The primary thrust of the development process involved integrating AltaVista search technology with Amazon.coms own information repositories. Amazon.com captures data from a variety of places, such as book distributors, book publishers, and content reviewers. That data is then reconciled and coalesced into a form that is indexed by the AltaVista software. The system uses a combination of full text retrieval and fielded text retrieval, creating a set of indexes tuned for the current content on the site.
One technique you might try is finding readers of similar books, contacting them through the reviewer profiles and sending them free copies of the book.
This has been done numerous times and it often generates tons of reviews for a book that otherwise wouldn't garner much attention. This can, of course be a positive or negative depending on the quality of the book. It may or may not help improve the search results. Pick Top 1000 reviewers because they are more likely to actally write a read the book and write a review. Get the author to sign the copies, maybe even send the email message from his or her account to the potential reviewers.
An example would be "17 Lies That Are Holding You Back & the Truth That Will Set You Free" by Steve Chandler. They sent out numerous copies, pulled in 60 reviews and managed to keep a 4.5 star rating.