Barry Eisler, in a Book Beast interview with Jason Pinter, goes into more detail about the thinking behind his news-making decision to turn down a $500,000, two-book deal from Minotaur to self-publish his books instead. He says, in part:
My tendency has been to focus too much on that big, seductive number. But to understand what the number really represents, you have to break it down. Start by taking out your agent's commission: your $500,000 is now $425,000. Then divide that $425,000 over the anticipated life of the contract, which is three years (execution, first hardback publication, second hardback publication, second paperback publication). That's about $142,000 a year. This is a more realistic way of looking at that $500,000.
But there's more. Some people have mistakenly argued that, for my move to make financial sense, I'll have to earn $142,000 a year for three years. But this is one time when you don't want to be comparing apples to apples. Because the question isn't whether I can make $425,000 in three years in self-publishing; the question is what happens regardless of when I hit that number. What happens whenever I hit that point is that I'll have "beaten" the contract, and then I'll go on beating it for the rest of my life. If I don't earn out the legacy contract, the only money I'll ever see from it is $142,000 per year for three years. Even if I do earn out, I'll only see 14.9% of each digital sale thereafter. But once I beat the contract in digital, even if it takes longer than three years, I go on earning 70% of each digital sale forever thereafter. And, as my friend Joe Konrath likes to point out, forever is a long time.
Ballantine managed to sell about 10,000 combined digital copies of my last two books at a $9.99 price point (a price point that was earning me $1.49 per unit sold, BTW) in the latest three-month period for which I have data. Call that 5000 of each book for three months, so 1,667 of each book per month. If I cut the Ballantine price in half and still can only move 1,667 units a month, at a $3.50 per unit royalty ($4.99 x 70% = $3.50), that's about $5,833 per month. But unlike paper books and digital sold at paper prices, low-priced digital books sell steadily, so it seemed to me that I could make about $70,000 per year, per book on my own. Assuming nothing changes and digital doesn't keep growing (and that would be crazy–Charles Cummings' critically acclaimed spy thriller The Trinity Six just sold three times as many digital copies as hardback in its first week), I should be able to make $140,000 a year for the two books I could have sold in a $425,000 legacy deal, instead. $70,000 for the first year, then $140,000 for each year thereafter, when I'll be selling two books instead of just one. So if I'm right about all this, and I'm pretty sure I am, I should be able to beat the contract about halfway through the fourth year. And again, all of that ignores the continued growth of digital, the way low-priced digital books reinforce sales of other such books, etc.
It's clear that Barry Eisler gave this decision a lot of very serious thought and his reasoning is strong. The article goes into much more detail, including his pricing plans for his books and his perspective on self-pub phenom Amanda Hocking taking a publishing deal. Fascinating stuff.