Merrill R. Chapman writes several interesting stories in his recent book, In Search of Stupidity: Over 20 Years of High-Tech Marketing Disasters. The book is organized as a series of case studies of computer companies that made bad decisions in analyzing their markets and deciding what products to create. The best entries are the ones from the 80's: MicroPro's mismanagement of the popular word processor Wordstar, IBM's brain-dead introduction of the IBM PC Jr., and Ashton-Tate antagonizing their developer community with dBase IV. When Merrill starts talking about more recent events, like the Internet boom, his analysis is weaker and the book reads like a conventional wisdom "I told you so".
The key takeaways are don't confuse the market by having multiple products that address one area, don't insult your customers, and don't solve problems that don't need to be solved. These are pretty obvious, but they've all been ignored in the past, and a lot of once-important companies aren't so important anymore.
My reaction to the dot.com parts of this book could be colored by the massive amount of dot.com commentary out there. There's a lot less written about the old marketing mistakes, so that material appears fresh; lots of people have written about the dot.com failure, so the material seems also-ran.