Yahoo is set to introduce on Monday a long-delayed upgrade to its Web search advertising system, setting the stage for a face off with Web search leader Google later this year.

The Internet media company is looking to encourage both new and existing corporate advertisers to start working with its rebuilt search marketing system in preparation for full introduction during the third quarter of 2006.

The highly anticipated redesign is the first big overhaul of the search advertising system since GoTo.com pioneered paid search advertising eight years ago. Yahoo paid $1.6 billion to buy the successor to GoTo, called Overture Systems, in 2003.

The system is a software-based marketing console to help corporate advertisers and advertising agency buyers understand how to target relevant advertising to Web users based on the keyword search terms consumers use to find information on Web sites.

Advertisers can instantly calculate the cost of acquiring new customers using Yahoo's data and analytic system. Buyers can also test for the most effective ads and make rapid changes.

The Yahoo ad system also offers geographical targeting based on databases of geographical terms it acquired when it bought London-based WhereonEarth last year. Thus Yahoo can infer when Web searchers mean Soho, London or Soho, New York.

Eventually, the system will also allow advertisers to buy ads across Yahoo, Google, Microsoft and other systems, allowing for comprehensive budgeting, something Google already offers.

The system will be rolled out in phases in order to lessen disruption to Yahoo's base of hundreds of thousands of customers, who need time to optimize the system to maximize the number of customers their ads may attract.

The first phase, which begins on Monday, is to deliver the core upgraded data and analytics platform. The second phase begins in the third quarter, when the full marketing campaign application becomes available to advertisers.

Future versions will give advertisers the ability to more precisely target Web advertisements based on demographic factors or online behavior.