The question is really a bit too broad to be capable of a direct answer. Things like choice of crossover frequency make and break speaker designs. You can do a paper/simulated design and think everything is good, but how it will sound is a whole other story.
The point here is that there is much more to it than designing an electric filter which (theoretically!) gives you flat frequency response. The crossover is a weak point in a speaker, because of the various odd phase and directionality anomalies that happen around the crossover point. Unfortunately our hearing tends to be most sensitive to these effects in the 1-4kHz range which is where xovers tend to want to be. (This is because we are evolved to hear speech, and most speech information is in this band.) You are slap bang in the middle of this.Some designers try to choose drive units to try to avoid this sensitive area.
Basically, to get really good results, you are looking at a lot of listening/measurement tests once you build the first prototype. These will tell you if you are somewhere close to correct.
In fact there are very good arguments that suggest that passive crossovers are a very sub-optimal solution anyway. Logically this makes sense : how did we end up trying to do subtle frequency tweaking at high power levels, and deliver the power to a very complex, low-ish impedance load? It doesn't make sense, from a system design perspective. Mostly its just done this way because it always has been. If you really want to tr to build a superior system - power amps are cheap, so consider powering each driver independently, and splitting frequencies at line level. You have far more options in terms of filter types, adjustment, and so on.