In 1997, Microsoft launched a concerted effort to build its own search engine. That effort, codenamed Yukon, failed. It appears to have never been opened to public use, although former Microsoft executives familiar with the program at the time declined to confirm.

Microsoft executives internally said they were determined to both out-perform and out-distribute the efforts of smaller companies then beginning to dominate the field. At the time, the search market had just broken into two camps: the well-distributed, such as Infoseek and Excite, and the well-constructed, such as Altavista and Hotbot. Microsoft knew it would have the best distribution; it just needed some decent original programming.

Details remain sketchy. The failure of the program may have had several different causes, or it might have just signified a change of strategy. An internal effort at the time was forcing all web initiatives to use Microsoft server software, which severely handicapped some high-performance applications. Microsoft was also at the time reportedly talking with several search engine firms about an acquisition, although valuations had already started to climb to internet bubble prices. Finally, Microsoft continued to fight the "curse of the new," the tendency for Microsoft-originated code to not perform as well as code subsequently lifted from the carcasses of acquired companies.

At this posting, Microsoft officials have not yet responded to questions seeking to determine if the new Microsoft search engine effort is related to past efforts.