It depends on your current level of chess understanding because it will mean working on different things. Estimate I'm coming up with now:
around 1000: You are out of the "opening" (and into the middlegame) once you seem to have completed chess opening principles https://www.chess.com/blog/KeSetoKaiba/opening-principles-again and the Rooks are "seeing each other" (all pieces developed off the backrank and the Rooks could recapture each others' square if something trades it off), then the "middlegame" here is mostly just making sure your pieces are not hanging (can be captured "for free") and to avoid basic tactics like forks, pins and skewers. While you are doing this, you are also trying to find tactics against your opponent to win material via forks, pins, skewers etc.
With more material, you have a better chance of winning the endgame - not guaranteed, but it makes winning easier if you have more stuff than the opponent 
around 1500: Everything from the section around 1000 still holds true, but now you are also focused more on improving your pieces (piece activity). If you aren't sure how to proceed, then often times, try relocating your "worst" piece to a more active place. By 1500 level, the player certainly understands the value of active pieces much more than the 1000 level player; "tactics flow from a superior position" (Fischer), so winning tactics are more likely to exist (or be created) in your favor if your pieces are more actively placed.
I'll include this in the "1500" category, but starting around 1600 chess.com rapid rating (maybe 1500-ish too) I started really getting introduced to the concept of weak squares, piece outposts and exploiting targets like an uncastled King, or attacking an isolated pawn. These are core to the middlegame for a 1500-ish player.
around 2000: (although I have to offer the disclaimer that I've not yet reached 2000 chess.com rating [1985 highest so far in rapid], but I've reached 2000+ rating on other site[s] and USCF Provisional rating of over 2000 before)
At this 2000-ish level, everything from the 1000 and 1500 level is still true, but also a much deeper understanding of the concepts the 1500 level mentioned. For example, although a 1500 might be able to recognize and utilize an outpost, a 2000 level player creates it, utilizes it and knows when it isn't worth pursuing - this requires a positional level of understanding the 1500 introduced to these ideas doesn't have yet. Everything is also slightly expanded too: 1500 seeing weak squares? 2000 can provoke weak squares in ways which they can make use of them etc.
The 2000 level also has a strong grasp on pawn structure considerations which both the 1000 and 1500 level players aren't even looking at! (I'd estimate 1800+ when most begin studying pawn structure concepts deeper than the fundamentals) For pawn structure I am talking a bit more advanced than isolated pawns, doubled pawns and passed pawns (although those are pawn structure considerations too). I'm speaking more on subjects like pawn majorities creating a passed pawn by force via structure, understanding elements of counterplay for both sides like the center attacking motifs of a hypermodern opening, or the pawn breaks in a jump formation.
The 2000 level middlegame plans seem to add a certain "positional" element which closely works with intuition - a GM doesn't need to calculate certain endgames because it is a theoretical win in the same way that a GM doesn't need to calculate much in a middlegame position because they know the pawn structure, or imbalances of the position, factors works in their favor.
Hopefully, this guideline I created is a good springboard to figure out where you are at and where you might want to dive into for the next step of improvement 
How to improve in the middle game? (chess)