Thanks to a partnership announced Friday, Yahoo's account executives will begin offering their clients a service designed to evaluate the offline sales impact of their online marketing programs.

"The tools we currently have in our set of analytics are focused on understanding the impact of online advertising on offline sales," said Michelle Madansky, Yahoo's vice president of research.

"But we haven't been able to help clients optimize their entire marketing mix. [With Marketing Management Analytics], we can help clients understand the relative return on investment of their entire marketing mix."

In October, Yahoo commissioned a study that showed that many shoppers did product research online, and then made the purchase at a store. For this group, brand -- and presumably brand advertising on sites such as Yahoo -- played an important role in the purchase decision.

Yahoo will offer assessment tools from Marketing Management Analytics (MMA) that it said can help marketers compare online programs on Yahoo with ad campaigns on other media, then give recommendations to optimize the media spend.

The assessment can include search marketing, as well as the many other advertising placements Yahoo offers, including rich media ads, sponsorships and a variety of banners, boxes and headers. Marketers using the MMA services can provide data from campaigns on other Web sites or e-mail campaigns, as well.

MMA uses marketing mix modeling to help companies plan, measure, validate, and optimize their marketing performance.

"We build mathematical models that represent the marketing landscape for a company or brand," said John Nardone, MMA's chief client officer. The methodology typically quantifies as many as 25 or 30 discreet variables that might influence a product's sales.

For example, for a financial services firm, MMA might include interest rates and housing sales in its analysis.

The analysis looks at historical data, and then makes forecasts for how well marketing might perform in the future. Nardone said that includes capturing the "non-linear behavior" of marketing in which any campaign can expect diminishing returns the longer it goes on.

"This lets us look across various marketing vehicles and see what's the marginal return for the next dollar. Dollars must be distributed in order to get the optimal mix," Nardone said.

Added Madansky, "Through modeling techniques, you can capture the synergy between online and offline media. We'll be able to say, for example, that the online channel works better when accompanied by television or print."

This article was adapted from InternetNews.com.