Information Architecture 2.0

To subscribe to an RSS feed, you need a feed reader such as Bloglines, shown in Figure 1. Though some browsers have built-in feed readers, stand-alone feed readers are also available as both desktop and Web applications. They all offer the same basic functionality: managing subscriptions, indicating new content, and displaying content.

The hope was that RSS feeds would let users cut through the clutter and go right for the content that interests them. Since there are just about as many feeds as there are Web sites, however, this hope was quickly dashed. RSS feeds help users stay up to date on their favorite sites, but do little to stem the flow of content.

Imagine you’re an information architect who is designing a new feed reader. You’ll face the following questions—among others:

• How should the feed reader facilitate finding new RSS feeds? Will the feed reader recommend additional feeds based on the current subscriptions?
• Should the feed reader filter the feeds, displaying new content from predefined categories?
• How much content does the feed reader need to display to help facilitate decisions about what to read?
• How should the feed reader store feeds?

Although you, as the information architect, are not working on strict taxonomies, you are considering essential IA problems—structuring information in an accessible way. While the structured, linear format of an RSS feed gives you some foundation, the feeds themselves are constantly and unpredictably changing as sites add new content. This means you are working with a corpus of data that’s constantly in flux.

The More They Want…

Sophisticated users, who have specific ideas about how to manage their information, present one set of challenges to the next generation of information architects. The ballooning amount of information offers a different set of challenges. More and more sites permit people to publish their own content, and individual publishing is a major source of the increase in the volume of online content.

While Weblogs may be the most obvious example of individual publishing, there are other online applications that have also contributed to this explosion. There are now Web services for managing information that people previously stored locally, on their own computers—events, bookmarks, photos, and contacts, among others. All of these Web services allow users to carve out virtual space for storing their information rather than hard disk space. One online application that serves this purpose is the wiki.

Wiki software is gradually gaining popularity as companies like WikiMedia, 37signals, and JotSpot find ways of improving upon the basic premise of a wiki to make them more accessible and useful. Like most new Web technologies, wikis are services that have no obvious real world analogue. The simplest description of a wiki is a set of Web pages that anyone can edit—though this definition loses some of the magic of wikis. Wikis provide simple editing tools such as those shown in Figure 2—allowing people to log in and create content. Although quite different from blogs, wikis lower the barrier to entry for Web publishing in much the same way that blogs do.

Wikis may present the information architect’s ultimate challenge: a virtual information space with no structure, to which anyone can contribute. Generally, the more successful wikis and collaboration spaces have some overriding structure and follow certain standards. Consider Wikipedia, perhaps the most well-known wiki. Wikipedia is an online encyclopedia whose authors and editors are volunteers from the online community. Yet, like a real encyclopedia, Wikipedia has strict content standards—what’s acceptable and what’s not—that their community enforces.

I don’t know if Wikipedia has an information architect. My guess is that they do not, given the underlying philosophy of the site. Still, Wikipedia illustrates the kinds of information architecture problems we face today. Even if there’s no one on their staff who calls herself an information architect, there are people in the community filling this role. This may suggest that there’s no value to having an information architect on staff, if community volunteers can easily fill the role. This model may work for Wikipedia, but businesses using wiki software may not have the luxury of waiting for content and categorization standards to emerge from the community or having those standards enforced ad hoc.

Digging a little deeper into Wikipedia, it’s possible to see standards emerging across similar entries. For example, take Wikipedia’s entries for Wesleyan University, my alma mater, shown in Figure 3, and Duke University, my wife’s alma mater. Both entries have sections on “History” and “Academics.” Both have sections on famous alumni—though Duke’s is called “Famous and Distinguished Alumni,” while Wesleyan’s is simply “Notable Alumni.”

The structures of these entries include elements that make them distinct, too. Duke has a section on “Athletics,” which is conspicuously absent from Wesleyan’s entry. On the other hand, Wesleyan’s entry includes a section on “Recent Activism,” but Duke’s does not. Figure 4 provides a side-by-side comparison of the tables of contents for the Duke and Wesleyan entries on Wikipedia.

Although Wikipedia has a taxonomy of subjects, the primary information retrieval mechanism appears to be search. The categories themselves are somewhat ad hoc and non-exclusive. For example, one of Duke’s categories is “Universities and Colleges in North Carolina.” Despite the apparent arbitrariness of the categories, the category pages seem to follow a particular standard, just as the individual entry pages do.

The point here is that establishing and using a standard structure enhances the value of community-contributed content. Flexibility in the standard structure of entries is necessary in order to highlight potential distinctions between otherwise similar entries—as in the case of “Athletics” or “Recent Activism”—and to allow the emergence of obscure categories—for example, “Universities and Colleges in North Carolina.” At the same time, the standard structure must be clear and easy to follow.

The traditional role of an information architect focuses on devising useful categories of content. Within a dynamic space like Wikipedia, an information architect would

• guide and coax contributors to create content using standard formats
• create a framework for eliciting and capturing content
• create a framework for displaying categories of content, but not necessarily the categories themselves

Wikis are Petri dishes for content and, since anyone can contribute content, they carry the risk of serious contamination. With a scope as big as Wikipedia’s, devising a standard structure for the site may seem like an impossible task for an information architect, and it is. But Web 2.0 calls for something else entirely: a more abstract scope. For Web 2.0, information architects must focus on the higher-level structures that create an overall information framework for a site.

The Less We Have To Work With

Individual publishing is only the beginning. The Web 2.0 paradigm shift has also introduced new sources of metadata. Typically the domain of specialists, metadata is information about information. Metadata supports various tasks like retrieval, administration, access, and more. Although allowing the consumers of information to contribute metadata is hardly a new idea, there are new implementations popping up online. Most notably, free tagging.

With free tagging, users have the opportunity to apply labels, or tags, to content in an unstructured way. Users are free to enter whatever labels they feel are appropriate. This approach is distinct from more traditional metadata approaches in at least two ways:

• The words used for tags do not come from an authoritative source like a controlled vocabulary.
• The tags can apply to any kind of data and do not relate to specific fields like author, topic, or source.

Free tagging is controversial. Some think it renders librarians and information architects obsolete, because there is no longer a need for formal categorization. At the other extreme, some insist that the metadata people create by free tagging is of no value. The moderate view is simply that free tagging and formal classification are two different techniques that address different problems.

Two typical examples of free tagging are del.icio.us, which allows people to assign tags to bookmarks, and Flickr, which allows people to tag photographs. The main uses of tags in these applications are to allow users to

• find previously viewed or contributed content again
• navigate the sites serendipitously

By tagging my bookmarks on del.icio.us with keywords that are meaningful to me, I can easily find them again. This saves me from having to remember how I found the resource in the first place—whether through search or navigation categories. Free tagging allows me to create bookmarks within the context of my work and my life.

Flickr provides good examples of serendipitous navigation. I’ve used keywords to tag my photographs on Flickr, as shown in Figure 5. By clicking a keyword, I can see photographs that other people have tagged using the same label. In this scenario, users are not necessarily interested in seeing every possible entry in the library of photos. The loose network of keywords on Flickr makes it easy for users to take a virtual journey, tripping over treasures, and enjoying the Web’s scenic view. Thomas Vander Wal coined the term folksonomy for this network of keywords.

While opinions on this approach vary wildly, free tagging does offer one important advantage in a Web environment where more and more content is community contributed. Users do not have to exert much mental energy to determine appropriate keywords for specific entries. When given a choice, users will shun creating metadata when there are too many constraints on what is permissible. Free tagging effectively removes the constraints of controlled vocabularies and metadata categories, lowering the barriers to entry for capturing useful information.

Herein lies the challenge to the information architect. Traditionally, metadata adds structure to otherwise unstructured content. Modern systems open the possibility of unstructured metadata and, in this context, information architects must

• determine whether their projects would benefit from lower barriers to metadata creation
• weigh the benefits of lower barriers against the potential costs of unstructured metadata
• design systems that effectively leverage unstructured metadata, imposing structure where there is no implicit structure

Conclusions

Three factors are creating new challenges for information architects that have emerged with Web 2.0:

• users expecting more control over information management
• large and dynamic information spaces
• unstructured metadata

What all three factors have in common is that they force information architects to think about underlying structures. Rather than developing concrete navigation categories or prioritizing certain content over other content, an information architect must think more abstractly. The domain of today’s information architect includes not only organizing information into categories, but principles of organization as well. Gone are the days when information architects could do a card sort alone to establish an information structure for a Web site.

The complexity of modern Web services demands the same level of rigor in their information architectures while, at the same time, requiring that we design information spaces on a more abstract level. With users continually contributing both content and metadata, information architects have much less raw data to work with up front. Instead, they must think about the structures and patterns that govern these dynamic information spaces without losing sight of their primary responsibility: ensuring people can find the information they need. 

Author acknowledgement—Special thanks to James Melzer for providing a critical peer review on an early draft of this article.