Does tagging lower costs of classification, but raise costs of finding information?

Gene Smith riffs on Ian Davis' take on why tagging is expensive. In a nutshell, the lower costs of classification are traded for higher costs in finding content. Gene makes much more sense of it than just that, though. At the end of the day, as with so many things, tagging is a great tool, but not a silver bullet.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Interesting But Missing Some Key Distinctions

"you end up with a lot of people paying a tiny cost to classify added to a lot of people paying a high price to discover."

As a librarian and web developer, I whole-heartedly believe that formal taxonomies, such as MeSH, reduce discovery costs for our users. But it is also clear to me that tagging systems will continue to play a key role in our users' personal information spaces. That is, tagging utilities generally serve as vehicles of personal re-discovery of information. This purpose fundamentally differs from that of something like MeSH, which serves to provide access to a body of knowledge for all users. My tags, of course, will be steeped in the context of my own information needs, and I would certainly not suggest that we use them to provide access to health sciences literature. They do, however, function very well for me as an efficient system of personal re-discovery.

Further, we cannot assume that all taxonomies will inherently reduce the work of users. I gave the example of MeSH, which is a truly advanced taxonomy. But think of your local office network drive. How findable, really, is your documentation? A poorly implemented taxonomy can be as bad or worse than tag soup because a large amount of valuable content may lie completely hidden under one branch of the taxonomy, whereas with tags a few useful items might slip through (ironically, this would happen because of a lack of standardization ;) ). Moreover, if I have the ability to tag relevant documents for myself, I would have an even better chance of retrieving it.

And besides, (here comes the kitchen sink part of this line of thought), even MeSH has some major problems. Try doing research in bacteriology, for example. Even after all that work, we are still somehow unable to provide reasonable and reliable access to knowledge in this area. More limitations of controlled vocabularies emerge: they are slow to deal with change and often leak at the seams, sometimes badly.

As with any discussion that involves formal systems of classification vs. informal systems, the key point is not so much "if" but "where" and "how." If we are to advance the conversation in this area, we need to begin speaking in terms of use-case scenarios rather than in terms of either/or. Those researchers in bacteriology, for example, might use a combination of controlled vocabularies and informal personal rediscovery (think CiteULike) to manage there research process.

Ahem, that being said, I did find the cost-benefit angle an intriguing one. Thanks for the post.


Rediscovery cost trends with tag volume

Hey Chad,

Thanks for your thoughtful post. One concern I see with the personal rediscovery angle is the sheer volume of personal tags over time - both number of tags and amount of tagged content eventually gets too big to easily refind - taking us to a place where tagging and search need to work together.