Other than the basic material count, there's piece activity, pawn structure, king safety and space.
So, maybe on the queenside black has isolated a and c pawns. These might prove very difficult to defend in an endgame. So something that seems harmless, like a trade of queens, may make the eval change quite a bit simply because of the pawns. More simply, near the end of an opening one player may have more pieces developed off the back rank.
With more space, you can maneuver more freely and the opponent will have trouble responding to the different threats you can generate. King safety is obvious.
Less obvious is when a minor piece is not very good. Sometimes nearly everything else is equal, but a knight or bishop can't find a good square or create any threats in the future. Then with good technical play it's as if that player is down a piece in the area that piece is no good. Sometimes it's a bishop and so the "area" you're going to use to win are the squares the bishop can't influence (the light or dark squares). Sometimes it's a knight that can't quite get to most of the queenside, so you can win over there.
How do chess evaluations work exactly? I know the signs that explain who is better or winning or ok but I want to know from where some evaluations come from. Some positions can look like it is equal, but maybe the engine can favor white over black for whatever reason.