Personally I find that improvement follows the conscious/subconscious competence cycle.

I’ll be exploring a new concept — maybe a new opening, integrating a new middle game concept, learning about a new pawn structure — and my rating will drop a bit as I’m not as smooth in using it and it’s not well oiled. Then I get better at it, and it slowly beds in as something I don’t have to think about, and my rating goes up to more than it was before.

I’d recommend not playing to destruction, however. If you want to improve, I’d recommend having a look at a few videos of GMs playing and explaining concepts. It’s like science — you wouldn’t expect to reach an understanding of relativity using the scientific method from scratch in one lifetime; you attend a few lectures to get the grounding you need to then build on. Happy to recommend specific resources if you’d like.

For me, each step seems to have corresponded with hearing a new piece of advice and gradually learning to apply it.

For example, "don't play hope chess" or "don't force your opponent to make a good move" were both things that I had to work at, but once they clicked, my competency felt a lot better. And each thing only made sense once I was at a point to understand it and ready to start applying it.

In my personal experience, playing the chess.com lower elo bots to improve is complete ass. They make mistakes players of the same rating very rarely make, just nonsensical moves. They also rarely punish blunders you make, so you're not learning much at all. The same goes for the game analysis tool. Sometimes it will flag you for an inaccuracy or mistake, and will feed you a 10 move line where you lose a pawn in the end. Useless for beginners and even intermediate players.

I improved the most by watching Aman Hambleton's habits series, followed by one of his many "speedruns" in a particular opening. Habits will teach you great fundamentals, and an opening (especially a "system") will show you more specific examples of where the pieces should go (in that opening).

Paired with daily puzzle solving, progress is quick. I recommend Lichess' puzzles, because they're unlimited and free. You can choose a specific topic to practice each day, for example, so you know what to look for, and then once or twice a week mix it all together to test whether you can recognize the patterns.

My buddy who was a 2300-ish player suggested the book "Chess The Easy Way" by Reuben Fine and I read it. It was written around the 1940s and is still a great book. "A Primer of Chess" by Capablanca is also great. And for total beginners, "Bobby Fischer Teaches Chess" is surprisingly helpful (it's almost entirely about back-rank checkmates).

Other than that, play a lot of slow OTB games. Don't play speed chess. Check every move carefully to make sure it's not a blunder. To do that, you have to spend time thinking at each move. That's not compatible with speed chess. Ancient advice is to literally sit on your hands while deciding a move. That is to slow down the impulse to move before you have completely thought it out.