Ian Sheperd's Article on Loudness Metering

It has nothing to do with pan laws. Don't change them mid-project, it'll screw up your mix. Your best bet using bundled plugins in SONAR is Adaptive Limiter, which conveniently displays LUFS values.

Other really good free options:
 
TBProAudio dpMeter II
 
http://www.tb-software.com/TBProAudio/dpmeter2.html
 
MLoudnessAnalyzer (in the MFreeFXBundle)
 
https://www.meldaproduction.com/MLoudnessAnalyzer
 

I think it depends a little on what streaming services you may be using. Apple for example like -16 LUFS. Apparently that is the level that the automatic soundtrack feature will strive for.  The level where it may do very little to adjust.
e.g. if a track is way louder than -16 LUFS then Apple will turn it down. And if its way lower than that then I assume it turns it up. Trying to get to the -16 LUFS reference.
 
-16 LUFS is very close to a K-14 master as long as the K-14master maintains much the same level all the way through.
 
Where things can get tricky is if a track is sitting at -14rms (or -16 LUFS) for say 2/3 rd of it but for a section then the overall level drops for say a 1/3 rd of the track. The LUFS reading will be lower then.
 
I think -18 LUFS might be a little soft.  As someone who is mastering tracks for people a -18 LUFS master would not be loud enough for them.
 
I have worked at K-14 for quite a long time and it seems now that it might be the perfect level to strive for.  The video and TV guys are wanting their soundtracks to be down at -23 LUFS which is also pretty close to a K-20 master as well. (Assuming overall constant level throughout) 
 

YouTube has settled on -11 LUFS. Spotify uses -14 (a recent change from -11). These are the target long-term values, above which they turn the material down. AFAIK none of them turn anything UP, only down. There would therefore seem to be no benefit in shooting for anything too far below those values.
 
Because I used the K-14 scale for years, it was natural to adopt -14 LUFS. But lately I've been targeting -11 because it works better for CDs. If played on YouTube it would not be lowered, and on other services it would only be turned down by 3 dB.
 
If you want to hear what a difference it makes, go to YouTube and listen to music videos. First listen without knowing its loudness and note whether it sounds clean or muddy, dynamic or flat. Then peek at the actual loudness value (right-click on the video and select "stats for nerds"). It'll be given as -X dB where X is the difference between the raw audio and the YouTube target loudness.
 
Here's one that was apparently mastered specifically for YouTube. It's one of a series, every one of which is mastered to the exact same loudness target. And it sounds lovely.
 

 
 

There are plug-ins that let you hear your mixes under the influences of what the streaming services ref level checking and adjustment might be doing.
 
https://www.nugenaudio.com/mastercheck-playout-loudness-and-dynamics-metering-aax-au-vst_41
 
For a track that is fairly even in level overall, the K system value is about 2 dB higher than the LUFS level. So a K 14 master is producing close to -16 LUFS.  Aiming for a -14 LUFS master is similar to K-12. Which Andy likes to master to.  It is still important to know where rms levels fit in and how they relate to LUFS levels.  Working with VU levels right from the start will set up you nice for a great target LUFS master.  I say dropping VU level reading is not good advice at all.  If used wisely it can set you up for a great sounding mix.  Nothing to stop you from using several measuring systems.  I tend to work with VU rms for a lot of the production. It is actually easier and faster. Are you going to fiddle around with LUFS readings on tracks for example? I end up with perfectly balanced loudness wise tracks for an album using rms readings. Then in mastering you can bring in extra reading systems. e.g. LUFS and I also like checking on a dynamic range meter (DR) as well. 
 
I work at the same level for all the mix production.  That is either -14 or -20 for me.  It is in mastering that I make the final change.  I think in time it will be the norm for mastering levels to be much more sensible.  I think -14 is an excellent level.  It has a good balance of loudness , dynamics and transients.  
 
The -23 LUFS ref level is also very versatile too.  If you make -23 the constant level for a soundtrack for a film say, it means a whopping 23 of headroom is available if you needed it. (They certainly achieved it in Blade Runner 2049!)  It is also very close to K-20.
 
The thing that makes LUFS interesting is the time over which the reading is taken. It tells you a lot more about the overall level of something. Over much greater time.  VU rms is still longer and slower than peak values but over much less time than LUFS.  They are all relevant here. Peak reading, rms reading and now LUFS reading. We need them all.