In a recent NetworkWorld piece, Gibbs wrote about the tagging meme, and where it apparently sits on the technology life cycle. No new insights for me there (but possibly fits the CEPP rule for others); I was involved in a number of knowledge management (KM) projects back in my Monsanto days (IAPL) [note to self: too many acronyms, hhh] and we hit many of the classic walls;
- CRM systems that failed because sales reps guard their customer intelligence
- Collaboration spaces that failed because corporate and international culture equated “asking for help” with “weakness”
- Document management systems that failed because arriving at consensus around a taxonomy is difficult
Ok, well, I guess failure is too harsh in all of those cases – it’s just that the ultimate deliverables got toned down a lot, because we were oversold on the value of things (like ubiquitous tagging) without appreciating the difficulty of implementation (ie. you say poe-tay-toe, I say poe-tah-toe).
Even in those days (mid-90’s), I never really understood why folks got all fired up about keywords for categorizing documents; Google and Yahoo and other online knowledge bases showed that searching for words and phrases was plenty effective. No need to browse tables of keywords looking for articles by category – most of the time, folks were looking for a needle in a haystack and needed the right water tank.
Besides, MS Office documents have supported document metadata for eons, but few even know it exists (try File, Properties) … how many Word documents have you received (or created!) where the Author was set to “Corporate User”?
I think the fundamental flaw here is the premise that tagging is done for other people’s benefit; my tagged content can be useful and relevant and helpful for you. The truth, however, is that it’s very difficult to control and/or predict how other people receive your message and understand your meaning. Again – why is Search often referred to as the next Killer App? Because the better engines pull commonalities from the content, and don’t rely solely on tags / categories.
McDonald writes about a personal bookmarking strategy and calls out some of these same ideas – a large chunk of the knowledge gathering and retention work we do is for our benefit, not for sharing. Nielsen has a nice article that calls out how a small percentage of users on a given website account for most of the activity, while the majority are lurkers, looking for information for themselves, not sharing. This is not a bad thing or a good thing, it’s just a thing; the foundation behind the Law of Large Numbers.
So why tag? Well, it’s actually a fairly common productivity device – the manila folder, the cool labeler that all the data center guys like to play with – it’s human nature, we just like to categorize. I’m actually a big fan of tagging, but for very selfish reasons; my tags become useful KM tools for me across all of the various sites, services, and tools that support the idea. “Selfish KM”, or “Web 1.9” – a slight step backward from the utopian view, maybe a bit more pragmatic.
I’ve launched into a mini-project to standardize tags for my del.icio.us links, blog content, and other stuff. A real hassle, for some of the reasons Gibbs calls out, especially when it comes to “syntax”; I use TiddlyWiki for my personal wiki, but it can’t handle tags made up of separate words, so I have to allow for CamelCase – ugh. I will, however, document my cross-platform (sic) tagging standards there, and slowly push them out to my other public and private knowledge bases.
I think the death of tagging has been greatly exaggerated, it’s just moving to a more pragmatic place in our KM toolbox.