The concept of an orbital directly derives from the concept of a free rotator in simple quantum mechanics. Before students learn about a free rotator, they typically learn about the harmonic (potentially including the aharmonic) oscillator and — first of all — the particle in a box. These are the three probably most fundamental quantum mechanical concepts which can be used as a first basis to understand everything else.
The particle in a box is a somewhat idealised case because the box is defined to have walls of infinite potential energy. Only in that case is the particle actually confined to the box. All other fundamental concepts in principle have a wave function contribution from $-\infty$ to $+\infty$ in all relevant coordinate axes. Which means that there is a nonzero probability of finding the particle anywhere in the coordinate system.
For the free rotator — which in principle gives us hydrogen orbitals — this means we need to define an arbitrary cutoff point if we want to display anything at all. Because 5 is a nice number, most of the time people choose a cutoff value of $95~\%$ and draw a line around the entire area (or volume) inside of which the electron has a $95~\%$ chance of appearing. Depending on your application, you might want to choose $90~\%$ (for less crowding) or $99~\%$ (to show more interactions) instead.
$100~\%$ cannot be used since that would put everything into an orbital and therefore the concept of an orbital would lose its usefulness. In general, an electron belonging to an atom somewhere in New Zealand has a nonzero probability of being found right here where I am sitting in Germany — and likewise a nonzero probability of being found at the other end of the Andromeda galaxy. (Note that while they are nonzero, they average out to zero at quite a large number of significant digits.)