Game design's kind of nebulous, but I'll try my best to talk about the design of a platformer prototype I made. 

Concept

I love physics, so I knew I wanted to make a smooth physics based game, and I also had the idea of a bullet hell platformer in my mind for quite some time, so I got to thinking, how would I put those two together? In the end I decided to go with an air-gun you could use to push yourself away from walls, and up walls. I wanted to player to get stuck in bullet hell situations and have incentives to move normally though, so I decided to limit the air-gun by giving it a charge, and when you're not using it it will recharge over time. As for what shoots the bullets though, I decided that I wanted them to be embedded in the level. At first I thought it would be cool to have enemies that bounce around like screensavers shoot bullets, and have the air-gun push them away. But I thought that would be a little bit too much for the player, and also would be too chaotic. So, turrets it was.

Play-Testing

I say play-testing, but it's more like play-test because only one person actually played it other than me. It became clear that the controls, or idea of the controls was unintuitive, especially the idea of hugging walls and pushing down against the wall to boost yourself up them.

But, I couldn't be bothered to make a tutorial. In the end it's no excuse to not have a tutorial or a little verbose wall of text explaining every microcosm of the game's mechanics, but still I couldn't be bothered. From what I gathered the idea did have potential though, it was pretty fun when I played it. The only thing holding back the idea is GDevelop lol.

Conclusion

Honestly though, I have no real clue why I make the decisions I make, I've made a few games already and for the most part my brain just says "This should be good",  and the more experience I've accumulated the better the ideas my brain generates. I tried to explain my thought process the best I could, but in the end the only way to learn how to design games is to design games, get an intuitive feel of what makes a good and fun game, and use a little bit of game design knowledge to analyse your intuitions. So in the end it comes to just making games. So go do it :).