I often read the suggestion to validate a business idea doing a marketing campaign (Adwords, Facebook...) and leading to a "sales" page for a product that doesn't exist yet and check if there actually are people willing to pay for it. Example article with this suggestion: http://www.fourhourworkweek.com/blog/2011/09/24/how-to-create-a-million-dollar-business-this-weekend-examples-appsumo-mint-chihuahuas/ I like this suggestion, but I don't like to let down those who click on an ad which is actually fake. I'd like to see some real-world examples of pages of this kind that worked though, to understand how you can do it without offending the potential customer.
In theory this is a good idea. The problem that I've found is that unless you already have a big following it's hard to really test out your idea this way. To get realistic and meaningful results you need a decent amount of traffic, and that can be hard to do for cheap if you are just starting out.
Several months back Hipster managed to create a lot of buzz with their landing page after being featured on TechCrunch. They got over 10K signups without even disclosing what their service/startup was about. However, that is rare.
As far as doing it without offending the potential customer, the key is to be honest. Don't create a fake ad, create a real ad. Make it obvious that the product doesn't exist yet and that you're just gauging interest. Keep the page simple and clean and have a place for people to sign up if they are interested. For an example, take a look at Splashtop's coming soon page. Also the TechCrunch article I linked to above has a screenshot of Hipster's landing page at the time it went viral.
There are a couple of sales landing pages for real products that I could suggest looking at but none that specifically ask "Would you buy this if we made it?" They can give you an idea as to how to at least format a page to achieve the results you want.
What you are really doing is something called A/B testing (which you can read up on here ) to figure out which features, functions and price points will get customer to convert.
Usually, A/B testing is done with a real product and real pages. It will be really hard to gage what a customer will want without actually having something to put in front of them.
Most won't really give you a "Yeah, I would buy that if you made it." Rather, it's more like "I bought it but would like to see these features."
I would suggest scoping your idea via a Minimal Viable Product (MVP) and go build it and then do your testing. I know this is more time, effort and money but the results will be a lot more meaningful.
I can't point you to any specific examples, but a suggestion might be to take people to a page about the product where they can see if the product is available in "their area." They could enter a zip code, after which they are taken to a page that says, "Sorry, but the product is not yet available in your area. Please enter your email if you'd like to be notified when it is." Or something along those lines. If it's not a physical product, you can use other derivations like signing up for a private beta. It will give you some indication of the level of interest without disclosing that it's just an idea.