The failure is mostly because there's no one left at Microsoft (or anywhere) to do the same research: gather and evaluate data over decades, make informed decisions, iterate on designs etc.