I've begun using Chat GPT to give me feedback on how to balance my thumbnails to be better and more attractive while representing my brand. Even when I've got a decent design this has 10x my ability to create new layouts with feedback on how to make slight adjustments to be design balanced. Here's an example of a thumbnail I tried to make on my own (First Image) then attempted to include some of its suggestions, but my words were blocking details, my branding while represented was too glaring, and the image I selected didn't stand out enough.
Version 2 in the middle, and Final version on 3.
I've shortened my thumbnail intrigue sentence, pushed up the contrast (Because the image was too dark), focused on the main character's expression by outlining her, and used an iconic shot from the film that any fan would recognize. The facial expressions are hopeless and scared despite the juxtaposed sentence on it to create a curiosity gap for people to click onto it.
I'm very hopeful for this thumbnail when this video releases in a couple of weeks.
BUT, here's the best part. I need to create THREE additional thumbnail layouts with this exact process of an entirely new image because then YOUTUBE will allow me to A/B/C test three thumbnails against each other to see which has the best click-through rate.
So wish me luck. I'm off to create three more.