I like the idea of a "Cost Per Influence" model, but as Ross Mayfield points out...
The meme being transmitted isn't necessarily the message of the ad.
Additionally, I'd think that a Cost Per Influence model would raise a number of problems that both advertisers and bloggers would have trouble dealing with:
- First of all, if two bloggers post a similar idea at the same time, who gets credit or responsibility for influence down the line? This is similar to a problem we already have in the affiliate marketing space. I think it would be a bigger headache here.
- Secondly, how is all of this going to be tracked? What if the chain is broken by bloggers who don't get picked up by the presumably Technorati-like tracking mechanism that would measure the proliferation of an idea or opinion? It could be a technological nightmare. Let's put it this way, it's taken nearly 10 years for the online advertising industry to define an ad view (and we're still not done getting publishers, adserving companies and advertisers to agree on and implement the definitions). Can you see what an uphill battle this might be from a technology perspective?
- The notion of people getting paid for influence might set us up for the kind of intellectual dishonesty that bloggers abhor. It wouldn't be long before the capitalists invade and start setting up "idea networks" where one blog simply parrots another for the sake of getting a paycheck.
Personally, while I like the idea of somehow rewarding the most influential bloggers for their influence, I think the model might be more trouble than it's worth. May I suggest a different direction?
Since online branding came back into fashion, companies like Dynamic Logic, Millward Brown and Factor TG have conducted online brand studies to gauge the branding effects of online ads. They could conduct similar studies in the blogosphere, recruiting exposed and control groups from areas where the idea has permeated. Online surveys could then determine how influential the idea was through already-established attitudinal metrics. What I like about this is that the methodology is pretty well established (Forbes.com uses essentially the same method to measure its performance against The Wall Street Journal on brand metrics and they'll refund your money if your campaign on Forbes doesn't outperform WSJ.)
How about adding a new twist in the survey? - asking respondents to reveal where they first heard about the idea. That might give rise to an influence-based performance bonus to influential bloggers.
BTW, I'm pinging my partner Jim Meskauskas on this one. He usually eats this kind of stuff up and maybe he'll drop in to contribute some ideas.