Sorting Data to Suit Yourself

The Bettmann Archive of 11 million photographs is kept secure in a limestone quarry 220 feet below the earth. The front room—a painted cavern—holds a couple of offices, scanning equipment, and, most important, the card catalog. In the back room, filing cabinets filled with photos stretch to the vanishing point. The Bettmann’s layout embodies the two basic orders of organization that have dominated our thinking for thousands of years. The back room represents first-order organization: the sorting and arrangement of objects themselves. The front room represents second-order organization: information about information (metadata), which is sorted in a different way. We encounter this dual structure in libraries, warehouses—even when catalog shopping. In the physical realm, it works. But as we digitize the world’s information, a third order of organization is emerging that’s undoing many of our old assumptions. In the digital age, we are coming to realize how much of what we do is shaped by the limitations of the physical world. No physical object can be in two places at once—so we classify and file documents based on the assumption that objects can go into one bucket only, with exceptions for the occasional cross-reference. But we can file each electronic object into dozens, or even hundreds, of buckets. That means traditional tree structures—organization charts, the Dewey decimal system, some knowledge management concept trees—are ill suited to the task of organizing digital information. Not only do traditional tree structures assume that objects hang from just one branch, they also value neatness and include exceptions only when they have to, flagging them with dotted lines. By contrast, lots of messy links enhances the value of digital information, providing context and guiding people to unexpected destinations. While the Dewey decimal system places books on related topics near one another, helping you find additional relevant volumes, messy webs of information awaken you to the far-flung, serendipitous connections made by bright (and not so bright) minds before you, spurring your own innovation. The biggest change wrought by the third order, however, is that owners of information (such as companies) no longer own the organization of that information. Users (such as employees and customers) can sort and organize information in any way that suits their needs. All over the Web, customers are getting used to sorting through long lists of products based on criteria that are important to them rather than browsing along a tree that some marketing person thought would be right for everyone. Accordingly, the rise of third-order organization changes the jobs of corporate information architects and knowledge managers. Their role is no longer to build trees that define the relationship of every bit of data in a company but to build enriched pools of data objects whose relationships to one another change constantly, depending on who is looking at them. For example, the uBio project, sponsored by the Woods Hole Oceanographic Institution, is creating a pool of information objects related to species. A marine biologist may assemble these classifications based on physical characteristics or gene sequence information. A restaurateur could sort fish and shellfish according to seasonal location and price. The same data pool allows for both. Within companies, employees are using third-order schemes to manage information about complex products that are frequently reused or recombined into new offerings. The British Broadcasting Corporation, for example, packages and distributes television and radio programs, movies, music, and Web sites created by producers all over the world, often combining them into new products. Each media type is described by different metadata; television programs, for example, might be broken down by scenes, articles by sections. The BBC is turning this vast assemblage of content into a pool of objects that users can sort, classify, and combine differently for each new project. That should greatly reduce the cost and labor of creating and delivering complex, integrated products. The third order is not without dangers. Companies lose yet another instrument of control over their customers. And if everyone creates organizational structures that reflect their immediate needs and individual ways of thinking, integrating information from dueling structures could get tough. But the benefits of letting users order their own data are compelling. Enabling customers to organize your information the way that works for them is a cool benefit today but will be a necessity tomorrow.

Sorting Data to Suit Yourself

Summary.

Partner Center