Where Tagging Works and Where Tagging Doesn’t Work – Search Engine Lowdown

Here are two posts on tagging, one Where Tagging Works and another where the author feels it doesn’t do well, Tagging Not Likely The Killer Solution For Search. Thanks to Amanda Watlington for pointing them out. Let’s start with the negative one. Danny Sullivan says the all the hype about tagging is, “ignoring the fact that we've had tagging on the web for going on 10 years, and the experience on the search side is that it can't be trusted.” He points out that Google has not picked it up, likley because of the past misuse.

Now to be fair, I heard from the foudner of del.icio.us at a Berkman meeting that it was not designed to develop folksonomies, “bottom up” key words. It was a social bookmarking tool, a way tha people can store and share links to stuff they feel is important. It was not designed to augment search by providing new key words.

The key word “web 2.0” gets 75,900,000 hits on Google but who goes beyond the first few pages. The list leads with the famous O’Reilly article. Del.icio.us said there were 0 results for the same term which seems weird, must be some error. So I tried “knowledge management” in del.icio.us and got 2272 tagged items and a list of related terms. In Google I got 73,800,000 hits. Only one item was in the top ten on both results, The Knowledge Management Resource Center. But to comapre the two tools is not really relevant as they have very different functions.

I guess I tend to agree with Danny Sulliivan about the tagging and search but that is not the original intention of tagging. If I want to search on a key word, I will still go to Google as the most efficient way. If I have the time to go exploring through multiple links and see the interrelations between key words, I might go to del.icio.us. However, if I want to set up a way to store and share links on a particular topic, I will use del.icio.us which I have done already in co-authoring an article.

Now where does it work according to Search Engine Lowdown. It works at a site, Millions of Games, because “it's narrowly focused on a specific topic and has a large degree of agreement among its user community.” Well this was exactly the situation when my co-author and I used del.icio.us to store secondary research for an article. I think that there will be some very focused uses of tagging emerging as people explore its possibility. The H20 Playlist created by Harvard’s Berkman Center is one example. enterprise tagging

Comments

[Lumpy]

I read all of those articles you mentioned even before I read your post. I recall finding many of the links in an earlier post(s) of yours.

I don't think that tagging will be a replace for search but, after I read this Slacker Manger Article

http://slackermanager.com/2005/12/the_several_hab.html

I found my self searching much, much less. I found myself learning much more as well. It even inspired me to post this

http://lumpyscorner.com/TidBits/?p=26

Search, Knowledge Management and tagging are related but each one is very different. I don't think tagging will replace search but I feel it fits into the concept of knowledge management much better than search. Searching is, maybe... not certain here, more for Data Mining.

Also, specialized search engines have existed for a good number of years. Tagging should work there because of a narrower scope and fewer key words. The down side of folksonomy is the dynamic nature of it. Are they feeds or RSS, Knowledge Management, KM or PKM? Will it still be called data mining ten years from now? For that matter, what was it called ten years ago?

Nonetheless, examining the tagging behavior of people, both individually and collectively, will certainly improve more general search. Folksonomy certainly help search engines create a better "find similar items" button.

[Bill Ives]

Lumpy - Thanks for your comments. I agree that tagging is more related to knowledge management than search. I had intuitively thought that way but you comment articulated it for me. Bill

[Stephanie Lemieux]

You may be interested in the discussion that ocurred on the Taxonomy Community of Practice (http://finance.groups.yahoo.com/group/TaxoCoP/) on the value of social tagging in a corporate setting, which is replicated on the community's wiki:
http://taxocop.wikispaces.com/Social+tagging

More specifically, as an example of where tagging does work, have a look at the sections regarding Raytheon's hydrid approach. They use the social tags to keep abreast of new vocabulary and enhance their more stable taxonomy.

Stephanie

[SteveEisner]

Hi Bill - I've been writing a series of articles related to this exact point. Tony Karrer's link to your article at http://elearningtech.blogspot.com/2006/05/tags-search-effectiveness-personal.html
helped me reframe my thinking a little regarding the relationship between tagging & search.
So here's a manual trackback to my own thoughts :)
http://www.ihol.org/blog/index.php/2006/05/23/tagging-vs-searching/

BTW while doing web research for my blog your space comes up a lot, thanks for all the useful articles!

[Bill Ives]

Thanks to Steve and Stephanie. I will be certain to follow up on your links. Bill

[Andy Havens]

I was dismayed that you coudln't find any del.icio.us "web2.0" tags... because I've got a bunch of things tagged thusly. So I went to del.icio.us and did a tag search for tags with "web2.0" and came up with...

nada. Very weird.

So I pulled one of my own tagged items with the "web2.0" tag, and then clicked on "view/all" at the top of the screen and got... a lot. Del.icio.us doesn't show total number found, but I listed 100 per page and stopped bonking after 12 pages worth.

So, there's a problem with the tag search function, apparently.

As to the meat of this post... search is one thing, tagging is another. Which goes without saying, but people forget it and try to use the two things towards the same means. The funny thing is, Google's search algorhithms try to automate a function that widespread tagging through services like del.icio.us is beginning to make available directly; the aggregation of attention.

Google ranks links very highly when determining which sites to list first in a search. Links represent a choice of attention; I link to your site, therefore, I have paid it attention. Problem is, a link is often undifferentiated. And so Google has to go through a bunch of equations to ring out the reason for the link; is it to a similar site? Was the link recent? Do other similar sites also link similarly? Is the link a posible "fake?" All kinds of guesses. And we all know we've done searches on Google on one kind of word pattern, only to have to try again and again for that special "combo" that will get you close to the type of stuff you want, that isn't polluted with "nearby" info or heavily advertised or versioned or poisoned info.

With tagging, yes... there are inaccuracies and vagueries. I use the tag, "book" for a site about books. You use, "books," or "read," or "reading" or "authors," or "fiction" or "toread" or "writing." All kinds of crud. But you get the SAME CRUD with search terms... just on the front end of the equation, rather than the back. It's just with tagging (at least on del.icio.us), you can "see the math."

What do I mean... I mean that if I use it to search for items tagged with "read," I can immediately see what other tags people have used along with "read;" books, literature, reference, reading, education, copyright, learning... etc. On Google... nothing helpful to even narrow (or expand, or differentiate) my search.

OK... But that's a really vague, wide term. Isn't it? What do we expect to get from "read?" Well, at del.icio.us, the first hit is for the site "What should I read next?" http://www.whatshouldireadnext.com/ Which is a pretty freakin' good site to come up with just on that one word. #2 is http://reader2.com/

What are Google's firs hits for a search on "read?" #1 is Today's news from MSNBC.com. Hunh? OK. I guess I would *read* that... but I read all news online. Why is MSNBC #1? #2 is http://www.readwriteweb.com/ Read/Write the web. #3 is BookCrossing, which is a free book-club. So that's about reading, OK.

As I go down the top 10 list for each, I find about 1/3 as many useful, "reading" related sites on Google.

Why? Because the word "read" doesn't necesarilly have much to do with the desire to perform the action. The sites most appropriate to that verb may have more to do with subjects and objects of its action.

I'm not saying tagging is the "death of search," but I've found that once you start using del.iciou.us and get used to the quirks and personality issues that come with referring to how *people* rank and judge sites rather than how a machine does it... you get not just better results, but deeper and more interesting and fun results. It's especially interesting to go through the tags of one user once you've found a really spot-on site from them. It's like finding a lost friend and then picking his pockets for notes.

[Bill Ives]

Andy - Great comment. Thansk for the time to do it. I agrre with what you said but I would still use Google for quick searches and del.icio.us when I have the time to explore a topic in depth and see what others are reading on it. There is another tool, iQuest at iquestglobal.com that combines search with social netwrok analysis. Google does a type of social ntwrok analysis but keeps that part secret to protect their algorithmns. iQuest does the search and keeps the social network part open. It is also much harder to try to trick it through link farms, etc. I should add that I have a relationship with the company but also feel it brings a new dimension to search. Bill

[John D. Lemke aka Lumpy]

First thanks for the nice words regarding my comment. Seems like my cocomments plugin might be missing some things for I have not gotten any data on the newer and good comments on this topic. (I just happened to notice traffic from this post reviewing server logs.) I was curious to see if it might be something to do with loggin into type key so, this time, I didn't log in.

I noticed a comment on the delicious tag and web2.0... I noticed a similar issue. I have been reading and digging a lot on this topic (tagging not web2.0) of late. Yes, there seem to be bugs. I would rather be optimistic on it. Things will catch up.

The whole mob rules mentality of folksonimies has one real thing going for it though. There is no mob unless enough agree. Search is directed at some group of people trying to please the searcher. As you pointed out, folksonomy is bottom up. The mob is telling others what they think the term should be.

I confess, as of now, I am a tagaholic. Search, in the end, is meant to find knowledge and information. after just a few weeks of using tags, networking and the inbox of delicious, I feel it is much more efficient than search.

There is also another major advantage of services such as delicious. One can easily create a novel tag to take advatage of the RSS feed to easily share information and knowledge. My hobby podcast at radiostatic.am is an example. We use the delicous service and a simple RSS to php script to generate entire pages that list the music we have played, the artists we have played and sites mentioned on the podcast.

I have been and will be writting more on this at my TidBits site. I honestly wish all of you would stop by and share your opinion on my posts. I am not here to create traffic... I want the feedback and conversation. We are living in a golden time for the sharing of information. I do not know where it is going to end up but it certainly is wonderful to be here to see and experience it!

Two more things...

Bill, Keep writing it is great stuff.

Steve and Stephanie, I am surfing off to your links next. Thanks everyone.