The problem for Webrox is that they were creating UWP applications that were largely mobile-focussed at a time when Microsoft was slamming the brakes on investment in Windows 10 Mobile - so it's really just a case of bad timing. There are still millions of W10M users out there, but there are never going to be hundreds of millions, let alone billions, i.e. a genuine mass market. So I wouldn't expect, given the YouTube client competition and even allowing for the unique 'cast' integration, sales of the likes of Tubecast to rise beyond a few thousand.
But the point that Stéphane makes is a valid one - on the Desktop (the principal SKU seen by most people in 2018), the Store is very much secondary to the average user experience. Whereas on Mobile, the app store is the place to go once a day to grab new 'stuff', on the Desktop most people set themselves up with their browser of choice (often Chrome, to Microsoft's chagrin), maybe grab a client for Spotify or iTunes from the Store or from the general web, maybe download the right version of Office, and then they're all set. With tunes playing and with a browser on-screen, just about everything they ever want to do is there, on the Internet. With a large display to work with, there's almost no need for 'apps'.
I've observed this in my own family, they just never - and I mean never - use the Store on their laptops. And I observe the same on my own desktop, an iMac, where I never, ever go into the Mac App Store. There's just no need, once I have the basics up and running. So this isn't just a Microsoft problem, it's just how things are in this browser-centric world.
There are subtle differences and capabilities, of course. iOS presents limitations for PWAs and these are only gradually easing, while on Windows 10 Mobile, the Edge engine was frozen in terms of capabilities at the '1709' OS branch, and the background 'service workers' don't actually err... work. But on the whole PWAs offer a way for a developer to simply concentrate on the main web site and then trivial work lets them wrap/package this in PWA form for every user on every device - a very efficient way to work.
In fact, following their introduction of a PWA for all (in the USA only, initially), Starbucks shared at Build 2018 that they "doubled the number of people who used the website to place orders each day, with desktop users now ordering at about the same rate as mobile users." So PWAs are very definitely the way forward for any 'service' functions on the Internet, where you're accessing information or interacting, and where you want a similar experience on any size screen and with any interface.
It's ironic that back in the earliest days of smartphones (2005?), I was posting articles like 'there's a bookmark for that', looking at accomplishing lots on the wider Internet using specific URLs in the first smartphone web browsers. And when Apple launched the iPhone back in 2007 they had 'web applications' for at least the first year, again using wrappers around general Internet content.
Then the world went app-crazy with the mass market adoption of smartphones for the best part of a decade and now we're back to where we began - with so much content in our apps being Internet-hosted that the simplest thing for developers is to again package the web, in the form of PWAs. The number of sites which default to serving up the PWA when you visit them in a phone browser is growing fast, I've noticed.
Whether it's Twitter or Starbucks or SkyScanner or wherever, PWAs aren't a flash in the pan, they're:
- a return to how Internet access started on phones
- a valid way to simplify a developer's workload
- a valid way to simplify a user's path to web content on the Desktop
- a way to utterly solve the issue of keeping users updated and safe (since updates largely happen on the 'back end')
- a way for even Windows 10 Mobile to (largely) join in the fun on the modern Internet
- a solution that answers so many problems
I'm rambling a little, of course, 'applications' often need to do things on a phone which are strictly local and which involve heavy user interaction, so traditional iOS, Android and - yes - Windows 10 UWP apps coded in C++ and F# and the like will continue to be needed. But you'll be seeing PWAs more and more. In the Stores, in browsers, and in the news.