I agree that reading MacLane is probably not the first thing you should do to understand category theory. In the beginning it's probably better to look at concrete examples of categorical concepts. When I encounter a new categorical concept, say coproducts, I try to figure out what it looks like in as many categories as possible (and whether it exists). The easiest category to start with is usually the category of sets, but after that I think it depends on what your background is. So what do a disjoint union, a direct sum, an intersection and a greatest common divisor all have in common? Well they're all coproducts in some category. Given a category it's also worthwile to figure out whether it makes sense to talk about an object with the desired properties. In the category of finite abelian groups you can't take an infinite direct sum of non-trivial groups, so you don't have all coproducts in this category. As in the rest of mathematics non-examples can be just as enlightening as examples (many much more so than the above non-example).

If you know some algebra I think working with the category of modules over a (commutative) ring can be very helpful, it certainly was for me. There are plenty of functors to look at; Hom, tensor, localization, derived functors etc.. You have limits and colimits, free objects, adjunctions (tensor-hom, free-forgetful). Thinking categorically about modules also prepares you for working with general abelian categories and homological algebra.