Super interesting but one thing bother me. If thumb up is 0.5 and label might not be applicable or at least optional. A thumbs up without applicable label will result in a much lower rating than a thumb up with some or all label. Then you average the total, the product with non applicable label will get 0.5 and the one with applicable label 0.8, 0.9 or 1. In the end what prevent people to only choose products with 0.9+ and ignore the rest?
Discussion
Certainly, people may only choose products that are 0.9+, but the point is that the *range* of good products in 0.5 to 1.0 rather than 4.5 to 5.0. Anything rated above 0.5 is something that is generally "good" and people may be more comfortable buying things in a lower range.
A big factor in how this plays out is how you *show* the rating to the end user. Showing stars or numbers may not be optimal. We need to be creative in how we convey this information so the sentiment is not lost as well. I don't have all the answers but we will get there with experimentation.