First-Layer Techniques
Forming a layer efficiently is a very daunting task, and certainly no one has it mastered.
This page will aim to explain some of the techniques that I've been using during L2L solving, but can also apply to LBL and Waterman.
The first thing I have to note here is just how important flexibility is.
If you keep one standard approach, no matter what it is, you're going to end up blinding yourself from easy alternative approaches. Not only do I mean this in a "don't always form your layer on yellow,"; but also "don't always form that way...or that way!"; I've found that it's best to keep a totally open mind about techniques. Noting that, here are a bunch of techniques that might help you if you decide to practice a method that would take use of these strategies:
Corners-first
Quite simply, this method solves four corners oriented/permutated correctly relative to each other, and then proceeds to solve the four edges while placing the center sticker. To train this, you should really just do a bunch of 2x2 solves. After getting your first (2x2) layer down to say 4-5 moves, decide how you're going to solve the rest. Although when I first did this I preferred to solve with the corners on D, as a bad habit from old Salvia practicing, I now solve the edges with the corners on left. I *strongly* suggest you to do the same. On that note, I also suggest you try solving your corners on left, to not deal with awkward rotations after wards.
Once you've completed the corners, it's obviously onto the edges and center. Usually, if none of my edges have already been solved, I'll try and solve an edge (or 2!) while placing the center, attempting to be efficient. After such, I suggest using sandwich-like techniques (DU to UL? U M2 U') to solve multiple edges at a time. The most painful part of this approach to FL solving is when you have L-slice edges already in the L-slice. Just practice these, and avoid them if possible.
Once you master all of that, be sure to try and solve edges during your corners. Crazy wide turns are your friends here. 
To learn more about this approach, please visit the main Waterman page.
Edges-first
although this feels very CFOPpy, which is deterring in some aspects, when a 3-move cross comes up, I really feel the need to take advantage of the situation. Once you've done the edges (I'm assuming the average user here can form a "cross"), do two edges at a time. I've thought about making algs for this, but don't really think it's necesarry - they are fairly intuitive to do two at a time.
1x2x3-first
For your 1x2x3, try to use ~7-9 moves. I'd check out some Roux documentation if I were you, since it's available, but I'll attempt to give you some tips.
Go slowly. Really go slowly. This block *can* be done in one look once you practice enough, but you really need to be efficient here when you can.
For these, I have 3 methods. Yes, methods for my submethods.
1- form a 1x2x2 and expand. 1x2x2 should be ~4 moves, and expansion likewise.
2 - form 2 1x1x2 mini-columns and then throw the edge/center in. Movecount should be the same as the above.
3 - get an edge attached to a center. Form two mini-columns as described above, without inserting them. Then insert them. I've seen that many people make the mistake of throwing these mini-columns as soon as they can, and this often increases move-count.
During this block, I suggest trying to get one corners done (UBL for me) as well. Then
Triangle-first
Many times, I find that making a triangle of sorts and then expanding the other three pieces, namely on D, is a very efficient way to solve. There are of course a few techniques to this approach, but I shall only explain one here.
The method I first attempted was to form a 1x2x2 block, and attach two corners onto that. The problem I find with this is that placing the final 2 edges and corner is inefficient and jumbled.
To avoid this jumbled solving, I recommend placing all three remaining corners at the same time after forming your 1x2x2 block. This will let you finish with a corners-first end-game. Typically, my solving at this point is done with just
Just do stuff
Once in a while, I get a little feeling that tells me to solve something that doesn't usually make sense. It works. If you get a feeling that tells you to make a 1x1x3 column, then do it. If you get that feeling a lot and it doesn't work out...that's the wrong feeling. There's not much I can say for this, besides it usually works out for me. Somehow.
I am definitely not a master at this, nor is anyone else as far as I'm concerned. I'm just expressing some ideas and hoping others will contribute their own.