TechCrunch has an article on yet another bomb from Y Combinator known as “AdPinion”. Well AdPinion, my opinion is that you suck and will fail.

There are a number of problems with this idea so let’s go through one-by-one:

1. Advertising is not music. AdPinion is not Pandora. Let’s be clear on this. People barely pay attention to ads. Let’s assume that on the average banner ad, click-through rates are 0.25%. If that’s the click-through rate, let’s assume that on a typical Web 2.0 site with a relatively tech-savvy demographic, maybe 10% even notice the ad. Most people in this demographic either have ad-block installed or don’t ever pay attention. How many of us see the TechCrunch ads on the top right hand side? I barely notice them and I think most of you don’t either.

2. No statistical significance. Let’s say that of the 0.25% that click the ad, a smaller portion are willing to engage with the ad to indicate whether they like it or not. Let’s say that 1/5th of them do. That’s now 0.05% of ad viewers. Now let’s assume that you need 1,000 opinions before you see any sort of statistical significance. That means you need 2,000,000 impressions. That is 2000 CPMs of an ad to determine any sort of statistical significance.

3. The chicken-and-egg problem is massive. A publisher has no reason to switch over unless there is a massive advertiser network (a la Google AdSense). Doubleclick (ostensibly Google) has far more advanced methods to target ads. The technology behind AdPinions is funded with a grand $20,000 from Y-Combinator? Give me a break. Google’s ad product team is huge and includes statisticians, engineers, marketing experts, etc. They didn’t create their model in 3 weeks living in ramen in a rat-infested Cambridge apartment.

The advertisers have no reason to use this either. Stupid ads might get some traction because users think they are funny. So what happens? All the advertisers that are getting screwed for having serious advertising leave your network. Bam! Where did all the insurance companies go? What about the mortgage firms? Trial lawyers? Consumer electronics stores? Auto companies? Hmm, all gone because of the stupid pricing model.

(We of course can’t expect TechCrunch to understand statistical significance since they think Y-Combinator is such a success. 3 hits out of 40? How is that a “success”? Oh wait, at TechCrunch, raising money is considered a success by itself.)

We’ll ignore that this is funded by Paul Graham and being shilled openly on his own personal web site with really crappy ads. I would think most ads on any Paul Graham essay should focus on problems that Y-Combinator types suffer from - dating (”Get out of your crappy North Beach apartment and get a girlfriend!”) or eating (”Stop eating ramen, try organic tofu!”).