Tagging offers a potentially powerful way for a company to organize information by making fresh content immediately searchable, letting users designate terms that make sense to them and providing users with a sense of ownership. This ability for tags to provide so much content-describing power for ordinary folks has given rise to the term “folksonomy,” as opposed to the more restrictive sounding “taxonomy.”
But like taxonomies, tags are all about finding data. “It’s another tool in the toolbox” for CIOs, says information architect Louis Rosenfeld. Rosenfeld notes that companies typically organize meta-data around attributes and values. Taxonomies often handle attributes well; a corporate library, for instance, can be organized quite well around book titles and authors. But if you want to search on a value—a book’s topic, say—things get harder. Searching through nontext, such as video, can also be a challenge.
Given their information density, Rosenfeld thinks intranets will be a prime testing ground for tagging at the corporate level. One company that has seen encouraging results using tags is IBM. “Tagging makes it easier for you to go back and find something,” says Maria Arbusto, IBM’s director for user experience who is responsible for how IBM presents its internal information, websites and applications to employees.
Arbusto says IBM is “still in the early days” of using the terms employees provide to improve discoverability. She says it has worked well in a pilot involving ThinkPlace, the intranet application IBM uses as an internal suggestion box for ideas the company should consider commercializing or developing and deploying to employees. In the system, employees can comment on the ideas and rate whether they should be pursued.
ThinkPlace originally classified ideas using terms from IBM’s official taxonomies for content such as industry and products. But “we observed the users and saw that the terms they used didn’t always match” the formal taxonomy, she says. So IBM created a way for users to enter keywords, or tags, that would be appended to the suggested terms from the formal taxonomy and thereby improve their ability to find relevant ideas. The results have been promising, says Arbusto. “You can see what your colleagues are interested in,” she says. “From a collaboration and knowledge-sharing perspective, that’s what’s neat about folksonomies.”
Tags are synonymous with the keywords familiar to anyone who has done a search either on the Internet or in a corporate content management system. Indeed, many current applications come with tagging tools that let users append descriptive terms to their documents. (The upcoming Microsoft Vista operating system will even include tagging as a part of its file system.) Another valuable element of tagging is that it works with any kind of file, including video or audio. Users can simply add descriptive terms such as “exterior,” “building,” “blue,” “quiet” and so on. This flexibility makes tagging “a very pragmatic technology”—simple to understand and use, says Andrew Jaquith, a Yankee Group analyst. No taxonomy can come up with every term employees might have for something. But with tagging, users gain the flexibility to work outside the taxonomy.
Thomas Vander Wal, who coined the term folksonomy and is founder and senior consultant for InfoCloud Solutions, says he thinks that in the next couple of years, companies will adopt tagging widely. “Having a folksonomy means you can use people to fill in the gaps in a taxonomy and track emergent vocabularies,” he notes.
Tagging for Dollars
Such flexibility is what is likely to make tagging useful to companies, says David Weinberger, a research fellow at Harvard Law School’s Berkman Center for Internet & Society. “It’s so easy, so cheap, and the benefits are immediate,” he says.
Tagging is already spreading quickly through the Web. The initial ideas are fairly old—the Bitzi online tagging tool (www.bitzi.com) has been around since the year 2000—but tagging as a phenomenon didn’t start until mid-2004, with the rise of sites such as Flickr, where users upload and tag photographs, and Del.icio.us, where people tag webpages. (Both Del.icio.us and Flickr were acquired by Yahoo in 2005.) Such sites have already shown that tags can be effective even if only small percentages of users adopt them. In fact, Caterina Fake, cofounder of Flickr and now director of technology development at Yahoo, says that a typical group probably needs fewer than 15 percent of its members tagging something for the tags to be useful in helping all members find things, and that number could be as low as 1 percent or 2 percent, depending on the size of the group.
Despite the amount of hype surrounding tagging, it pays to remember that it is still in its very early stages. “We’re in the voyeurism phase here,” says Greg Blonder, a venture capitalist at Morgenthaler Ventures. And he doesn’t believe all the early tagging leaders will stick around (he tags things like Flickr and Del.icio.us as “fads”).
But Blonder does think tagging is useful. Morgenthaler has even invested in a company called Digital Railroad, which takes advantage of already-defined tags used in commercial photography to help photographers sell their images.
Tagging is also young enough that serious disagreements still exist within the tagging community about something so basic as how to use it. For instance, should museums let people tag a painting “beagle” if it in fact depicts a dalmatian?
On a more practical level, “Tags really don’t fit well with the way corporations tend to organize information,” notes Jaquith. “Organizations like file folders, nested hierarchies and those ways of classifying things.” But, he says, few companies do a good job of organizing information in this way, and some may find tags offer a more commonsense approach.
Can You Tag Your Enterprise?
No one sees companies abandoning workflow tools, however, in favor of tagging alone. For one thing, most current tagging tools are not available commercially—though there is interest. Chris Fralic, vice president of business development at Del.icio.us, says that a number of companies had contacted the company before it was purchased by Yahoo, looking for a corporate version of its tagging tools.
Yahoo offers Del.icio.us and Flickr and has some beta offerings in the works, and Del.icio.us has inspired other things that might help CIOs. For instance, Jonathan Feinberg, an advisory software engineer at IBM, saw Del.icio.us and decided he wanted a version for the host of bookmarks he has on IBM’s intranet, so he built a program he calls Dogear.
It functioned so well that IBM’s CIO Bob Greenberg designated Dogear as part of the company’s Technology Adoption Program, which IBM uses to help it leverage good ideas from research. Dogear was opened for use across IBM in November, and a mere 1,235 of IBM’s 329,000 employees have logged in to the tool more than once.
But Feinberg and his manager, David Millen, an IBM Research scientist, have already refined Dogear, giving it privacy designations (for instance, for those bookmarks people want to keep to themselves). And they’ve added Really Simple Syndication to let people know when content has been “dogeared.” What the company doesn’t know is how many users you need to make Dogear worthwhile. That will become clearer in 2006, says Millen, as several IBM customers are expected to try Dogear.
No matter the company’s size, there will be management issues. Yahoo’s Fake says the tendency within a company is to think that social software is driven by users and thus self-managing, but this is often not the case. “Somebody needs to take responsibility for [the systems], and that’s something companies don’t understand,” Fake says. “They need a community manager.”
Fake says all sorts of companies will find that tagging helps them define their actual culture. She talks about the value of the “cool lamp” test—if you type “cool lamp” into Google, the first thing that comes up is a lava lamp, which to her is decidedly uncool. But in a group she’s involved with on the tag-driven site Yahoo MyWeb, typing in “cool lamp” generated results that she thought were very cool indeed. Similar technology could prove very valuable in corporations, for instance, where people in research and others in sales might have very different opinions about, say, what kinds of “product details” are interesting, with one group concerned more about material lists and manufacturing constraints and the other interested in profit margins and sales incentives. Companies could also test product concepts by letting users tag them and see what terms they use.
Whether tagging is a poor man’s database search tool, a quick-and-dirty search tool on an intranet or an inherent, ground-up part of corporate collaboration doesn’t really matter. What does matter is that users do most of the work, and both they and their corporations get the benefit.