N ideal time to use marketing

Kalec's answer has some good general points but misses the actual question a bit, so here goes:

First, all features improve the game if they're included. Some are mutually exclusive, though - you can't have stereo and surround sound, for example. Obviously in a case like that surround sound would be a better boost to game quality than stereo.

BUT: Each feature costs development time to implement, and you only have a limited amount of it. If the amount of features you're pushing into a given field requires more dev time than you've allocated, the overall quality of that field decreases -- you'll see a percentile value behind the field name. When that happens, you'll have to pick and choose which features are really important, or allocate more time to that field.

Bigger games have more dev time, so can fit more features simultaneously.

Having a game engine gives you a quality boost in the features that you implement into your game. So say your game engine has 3D graphics V1 and you use that in your game, your graphics are going to produce more design points (and possibly tech depending on the stat, I'm not sure), resulting in a higher quality game overall. These will increase your ratings and sales.

It's like software projects in the real world. If you have a framework to work with from the start, you don't have to spend as much time on the back end and you can focus on the fine details.

Game Engine Parts: There are some features that you need to have in an engine to use in your games.

I would recommend creating a game engine as soon as you can. There really isn't a reason not to if you are producing profitable games. It doesn't take a lot of time for a simple 2D linear story engine, and you'll see the rewards immediately.