- cross-posted to:
- technews@radiation.party
- cross-posted to:
- technews@radiation.party
There is a discussion on Hacker News, but feel free to comment here as well.
You must log in or register to comment.
Maybe they could use NPS (Network Promoter Score) style calculation? Translated to a 5-star system, it would mean taking the % of 5-star ratings, and subtracting from it the % of 3-star or below ratings.
It’s kind of harsh, but it’s more accurate than what they seem to be trying now.
But this doesn’t solve all the gaming of review scores by sellers.
Or just have three icons of angry, sad/meh, and happy. Do away with having to explain between 0 - 10 or 0 - 5 as rank with terms of satisfaction.
Regardless of rating and score calculation the same issues comes from people not understanding blanketed opinion. When you have several categories to rank or comment on then that is when you get a better gauge. Amazon just took the simplistic route.
I always end up having to go to the full reviews page anyway just to look out for what I like to call the C curve. A bunch of 4 and 5 stars, a few 3, then a big spike at 1 and 2 stars. Usually means that there’s a high failure rate and that a bunch of users got lucky with ones that work or they were fake. This shouldn’t affect much of anything if you’re always verifying, but most customers won’t.