A working creative testing system moves in a fixed loop: define an angle, the argument you're making to the viewer, produce a handful of variants of that angle, launch them with enough budget to reach statistical signal, then kill or scale based on hook rate, CTR, and downstream CPL — not gut feel.
The discipline is in documenting what wins and why by angle, not just by individual ad, so the next testing cycle starts from a hypothesis instead of a blank page. Most accounts that plateau on creative performance are running one-off tests instead of a system.