The author left out one major reason textbooks are so expensive-- course adoption lists (i.e. which books go with which courses) are kept private or disorganized so that students or businesses can't offer rival "Search by Course" type sites without scraping the local bookstore. As a result students go to the bookstore which knows their requirements unlike Amazon.com.

I open sourced a course data scraper a while ago to help solve this problem: https://github.com/bsgreenb/Open-Textbooks