* We've been seeing click through of 5-20X times higher comparing existing ad units on http://www.phonemyphone.com with the same ad unit delivered in an emptyspacead. So far so good :P.
* About the relevancy of ads we serve: First off, we're not an ad network. We do work with ad networks to serve their ad units in website's empty space. Secondly the publisher is free to create their own text and image ads as well as ShoppingAds which can be keyword targeted to the site's content.
* The core of the product is the patent pending ad format we've developed. I think finding a bunch of advertisers and serving relevant ads is a hard problem I'd rather let others solve. Hats off to the folks at AdPinion (http://www.adpinion.com) for being the YC company to step up and do this.
> "About the relevancy of ads we serve" - I missed that one ;-) But I think that it is definitely the right decision to focus on what you do well, and do only that. Especially if you are only one founder and thus have limited resources.
A word about patents. I have done two startups, both based on patents I have. One of the things I would have done differently is to not focus so much on the technology that we happened to have a patent on, but go where the market leads. It is easy to become thrilled by a patent, but in my experience the only ones that care about it are investors. It's a great thing to have, but don't base your business on it. Just saying this because I noticed you mentioned it a few times in the comments.
Also you should watch out for your brand - there have been many examples of obnoxious ad models that tend to turn people off. A good way to nurture the brand is to serve relevant ads, and in this regard of course you have a formidable competitor in Google.
The reason Google is doing so well is exactly because they are non-obtrusive and relevent.
I would work hard with the non-obtrusivenss, and think about how you can serve up relevant ads. The last one is of course extremely difficult for a startup because you need a lot of data, a lot of customers and somebody with a good knack of algorithms.
But all the best of luck :-)
Just thoughts off the top of my head....