Before you open the editor: does this need Web3?

Run your idea through these six questions. If none of them is a yes, use a database and a normal backend: you will get further, and a judge or an investor will value it more.

1. Are there several parties who should not have to trust a single one?
2. Is there value or ownership that has to change hands?
3. Does independent verifiability make the product better?
4. Does a payment that settles by itself, globally, in seconds, change the product?
5. Are there rules that should execute on their own, with nobody approving them?
6. Does an open network, where anyone can plug in, make it better?

If a single one is a yes: that is the part that goes onchain. Everything else stays in Web2. Write it in one sentence before you code:

"Part X of our product goes onchain because Y."

If it does not fit in one sentence, there is still thinking to do.

The three ways to build

	What it is	When	Examples
Web2 native	Everything on your usual stack. No wallets, no chain	No question was a yes	A CRM, a services marketplace, most software
Web3 hybrid	Web2 for 90% (interface, data, logic); onchain only the piece that needs it: the payment, the proof, the ownership	One or two questions were a yes	Agrotoken, Collector Crypt, an API that charges per request
Web3 native	State lives onchain; the app is a window onto a public program anyone can call	Composability is the product	Jupiter, Drift, Kamino: DeFi, markets, protocols

Most products with users are hybrids.