I will give you some insight into what has helped me. Look for weaknesses in your opponent's position, and make sure you're tactically aware. A combination of those two things will seriously help you improve. Also, talk to yourself about the position. I mean out loud. It's similar to the rubber ducking method that programmers use. I hope this helps you.
I only consider myself somewhat qualified to answer this because I have increased my classical rating by about 200 over the last year, cracking 1900 for the first time this weekend.
The most important thing is for you to be patient. It takes time both to learn chess concepts and it put what you learn into execution. This game is not for those with a short attention span or who are after immediate gratification.
Play slower time controls. I think longer classical games are the best measure of strength. My favorite time controls are 60+30, 90+0 and 45+45. Rapid is still kind of fast, I think, but definitely better than blitz or bullet, which I still do but not at the expense of longer games. Correspondence is also useful for things like playing openings you are just learning.
As has already been said, tactics and endgames are most important. I am focusing on endgames now, but I wish I had more earlier. Endgames teach you how to calculate and they demonstrate the full capabilities of each piece.
I am honestly kind of surprised I was able to reach as high a rating as I had not being very knowledgable about endgames. I think the reason for this is that at my level and below it is fairly common for either a blunder or something tactical to occur in the middle game or even the opening that results in a material advantage for one side. In games like this you either won't reach an endgame or you will reach an endgame that is so one-sided that doesn't require much preparation or accuracy. At higher levels I feel it is more common to reach endgames where one side has an advantage but it is a fragile advantage requiring a high level of accuracy. I don't think I can expect to reach much higher a level than I am at now without more advanced endgame proficiency and preparation.
Openings may not be quite as important but they still are important. Don't rely strictly on the engine in your opening preparation. When it comes to tactics and endgames, the engine is always right but in opening there will be times when it will recommend a particular move but there are in fact several or many reasonable alternatives. Also, it can hinder your preparation for some openings because it will tell you to expect moves in some openings that players who play that opening simply will not play. The Falkbeer Countergambit in Kings Gambit is one example of this. Also, at one point the engine told me that it was an "inaccuracy" to accept the Queen's Gambit, which is nonsense since QGA is fine line for Black that has a time-honored history of top level play.
I hope this helps. I hope to soon be telling you what got me to 2000. :-)
Well, in my opinion bullet resembles more an ego-shooter, so not sure what that rating means. Probably you have to subtract a couple of hundreds to compare it with some offline rating. If you want to improve at chess you should digest some serious chess content.
Nowadays the forums are full with people who wanna achieve quickly everything (see title). No, it takes time. Playing bullet in the internet, what‘s that?
Reading books is good though, analyzing your mistakes is even more important. Hopping to the next bullet game, no. Your choice.
Sorry but that is not possible. You are playing just two years, ten more years and you will see the improvement. I promise!
There is also an older discussion in this topic worth reading: