Gladwell's article reminds me yet once again why book review-related systems, whether they be Amazon star ratings or what Goodreads employs or even those 5, 10, or even 40 point scales utilized by some online reviewers, ought to be taken with quite a few grains of salt. One size-fits-all systems are ill-suited for heterogeneous entities, whether they be types of cars, university core subject foci, etc. This is especially true for books, whose elements often do not fit into nice pre-fab squares (one only had to look at the interminable debates over how to define "fantasy" and "science fiction," not to mention other genres and subgenres to see the futility on display there).
It has always baffled me to see many readers comment that they value ranking systems, considering how obviously flawed they are going to be. Can one rate, for example, a work by Rabelais with one by Sartre? Or a book written by Gene Wolfe compared to one by Karen Tei Yamashita? Should one even attempt to "rank" at all when doing a review?
I would argue that due to the wide variance between stories that having anything approaching a single-entity review system, particularly that which employs "rankings," is going to be a pointless exercise that detracts from other, more important aspects of reviews, which deal with the reviewer's interactions with the text and how well engaged that reader was with that particular text. Comparing Umberto Eco to George R.R. Martin makes as much sense as comparing a piranha to a cobra. Both might bite you in the ass and kill you, but the methodologies of each are too dissimilar to rank one higher or lower than the other.