alternatively, you can do what i did and make one shit layout and bang your head against it until it's not shit anymore. only takes 2-3 years or so
Dont instantly open hammer
Make drawings first.
Drawing might be inaccurate but is able to give more ideas on what you want in the map and how you potentialy could connect them.
Once you have a good idea dont start building it yet, give it some rest. I had some ideas before that once i did actualy think about it didnt realy make sense and would be a hell to make. By simply waiting you will still think of the map and improve it to a point you cant realy get more ideas.
Just think of problems like sightlines, height variety (which you cant decently draw), balance problems, complex areas (which you should avoid). Think of what an brainless person would do if playing the map (follow straight lines. dont look around corners - that means over 90 degree corners - for other paths). Try to get the map to a simple yet interesting shape.
If you are unsure of an idea (if its possible in hammer) try it out first, but only that part and dont make it realy detailed... keep it to experimenting only.
Once your idea phase gets truly stuck start mapping, you got a good blueprint of the map which in your mind already survived the theory test and should have solved most issues.
Intercept was thought about 4 months before i even started mapping, just to solve the issue of giving it a nice shape, well thought out options and so on. In the meantime i simply did have some drawings to be sure im not forgetting things and notice other issues. And even with that thinking the map ended up about 70% diffirent than the drawings, but its key parts to prevent issues remained.
In tf2 i only made a few bad maps (because i didnt realy think and just started mapping - for example checkmate, invasion, donut and ramparts) but the ones i did plan, ended up quite good (skullcove and intercept)