Your first mobile app probably should not be an app
App stores add distribution friction, review cycles and platform work. Validate the idea somewhere cheaper first.
Wanting an app is usually a proxy for something else: wanting to be on someone's home screen, wanting notifications, wanting to feel like a real product.
What an app actually costs
Two codebases, or one framework with two sets of quirks
Store review on every release, on someone else's schedule
A download before anyone can try anything
Users on old versions you cannot upgrade remotely
None of that is prohibitive. It is simply a lot of overhead to carry while you are still finding out whether people want the thing.
When an app is genuinely right
Reach for a native app when you need something the web cannot give you: sustained background location, deep hardware access, reliable offline use, or performance that a browser will not deliver.
Push notifications alone are rarely enough of a reason any more, and "our competitors have one" never was.
The cheaper first step
A fast, installable web experience validates the idea with real users and no store gatekeeping. If it succeeds and the constraints start to bite, you will build the native app with a far clearer specification than you could have written beforehand.