| |
By Fran Alexander
 Folksonomies: Business Use
Written by Fran Alexander.
Folksonomies have become very popular on content-sharing websites such as Flickr and Delicious, with their advocates claiming that they provide a free content organisation system that enables anybody to find anything any way they want. However, folksonomies only work well under certain conditions, and not in all business contexts.
Precision and recall
Folksonomies are a cheap way of getting metadata added to content, but they are cheap because there is no quality control, meaning that the metadata will not necessarily be useful, sensible or consistent. So folksonomic metadata is unsuitable if precision and recall are important, or when accurate item management is needed.
Websites like Flickr do not have to worry about precision and recall - if some of the photos on Flickr are never found by anyone else, nobody cares. No-one is going to stop using Flickr simply because erroneous, humorous or malicious tags lead to irrelevant search results. No-one is going to demand that Flickr produces a comprehensive and definitive list of every single photo it holds of a particular subject.
In contrast, precision and recall are important in many business settings. A folksonomy would make a poor stock control or digital asset management system. If two people tag the same component with different names, a stock controller needs to check both names every time they assess stock levels. They would need a list of all the tags everybody had used for each item, and the process of repeatedly compiling and checking such a list would be far less efficient than imposing a fixed system like a controlled vocabulary in the first place. Barcodes are not fun or user-friendly, but they are an extremely precise and accurate way of tracking items across all the companies in a supply chain.
Customers are not stock controllers, however, so a folksonomy that sits alongside the stock control system, for tagging products on a shopping website, for example, would be a positive feature. Folksonomies help customers personalise their shopping, allow them to use their favourite words, and possibly even help them get more involved in the shopping experience. Such tags are no use for processes like stock control, but might provide interesting insights into customers' viewpoints for marketing purposes.
If comprehensive recall is important, folksonomies alone are unreliable. For example, an accountant wanting to be sure they have checked all relevant financial legislation for a client should demand a more thorough search than simply hoping that someone happened to tag up the relevant documents with an obvious term. However, staff using an Intranet (rather like external customers) may appreciate being able to add their own personal folksonomic notes to documents. Provided that the folksonomic tags are clearly differentiated from "official" tags, users can choose from the authoritative taxonomy or the loose folksonomy, depending on their own priorities and concerns.
Business contexts where precision and recall are not important tend to be in less process-critical areas or where individual content items are not business critical, such as wikis, blogs, or corporate social networking sites.
Characteristics of folksonomies and taxonomies
Folksonomy
|
Taxonomy
|
anarchic
|
authoritative
|
cheap
|
costly
|
difficult for subject novices
|
good for teaching
|
easy to implement
|
requires trained staff to implement
|
fluid
|
stable
|
fun
|
formal
|
good for customers
|
good for specific job functions
|
haphazard recall
|
comprehensive recall
|
innovative vocabulary
|
established vocabulary
|
likely to get irrelevant results
|
less likely to get irrelevant results
|
messy
|
controlled
|
order expected to emerge
|
order imposed from above
|
personal viewpoints
|
corporate viewpoint
|
prone to error, malice, humour
|
quality controlled
|
responsive to new terms and changes
|
slow to respond to new terms and changes
|
supports niche views
|
supports consensus view
|
term-driven navigation
|
browsable navigation
|
unreliable for asset management
|
reliable for asset management
| It is important to consider what sort of tagging will feed a folksonomy and how people will perceive the benefits of participating. A folksonomy will work better when authors tag their own work, but less well when the potential taggers do not know much about the subject matter and do not have a clear idea of what content is contained in a repository. If the content is being used for training purposes or to aid knowledge discovery, a taxonomy can help inexperienced users gain an overview of what is there.
Folksonomies tend to be mixtures of self-orientated tags and truly social tags. Self-orientated tags are not primarily intended for anyone else to read. They include judgements (cool, interesting) or personal notes (my dog, Dad's birthday) that are great for the individual tagger, but typically not very useful to other people. Social taggers want their content to be found (people posting videos to You Tube generally want other people to watch them), so make an effort to use tags that are clear and appealing to a broad audience. However, because social tagging is difficult to do well, people do not tend to bother unless they can see some sort of reward for making the effort. So, people will put time and effort into making sure their LinkedIn profile is tagged well and is therefore findable by potential employers, but they do not generally bother to tag books on Amazon.
When self-promotion is the motivation, one tagging hazard is that the taggers will simply label their content with whatever seems to be the most popular tag at the time, to try to force their content to appear high up in search results. Such misleading tags spoil the usefulness of the folksonomy for others. To remove such spurious tags requires staff intervention, which is costly, and raises the question of whether it is sensible to employ 'tag police' to tidy up a folksonomy instead of employing indexers to produce a high-quality set of metadata in the first place.
However, without some motivation, there may not be a big enough pool of prolific taggers to fuel a folksonomy. In many companies, most content is created for the people who need to see it as part of a workflow (e.g. management reports) by authors who neither want nor expect it to reach a larger audience. Folksonomy evangelists claim that free tagging is so quick and easy it is practically effortless, but it is not completely effortless, and even the tiniest amount of 'cognitive friction' - having to think for a second or two about a tag or typing a few extra characters - is enough to put busy people off.
Appropriate rewards to encourage tagging will depend on the type of content and the corporate culture. If people feel that they are spending a lot of time tagging, but others are benefiting more than them, or that they are adding helpful tags and other people are adding self-promotional or silly tags, they may become disillusioned and the folksonomy will stagnate.
Different viewpoints
Folksonomies are promoted as a way of allowing for the expression of diverse viewpoints, but this is neither guaranteed nor predictable. If the taggers have very different outlooks, vocabularies and experience, a communal perspective may not emerge. Tag coverage may be patchy and the same tag may have been used to mean many different things. Conversely, a single dominating consensus may overwhelm any diversity of views.
When folksonomies become very large, what amounts to proto-taxonomic hierarchical groupings may either emerge or need to imposed to improve usability. For example, Flickr has added clusters, or categories, to make its huge volume of tags more navigable. With small groups of taggers, idiosyncrasies are preserved, but if those idiosyncrasies are unhelpful, they will have a disproportionate effect (e.g. a lone but prolific tagger who cannot spell).
Small groups can produce useful folksonomies if they are working in a niche, emerging, or rapidly changing subject area. Such groups are often committed to developing their own specialised indexing resources and find that more formal systems are slow to adapt. This means that folksonomies can work very well within innovative research groups or small project teams. In this context especially, folksonomies are a wonderful resource for taxonomists - suggesting new terms and highlighting areas that may need attention in formal systems.
Best of both worlds
The purpose of any knowledge organisation system should be to make it easy for people to find things, and so the more options and flexibility provided, the better. Folksonomies are fluid, responsive, and encourage personal expression, so for niche groups, customers, non-critical content, and for fun, they have much to offer.
They are an 'anything is better than nothing' option if there is simply no budget, especially if content is in an otherwise hard to search format, such as audio or video. As a supplement to authority-based systems, such as taxonomies, folksonomies are a useful tool with some unique properties, but there are many business contexts where they are too unreliable to replace formal systems altogether.
Suggested contexts for using folksonomies and taxonomies
Function
|
Folksonomy/
personalisation
|
Taxonomy/
controlled vocabulary
|
asset management
|
no
|
yes
|
corporate communications
|
yes
|
yes (for navigation)
|
customer-facing sites
|
yes
|
yes (for navigation)
|
fun content
|
yes
|
maybe
|
help and technical support
|
yes
|
yes (for knowledge discovery)
|
high-value content
|
no
|
yes
|
legal documents
|
no
|
yes
|
must-find content
|
no
|
yes
|
niche groups
|
yes
|
maybe
|
records management
|
no
|
yes
|
self-contained project teams
|
yes
|
maybe
|
social networks
|
yes
|
maybe
|
stock control
|
no
|
yes
|
training resources
|
yes
|
yes (for knowledge discovery)
|
user-generated content
|
yes
|
maybe
|
wikis
|
yes
|
maybe
|
By Fran Alexander
Fran Alexander's interest in taxonomies, folksonomies, thesauri,
controlled vocabularies, and knowledge organisation evolved through her
work as an editor of dictionaries, encyclopaedias, thesauri, and
almanacs. In 2001, she joined Keesing's Worldwide, where she managed
the creation of the new online news archive, incorporating both
folksonomic and taxonomic features, before becoming editorial director
in 2007. She is currently studying for a Masters of Research degree at
the Department of Information Studies at University College London and
blogging at http://www.vocabcontrol.com
FUMSI articles by Fran Alexander »
![[Get Copyright Permissions]](http://license.icopyright.net/images/icopy-w.gif "[Get Copyright Permissions]")
Click here for copyright permissions!
Copyright 2010 Free Pint Limited
Related articles:
You may also be interested in:
|
