Did you forget searching for the app (typing + click) and the chance of failing to find the app on the search result page? That is a source of frustration for users. Which app is the one they’re looking for?
Let’s really be generous for the moment and consider the best case for native:
If you have a landing page for the app on the web you can circumvent the search and go straight to install, meaning, click for store URL, click install, wait for download (can take minutes), find app icon, click to open. So even best case scenario from a landing page, 3 click minimum (assuming it’s not an older app still asking for permissions before install, in which case, 4 clicks, and two of them are scary) vs, you’ve landed in the app and you can start using it right away. Zero scary choices prior to test drive.
A huge number of users download an app and then never open it. (See activation rate link in OP).
As for the friction you mentioned about logging in, etc.. exactly the same friction exists for native apps. Developers can decide whether to require login up-front, only to save work, etc…
As for whether or not the install prompt will appear, PWAs are supported by 50% of mobile devices, globally. That’s much larger than global iOS market share (Android has 86%, remember), and the fastest growing segment.
In other words, it’s already a really huge install base. I’ve been playing with a lot of PWAs recently, and seen the add to homescreen prompt a lot. It’s my new best friend. I already have six PWAs on my home screen.
Guess how many native apps I’ve installed in that time…
Granted, I’m an early adopter actively seeking out PWAs to try. There are currently a small number of PWAs out there, but as more developers adopt it, my money says a lot of users will embrace them, and pretty soon, dealing with the app store will feel like a clunky, obsolete inconvenience to users.