Yup. Well said. As a Windows Phone user, I realize it may have been self-serving, but it came from my role as a business and product development strategist, not a user: when you let products die, there is a cost to all future products, because the market won't trust you. Now, it may still be the right move. If virtually no one is using it or it's losing money, then killing it is better that continually pouring money down a hole. But before choosing to deprioritize or abandon a product, consider the new costs that creates by making it more expensive to launch everything in the future. In some cases, that will change the conclusion to proceed or cancel.
Consider, if MS had fought tooth and nail to make Windows Phone succeed. Even if it had failed after that, when they released the Duo, users would know that MS would at least give the product its all. Customers would be willing to take a chance, confident it would get better over time (like what CDPR has done with Cyberpunk 2077, albeit for software rather than hardware). Instead, users were hesitant to trust MS with the DUO, standing on the sidelines waiting and maybe hoping, but not willing to buy.
That is the cost of the way MS handled Windows Phone. The failure of the Duo is not just MS repeating the same approach, but it's actually, in part a CONSEQUENCE of how they handled Windows Phone (and Kinect, Zune, and, to a lesser extent, because it was never a major product, Band).
For MS to succeed with these products, it's more important to demonstrate a willingness and commitment than it is to actually get it perfect out of the gate. PROVE to customers that even if it's flawed, you'll always keep working to make it right. That engenders trust and then customers will spend money.
Ironically, in the Bill Gates days, this is exactly what made MS successful. They would often release crappy products, but they kept at it like their life depended on it until they got it right. That's how Word took over from Word Perfect, Excel from Lotus 123, IE from Netscape, Windows NT -> XP over Solaris and other *nix for user workstations etc. All of those started as feeble versions that the MS of today would abandon for lack of customer/user interest.