For an equirectangular map, you can measure distances exactly up and down or exactly at the equator. Any angle other the up/down or left/right at the equator will cause distance distortions. Area and angle are always distorted for that projection, but the distortions gets smaller as you get closer to the equator (the horizontal scaling factor is 1/cos(latitude) for that projection).
The smaller an area that is shown on a map, the less spherical distortion will matter. A nice whole-world map will give you a good overview, while regional maps will make it easier to measure. Mollweide is a nice projection, but distortion increases as you move away from the center of projection. An interrupted map allows you to have multiple centers of projection and to place those centers of projection on places of interest to minimize distortion in those areas.
As an example, the below images show an equirectangular projection, a whole-world Mollweide, and an interrupted Mollweide (I picked interruption centers for convenience of editing rather than for any useful properties). You can see how the distortions are reduced in the interrupted version.
Untitled-1.jpg
For what you're trying to do, I recommend making a set of maps that cover your primary areas of interest. If you use a grid for each area that is consistent with a global grid, then you should be able to use your regional maps as working maps for measurements and reproject any changes to those maps back to a whole-world map that functions as a master index for your maps. It's not the most convenient system, but one of the reasons why there are so many kinds of projections and scales of maps is that they all serve different purposes.
For a lot of purposes on the scale of small continents, the Lambert Conformal Conic projection is a good enough choice.