(When did you hear about a startup that does a meaningful R&D?)
The easiest one are B2C:
1) Amazon - selling books.
2) Air B&B - renting rooms.
3) Uber - renting cars (advanced taxi).
4) Facebook - a database (My SQL) for Social Network.
But even "hard" tech startup:
1) Google - research done at Stanford.
2) Databricks - research done at Berkley.
3) Microsoft - No research at all, founders drop from college, Selling interpreters.
5) AWS - Research done at Cornell. Selling VM
6) VMWare - research done at Stanford.
Only today, we see startups (e.g. Cruise) that are trying to tackle autonomous driving, which is still an R&D problem, and see where they are.
As for being successful.
There are two risks in a startup - execution risk and market risk.
With boring, you take out execution risk.
For market risk, the main thing is timing (I.e. most of the market risk is outside your control).
You must be in the right market, at exactly the right time.
Hence you must start the project 2 years in advance and hope that the market will become "hot" when you finish.