Most early ideas do not need a native app. They need to be used, and a website gets you used faster: nothing to install, it works on every device, and you can change it the moment you learn something. Reach for an app when the idea genuinely needs what only an app can give, not because an app feels more like a real product.
The real question is not "app or website". It is whether the thing needs to live on the home screen and use the phone itself. If you need the camera, real offline use, push notifications people actually act on, or daily habitual return, a native app earns its cost. If you mainly need people to find you, understand you, and get in touch or buy, a website wins on speed and price.
Web is cheaper mostly because there is one of everything. One codebase instead of two for iPhone and Android, no app-store review sitting between you and a fix, and no waiting on Apple or Google to approve an update. While you are still learning what people want, being able to change the thing the same afternoon is worth more than almost any native feature.
A lot of what people mean by "app" is available on the web now. A modern web app can be added to the home screen, work offline to a degree, send notifications on Android, and feel fast and full-screen. It is not identical to native, but for many products it closes the gap enough that you can prove the idea on the web first and decide about native later, once real users are telling you what they actually miss.
Native is genuinely worth it when the experience depends on the device: heavy camera or sensor use, tight offline reliability, work that runs in the background, deep ties into health or payments, or when notifications are the core of the habit you are building. If that is your product, build the app and do it properly. If it is not, an app is mostly a tax you are paying to look serious.
If you will likely need both in time, still start with the web. It is where people find you from a search or a shared link, and it stays your marketing home even after the app exists. Do not count on the app store for discovery either; most apps are found because someone already heard of the product, usually from a website. So the site is rarely wasted, even in an app-first future.
Write down the one thing a user must do for you to have a business. If a good website lets them do it, start there. You can always build the app once the demand is real, and you will build a better one for having waited and watched how people actually use the web version first.