Building the dot-gov retail directory

Measuring the size of the federal government's online business was not an easy task for the Federal Computer Week research team. Simply surfing government Web pages was out of the question; agencies operate thousands of Web sites with an estimated 37 million pages, according to the FirstGov office at the General Services Administration.

To find the government Web sites that sell goods and services to the public, the research team hired a computer programmer to write a "spider" program that would search the Internet for specific words and retrieve relevant documents containing those terms. The spider program searched for terms such as publications, maps, auction, surplus, credit card, books, ordering information, CD-ROMs and gifts, and returned more than 2,000 entries. That's when the real work started as the team began verifying, one at a time, if those Web sites sold anything to the public.

For a Web site to be included in the directory of government online retail sites, the team applied several criteria. For example, the public must be able to begin a transaction online by clicking on a link, printing a form or finding a phone number. In divisions where regional offices of federal agencies maintained their own sites and sold items from them, the FCW team listed those addresses separately from the main agency address.

The listing makes special note of sites that sell merchandise only to government employees or active-duty service members. Sites were excluded that referred users to a private vendor at a different Internet address. Sites that only offered free items weren't listed, nor were sites that referred buyers to a commercial Web page where a private company or individual kept the revenue.

We also limited the directory listings in several specific ways:

When a site referred buyers to the Government Printing Office or National Technical Information Service, it was not listed. The GPO and NTIS sites, which are listed separately, sell thousands of items on their own. The directory includes mainly .gov and .mil domain names, with some exceptions: USPS.com, the domain name chosen by the U.S. Postal Service, and MuseumCompany. com, the private-sector com.pany with which the Smithsonian Institution has contracted for all its museum sales. Some presidential library stores and organizations that sell items for the national parks use .org domain names. The nationalacademies.org site was included because Congress established the research entities it represents, and they are funded primarily by grants from federal agencies.

Calib.com, the domain name for the sites of the National Adoption Information Clearinghouse and National Clearinghouse on Child Abuse and Neglect, and Bid4Assets.com, an online auction site that sells government surplus items for GSA and other agencies, represent the private-sector contractors that maintain the government sites.

The National World War II Memorial site uses a .com domain name because "outside the Beltway, .gov is still somewhat obscure," according to Dick Couture, the memorial's director of donor marketing and data management.