Ask HN: How to validate people's willingness to buy without breaking promises?

I recently read an article [1] that was about how to validate a business idea on the cheap (both time and money). The article was well-written, and the business this individual started seems to be going well.

In a nutshell, the author was seeking to answer two questions. The first was trying to assess whether there was a potential market for the author's idea. The second question was seeking an answer was: "Will people pay enough to make this a meaningful business?" It is how the author goes about answering the second question that concerns me here.

The author put up a basic web page with a simple web site with a button to order the product in question. The button actually sent the potential customer into a purchase work-flow which ends up charging their credit card. The author follows-up with an email indicating that the product is not available and offers a full refund. Some "customers" request this, but the majority do not. Eventually, the author ends up refunding all of the "customers" since the author is not able to deliver on the promise of fulfilling the order in the time frame specified in the follow-up email. I do not think the author was intentionally misleading the "customers" on this point. I think the author did not realize how much work would be involved in actually fulfilling that promise. It comes out later in the article that an unspecified amount of money is spent over a time frame approaching a year before they were able to deliver on the initial promise. (I do not know whether the business was able to deliver on the promise to the original "customers".)

Is there a way to do an experiment on the cheap, that does reliably validate the business without misleading potential customers in the process?

[1] I debated about whether I should link to the article. I have decided not to. Really, this article could have been written (and has been written) by any number of folks about any number of businesses.