I guess this sort of depends on the sample set. If it was fed all webby award winners and sites from dribbble then scoring a 51 would be somewhat amazing. An exam has an absolute range given an absolute understood input. I suppose the weird thing here is that we don’t know the inputs. If mainly HN and UX conscious forms have been using the system then 51 is also good.

I don’t have an answer here but it got me thinking. Thanks for this demo! My company site scored a 72 and I might be happy about that.