Sean Ammirati wrote about the problem of competing with Google, and how an open ad network could be the mechanism to do so. His post at Read/WriteWeb noted Google's potency in online advertising, and where they could be vulnerable:

I believe in one area Google is more vulnerable than analysts appreciate. More than one-third of Google's revenue is completely out of their control. To be more specific, in their most recent quarterly results 37% or $1.35 Billion of revenue came from advertising delivered on other sites.

At face value, the vulnerability looks like a prime opportunity for someone to swoop in and grab some market share. Ammirati's proposal for an open ad network would "force the industry to make money by selling services around the network, instead of taking a percentage of the revenue on the network."

Scratch a little at the surface, and the blood flows out of this idea like it's a victim in a Peckinpah film. There are a couple of reasons why organized efforts out of Yahoo and other places haven't eroded Google's network.

Reason number one is Google's algorithms that match ads to context for AdSense publishers. Yahoo's capitulation to relevance as a determining factor in ad placement in its new Panama ad server can be called Exhibit A.

When Ask.com announced its plans to enter the contextual ad market, they disclosed a novel approach to ad relevancy. Site publishers displaying their ads will be able to choose how to weight what appears on a sliding scale between relevance to context and how much the ads are worth.

The worth of ads brings us to reason number two. That $1.35 billion, 37 percent of Google's ad revenue that is at risk? It's off by $1.05 billion, which is the amount Google paid back to those partner network sites for the quarter.

A successful threat that could wipe out Google's 37 percent would really just cut out $300 million in revenue. Publishers who are getting a piece of the $1.05 billion payback would need to see a better deal to make any kind of switch.

That could certainly happen. Today, it just doesn't look very likely.