I'd argue that if the data is that specific to the app (e.g. which trees and courses of Duolingo you've completed), then if the app dies, the data is irrelevant. Either the data is useful outside the context of the app, in which case it tends to have standard formats (.png for images, .md for formatted text, etc), or the data is not useful outside of the context of the app (e.g. Duolingo) in which case the data is no longer of value and can be discarded, as no-one is going to be able to offer the exact same courses and trees and in-app cosmetics and gems etc etc etc.
It's true that the limited period of time between end of life and servers going offline may be a problem, but that tends to be a long period of time (3-6 months in general from the services I've used that have shutdown).
As I said, the examples you chose are for the simplest use cases, and fungible applications/features for which the raw data is pretty much the whole thing and having a database dump is enough.
Valuable applications more often than not simplify complex things or do hard things behind the scenes. Their value is in the workflow/experience or the processing that takes place on the data, including APIs and integrations to unlock users' creativity.
>It's true that the limited period of time between end of life and servers going offline may be a problem, but that tends to be a long period of time (3-6 months in general from the services I've used that have shutdown).
Again, have you used these as an individual relying on them lightly, or heavily, or as a business/organization where your work relied on these? Have you used them as a user, or as the person who was involved in making the purchasing decision for the whole team/organization?
The only example you gave for app such as you describe was Duolingo, for which I pointed out that the data is useless outside the context of the app.
For business applications, the conversation may be different, but for the consumer ("average non-techie" as the original comment said) I believe what I said holds true.
It's true that the limited period of time between end of life and servers going offline may be a problem, but that tends to be a long period of time (3-6 months in general from the services I've used that have shutdown).