The greatest power of REW is comparing changes before and after. This applies to speaker positioning, tweaks to gear or EQ, room acoustics, etc. In order for your comparisons to be valid, you need the testing to be identical before and after. I either leave everything set up exactly, or recalibrate each setup.
Testing speakers in the near field, your best graphics will be frequency response and distortion. Use the FR window that lets you show multiple measurements on one chart and apply some smoothing to each measurement (I like 1/6 octave). For distortion, compare THD but also look at individual harmonics, particularly the 2nd and 3rd. Try to find where harmonic distortion is rising above the noise floor.
Testing room response requires you to put the mic where your head would normally be during listening. Frequency response is useful here, but the most powerful tool is the waterfall to show reverb time at various frequencies. This will clearly show room modes and anomalies; your goal is to minimize reverb and make the curve as smooth as possible. Room acoustics are where you have the most opportunity for improvement. Any decent gear can reproduce audio to +/- 3db across the spectrum in an anechoic environment, but room acoustics often vary +/-30db or more with lots of time-based distortion that can ring audibly for a second or more.
