As a former Apple guy, it pains me a bit to say this, but I’m coming to believe that the whole “It Just Works” thing is a temporary illusion.
During the years I have tried out, installed for other people, and contributed to an insane amount of distros.
Do you want to know the truth?
Most people value convenience a lot. And making something good convenient is part of the job.
Windows and Apple are widely used, and not Linux, for their own merits. Because they are “convenient” to use from the business perspective (even when from the use perspective they are pretty nasty).
You just purchase a computer and you get something that “works”, that everyone knows, and it’s standard.
The question is that most software either chooses one way or the other. To be either easy, or powerful and flexible. As if these were mutually excluding.
But there is a third option, which is similar to the KDE motto: simple by default, but flexible when needed.
I made Zenned based on this idea.
The point here is that convenience over time is a relevant metric too.
Something that’s very convenient up-front but then becomes less convenient later isn’t actually as convenient as one might think. And vice versa. It’s like how an incandescent lightbulb is cheaper than an LED bulb but costs much more to run over its lifetime, and that lifetime is shorter, too.
I’m advocating that we advertise the total package, basically.
in engineering we call this a cost of ownership calculation where you count up front cost separately from ongoing costs to get a fuller picture of what it “costs” to own and operate something.
My understanding is that products usually have two variants: retail and sophisticated.
The retail person is the usual person. They liked to have the product, but they aren’t that passionate about it to investigate it to the detail. They base their decisions mostly on impressions and recommendations.
The sophisticated person will investigate in detail what suits them the most, no matter the time it takes.
The question is: being a non sophisticated client for most things you consume makes sense, because there isn’t that much time available.
Sadly companies usually take advantage of this, and make the products against their users best interest in secret.
For example a shampoo made for a professional hairdresser may leave the hair very well conditioned for a long time.
But a shampoo made for the consumer may leave the hair very well conditioned for a short period of time, encouraging the client to constantly wash their hair with it.
Same with computers. For example HP computers for the enterprise are highly durable, while those for the consumer are brittle.
Both Microsoft and Apple optimize for convenience in the short term.
I worked in IT supporting hundreds of Mac users in an academic environment for over 20 years and the whole Apple’s “It Just Works” was always just marketing baloney as we were quite busy with things that didn’t just work. That’s how I got a strong dislike of macOS and any UI/UX that even remotely resembles it. Kind of work related PTSD, I suppose.
I don’t believe “It Just Works” is achievable, actually. Some tech works better than others but nothing works perfectly all the time. And anything that appears pretty, convenient, friendly and “nice” to the user actually comes at a cost of significant complexity under the hood that often compromises stability.
Which is funny, because while I get that, I still prefer nice, pretty, friendly GUI over CLI any day 
I realised this when I was deeply embedded into the MS ecosystem.
I wanted to create an application that worked on XBOX, Windows Phone, and Windows (Home, Professional, Enterprise, Server, and IoT) via their seemingly brilliant UWP framework. I leant some XAML and C# inside VS 2019, then spent £50 on a developer registration to submit to the Microsoft Store.
The problems arrived when:
I realised that this “universal” software was literally incapable of performing basic actions, like a shutdown! [1] It was more akin to AOSP’s sandbox, than Flatpak’s.
A friend replaced his old x86 Macbook’s OS with shiny, new Windows Server 2019. Turns out that the damn thing didn’t include the Microsoft Store by default, despite MS itself recommending that all new software be delivered via MSIX. That was meant to be my primary deployment target.
I could install the Store regardless, but it was an unsupported delivery mechanism. I felt a little miffed, considering I’d paid to use MS’s infrastructure. “Oh well, no worries. I can deliver updates via an alternative mechanism, like Scoop.”
Soon afterward, my deployment targets nigh disappeared. WinUI3 shall never be integrated into UWP, which means that I’ll never be able to target XBOX unless I stay on the deprecated UWP XAML, which is ugly and really, really basic. Windows Phone, to me back then, seemed to be steadily improving, yet was nuked from orbit.
Basically, MS told me that I could get an end-to-end development solution for all their deployment targets with UWP via Visual Studio. Now, my XBOX ONE S sits unused, and I get better cross-platform support from Python 3.12 and Qt 6 than I ever did with C#.
I’ve myriad additional examples where MS 365 (Office) incorrectly parses files that LibreOffice doesn’t, and absurdly prominent deficiencies in the ecosystem, like how Outlook’s mobile applications can generate a CalDAV synchronisation loop (across OSes).