The assumption here I think is that a book at $9.99 might attract people who wouldn't buy it at $14.99. I'm probably one of those people, particularly with fiction books (computer science books/math books/etc are another issue). There are so many choices that are far below the $14.99 price, that unless it's from a favorite author I'd happily pass up the higher price for another book, or another form of media (like a game). It doesn't matter what other retailers offer the price at, $14.99 is simply too high for an e-book, and I only read fiction books as e-books. I might be the minority, but given the expertise of Amazon at selling books, I think that Amazon probably has a valid point.
You're right though, that it doesn't completely prove the point, but I think it definitely makes a point. People who wouldn't buy the e-book at 14.99 are probably unlikely to go out and buy the same book from another retailer for 14.99. The question is if it really increases the market share of the book, or if instead lowering the price on amazon makes everyone who would buy the book would just go to amazon for the cheaper price. (Given that e-books cost nearly nothing to distribute, I would think that the price drop would pretty easily pay for itself though)
You may be one of those people (where a price decrease would induce a purchase), but your personal value judgement is not relevant here. Personally, I don't read fiction, so the value of a fiction book is $0. Does this mean that the appropriate price for ebooks is $0? Obviously not.
Amazon has put forth a number that is an aggregate.. it takes into account your value judgement, mine, and 100,000 other peoples judgement of the value of an ebook.
Which reminds of a 3rd case that could affect Amazon's test: people do not purchase a generic ebook. They purchase a specific ebook, and the value of that ebook very much depends on the content. A poorly written book is worth less than a well written book.. and a book on a popular subject is worth more than a book on an unpopular subject.
If the book (or books) Amazon tested were more mispriced than the typical book (so they showed a larger increase in sales because $10 was closer to the actual value of the book.. whereas maybe $12 (or $20) represents the value of the average book).. then when all books are lowered in price, the result may be a decline for Hachett and the authors.
So I've put forth 3 cases that may affect Amazon's test. If just 2 in 13 people are affected by ANY of these 3 cases, then Amazon's conclusions are wrong, and the net result would be a decrease in revenue.
You're right though, that it doesn't completely prove the point, but I think it definitely makes a point. People who wouldn't buy the e-book at 14.99 are probably unlikely to go out and buy the same book from another retailer for 14.99. The question is if it really increases the market share of the book, or if instead lowering the price on amazon makes everyone who would buy the book would just go to amazon for the cheaper price. (Given that e-books cost nearly nothing to distribute, I would think that the price drop would pretty easily pay for itself though)