The current issue and full text archive of this journal is available at

www.emeraldinsight.com/1741-0398.htm

Retrieving
Retrieving relevant information: relevant
traditional file systems versus information
tagging
79
Thomas W. Jackson and Stephen Smith
Department of Information Science, Loughborough University, Loughborough, Received 21 January 2011
UK Revised 21 March 2011
10 May 2011
Accepted 10 May 2011

Abstract
Purpose – The aim is to determine, in a business context, if tagging is a more effective method of
discovering relevant information when compared to traditional hierarchical filing systems.
Design/methodology/approach – A five-step interpretive hybrid approach of using both a focus
group, questionnaires and SWOT analysis was used to test the proof of concept of tagging files
compared to a traditional hierarchical filing system. The approach taken was chosen because of the
difficulties and tradeoffs that had to be made between the number of champions and people available
to take part in the research; the time that they could allow; and because transcription or recording of
the participants was not permitted. The participants were encouraged to use the questionnaires and
the SWOT analysis to record their thoughts anonymously whilst the focus groups allowed elaboration
and discussion to help understand the true feelings and thoughts of the group collaboratively.
Findings – Traditional hierarchical filing systems can lead to the retrieval of irrelevant information,
or to none at all, even though the information exists. The study shows that tagging could provide a
cost-effective solution by providing a better structured filing system that can help reduce duplication
and the retrieval of irrelevant information.
Research limitations/implications – One limitation of the study was the limited number of
participants from just one organisation. Thus, generalisation of the results of this study to the wider
population must be done with great care.
Practical implications – Organisations should evaluate the functionality of their chosen operating
system and information store software in light of the potential benefits offered by tagging, and costly
limitations of traditional file stores.
Originality/value – The paper contributes to the information retrieval and information overload
literature by studying the effect tagging files has on an organisation. It provides an insight to the
future of filing systems for management and triggers future empirical work into reducing information
overload in the workplace.
Keywords Tagging, Information overload, Information storage, Information retrieval, Semantics,
Information management, Classification schemes
Paper type Research paper

1. Introduction
Searching for relevant files within an organisation has been a source of problems for
employees for a number of years (Kelly et al., 2010; Kobayashi et al., 2006). For this
reason many companies invest substantial amounts of money in searchable portals Journal of Enterprise Information
and document storage facilities. Ineffective searches and wasting time looking for Management
Vol. 25 No. 1, 2012
information has been said to cost up to 10 per cent of an employee’s time (Dubie, 2006). pp. 79-93
Many organisations employ highly sophisticated search engines to search the full text q Emerald Group Publishing Limited
1741-0398
of all of the files stored within their systems (Smith, 2010). However, many search DOI 10.1108/17410391211192170
JEIM systems cannot retrieve documents that have not already been indexed by that search
25,1 system such as those within a user’s corporate network profile. In addition, web-based
search systems, such as Google or even a corporate intranet search engine do not help
users locate files stored on their local computers. The user can resort to using their
operating systems search facility but this suffers from the same problems as the web
based search systems (Smith, 2010). Using such search engines also requires users to
80 visit the search page and wait for results to be returned. Further to this, if the user
wishes to browse for a file whilst continuously narrowing down their search as they go,
they are restricted to a traditional hierarchical file structure. Ultimately there are a
number of barriers for users to overcome when attempting to find information using
traditional search approaches (Smith, 2010). This increases the difficulty in discovering
relevant information and can lead to an inability to find what the user is looking for,
which can also be compounded by task uncertainty (Nunez and Giachetti, 2009).
To illustrate the severity of this problem Nelson (1994) references Naisbitt, arguing
that: “Inundated with technical data, some scientists claim it takes less time to do an
experiment than to find out whether or not it has been done before.” It is not surprising
that in a world that is increasingly dominated by computer-mediated communication
systems, it appears that the volume and pace of information can become overwhelming
as shown by even earlier studies in the 1980s (Hiltz and Turoff, 1985; Kerr and Hiltz,
1982). Further studies have gone on to define information overload and the impact it
has. For example in the 1990s Meglio and Kleiner looked at ways of managing
information overload. They looked at time management, communication and the idea
that individuals can reduce information overload as they are part of the problem
(Meglio and Kleiner, 1990). The research showed that users of information all
contribute towards information overload and if these issues can be realised a conscious
effort can be made to reduce the overload, thus improving the effectiveness of
communication. A decade later both Farhoomand and Drury showed that information
overload is still a problem and it affects managers in organisations on a daily basis and
can have dramatic effects within an organisation (Farhoomand and Drury, 2002; Kirsh,
2000). Continuing the journey on the timeline, a decade later in 2010, Soucek and Moser
reported that information overload causes problems in email communication within
the workplace and that the issue of information overload still needs to be addressed
and training might provide a potential solution (Soucek and Moser, 2010). Other
research has taken more of a holistic view in identifying the factors surrounding
information overload, and in particular the main factors contributing to
technology-based productivity losses through information overload, communication
overload, and system feature overload (Karr-Wisniewski and Lu, 2010). To help
visualise the sources of information and potential overload an employee might
encounter, the authors have created a conceptual map of the sources of information
that an employee can access to retrieve information and knowledge as shown by
Figure 1. This study investigates a solution to reducing the information overload
problem in organisations by focusing on improving the filing and retrieving of
electronic documents (denoted by the red).
In the problem domain of information storage and retrieval, the main research
question of this research is, can tagging files save employees time retrieving
information compared to a traditional hierarchical file system. An enterprising
innovation to aid in the storage and retrieval of business information is critical in
 Retrieving
relevant
information

81

Figure 1.
How an employee locates
information/knowledge
sources

generating greater cost effectiveness and increased performance of organisations. It

will also provide organisations with the “best practice” when it comes to information
storage and retrieval. As highlighted by Rosacker and Rosacker (2010) it is becoming
critically important for managers to understand “best business practices” so that these
successful techniques can be applied appropriately to enhance and refine operational
practices.
The paper starts by looking at the traditional use of tagging and investigates its
potential benefits in the retrieval of relevant information within an organisation. It then
outlines the methodology used to evaluate tagging as an alternative to the traditional
filing and retrieval systems. The next section describes of the proof of concept called
TagDav, which is a tag based system developed by the authors to enable a comparison
to be made against traditional storage and retrieval systems. The results and research
synthesis form the next section of the paper, and the paper finishes with a conclusion
and limitations of the research.

2. Discovering relevant information

The current directory structure found on many computers, along with those used by
shared document systems, require a user to place a file into a folder, the file is then
retrieved by navigating to the same folder. Placing files into a system whereby the user
was guaranteed to retrieve the file that they wished to find would be extremely
difficult. The user may even have to add the file a number of different times in different
directories to ensure that it was found. Whilst one possible solution is to make use of
one of the existing corporate search systems, another possible solution is to replace the
file system with one that makes use of tagging.
The task of discovering relevant information has seen specific attention within the
field of the internet where the corpus of information available is massive (De Kunder,
JEIM 2011). Over time the internet has moved from its traditional text and hyperlink driven
25,1 state to a frenzy of media rich web sites containing masses of information for, often,
millions of users. Within this transition some interesting sites and approaches to
discovering information have emerged. Flickr and Delicious provide two interesting
case studies.
Flickr is an online photo sharing web site. It allows users to upload their
82 photographs for other people to see. Delicious allows users to bookmark web pages and
then find their own and other users’ bookmarks when needed. To quantify the volume
of information that is stored in these systems, the Flickr web site receives thousands of
new photos per minute (Flickr, 2006) and a more recent statistic shows that on average
4,602 images are uploaded every minute (Lux, 2010). All of these photos are stored and
retrievable by its members. Delicious is also a high traffic site; it received 150,000 posts
per day in June 2008 (Keller, 2008). Instead of using traditional methods and categories
to allow users to find photos or bookmarks, they make use of tagging. Tagging offers a
strong alternative to traditional hierarchical structures and has allowed many
web-based systems to dispense with manually created categories for users to place
their content into.
Tagging has proven extremely popular, but the literature has also shown that there
are a number of in use issues that need to be considered. These have been summarised
with some examples:
.
Single use tags – or tags which have not been used before should be avoided
unless necessary. Delicious offer a number of recommended tags as this helps to
re-use tags that have already been used before. Some systems also show the user
how many times they have used a tag before, as they are adding a tag to an item.
.
Pluralised or singularised versions of words – The unstructured nature of
tagging can lead to issues when items are tagged with either singular or plural
versions of a tag (Mathes, 2004). In some cases a user may use the tag “apple”
and in others “apples”. It is clear that whilst “apple” could describe both the fruit
and the computer manufacturer, the word “apples“ would not describe the
company. A decision should be taken on whether to allow both or just one, if only
one is used then the singular form is recommended.
.
Spelling mistakes – can cause the creation of a new tag, but unnecessary one,
and make a tag of no benefit.
.
Personal tags – users freely choose the tags and tagging can be seen as “subject
indexing without a controlled vocabulary” (Hayman and Lothian, 2007). In
addition, when items are tagged by multiple parties, during collaborative
tagging there is a higher likelihood that different tags may be chosen (Golder and
Huberman, 2006). This should be avoided and could be prevented by prefixing to
allow for a period of distinction.
.
Spacing and capitalisation should also be considered. Delicious makes the
recommendation that users do not use spaces at all within tags to ensure
optimised efficiency and effectiveness.
. Synonyms – words used may also be synonyms of one another, for example,
whilst one person describes a resource with the word “person”, another may use
the word “human” (Golder and Huberman, 2006). Including as many synonyms
as possible can help to reduce the issues associated with only entering one.
 .
More tags – entering a larger number of tags may help to improve search Retrieving
efficiency and actually allow more focused searches if necessary, preventing the relevant
problems caused by differing granularities of search.
information
For an organisation aiming to implement tagging as a potential solution to aid in the
discovery of information, these issues highlighted by the literature should be
considered. The next section details the methodology that will be used to determine if a 83
tag based system is more effective for storing and retrieving information than a
traditional hierarchical filing system.

3. Methods
An interpretive philosophy was chosen as the underlying framework and the research
adopted a systems development methodology that recognises three main stages:
concept development, system building, and system evaluation (Burstein, 2002). In
terms of system evaluation a system-centred approach was taken. Wang and
Forgionne’s (2008) comprehensive approach was considered but rejected due to the
time constraints of the research project.
The authors developed a proof of concept, called TagDav, to enable users to tag the
files they use. Details of the system are described in the “Proof of Concept – TagDav”
section. A number of different methods were used to test TagDav at SoftwareCo, which
are detailed later in this section. SoftwareCo, is one of the largest software
organisations in the world and is in the top ten of all of the major software rankings
including the Forbes 2000 and Research Foundation’s top 100. The organisation
employs over 50,000 people in over 50 countries. The company develops a range of
software solutions in house and its products are used globally. The SoftwareCo
department that participated in this work was one of the rapid development and value
prototyping divisions. The department’s employees are highly skilled within their
respective field and the department has a very unique structure. The department has
attempted to create an environment specifically suited to rapid application
development, testing and deployment. The department aims to have only a limited
hierarchical structure, with all members of the department seen as equals and
interacting with each other to take advantage of their respective skills. The nine
participants for this research study were selected by one of the champions within the
organisation, who selected on the criteria of someone who would give valuable input,
and their knowledge of tagging (experienced and inexperienced). The nine participants
were also chosen to include the most varied views including those who would be
expected to be advocates of such a system and those expected to be against it.
Figure 2 shows the approaches used to determine the effectiveness of a tag-based
filing system:

Figure 2.
The assessment of tag
based file storage
JEIM .
Step one – an eight question questionnaire was given to the nine participants to
25,1 assess the barriers to tagging, before they were given a demonstration of
TagDav.
.
Step two – a demonstration of TagDav was given to the employees at
SoftwareCo within the focus group environment.
.
Step three – the second questionnaire contained 16 questions and focused on the
84 way that users currently store and retrieve files and questions about the TagDav
system. For example, how easy it is to use and whether they felt that training
would be necessary in order to use the system effectively. The aim of these
questions was to establish if the participants had a common system, or method of
working, or if they all used different filing schemes.
.
Step four – the focus group consisted of nine members from the SoftwareCo. A
focus group was chosen to enable an open discussion to enable participants to
cover a large range of aspects that are relevant to them as a group.
.
Step five – participants were provided with a matrix on which they could record
their thoughts within four key areas, strengths, weaknesses, opportunities and
potential risks that might be introduced by the use of the tag based approach to
document retrieval.
The study took a hybrid approach using both focus groups and questionnaires, which
is similar to, but did not follow, a Delphi style approach. The approach taken was
chosen because of the difficulties and tradeoffs that had to be made:
.
between the number of champions and people available to take part in the
research;
.
between the time that they could allow; and
.
because transcription or recording of the participants was not permitted by
SoftwareCo.
The participants were allowed to use the questionnaires and the SWOT analysis to
record their thoughts anonymously whilst the focus groups allowed elaboration and
discussion to help understand the true feelings and thoughts of the group
collaboratively. Although transcription or recording was not allowed by the
organisation a number of notes and “sound bites” were taken that were approved
by the participants.

4. Proof of concept – TagDav

TagDav is short for Tag-based Distributed Authoring and Versioning. It was
developed using Ruby on Rails by the authors. A server was developed that would
allow users to mount the folders of the server as if they were folders within the
computer, as traditional network attached storage (NAS) devices would, but used a
tag-oriented navigation method rather than the traditional hierarchical approach. The
benefit of integrating with the operating system is that the files will appear as if they
are stored on the local computer. This removes the need for the user to open a web
browser and perform a search. It would also allow the user to access these files from
inside any application that was running on the computer instead of having to
download the file first. The ultimate benefit is that the files would appear to be stored
on the computer like any other file in an attempt to allow users to improve their ability Retrieving
to find relevant information whilst remaining in a familiar environment. relevant
The server was developed in Ruby on Rails, due to the familiarity of the authors
with this programming framework. The server was used to create a virtual file system. information
The server would not retrieve files directly from a given directory but would work in
conjunction with a database. Storing the files in conjunction with a database allows
additional information to be stored with each file such as the tags that were used to tag 85
that file. As the tags were also all stored in a database, it would be a very quick process
to create a list of all of the tags used by files stored within the system making the
system extremely responsive and prevent the delays that are associated with
alternative systems such as searching the full text of a document. The system would
also have a much lower storage requirement than traditional search based approaches
as indexes would not be required in this method.
The server would not store multiple versions of the files but would simply store one
version of the file and make use of the database to allow the server to generate a virtual
structure that is sent to the client. The database would maintain a list of files and the
tags associated with that file. Each of the tags would then appear as a “virtual
directory” within the root folder. When the user chose that directory, the tag would be
used to find all files that contained that tag and display those files. Along with those
files, all of the other tags that were used by the files shown would also be present as
directories to allow the user to further refine their search and narrow the list of files.
This enables the user to choose as many tags as they like until the files that have all of
these tags are listed within the directory. The files would be refined based upon the
tags that are associated with those files rather than where they had been stored.
Although to the user it would appear they were navigating a list of files they were
actually just communicating with the server that was dynamically creating the folders
they would see. An additional benefit of this method is that users may browse for the
files that they wish to find, refining their search by clicking on directories, or tags, as
they went. This is a strong contrast to search-based approaches where search terms
must be entered before any results are displayed. The system also allows the user to
quickly narrow down their search using the directories to represent tags when they are
searching for something specific. Instead of using a pre-defined directory structure the
directory structure is automatically created based upon the tags of the file. Although
the user is essentially browsing a directory structure, the directory structure is
continually narrowing down the list of files that contain those tags to allow the user to
search for the file.
Table I shows the three files that were uploaded into the system and the tags that
were applied. The files in Table I provide an example of how the files would appear
with their respective tags to the user.
Figure 3 shows the TagDav system in the root folder. All of the tags used are listed
so that the user may begin their search. Although it may seem like there would be an

Name Tags

File_Tagged_One_and_Both tag_one, tag_both Table I.

File_Tagged_Two_and_Both tag_two, tag_both Example files in the
File_Tagged_One_Only tag_one TagDav system
JEIM
25,1

86

Figure 3.
TagDav with close folders

issue if there were thousands of tags entered into the system this would not be the case.
The tags would be ordered alphabetically and could quickly be found. In many
operating systems simply beginning to type the directory name would highlight that
directory and in this case the tag. Once one tag had been chosen, the list of remaining
tags would be significantly reduced. The process could continue until the desired file
was found. This root folder also has the option to show all of the files that exist in the
system but for simplicity it is not shown in Figure 3. The tags from Table I can be seen
as directories within the file system that contain the files.
It is then possible to expand and navigate within the folders to see the files that are
tagged by that content. Figure 4 shows the tag “tag_both” expanded as a folder

Figure 4.
TagDav with one folder or
tag expanded
showing the two files within this folder and also the two tags that those files are also Retrieving
tagged with.
The tags may be combined to further refine the search within the folders. Figure 5
relevant
shows the full expansion of the example files and their folders. information
Although in Figure 5 there are nine files listed there only actually three physical
files stored within the system. Each file is only stored within the system once and each
of the files shown in Figure 5 is actually a virtual reference to a file that will be 87
retrieved at run time should the user wish to actually download or open this file. In
order to work with this file, the user simply has to click the file as they would if the file
were in any traditional file system.

5. Results and analysis

5.1 Knowledge of tagging
The results from step one showed a large variance when it came to tagging. Four out of
the nine employees had never tagged content before. Of the five members of the group
that had tagged content before, there was a wide variance in the answers given to the
initial questionnaire regarding how they use tags. This section is based upon the five
users that had previously tagged content.
When asked how many tags they use on average to tag content, the employees were
quite close to the average found in the study by Smith (2010). His research showed an
average of 3.3 tags used per user. This study showed that one participant (20 per cent)
used just one to two tags and only three to four tags were used by the other four
participants (80 per cent). This highlights many users who do not use as many tags as
they perhaps should. Two (40 per cent) re-used them always and two (40 per cent)
sometimes, with one participant (20 per cent) rarely re-using existing tags. The figures
were worse when the tags were created by another person. One participant (20 per cent)
always re-used tags used by others, one (20 per cent) sometimes re-used them, and the
other three (60 per cent) rarely re-used the tags, one (20 per cent) used a mixture and
one (20 per cent) used plural. Again, there was no consensus over spaces, with two

Figure 5.
Fully expanded view
JEIM users (40 per cent) stating that they used them, three (60 per cent) stating that they did
25,1 not and the other using them and some not. Only one participant (20 per cent) used
synonyms of a word when creating tags in order to help them to retrieve the content in
future.
During the focus group participants said that they were not aware of the issues
associated with tagging. Many of the participants had only thought of tagging as being
88 of use for them personally and not considered that other people would be using these
tags to retrieve content. The participants also commented that many different systems
ask users to tag differently or do not give any guidelines as to how tags should be
entered into the system. All participants agreed that with the correct training and if all
participants were to make use of the same tagging scheme, then the benefits of tagging
would be far greater.
The questionnaire helped to identify the way that users were currently tagging and
through discussions during the focus group it was possible to expand upon the
potential issues with tagging. One of the group members stated “although there are
barriers towards tagging if they can be overcome then this could actually lead to
increased consistency, ease of finding docs/data” with another user stating that “after
training tagging might present a powerful option but that they were not aware of all of
the potential problems”. These issues highlight the need for training but also for a
system to remind users of the benefits of considering these options when tagging
systems are in place.

5.2 Assessing TagDav

To assess the potential of the TagDav system to help users obtain relevant
information, the second questionnaire (step three) was given to the group of nine
SoftwareCo employees. Some questions were left blank by some participants, and in
these cases only the completed questions were used in the results. This explains the
odd percentages for some of the results as the total number might not have been out of
nine participants. Once the members had answered the questions in the questionnaire,
a group discussion was held centred around the benefits and weaknesses of the system
and any potential it may have (step four) to reduce information overload.
When asked if they had difficulty finding files they have created because they did
not know which directory they have been placed in, 33 per cent of participants stated
often, 44 per cent sometimes had a problem but 22 per cent of participants rarely
experienced a problem. On locating documents, 22 per cent stated that they always
resorted to using the search facility of a file system, 22 per cent often did and 44 per
cent sometimes resorted to using the search facility, with 11 per cent never using it. Of
the eight participants that did use the search facility only one felt that it was rarely
sufficient to find the files they required. One of the participants always placed the same
file in more than one location or created symbolic links in order to make it easier to find.
Three of the participants sometimes did and one rarely did, with four participants
never doing this. This does highlight that 55 per cent of the participants, have at some
point, created a duplicate file in order to ensure it can be found in the future. Copying a
file to another location does carry a high level of unnecessary risk, if for example, the
file is updated in only one place. When participants were asked how they store their
files every response was different. Some stored files by project, some by topic, some by
year and some used completely different schemes entirely. With all of these different
storage methods, a collaborative file store could easily become quite difficult to use. Retrieving
There would be duplication of files and it could be difficult to find a file if it is not relevant
where it is expected to be stored. The responses confirm that there are a number of
issues associated with traditional file systems and storage of documents. information
The following responses relate to the TagDav system and its potential to improve
upon these issues. After using the system, participants were asked whether they felt
that tagging files would help them to find the files and prevent the need for them to 89
have to resort to using search engines. The majority said (67 per cent) that they
definitely felt the system would help them find files they were looking for without the
use of a search engine and 11 per cent stated that it often would. Two of the
participants felt that the system could help sometimes. When asked whether tagging
files would save time when it came to the retrieval of files, 78 per cent felt that it
definitely could and 22 per cent of the participants felt that it would often be possible.
The participants indicated that they would use this system to replace their current
way of filling and retrieval with two of the participants (22 per cent) saying they would
definitely use the system, 44 per cent stated that they would make the switch and 22
per cent stated that they would use the system sometimes. All of the participants felt
that the system had benefits over traditional directory based systems and all but one
(11 per cent) of the participants felt that browsing a file structure was harder than
uploading a file and tagging it. Participants were also asked whether they felt they
could save time using this system and all participants agreed they could. When asked
how long they could save each day using this system, participants were allowed to
enter any value of their choice, as no pre-defined options were offered. Answers ranged
from ten minutes to 120 minutes, but on average participants felt that they could save
40 minutes per day through using the TagDav.
One important factor to note is all of the participants felt training would be required
before it was used. Although most of the participants (56 per cent) felt it would only be
necessary sometimes, three (33 per cent) of the participants felt training would always
be required and one (11 per cent) felt that it would often be required before the system
could be used.
In step five, the participants were asked to record their thoughts about the
strengths, weaknesses, opportunities and threats of TagDav. Most of the participants
were impressed by the way the system allowed them to find a document. The system
was said to be very easy to use and was praised for its ability to allow participants to
find files from any number of approaches or viewpoints “files can be found from a
number of different approaches and multiple file representations will appear making it
easier to drill down and find documents”. One participant also stated that it would
allow users to “reach consensus about file system structure” the users also appreciated
the ability to create a “reduction in storage space required” as files would not need to be
saved in multiple locations. A number of interesting weaknesses and worries were also
raised. One of the worries was that a file with no tags would be “lost in the system” and
that if enough tags were not used then the file may be difficult to find. The key thought
from participants was that the system would have to be used with a lot of care and
“taken seriously” from the start and if it were not, there would be a risk of losing files or
having an “unstructured mess”. The participants felt training would certainly help, but
they still had concerns about the consistency that would occur if everyone using the
system did not fully understand the potential of the system and how to correctly tag
JEIM the files. The other worry was that some form of consistency would be required in
25,1 order to ensure that the tags used by some users would be retrievable by others. Some
participants also felt that once there were too many tags it might be difficult to drill
down and find the files that were required. However, as a group it was felt that this was
“more of a benefit then a disadvantage, normally you would not have the option of
drilling down further”. One of the participants also raised the issue of the time it would
90 take to store the documents, however, as a group they were not concerned about the
amount of time it would take to store the documents as in many cases “it takes ages to
find the directory to save something in now”.

5.3 Research synthesis

Documents stored on a computer can often be extremely difficult to find using
traditional systems and users filing documents are forced to consider where they place
files in order to be able to retrieve them in future. Nowadays, in some organisations the
consideration given to storing documents seems to be increasing, for example, users
having to determine how a document should be classified, categorised, named, expiry
date, and storage location. The problem is further compounded by the lack of training
provided to employees on how to store and retrieve information or the lack of
communication of potential best practices to improve operational practices (Rosacker
and Rosacker, 2010). Meglio and Kleiner’s (1990) research showed users of information
all contribute towards information overload. The results in this paper confirm this
finding by the wide variety of different filing methods users use to store their files,
which leads to issues retrieving relevant information. However, if training is provided
to overcome these issues, research has shown that the positive effects that training
might have had are considerable reduced 30 days after the training was provided
( Jackson and Culjak, 2006). Therefore any system that is introduced to replace the
current facility to store and retrieval information, needs to be proven to be more
efficient so users can see the benefit of change; easy to use with a small amount of
training; and have computer assisted structured rules base to ensure filing is completed
correctly. The research in this paper provided the basis from which the rules can be
constructed, namely, no single use tags – or tags which have not been used before
should be avoided unless necessary; decision should be made on if pluralised or
singularised versions of words; inclusion of a spell checker; removal of personal tags;
provide guidelines on spacing and capitalisation; inclusion of synonyms; and promote
the use of multiple tags. These rules can also help overcome some of the employees
concerns raised in step five of this research, regarding creating an unstructured filing
system using only tags, by introducing a computer assisted rules base helper at the
point of storing or retrieving information.
In terms of information retrieval performance, Dubie’s (2006) research showed
ineffective searches and wasting time looking for information costs up to 10 per cent of
an employee’s time. The results from this study confirm Dubie’s findings that there is
inefficiency in the system. By introducing a new way of storing and retrieving
information employees could save on average 40 minutes per day using TagDav. This
amount of time also includes the increased time employees would have to spend
tagging files to be stored. Therefore TagDav could save up to 8 per cent of an
employee’s time by effective searchers based on an 8 hour working day, providing an
80 per cent efficiency improvement based on Dubie’s research figures.
 Overall, as a proof of concept, this research has shown that tagging is more effective Retrieving
than a traditional filing system, and it has both added to and confirmed the wider relevant
findings found in the literature surrounding effective searching. In particular, this
research has added to Meglio and Kleiner’s (1990) research by showing that the wide information
variety of different filing methods users use to store their files leads to issues retrieving
relevant information. In addition, information retrieval performance can be enhanced
by using tagging system and this alternative approach capitalises on the deficiencies of 91
a traditional filing system (Dubie, 2006), which leads to a more cost-effective approach
to filing and retrieving relevant information.

6. Conclusions and limitations

The study examines the role of tagging, in a business context, on discovering relevant
information when compared to traditional hierarchical filing systems. Within the
broad problem domain of information storage and retrieval, the main research question
of this research was: “Can tagging files save employees time retrieving information
compared to a traditional hierarchical file system?”. An interpretive philosophy was
chosen as the underlying research framework with focus on literature from information
retrieval and information overload theory. The study has shown that users use a wide
variety of different filing methods to store their files, with every participant providing
a different answer, for example, storing them by content type, project, date customer
and many others. In trying to improve information storage and retrieval this study
introduced TagDAV, a proof of concept tag-based file system, to replace traditional
hierarchical file structures that could be used by both individuals and groups. The
results were very positive towards the new method of filing information as:
.
67 per cent of the participants said the system would definitely help them find
files they were looking for without the use of a search engine;
.
78 per cent of the participants said it definitely could save time when retrieving
files;
.
on average they would save 40 minutes per day using the TagDav, even though
more time would have to be spent upfront tagging files to be stored; and
.
a system such as TagDav could make up to an 80 per cent efficiency saving in
employees efforts to store and retrieve information.

The practical implications of this research are that information rich enterprises, similar
to the one studied in this research, should evaluate the functionality of their chosen
operating system and information store software in light of the potential benefits
offered by tagging, and costly limitations of traditional file stores. The evaluation will
provide information rich enterprises with the “best practice” when it comes to
information storage and retrieval. It can be concluded that the TagDav approach
provides a mechanism to aid in the storage and retrieval of business information which
can lead to generating greater cost effectiveness and increased performance of
information rich enterprises.
A limitation of the study is TagDav was a proof of concept and the employees were
not able to use it on a daily basis to see if it really made a difference to their information
capabilities. A further limitation is the number of participants and any further research
in this area would benefit from a larger group of participants. Overall, the study has
JEIM shown that traditional filing systems should be reviewed in the light of the potential
25,1 benefits offered by tagging and the costly limitations of traditional file stores.

References
Burstein, F. (2002), “System development in information systems research”, in Williamson, K. (Ed.),
Research Methods for Students, Academics and Professionals: Information Management and
92 Systems, 2nd ed., Centre for Information Studies, Charles Sturt University, Wagga Wagga,
pp. 147-58.
De Kunder, M. (2011), “The size of the world wide web”, available at: www.WorldWideWebSize.
com (accessed 10 May 2011).
Dubie, D. (2006), “Time spent searching cuts into company productivity”, Network World,
available at: http://www.networkworld.com/news/2006/102006-search-cuts-productivity.
html (accessed 10 May 2010).
Farhoomand, A.F. and Drury, D.H. (2002), “Managerial information overload”, Communications
of the ACM, Vol. 45 No. 10, p. 127.
Flickr (2006), available at:www.flickr.com/photos/tags/ (accessed April 2006)
Golder, S.A. and Huberman, B.A. (2006), “Usage patterns of collaborative tagging systems”,
Journal of Information Science, Vol. 32 No. 2, p. 198.
Hayman, S. and Lothian, N. (2007), “Taxonomy directed folksonomies”, available at: http://
archive.ifla.org/IV/ifla73/papers/157-Hayman_Lothian-en.pdf (accessed January 2011).
Hiltz, S.R. and Turoff, M. (1985), “Structuring computer-mediated communication systems to
avoid information overload”, Communications of the ACM, Vol. 28 No. 7, pp. 680-9.
Jackson, T.W. and Culjak, G. (2006), “Can seminar and computer-based training improve the
effectiveness of electronic mail communication within the workplace?”, in Spencer, S. and
Jenkins, A. (Eds), Proceedings of the 17th Australasian Conference on Information Systems.
Karr-Wisniewski, P. and Lu, Y. (2010), “When more is too much: operationalizing technology
overload and exploring its impact on knowledge worker productivity”, Computers in
Human Behavior, Vol. 26 No. 5, pp. 1061-72.
Keller, P. (2008), “deli.ckoma.net Del.icio.us stats”, available at: http://deli.ckoma.net/stats#tab_
tags (accessed December 2008).
Kelly, D., Fu, X. and Shah, C. (2010), “Effects of position and number of relevant documents
retrieved on users evaluations of system performance”, ACM Transactions on Information
Systems (TOIS), Vol. 28 No. 2, pp. 1-29.
Kerr, E. and Hiltz, S. (1982), Computer-mediated Communication Systems: Status and Evaluation,
Academic Press, New York, NY.
Kirsh, D. (2000), “A few thoughts on cognitive overload”, Intellectica, Vol. 1 No. 30, pp. 19-51.
Kobayashi, T., Misue, K., Shizuki, B. and Tanaka, J. (2006), “Information gathering support
interface by the overview presentation of web search results”, Proceedings of the Asia
Pacific Symposium on Information Visualisation, Vol. 60, pp. 103-8.
Lux, M. (2010), “Flickr uploads per minute – mean uploads per hour”, available at: www.
semanticmetadata.net/2010/03/12/flickr-uploads-per-minute-mean-uploads-per-hour/
(accessed 10 April 2011).
Mathes, A. (2004), “Folksonomies-cooperative classification and communication through shared
metadata”, paper presented at Computer Mediated Communication, LIS590CMC (Doctoral
Seminar), Graduate School of Library and Information Science, University of Illinois
Urbana-Champaign, Urbana, IL, December.
Meglio, C. and Kleiner, B. (1990), “Managing information overload”, Industrial Management Retrieving
& Data Systems, Vol. 90 No. 1, pp. 23-5.
Nelson, M.R. (1994), “We have the information you want, but getting it will cost you!: held
relevant
hostage by information overload”, Crossroads, Vol. 1 No. 1, pp. 11-15. information
Nunez, A.N. and Giachetti, R.E. (2009), “Quantifying coordination work as a function of the task
uncertainty and interdependence”, Journal of Enterprise Information Management, Vol. 22
No. 3, pp. 361-76. 93
Rosacker, K. and Rosacker, R. (2010), “Information technology project management within public
sector organizations”, Journal of Enterprise Information Management, Vol. 23 No. 5,
pp. 587-94.
Smith, S. (2010), “Reducing information overload by optimising information retrieval
approaches”, PhD thesis, Loughborough University, Loughborough.
Soucek, R. and Moser, K. (2010), “Coping with information overload in email communication:
evaluation of a training intervention”, Computers in Human Behavior, Vol. 26 No. 6,
pp. 1458-66.
Wang, Y. and Forgionne, G. (2008), “Testing a decision-theoretic approach to the evaluation of
information retrieval systems”, Journal of Information Science, Vol. 34 No. 6, pp. 861-76.

Corresponding author
Thomas W. Jackson can be contacted at: t.w.jackson@lboro.ac.uk

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Or visit our web site for further details: www.emeraldinsight.com/reprints