So, I'm confused. I can see that people clicking through AdWords allows you to gauge interest. But having failed to find what they were actually looking for, people actually stick around and fill surveys about it? Really? There must be a giant selection effect in who decides to actually go through with it. I can imagine it ending up being quite misleading.

If it's presented as "We're building this product - tell us what you want it for" it would tend to select the people who are really keen. That's the kind you want to start with - but you're right, it is a subset, probably a very small one. So maybe this is better for operationally getting started, rather than measuring market size.

A related idea is to build less of the product, eg. the backend is run manually (ie. you are the backend). I like this one better, because it seems dishonest to me to say "buy our great app!" and then not have it. But I guess you can word it as a request for interest, like the HospitalVille example.

Question: has Zyngna actually been successful with this technique? (I've only heard of Farmville).

Tim Ferris describes a similar idea in the Four-Hour Work Week, only a little more underhandedly - he has them fill in the order form, click the link as if to proceed to the payment page, only to present them with a page saying "Sorry, due to unexpected demand, we can't take your order right now." The customer believes that they are actually buying something, when they're actually just filling in a survey.

Like I say, it's underhanded, but at least you can have a little more faith in the results.

tl;dr: Before developing a product, test the idea by creating a landing page with a feedback mechanism (survey, email notification, newsletter, etc.) then setup an AdWords account. It's a recycled idea that he has re-branded as "Ghetto Testing".

What he leaves out is that this is also a good way to gauge the cost of acquisition. I've seen several products developed that never advertise on AdWords because they discovered after development that the keywords were too expensive to bid on.

Who came up with that name? It's terrible. I'm not offended, but I have to say, I think of way worse things when I hear "ghetto testing".

The old name for this was "market research".

Innovative? This is just another name for vaporware, launch a small ad campaign, see how it goes, discard failures brutally. It is also incredible offputting to customers (who actually aren't complete idiots and can remember a brand). I'm sure anyone who has googled around for a solution to something knows the feeling "We have a product that matches what you want to do completely! Click here! Oh btw, by 'have' we mean 'don't have but will. Perhaps". Yeah, next time you offer something real, there's a good possibility may say "Suuure you do".

Asking users is one thing, but all out trusting them is another.

"If I’d asked people what they wanted, they would have said faster horses" -- Henry Ford

In practice, this can be very expensive. If you need to pay a couple bucks a click, you could be in for thousands of dollars before you get meaningful data. It can also take a long time. You may as well build a demo and test that.

This is a tactic better suited to free traffic, which makes it tough to pull off from a cold start.

I'm guessing this was made by someone who has never been to the ghetto.

Ghetto Testing? (pause for effect) Really!!