Well you can run audyssey, measure, try toe in, run audyssey, measure and compare... However the general rule of thumb is use REW and the tools at hand (placement, manual EQ, treatment placement, toe in, and such) to get the best frequency response and other measurements possible with audyssey off. Then run audyssey and verify yup that's the best end result yet. There are situations and folks that recommend manual EQ after audyssey though. Which ones right for you? Easy answer, measure.
One common thing to do with REW is tweak the transition range between subs and mains. One way to go about it... you do a audyssey min run to get the distance and level recommendations ballpark right, then turn audyssey off and tweak distance / phase while doing multiple measurements to get the best possible, then re-run audyssey. After the full up run you put the distance and levels back to what you determined was actually best (audyssey is generally pretty good at this but you can make some improvements)...always do an end measurement and check, is that better then the previous best? Do I really like the sound better is an important question to but in general if it measures better it sounds better and if it doesn't commonly it just takes a little time getting used to it and you will usually realize it does sound better. but if not, you're the one listening to it and it's about what you like.
Hmm contemplating new speaker X, or amp y, or cable Z, or component B... put it in your system, break out REW and measure. Is it better? is the improvement worth the cost? Do speakers matter? Do cables matter? What about amps matters? How much is this dac helping over the one in my AVR?
Keep toying with REW and reading. It is an invaluable tool, i enjoy doing the tweaking, and can't imagine going back to no measurements throwing money at "upgrades" fairly blindly mostly based on the "helpful" advice of the "experts" at the high end store who would never steer me wrong even though their salary depended on it.
Don't look at it as a set it up run it and tweak for an hour or three, ok done and the system stays like that forever. Experiment and try different stuff. Keep the old measurements and make good descriptions for comparisons and better ideas what to try next (and what you should perhaps go back to)