Nadella hammered the nail in the coffin, but because they couldn't get the Business Model straight for W10M or any mobile strategy. They kept an antiquated licensing model, while not seeking to focus on high end phones with a pre-established app partnership / development prior to release. This was Apple's model. Alternatively, they did not make RT / W10M open for OEM modification and deployment to then use as a way for the Store to be the access portal to their, and other's, software services (with some key app partnerships needed -I.e., Facebook, Snapchat, etc.). This was Google's model. Instead, they kept the OS proprietary like Apple and sought to license and did not set up enough apps in the ecosystem prior to release. This meant that OEM's only could differentiate on phone parts and not other software - device features (like Android, c.f., Samsung) and then did not have the apps people wanted lined up for use upon launch. So, this relegates W10M to low cost phones in price sensitive markets with little market penetration in the high margin markets and, hence, no overall popularity in the more profitable markets.
If Nadella made this key change with W10M, coupled with a push to Enterprise to supplement Azure, it may have had a potential to get up to 10%of the market, maybe a little more, which would great Mindshare in the public and further lead development to their UWP strategy and provide a means to protect their Enterprise vs Apple and Android threat and, more importantly, AWS threats.