When people are on a content page, they’re generally looking for the content directly on that page – and the only ads that work are serendipitous ones. I’m not sure that Google’s attempt to move their ads to content pages will work quite as well as they’re hoping. The clickthrough rates are likely to be insanely low, which won’t make anyone happy.

I haven’t paid much attention to the approach that goole is taking in setting up their content-based text ads: is it really a shotgun approach of just serving up ads in their unobtrusive text format without any real targeting?

If that’s the case, then I’m with you — it’s not going to perform well, at all, especially not over the long term (once people get over the “this kind of ad is good” feeling). That would seem like an odd approach for google to take, though, when one of their biggest strengths is the tools that allow them to evaluate information without human intervention.

Were I trying to build this as a saleable product, I’d be evaluating each page, serving ads based upon the content, and using clickthroughs as a feedback mechanism to determine which ads are succesful with content (apparently) on a given topic.

If you put that together with knowing what source is requesting the ad (for some more general information on what the population viewing the ad will be), I would think that you could target ads pretty much as effectively as you could for searches.

The down side would be that you’re very likely to be displaying the same ads over and over within a given day (because it’s the same story, rather than millions of different searches), which makes it harder to sell ad space; and further you’re likely to show the same sorts of ads to the same people on a given site, over and over…if I check the same sorts of stories on a news site each day, I’ll probably start seeing the same ads, which become less and less appealing until they become totally invisible.