On the surface the conditions even almost seem ideal - you have an existing oligopoly where each player has huge overheads and then on top of this add huge markups. You have a somewhat captive market (the students are entirely captive, while the professors are generally pretty free to choose the books that they want), and a product that's targeted towards younger, educated people (people to whom technology isn't a huge burden).
The only thing that's really standing in the way of someone shaking things up is in convincing the professors that your option is better than whatever's being offered by the incumbents. I really, really hope that at some point soon someone takes on the publishers - with a serious product, they'd have the potential to win, and more importantly, to do something useful for making education more accessible.
Textbooks for introductory courses don't change much between editions. The pagination may change slightly, as might the odd illustration, and that is that. What needs to happen is that the students need to get off their collective asses and buy the previous edition of the text from their colleagues from the year before. Nothing else will bring prices down. It's not likely to happen, considering just how spoonfed students at your average college are.
My own personal experiences have been with archaeology texts where the chapter-end studies were replaced with a different culture, physics texts where the figures within the questions were changed and maths texts where the questions were the same but reordered. Yeah, you can learn the same stuff from the textbook's main content, but if you get assigned a list of questions to complete and those questions don't match with this years' then buying the previous textbook is essentially useless. In my later years, the homework was never "do these questions from the book", so the texts dated less.