views from a
practical perspective with regards to use of WA
in academic libraries. This paper also points
towards the elements of WA use in academic
libraries that can hopefully lead to designing
of a WA package that solely caters to
academic libraries.
2. Literature Review
Studies have acknowledged that libraries need
to be more holistic in their decision making
including the user (Nicholson, 2004; Saraf
and Mezbah-ul-Islam, 2002). Decker and
Hermelbracht (2006) used conjoint analysis
to evaluate library service goals, thus stressing
on the need for libraries to make the right
decisions. Kettunen (2007) evaluated the
strategic plans of a library consortium from
four perspectives— customer, financial, internal
process, and learning. Beck (2003) found that
libraries have recognized the need for new data
measures that would help in the evaluation of the
quality of their services. He also emphasized the
importance of the users in the decision making
117
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
process. Hiller and Self (2004)also discussed
the importance of libraries as user or customer
centered organizations and their need for data
that could be used for decision making rather
than subjective impressions and opinions.
Additional studies have identified the importance
of the library user’s perceptions and use of
information sources on library management
decisions (Kim, 2006a; Kim, 2006b).
Eldredge (2006) discussed the need for a
systematic process of evaluation for libraries,
such as Evidence Based Librarianship (EBL)
to increase library decision-making efficiency.
Nicholson (2006a) noted the lack of appropriate
research and literature that could be used by
libraries as evidence. Moreover, the time taken
to collect evidences sometimes results in fewer
publications, thus reducing the influence of
traditional EBL. Booth (2009) performed an
extensive literature review to evaluate the gap
between evidence-based research and evidence-
based practice. He suggested performing more
studies that could help apply research into
practice within the framework of EBL.
Lakos (2007), in a study, discussed that
many libraries plan and make decisions for
data use, but they do not actually use the data
systematically or effectively. His interviews
with library directors revealed that they were
not satisfied with their data access needed for
decision-making. Also, library staff seemed
to be reluctant to provide the requested data.
There was a negative reaction and distrust on
their part to the use of statistics in general. This
seemed to be based on a number of factors, viz.,
lack of analytic skills, lack of interest, and lack
of time. The directors believed that staff reliance
on intuition and accepted assumptions would
be difficult to change, though they accepted the
notion of collecting available data and using the
benchmarking studies.
In recent time, studies have illustrated the use
of WA for library decision-making. Betty (2009)
explored the use of WA for academic libraries
especially when libraries have an increasing surge
of “homegrown” collections. He also suggested
some best practices for libraries. Herold (2010)
also analysed the use of WA to study the users of
a digital archival image collection within a library.
The author used WA to look at the users who
access it, how many of them were from other
regions of the world outside the campus, what all
items they accessed, etc. Loftus (2012) also showed
the use of WA in continuous tracking of health
science libraries. He used other methods also to
gather user feedback, such as through surveys,
targeted interviews, stakeholder analysis, etc.
He stressed on the use of WA as a supplemental
tool. Waller (2009) also explored the use of WA
for a public library website. Turner (2010) in his
guidelines for website evaluation for libraries
using Google Analytics (GA) recommended that
libraries should have KPIs (Key Performance
Indicators) to help them in more directed
decision-making. These KPIs will prove more
beneficial for them towards meeting their goals.
Further, Arendt and Wagner (2010) studied
an academic library website and recommended
the use of Google Analytics over data logs. They
also found that some library decisions were
not implemented despite having support from
analytics (e.g., in spite of the low usage of certain
prominent links on the home page, no changes
were made to the presence of these link and the
top 20 content was not made prominent). The
authors noted the limitations of interpreting
analytics reports. They felt a need for complete
implementation of the WA tool that would
permit a full data interpretation.
Jansen (2009) pointed out the importance
of WA in collecting behavioural data and the
importance of linking investment returns and
website goals to support actionable outcomes.
He also provided a list of guidelines for choosing a
WA solution and a set of factors, such as the WA’s
features, investment, ownership policies, and
integration with other evaluation tools,
for consideration.
On one hand libraries have been urged to
use WA for academic library websites, on the
Anindita Paul and Sanda Erdelez
118
World Digital Libraries 6(2): 115–132
e-resources
& online
services
other hand businesses which have been using
WA for some time have been asked not to rely
solely on it. Weischedel and Huizingh (2006), in
their case study of a major US player in the IT
industry, observed a strong preference for the
quantitative data delivered by analytics tools as
opposed to user feedback by managers in the
company. The authors also noted that there
was a discomfort in using WA for answering
“why” and “how” questions and promoted the
use of complementary qualitative studies. It
is not adequate to depend solely on WA data
to understand the user’s views. Interpretation
of the WA data needs to be supplemented
with feedback data, such as user interviews or
surveys as indicated by Loftus (2012) for better
understanding of the results.
Individual privacy of users is a major
concern with WA. Tracking user’s activities on
the website is against the privacy policy of the
libraries. Though there are efforts to conceal
the user identities when tracking user data
(Akkus, Chen, Hardt et al. 2012), it is still not
reliable. Nicholson (2006b) acknowledged the
importance of maintaining the user’s privacy. He
recommended evaluating the overall data of the
user rather than log items of individual users that
are identifiable. One of the benefits of most WA
tools is that aggregation conceals individual data.
As the literature indicates, there is a need
for an advanced, yet uncomplicated tool to
support library decision-making. There is a lack
of a better understanding of how to use WA
effectively in a library setting. This paper has
tried to address this issue partly by looking at
the concerns of the library committee, which
will help in understanding the ways committee
has used WA in the past, plans to use WA
in future, and the barriers to a full-fledged
use of WA. Apart from the reflections of the
library committee on their use of WA, other
emergent findings from the interview obtained
from the reaction of the committee to the WA
reports were also highlighted. The authors
hope that these reflections and findings from
the committee’s interview will further help to
develop a better understanding of the aspects of
WA for library decision-making.
3. Research Setting
The study was conducted in an academic
library of a mid-western American university
(hereinafter referred to as “the Library”) in 2008.
The Library had implemented Google Analytics
(a free web analytics tool by Google) on its
website, but the data provided by it had never
been adequately utilized, though there were
instances of some application of the inferences
from the reports (e.g., reorganizing hyperlinks).
The Library’s management was eager to use
GA for decision-making and wanted assistance
in this process. The Library formed a web
advisory group responsible for the website. The
advisory group was further divided into three
committees— the web usability committee, the
content coordination group, and the design
group. The eight members of the web usability
committee comprising members from the main
and branch libraries, were sent e-mail invites to
participate in this study as they were supposed to
use GA the most out of the three committees.
4. Method
The researcher was granted access to GA reports
of one semester’s use of the Library website.
These reports were reviewed and then included
in a presentation. Based on this, an interactive
group interview was conducted with the Library’s
web usability committee. An invitation for
participation was sent to the eight committee
members for an interactive group interview
conducted within the Library premises. Out
of eight, six (one male, five females) members
attended the interview. Only three of the six
participants had previously used WA. As the
three non-users were not comfortable about
discussing WA, they were occasionally prompted
for feedback and were probed for follow-up
questions. The most junior respondent had
eight years of work experience, while others had
119
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
12–35 years of experience. All participants held
positions of responsibility within the Library,
including head of information services in a
branch library, library information specialist,
catalogue librarian, head-catalogue management
unit, engineering librarian, and subject
bibliography/reference specialist.
4.1 Data Collection
The interview centered on a presentation of WA
reports drawn from the previous semester, which
provided a context for the discussion and a focus
for the committee’s attention. The interactive
group interview technique (Patton, 2001) was
used to engage participants in the discussion
and to promote specific objectives to generate
an open-ended conversation with respondents
to collect their views on the WA implementation
for the Library, and to get their perceptions on
the use of analytics reports for decision-making
within the Library. The question guide used for
the interview is presented in Box 1. The questions
were geared towards prompting the participants
to reflect on the use of WA in the past as well as
on current decision-making processes.
The sessions were recorded using Morae
software by TechSmith (http://www.techsmith.
com/morae.html). Morae captured the monitor
screen as well as the actions performed on the
computer. Along with the screen capture, any
facial expressions of the user also get captured.
In this study, Morae captured the computer
screen that displayed the Google Analytics
report to the committee and also recorded the
view of the group seated around the table. This
data collection set up helped in capturing the
contextual information along with the animated
discussion captured using the video and audio.
4.2 Data Analysis
The researcher assembled the raw data from the
interview, transcribed the user feedback, analyzed
the transcripts, and kept extensive records
from the interview. Patton (2001) explained
the importance of comprehensive records that
systematically arrange voluminous data into a
Box 1: Guide questions that were used to interview the Library web usability committee
PWhen was the last time any changes to web analytics tracking code were made and why?
PWhen was the last time a web analytics report was generated and/or was viewed?
PWhat other uses, if any, of web analytics have been made by the Library in the past?
PWhat are the various metrics of web analytics that have been considered for use and how they
been interpreted in the past?
PHas the Library implemented any of the web analytics findings?
PWhat are the typical activities that are carried out by the web administrator on the Library
website?
PDoes the web analytics report answer the questions that are asked by the Library
management regarding the electronic services provided by the Library?
PDoes [web] analytics report provide the required evidence for implementing any changes to
the website?
PWhat, according to you, are the weaknesses of the report?
PWhat are the deficiencies that can be corrected by web analytics implementation (in other
words, what tracking codes can be changed in the analytics implementation)?
Anindita Paul and Sanda Erdelez
120
World Digital Libraries 6(2): 115–132
single primary resource package. The transcript
of the group interview was reduced into simple
sentences and analyzed. The researcher then
identified statements for clarification with
the interview participants. Some of these
clarifications were directed to the Library’s use
of GA, while the others were used to resolve the
meaning of certain statements. For example, one
respondent’s statement, “We have used [WA] to
see what the higher [rates] are per visit. What else
the users are coming through to see, so we could
prioritize resources and make sure they don’t
have any problems,” was rephrased as “Libraries
have used analytics to see what are the visit
rates, what parts of the website have higher page
views, what are the top areas that the user visits
on the website. Libraries would like to know
what problems do the users face and try to solve
those.” A follow-up question was— what specific
metrics were you referring to by higher rates
per visit?
The qualitative analysis consisted of content
analysis with open coding followed by axial-
coding. Strauss and Corbin (1990) defined axial
coding as a set of procedures whereby data are
put back together in new ways after open coding
by making connections between categories by
utilizing a coding paradigm involving conditions,
context, action/interactional strategies and
consequences. The researcher formulated themes
and sub-themes with the respondent statements.
Table 1 provides an example of an analysis table
containing themes, sub-themes, statements, and
underlined dimensions.
5. Findings and Discussions
Analysis of the interview transcripts revealed four
major themes—decisions about services, decisions
about web design, library’s interest in user
behaviour, and challenges and opportunities of
WA use. The first two themes are direct outcomes
of the use of GA by the web committee in the past.
These two themes have direct implications on the
understanding of the current use of analytics by
the academic library. The third theme relates to
the observations and inferences drawn from the
interview session that indicated the committee’s
proneness towards using GA for the Library.
This theme involved looking at the non-verbal
and verbal cues of the committee when assessing
the reports presented to them. This theme was
important and helped to understand certain
subjective elements of the Library’s inclination
towards using WA in the future. The fourth theme
Table 1: Analysis table showing example of the elements of analysis—themes, sub-themes, statements, and
underlined dimensions
Themes Sub-themes Statements rephrased Explanation
Decisions
about
services
Providing for
specific user needs
The library has used the search function in
analytics to look at specific link visits (R 5.4 user
behavior, service) Libraries tend
to make service
related decision
based on
information on
how users use
the different links
presented on their
website
Decisions
about
services
Providing for
specific user needs
It might or might not be necessary to link pages
from the gateway as users could be bookmarking
certain pages (R6.10 user behavior, services)
Decisions
about
services
Providing for
specific user needs
Any information on where the users are directly
going to (probably by bookmarking) would help
libraries to place links in appropriate places (R3.7
Services)
Decisions
about
services
Providing for
specific user needs
Libraries would like to know which links are
redundant and hence get rid of those links (R
6.11)
121
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
relates to the opportunities and challenges that the
participants felt would affect their use of analytics.
All the themes inform about the participants’
view on GA for Library use. The participants
also indicated the current deficiencies of GA for
the Library’s purposes, thereby suggesting how
the future use of WA tool for the Library can be
enhanced specifically by improving upon the
pointed deficiencies.
5.1 Decisions about Services
The theme Decisions about Services emerged from
decision-making in order to improve services
provided to patrons. It is critical for the Library
to provide its patrons the required information
through the Library’s resources. The team
formed by the Library web usability committee
ensures the smooth operation of the Library
website and makes effective decisions. Themes
and sub-themes associated with decisions about
services follow.
5.1.1. Supporting staffing decisions
The committee emphasized the importance of
knowing the time of day the website was used
more frequently. This would help the Library to
support the users’ activity on the website. As a
part of ensuring proper services, it is important
to know the time of day when they experience
high traffic. Respondent six stated about the use
of analytics to support staffing needs:
…If we were seeing big uptakes and if
it, say, dropped off at certain time of the
evening that will help us make a case
either for or against expanding reference
desk staffing hours. Certain times…
maybe we need to get some chat 24/7
[as] these people are using our webpage
around the clock and somebody can
contact the reference.
5.1.2. Providing compatible resources
The respondents mentioned the importance
of knowing users’ computer preferences. Since
online visitors use different technology to
access the Library resources, knowledge of the
technology can help the Library services to follow
these specifications. The constant upgrades in
technology further enhance the need for keeping
a track of whether the online visitor machines
are compatible with the Library services. One
respondent stated that the compatibility with
users’ browsers was a major concern because
one of the branch libraries has resources that are
mostly audio clips, and Firefox users might have
issues.
It is also important that the Library should
be aware of the screen resolution of users’
computers. One respondent gave an instance
when the Library had concerns about an
image that was 600 X 800 pixels that was not
compatible with bigger monitors accessed by
users.
5.1.3. Offering useful subscriptions to
electronic journals
The Library provides information services
by subscribing to e-journals. The participants
mentioned that due to the Library operations
in the past, most of the subscription decisions
have been made without usage evidence. Some
of these decisions about e-journal use have
been made a long time ago, and the Library’s
management usually adheres to these past
decisions.
The application of analytics on the Library
website provided a way to understand how
different e-journals are being used. According
to one respondent, analytics implementation
can help in decision-making regarding the
journal subscription. The participants agreed
that analytics can provide detailed information
on the visitors’ activities on e-journal usage.
This detailed information can complement the
data provided by the vendors on e-journal use.
Also, analytics reports are generated from the
library and hence, are trustworthy. However, the
respondents expressed their concern about the
noise in analytics reports.
Anindita Paul and Sanda Erdelez
122
World Digital Libraries 6(2): 115–132
5.1.4. Providing a satisfying search
experience
Most of the online visitors accessing the Library
website are accustomed to get information quickly
using Google and other search engines. The
Library web usability committee was aware that
users prefer to search rather than browse through
the different Library website links. The Library
provides a search box on its website that searches
across the Library gateway. Apart from this, the
Library also provides a menu-based search option
that allows visitors to find information across
the Library’s collection. To search for a specific
article, database, or a journal, there is a search box
for each— “Find database,” “Search for articles
on a topic,” and “Find specific article or journal.”
The menu-based search options in the Library
website further led the interview respondents
to wonder how users are accessing the different
menu options while conducting a search. One
respondent stated - “…What kinds of searches
people were putting in and whether they were
correctly changing the drop down menu. If they were
searching for what looked like an article title or book
title, then were they doing their appropriate kind of
search for the kind of search box they were using?”
The participants found that understanding
the different access terms, which visitors use to
find the information, was useful. The Library’s
subscription databases allow the Library to
create aliases for accessing the databases. Aliases
enhance the use of databases when the visitors
employ alternative terms or misspelled search
terms to access a database. Respondent five
mentioned that the access terms that the website
visitor may use with the Library website as shown
by analytics can be applied as database aliases.
The respondents wanted to know how many
users are accessing the Library website from a
search engine, such as Google and what search
terms they use. The search terms entered can be
used to make changes in the Library’s keyword
search. For example, Respondent six stated:
It [has] been very useful for us to see
what terms people are putting in to
the search engine on our web page for
a couple of reasons— one, to see the
spellings of various databases, so we
could incorporate the spelling in Library
databases to bring the search directly to
what they are looking for. So we were
able to do an alias and were redirected
to get people seamlessly into that page.
Another [reason was] to see what kinds
of searches people were putting in, and
whether they were correctly changing
the drop down menu to pull up [the
searches]. If they were searching for an
article title or book title, were they doing
their appropriate kind of search for the
kind of search box they were using…
[Also] which kind of searches were done
most often and what kind of choices
were…pulled up most.
5.1.5. Providing for specific user needs
One respondent stated that it would help to
know whether online users were coming from
the Library staff machines or from outside
the Library. The respondents were interested
in knowing which of the Library visitors were
students and were they behaving differently than
the faculty. Knowing the profile of the specific
users and if they were staff, students, or faculty,
would have supported the user specific decisions
of the Library, such as if undergraduate students
need resources more than graduate students.
Respondent six also stated that the library would
like to know whether its users are coming from
off-campus or on-campus. This would also help
to identify staff visits. Identifying user visits from
staff would provide a clear picture of the usage
of the Library resources. GA could be used to
segregate the visits based on the profile of the
visitors (such as, students, staff or faculty) from
the time they start browsing. Alternatively, the IP
address of some machines internal to the Library
or branch libraries that belong to the internal
staff could be filtered from the reports. However,
earlier a very basic form of GA was used in the
123
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
Library and so the above-mentioned settings
were not implemented.
5.2 Decisions about Web Design
This theme looks at the use of GA by the Library
for improving its website design. There are
different pathways of information resources on
the Library’s website, such as catalogues and
research databases, some of which have limited
access to Library users only. There are also
different types of information provided about
the Library, such as hours of access, different
workshops held, copy services, and online chat
references. The participants wanted to know
that how visitors successfully accomplished the
tasks for which they visited the website, so that
the design of the site could be improved and its
content restructured for improved navigation.
The sub-themes under this topic follow.
5.2.1 Continuity of users’ actions
Information about whether the user has moved
forward or abandoned the website would
have helped in making changes to the website
structure. Participants would have liked to see
the continuity of users’ actions between pages, as
stated by Respondent three, “Seeing one page at a
time is useful, but seeing one thing after the other is
useful some times.” In other words, continuity of
users’ actions when visiting the Library website
was considered useful for the Library’s purposes.
When users spend hours on a page it would be
informative if the data suggested that the user
was lost and then left the page. Analytics’ report
showing the relationship between the time-
duration a user spent on a page and the exits
from those pages would be beneficial for the
Library.
Users’ paths of access are important for the
Library to guide their movement through the
pages. Respondent three stated that the Library
is interested in looking at whether people went
from a page to where they would like them to
go, where they thought they should go, or they
left the website without clicking on any link. The
participants indicated that it would have been
useful for them if the analytics presented the full
path to the Library resources.
5.2.2 Comparison of link usage
In some cases, when the Library wanted to know
about user activity on a specific link, it had to
search for that link from the content report in
analytics to get specific metrics for that page.
Analytics provided an option to compare a page
metrics with the website metrics. However, the
participants mentioned that a comparison of the
usage of the links leading to different resources
would be useful for the Library, as comparison
indicates preferences of one resource over the
other. The Library has taken steps to improve the
visibility of the less frequently used sections of
the website. One respondent mentioned that the
Library has used analytics in the past mostly for
moving links based on the clicks they received on
each link.
5.2.3 Reorganizing links to improve
navigation
Over time, the Library website has accumulated
many web links across different pages. A
major concern for navigation through the
Library’s pages is that the users can get lost
or confused. Respondent three stated, based
on past experience, that most of the users got
lost because of the presence of multiple links
and eventually found themselves on the same
page where they started. The Library website
also had multiple hyperlinks that led to same
pages. These links were often labeled differently.
Respondent six stated that, “…We may have
kinds of redundant links on our website. One [says]
search for articles on a topic and another for [says]
databases.” However, the destination pages for
both these links were same. In order to keep
an account on link which visitor used to go to
the next page, analytics provided an option to
tag these links. Tagging links allowed analytics
to track the different links that were used to
access the same destination page. However,
Anindita Paul and Sanda Erdelez
124
World Digital Libraries 6(2): 115–132
the presence of many developers handling the
Library website without following any naming
convention has resulted in the current state of
the website where many redundant links exist
and some of which may lead to the same page.
5.2.4 Direct entries to the web pages
The participants were curious to know how
users adapted the bookmark feature of the
browser to access the Library pages. If a
visitor has directly entered a page through a
bookmarked link, it would mean that the user
visited that webpage frequently or they found
it difficult to access the page by browsing and
hence, preferred to bookmark it. Respondent
three stated, “You know that people are going from
here to there, whether they find it a whole another
story.” Information on direct entries to particular
pages would help the Library make decisions
about structuring the navigation pathways to the
resources it provides. WA data on direct entries
to a page can be utilized for this.
5.3 The Library’s Interest in User
Behaviour
This theme further provides insights on the
committee’s inclination towards GA use by
subjective interpretations of the participants’
feedback and behavioural cues as captured
through the video during the interactive
interview. The participants saw the Library’s
analytics implementation as a means to get a
bigger picture of how their users interact with
the Library website when no immediate decisions
were needed.
The participants expressed interest in the use
of GA to understand the use of branch library
resources, visitor activities, shifts in users’
behaviour, and entry points to the website.
The Library provides various resources for its
users within the main Library, as well as across
different branches of the Library that were
established under specific academic departments.
The distribution of resources across the Library
and its branches made the respondents eager to
know the usage of the Library’s resources and its
relation to the different branches, for example,
the number of users who came from the branch
websites to use the resources that are under the
main website. Respondent five mentioned that
it was interesting to see that which branches sent
the most traffic to online resources. Information
on usage of a particular branch website of the
Library would also benefit the branch gateways
and would help the respondents to know how
users access the resources available through the
subsidiary gateways. The participants found
reports that enabled comparison of users’
behaviour on the website of interest. One
example was the comparison report between the
website and the database usage. The respondents
noticed that the database use pattern was similar
to the website use pattern, indicating that
database usage contributes considerably to the
total website usage.
The participants were concerned about the
way users browse for information on the Library
website. They would like to be aware of any
shift in the users’ access behavior, such as if the
catalogue use has decreased from previous years
and how that might be related to the usage of
other resources. Respondent three was curious
to know, “[whether] there is a shift of usage from
catalogue to some other medium within the Library
website, if at all.”
Since, there is no one definite path for the
online visitor to enter the Library website,
the participants were curious about the paths
the visitors take most often. Respondent one
stated her curiosity, “Where people are coming
from, whether they are coming from the gateway
or whether doing an organic search.” As stated
before, the knowledge of the visitors’ entry
whether coming from off-campus or on-campus
was of interest as well. Also, information on
whether users were coming to branch websites
through the main website or were being
detracted from other websites was of interest to
the participants. Use of links, as discussed before,
was rather a discrete way of looking at visitors’
125
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
activities. However, the participants wanted
to understand visitors’ overall intentions and
actions on the Library website. These findings
indicate that there is an inclination for further
use of WA for the Library, but the way to achieve
it is not very clear.
5.4 Challenges of Web Analytics Use for
the Library
During the interview with the committee,
certain opportunities of GA implementation
in libraries as well as the various challenges
faced in implementing some of the results from
the analytics reports were discussed. Various
sub-themes that came out of the analysis are
described in the following paragraphs.
5.4.1 Privacy of users’ data
Web analytics data are IP-based data and are
collected through scripts included in the Library
website’s source code. The reports provided are
aggregated without any reference to the visitors’
IP address. Though the respondents would like
to get a detailed account of the visitors’ activity
on the website, they feared that the privacy of the
visitors was going to be violated. For example,
respondent six stated that the users would not
like it “if spammers and crawlers acquire the
email addresses of [the Library website] users.”
A tradeoff between visitor privacy and visitor
detail, which either meant missed opportunities
or violation of privacy was of prime concern.
5.4.2 Measuring usage
The Library provides facilities that allow visitors
to access the website from within the Library
premises. The users could also access the
Library’s website through computer labs present
in the campus, through their own laptops, or
through computers in their offices and homes.
The Library’s internal staff visits the Library
website more often to perform regular tasks
related to their job. Also, sometimes the same
users access the Library website from different
computers. At other times, different users
access the website from the same computer.
It is difficult to get an account of the unique
users to the Library website. Hence, interpreting
the reports based on the visitor metrics may
not always be accurate. Another challenge is
to interpret the usage metrics, such as bounce
rate and exits. E-commerce websites view high
bounce rate and high exits as detrimental as this
would mean that more visitors are leaving the
website. However, for libraries the web pages
that have external pointers to direct the users to
appropriate resources, bounce rates and exits
may not be detrimental. Hence, a high bounce
and exit rate for the Library may not be much of
a concern.
5.4.3 Analytics implementation and
subscription to proprietary systems
The Library provides services from proprietary
systems. The visitors were directed from the
Library website to these proprietary systems.
Since these proprietary systems were external to
the Library, it could not implement analytics to
measure the usage of these external systems. One
respondent mentioned that the databases they
subscribe provide usage statistics to the Library.
The Library had to rely on these reports for any
decision-making. Hence, the Library is unable
to draw a clear picture of the usage of all the
services provided by it through the website. It has
to rely partially on the proprietor.
5.4.4 Evidence-based decision-making in
libraries
The participants recognized the potential
use of analytics in providing evidence to
support important decisions. For instance, one
respondent stated that in case a decision was
made on dropping the link to the Library website
from a referral page internal to the university,
the Library could use the WA reports to defend
its space on the referral page. Moreover, use of
journals and databases could be supported by
the number of users visiting these resources. A
further understanding of what the visitor was
Anindita Paul and Sanda Erdelez
126
World Digital Libraries 6(2): 115–132
able to access and use could also be made by
looking at the compatibility of users’ computers
with the services and features provided by the
Library. This facility is an opportunity for the
Library to make good use of WA implementation
on their website.
5.4.5 Library’s disregard of analytics use
On one hand, analytics provided useful evidence
for supporting Library decisions. On the other
hand, these evidences tare likely to be overlooked
because of the pre-existing policies. One
respondent stated that,
“…We may have various kinds of
redundant links on our website, one goes
to the search for articles on a topic and
another one goes to databases. Now we
know that this [is] intuitive of people who
go to databases, go to search for articles
on a topic and … probably with some
comparative numbers of which one of
those links attracts the most clicks. But
basically…for political reasons we are
going to keep both of them on there.”
Another instance of a Library decision
that disregards analytics figures was the
‘Announcement’ section that had to be present
even if it does not capture many clicks.
5.4.6 Marketing library services
Businesses prefer commercial websites to be
accessed through many referral sites. However,
this study’s participants preferred to refrain
from promoting the Library gateway to the
referral websites, especially to the external ones,
such as Factiva that sent traffic to the Library
website. The participants were concerned
that any promotion to the referral sites would
perhaps give these sites marketing ideas and
the Library may have to incur a cost for the
visits they receive through these referrals. One
respondent explained that if the Library tries
to promote their website to an outside party, it
gives marketing ideas to the sites they are not
directly linked to. However, the participants
favoured the usefulness of the referral reports on
visits from internal referral websites as this could
help support situations when other departments
decided to remove the Library website link from
the referral pages that reflect in the report.
6. Discussion
The study revealed the various ways to use
web analytics by the libraries although it is
not adequately used in the Library. Manuel,
Dearnley, and Walton (2010) found that
almost 60 per cent of libraries were struggling
to use analytics for decision-making. This
section discusses the study findings under three
headings—current use of web analytics, areas
where web analytics use can be extended, and
factors that affect web analytics use.
6.1 Current Use of Web Analytics in
Libraries
Though libraries are increasingly adopting
web analytics for their purposes, there is some
degree of reluctance in the full-fledged use of
analytics in academic libraries. In this study,
the Library’s use of Google Analytics seemed to
be very limited of all the capabilities provided
by the tool. The analytics reports were used
mostly for improving the service and the website
design. Understanding compatibility of the
user’s machines with that of the technology
used to avail services of the library could help in
improving the services provided by the Library.
Though studies (Betty, 2009; Black, 2009) have
suggested the use of web analytics to check
compatibility of technology of the visitor’s
machines so that it could be used for planning
in future, such a use was yet to establish. In this
study also, the Library had used this feature only
once or twice.
Use of keywords report to understand the
various keywords used by the visitors to access
the website was another important utility of the
tool that found use in the current study. The
keywords were used to improve the database
127
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
services by putting aliases wherever necessary
for easier access to information provided on
the website. The databases used in libraries can
have names which can be confusing to spell
correctly, for example, the database Psycinfo may
be mistakenly spelled as Psyc‘h’info. Sometimes
the user base of academic libraries involves
international students and hence putting aliases
can help these students who could possibly
misspell some of these database names.
Libraries have increasingly used analytics for
rearranging website links (Manuel, Dearnley,
and Walton, 2010; Loftus, 2012; Black, 2009).
The participants of this study extensively used
and speculated more use of Google Analytics
for improving the website design. They used
the “Top Content” reports for rearranging the
different links present within the website based
on the hit each of these links got and the need to
push each link. There are many other prevalent
uses of web analytics in academic libraries, which
was not reported when the study was conducted.
Reports on visits to library resources could be
useful for impact studies of funded projects.
Also, origin of the visits could help in studying
region wise impact on library’s resources.
Therefore, there is much more that needs
to be done in terms of increasing the current
usage of web analytics for the library purposes.
As mentioned before, proponents of web
analytics have urged libraries to utilize the
benefits offered by the tool, though there is a
certain lag in understanding the context of the
visitors. However, it cannot be denied that the
huge amount of data captured by library website
need to be interpreted well to enhance user
friendliness of the library websites. Additionally,
analytics can provide ‘behavioural evidence’
that can complement the real users’ intentions
captured using a qualitative study.
6.2 Areas where Web Analytics is Yet
to be Used
This study revealed the need for an extensive use
of web analytics for libraries. The participants
were aware of the potential that analytics have,
but there was no urgency to use it. In future
web analytics can be used in branch libraries as
well. Comparative reports between the main and
branch library can inform the relative stake of the
branch library. Though the branches of a library
could have their own analytics implementation
that informs them about their website usage, any
information about the use of branch resources
can be useful. It will also be useful if libraries
consider implementing one complete analytics
solution for the entire library system which
will enable a better exploration of the reports
across all subsidiary websites. This will enable
centralized management of data. However, issues
such as any internal rivalry between the branches
and the main library may lead to resistance in
sharing website usage information with other
branches.
The study emphasized on the maximum use
of analytics to understand the importance of the
Library resources. One way to know this is to
look at the traffic paths of access to the website
resources. The visitors could access the web pages
directly or they could browse from the home
page of the website. A direct visit could be made
either by directly typing the URL or clicking on
a bookmark. The demand for a website is more
if a visitor has bookmarked it, so that they don’t
have to remember the website URL. Such a habit
of bookmarking can be used to understand how
content needs to be organized throughout the
website. Alternatively, visits through the search
feature can inform about the information needs
of the visitors by looking at their search terms.
Most of the visitors access the information
quickly using search. “Googl-ing” information
has become the trend. Finch (2010) indicated the
importance of search report to enhance website
visibility. Academic libraries need to make better
use of the search reports for better information
experience by the visitor. As mentioned before,
the Library has used information to improve
searching for database. However, more needs to
be done in terms of Search Engine Optimization
Anindita Paul and Sanda Erdelez
128
World Digital Libraries 6(2): 115–132
as now visitors rely on searching techniques to
get information quickly and easily. The ‘Site
Search’ feature of Google Analytics that enables
capturing of the terms used by a visitor on the
website was not set. Not many studies have
specifically discussed the use of Google Analytics
to improve the internal search function or have
looked at the internal search keywords reports
that were generated in the analytics reports and
related interpretations of it.
Referring site is another option to access
the website apart from directly visiting the
website and searching their way to the website.
Businesses can make good use of such a feature
to increase the traffic in their website, but
libraries operate differently. Libraries are not
concerned that which websites are sending
traffic to their website. They are only concerned
to know that visitors are coming from which
internal sites of the university system and would
like to use those to vouch for support to retain
the links on these internal websites. Hence, the
traffic sources report can be utilized to enhance
traffic to the website and allow user-friendly
seamless access path to information on the
website.
Library websites provide a lot of information
about the service, about the library hours,
any new features added, maintenance related
announcements, etc. Website visitors may
overlook such information if it is not very
obvious or if it is cluttered amidst a lot of
text and links. The respondents of the study
observed a feature of Google Analytics called
“site overlay”1, which was considered a user-
friendly way of visually identifying issues within
the website. Again, the usefulness of site overlay
is not properly emphasized in the literature. Site
overlay could be helpful for a visual presentation
of data superimposed on the webpage. Such a
feature is especially useful if the management
would like to have a visual representation of the
data filtered on the basis of links present on
the page.
6.3 Factors Affecting Web Analytics Use
in Libraries
This section identifies the factors affecting
web analytics implementation and their use
in library decision-making. Clear and concise
representation of URLs without making it too
long is one of the factors that can lead to possible
interpretation of analytics reports. Usually
library websites are multi-purpose in terms
of the multiple things that the visitor could
do on it, such as visiting database, searching
articles, checking e-journals, etc. Sometimes
there could be redundant links. Redundancy
and ineffective naming of links on the website
can lead to confusion, such as multiple links
may end up on the same target page or the same
page may be represented by multiple links. It is
hence important that libraries form a naming
convention for the links in the website that
relates to the content of the page. Interpretation
of the reports by looking at the URLs is easier
when the URLs reflect the page it represents.
Another factor affecting web analytics use
is the multiple users and their intentions and
motivations to use the library website. Users
of academic libraries have varied profiles and
motivations for which they visit the website. An
academic library user is frequently involved in
collaborative information seeking and sharing
via multiple computers, and the faculty and
librarians may have considerable influences on
student behaviour. Metrics of user activities
must be interpreted while keeping in mind
the various contexts of users’ actions and the
population that the website serves. Jansen
(2009) emphasized the importance of context
and population in web analytics research.
Further, libraries have many pages offering
multiple resources and services across its
1The “site overlay” feature of Google analytics has been replaced by “in-page analytics”. See http://searchengineland.com/
google-analytics-goodbye-site-overlay-hello-in-page-analytics-53060
129
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
many branches. As a result, interpreting user
activity by assigning meaning to the metrics is
very complicated. More details in the analytics
reports, rather than disjointed figures can
make interpretation of the metrics easier. More
studies need to focus on the various metrics
that should be further developed based on the
library’s needs.
External environment can be another factor
that determines web analytics use, such as high
technological volatility determines visitors’
changing information habits. Web analytics
should understand the changing trends in user
behaviour on the website or any major shifts
in users’ access behaviour within the website,
such as any shifts from searching of information
to clicking through the various links to the
information or shifts in usage of the OPAC
to the database. Web analytics reports on the
keywords used to access the library website can
also help to understand the visitors’ information
seeking behaviour, specifically their information
needs. Waller (2009) urged libraries to realize
the untapped potential of analytics tool for
interpreting visitor behaviour.
Library management supports a holistic
use of web analytics. Employee motivation
to use analytics reports is also an important
factor for a successful implementation of
web analytics in academic libraries. Manuel,
Dearnley, and Walton (2010) also pointed out
that continuous improvement for websites need
senior management support. There is a tendency
of libraries to stick to their traditional ways
of operating, hence creating much resistance
to any new technology. Proper guidelines for
implementation of web analytics for libraries
need to be provided. Libraries also need to
be cautious about where to draw the line
when it comes to using analytics for decision-
making. More studies are required on web
analytics use for academic libraries that will
help in identifying best practices for libraries.
Eventually, a web analytics tool solely for the
purpose of libraries can be useful. Additionally,
benchmarks of good performance could be
established. Such a tool can further be enhanced
according to the evolving needs of the libraries.
7. Conclusion
This case study attempted to identify the usage
of web analytics as a tool on the academic library
website. An additional contribution of this study
is that it looks at the usage of web analytics from
the perspective of the web usability committee
formed to take decisions on improving the
library website. The committee’s reflections
were collected in an interactive group interview,
where the respondents were shown the Google
Analytics reports for the past semester and the
related discussions. Qualitative analyses of the
reflections were done. Four themes emerged—
decisions about services, decisions about web
design, library’s interest in user-behaviour, and
challenges of web analytics use for the library.
The decisions about services and decisions about
web design brought forth aspects of analytics
that are identified as useful for making decisions
about library services and library’s web design,
respectively, as per the opinion of the committee.
It also explained the detailed aspects of the
committee’s inclination to use analytics for
interesting behavioural elements of the visitors of
their website. Finally, challenges of web analytics
use for the library include the management,
policy, and other related factors that pose a
challenge to a full-fledged use of analytics.
Discussions point out the aspects of the current
use of web analytics, areas where web analytics
has not been used yet, and factors that affect web
analytics use in academic libraries.
Findings in this study are based on
investigations in one American based academic
library where GA was implemented. Though the
analytics solution had much functionality, it was
not fully utilized. Future studies could look at
the human information behaviour perspective
that inquires how analytics can be useful as a tool
for learning about users’ information, seeking
process at a broader level. Google Analytics is
Anindita Paul and Sanda Erdelez
130
World Digital Libraries 6(2): 115–132
constantly upgrading its features. This study was
conducted over a period of two years, during
which many features were added. More studies
involving the latest analytics solutions in a library
setting are required across the world. World’s
perspective to the use of web analytics for
academic libraries also needs to be considered.
The view of developing countries on the use of
WA in the library decision-making is also an
important point of consideration.
References
Akkus I E, Chen R, Hardt M, Francis P, and Gehrke J (2012), “Non-tracking web analytics”, Paper
presented at the Proceedings of the 2012 ACM conference on Computer and communications security. pp
687–698
Arendt J and Wagner C (2010), “Beyond description: Converting web site usage statistics into
concrete site improvement ideas”, Journal of Web Librarianship 4(1): 37–54
Beck S J (2003), “Making informed decisions: The implications of assessment”, A presentation,
Association of College and Research Libraries, 11th Annual Conference
Betty P (2009), “Assessing homegrown library collections: Using google analytics to track use of
screencasts and flash-based learning objects”, Journal of Electronic Resources Librarianship 21(1): 75–92
Black E L (2009), “Web analytics: A picture of the academic library web site user”, Journal of Web
Librarianship 3(1): 3–14
Booth A (2009), “A bridge too far? Stepping stones for evidence based practice in an academic
context”, New Review of Academic Librarianship 15(1): 3–34
Bose R (2009), “Advanced analytics: Opportunities and challenges”, Industrial Management and Data
Systems 109(2): 155–172
Cooper J D and May A (2009), “Library 2.0 at a small campus library”, Technical Services Quarterly 26
(2): 89–95
Covey D T (2002), “Usage and usability assessment: Library practices and concerns”, Report funded
by Digital Library Federation and Council on Library and Information Resources. 10–13 Charlotte,
North Carolina.
Decker R and Hermelbracht A (2006), “Planning and evaluation of new academic library services by
means of web-based conjoint analysis”, The Journal of Academic Librarianship 32(6): 560–562
Eldredge J (2006), “Evidence-based librarianship: The EBL process”, Library Hi Tech 24(3): 341–354
Fang W (2007), “Using google analytics for improving library website content and design: A case
study”, Library Philosophy and Practice—LPP Special Issue on Libraries and Google. Retrieved on
September 27, 2007 from http://www.webpages.uidaho.edu/~mbolin/fang.htm
Fang W and Crawford M E (2008), “Measuring law library catalog web site usability: A web analytic
approach”, Journal of Web Librarianship 2(2): 287–306
Farney T A (2011), “Click analytics: Visualizing website use data”, Information Technology and
Libraries 30(3): 1–23
Finch J L (2010), “Measuring effectiveness in a virtual library”, Journal of Web Librarianship 4(1):
63–78
131
World Digital Libraries 6(2): 115–132
Implementation and Use of Web Analytics for Academic Library Websites
Herold I M H (2010), “Digital archival image collections: Who are the users?”, Behavioral & Social
Sciences Librarian 29(4): 267–282
Hiller S and Self J (2004), “From measurement to management: Using data wisely for planning and
decision making”, Library Trends 53(1): 129–55
Jacoby G A and Luqi L (2007), “Intranet model and metrics: Measuring intranet overall value
contributions based on a corporation‘s critical business requirements”, Communications of the ACM
50(2): 43–50
Jansen B J (2009), “Understanding user—Web Interactions via web analytics”, Edited by
G Marchionini, Synthesis Lectures on Information Concepts, Retrieval, and Services, Morgan & Claypool
Publishers
Kettunen J (2007), “The strategic evaluation of academic libraries”, Library Hi Tech 25(3): 409–421
Kim J-A (2006a), “Capturing metrics for undergraduate usage of subscription databases”, Online
30(3): 32-39
Kim J-A (2006b), “Toward an understanding of web-based subscription database acceptance”, Journal
of the American Society of Information Science and Technology 57(13): 1715–1728
Lakos A (2007), “Evidence-based library management: The leadership challenge”, Portal: Libraries and
the Academy 7(4): 431–450
Loftus W (2012), “Demonstrating success: Web analytics and continuous improvement”, Journal of
Web Librarianship 6(1): 45–55
Manuel S, Dearnley J, and Walton G (2010), “Strategic development of UK academic library websites:
A survey of East Midlands university libraries”, Journal of Librarianship and Information Science 42(2):
147–155
Nicholson S (2004), “A conceptual framework for the holistic measurement and cumulative
evaluation of library services”, Proceedings of the 67th ASIST Annual Meeting 41: 496–506
Nicholson S (2006a), “Approaching librarianship from the data: Using bibliomining for evidence-
based librarianship”, Library Hi-Tech 24(3): 369–375
Nicholson S (2006b), “The basis for bibliomining: Frameworks for bringing together usage-based data
mining and bibliometrics through data warehousing in digital library services”, Information Processing
& Management 42(3): 785–804
Patton M Q (Ed.) (2001), Qualitative Research & Evaluation Methods (3rd ed.). USA: Sage
Publications, Inc. pp 127
Saraf V and Islam M M-U (2002), “Measuring library effectiveness: A holistic approach”, Journal of
Library and Information Science 27(2): 81–105
Sen A, Dacin P A, and Pattichis C (2006), “Current trends in web data analysis”, Communications of
the ACM 49(11): 85–91
Srinivasan S, Amir A, Deshpande P, and Zbarsky V (2004), “Grammar-based task analysis of
web logs”, Proceedings of the 13th ACM International Conference on Information and Knowledge
Management, NY: ACM Press. pp 244–245
Anindita Paul and Sanda Erdelez
132
World Digital Libraries 6(2): 115–132
Strauss A and Corbin J (1990), “Basics of qualitative research”, Grounded theory procedures and
techniques Newbury Park, CA: Sage Publications
Turner S J (2010), “Website statistics 2.0: Using google analytics to measure library website
effectiveness”, Technical Services Quarterly 27(3): 261–278
Waller V (2009), “How do virtual visitors get to the library?”, The Electronic Library 27(5): 815–830
Weischedel B and Huizingh E K (2006), “Website optimization with web metrics: A case study”,
Paper presented at the 8th International Conference on Electronic Commerce: The New E-commerce:
Innovations for Conquering Current Barriers, Obstacles and Limitations to Conducting Successful
Business on the Internet, pp 463–470
CopyrightofWorldDigitalLibrariesisthepropertyofTERIPressanditscontentmaynotbe
copiedoremailedtomultiplesitesorpostedtoalistservwithoutthecopyrightholder's
expresswrittenpermission.However,usersmayprint,download,oremailarticlesfor
individualuse.