Yes. For like a simple naive Bayes, you can say we start at 0, a user vote is worth +3, a merchant vote is +5, we decay by 0.1 every month, publish if we hit +1

Yes. For like a simple naive Bayes, you can say we start at 0, a user vote is worth +3, a merchant vote is +5, we decay by 0.1 every month, publish if we hit +13 (95%), un-publish at +10 (90%), etc. etc. It's the same math as logits, but the scale's a bit nicer.