A categorical quotient of a variety $X$ acted by $G$ is a morphism $f: X \to Y$ constant on $G$-orbits such that every $h: X \to Z$ constant on $G$-orbits factors through it.
Mumford on the page 5 of Geometric Invariant Theory states that if $X$ is
- reduced,
- connected,
- irreducible,
- locally integral,
- or locally integral and normal,
then $Y=X//G$ also is. Is it really obvious from the universal property (and why)?
First, let me make your statement a bit more precise. Saying the map factors through it (in your first line), while correct, should be there exists a unique morphism $g:Y\to Z$ etc.
Now, let me do this for 1. If $Y$ is not reduced, then let $Z=Y_{red}$. So, Then it is clear that $f$ factors through $Z$ and then you have $h:X\to Z$ and thus $g:Y\to Z$. Now, we have two maps $Id:Y\to Y, g:Y\to Z\subset Y$, and so by uniqueness, $Id=g$ as morphisms and then easy to see that $Y=Z$.