I think it’s difficult for buyers to understand the impact that ratings have on a seller’s account. Truly terrible sellers that steal, scam, bully, or produce crappy work should get low ratings and fall out of the search results so that great sellers can rise to the top. The problem is that a lot of buyers think a three or four stars review is a good review, and the way reviews are weighted, these can be really /not/ good reviews. Anything less than five stars hurts your percentage and therefore hurts your account and makes it even more difficult to make sales.
When a seller produces work that they feel meets the promise their gig made and they are given a three star “Satisfactory” review, they are understandably upset. That’s considered a positive review, but it’s also a 60%. If you’re a new seller who has gotten ten five star reviews and then one three star review, your 100% just dropped to like 96%. That feels and looks like a huge penalty when you thought you produced good work, or, at the very least, work that 100% met the expectations of your gig.
The biggest problem that I see is that the rating criteria are subjective. Or, at least, many buyers use them that way. They rate not how closely the delivery matched the promise made by the gig, but how much they liked the work. Something could match the instructions and the gig’s requirements 100%, but if the buyer thinks the work is “blah,” he can give a 60%. And whatever, maybe buyers should rate based on how much they liked the work. That’s just not what the rating criteria says the buyer should be rating.
The thumbs system, I think, made this a lot easier. Instead of having to decide between essentially fifteen different choices when rating a seller (three sets of five stars), the buyer had one choice: Thumbs up or thumbs down? Was this okay, fine, good, great? Or was it terrible?
So yeah, I don’t think anyone is going to argue with you that the rating system is broken.