Nothing wrong in that. But in research we have a thing called survey the already existing literature on the problem and checking what worked or didn't. If you want to learn by doing mistakes of your own, be my guest. Just know that it is expensive. Also if you aren't thinking that hard about your startup ideas, then most probably the VCs aren't either which means you get some free money to play around with and make mistakes as long as the music last. That is also a thing.