Yahoo will launch its long anticipated Panama ad serving system today. The Panama system is similar to Google's AdSense and Microsoft's AdCenter. The key improvement is that ads will be placed based on expected Click Through rates (CTR) and revenue yield, instead of just placing the highest bid ad at the top. Google and Microsoft figured out a long time ago that the highest bid ad may not attract the most clicks or generate the most revenue for Google.
The New York Times says;
Just about everyone inside and outside Yahoo expects the system to generate more revenue for the company, but no one knows exactly how much more. Last month, Yahoo cautioned investors not to expect the financial impact of Panama to show up until the second half of the year.
That is in part because as the system is introduced, some advertisers will end up paying less and some more for each click, as ads vary in how likely they are to attract a click from a searcher. Search marketing experts say that in general, the well-known brands will get better placement for their ads at lower bids, because people are more likely to click on them. The reverse will be true for lesser-known brands.
Will Panama make a huge difference for Yahoo? No, I don't think so. It will improve the revenue per search from US$.025 to maybe 4 cents. By comparison, I have seen estimates that Google gets as much as 9 to 10 cents per search.
The problem is Yahoo isn't generating enough search traffic. Most of Yahoo's traffic comes from Yahoo Mail (33%) and Home page visits (32%) , both of which can not be easily targeted because there is no user indication of interest like a search keyword. Yahoo generates less than 11% of its traffic from search while Google 88% from search. See this chart from "Why Yahoo's Panama won't be enough" that illustrates the difference in traffic at Yahoo and Google. 
Danny Sullivan's SearchEngineLand provides some more detail about how the Panama system works. I remember back in my AltaVista days when we changed our algorithm to mimic Google's PageRank system. We used all the same attributes to rank relevance, but the secret sauce is in how much weight you give to each one.
Algorithms are like a recipe for for cheesecake...they may have all the same ingredients but they sure can taste and look very different. Its all in the cook's secret technique...not the ingredients. Does Yahoo have a winner here? To continue the cooking analogy...The proof is in the pudding.
