I recently did a survey of 25 magazines that currently solicit short stories for publication. These range from well-known magazines such as Asimov's Science Fiction, Harpers, and Women's World, to smaller 'zines offering with a religious, local, or literary theme.
The results are as follows. First, the odds of a story getting accepted by any of these magazines? Only 3.3 percent. Then, if a story is accepted, it's an average* wait of 9.5 months before the story is published. And finally, there's the payment; about $173 per 3000-word story.
Contrast that to a short story such as Seven Lives to Repay Our Country, or Happily Ever After. In each case, they were 100% certain to be published, and there was no wait time. The payment received is 35 cents per copy sold, and reviewing my sales figures for both books, the sales figures show that I can expect to sell 5-15 copies per month, with little or no effort at promotion.
This means in the 9.5 months I'd be waiting for a story submitted to a traditional magazine to be printed, I'll earn around $32 - about 20% of the what I would have made by selling the story. Within 3-4 years, I'll have earned more than I would have made by publishing through a traditional magazine, while still retaining all rights to the story.
So, from my perspective, 99 cents is a perfectly reasonable price for a short story, from a business math/statistics perspective. It's also a sum that most people are used to paying for a single "chunk" of entertainment value - a song download or a movie rental.
On the other hand, let's say I write a longer book, or aggregate several short stories into a collection. In that case, I would never considering selling such a work for 99 cents, because it would have the effects of causing readers to ask "Why should I buy one short story for 99 cents, when I can buy a book with 4 stories by the same author for the same price?"
For novels, story collections, and how-to guides, I believe the $2.99 price point is perfectly reasonable for self-published authors; and I do believe there's a maximum level beyond which e-books should not be priced, as well. The cost of electronic distribution is extremely small, so why should I pay $14.99 or more for an e-book? I believe that pricing e-books between 99 cents (for short works) and $2.99 - $9.99 for longer works would give customers great value, and still allow best-selling authors to be rewarded at a greater rate for the same word count as a less notable writer. I also believe Amazon should do away with the $2.99 price point to determine royalties (books below $2.99 only net their authors 35% in royalties, whereas those priced at or above yield a 70% payment.)
I would be interested to hear your thoughts, and if you know any writers who might be interested in the mathematical analysis shown above, please share this post with them. As always, word of mouth being the best advertising, I'd ask that you Tweet, Like, etc - and I'll be sure to return the favor!
*The term "average" used here actually refers to the "median", which is the "average of the average," and adjusts the output to reduce the effect of "outliers" - very high or very low data points. Sorry, fellow math geeks, but I used the term "average" because more people will understand it in that light.