It is sad to see less than one review per 1000 downloads, but I can understand it given the audience.

I have looked extensively at possible publishing models, and there are a few techniques that might work at BDSMLibrary. One is to reward prolific reviewers with privileges on the web site, for example give them some 'n' number of membership months. To make sure the reviews are of reasonable quality, you can have a grading function like Amazon (3 out of 5 readers found this helpful) or simply correlate a reviewers score with the median score.

You could also make grading a work really easy, like at Netflix.com where simply clicking on 1-5 stars automatically adds that scoring to the overall scoring. Many people don't want to go through the full review form while they might simply click a 1-5 (or 1-10) star rating. Some safeguards might make sense to prevent fake votes. There might even be two types of ratings: one by reviewers who are going for incentives and one by general readers who only want to spend a clicks worth of time.

The failed iPublish.com experiment required authors who submitted a work to review at least 3 other works. But requiring reviews leads to another problem. Erotica is an extremely personal experience with amazingly diverse "hot buttons" for readers. What is someone's perfect fantasy might be another person's worst reading experience. I would hate to see a story downgraded because the reviewer didn't find the fetish up their alley even though the technical aspects rocked.

As an aside, I really appreciate reviewers like hypen666 who comment on dozens of works. He's a statistical anomaly. You could do a sweepstakes with each review getting an entry.