Then we disagree on what a "Top Seller" is and how it should be calculated. Under your scheme, a vendor could put up 10 crappy modules and "sell" one of each to his buddy (or an alias) who then upon giving them all a 5-star review will make him a "Top Seller". Not everyone reviews vendors so the data collected is a sampling at best and could easily be skewed. By using hard data, such as number of sales in the last 12 months, the system should be more accurate. As see more for your assertion that I would exclude modules that were created some time ago, you have misunderstood. I'm well aware that some very good modules have been kept (or been designed to be) in very good working condition and are very popular... so should be included in calculations. My suggestion was to allow vendors with older modules that get few sales to have the option to keep them listed but tagged as "older" (or some appropriate wording) so they can be ignored in calculations *IF* such calculations are based on a group of modules and their combined sales. Maybe what is need is both a rating system based solely on hard facts and a separate one based on buyer feedback. However, I would suggest anything based on buyer feedback require a healthy minimum of input before any kind of assertion is concluded from the feedback... thus eliminating the possibility that a single 5-star review for a product could allow it to be translated into label of "Top Seller". Plus the age of the review should be taking into account since I have found cases where vendors started out, some years ago, getting great reviews then appear to have stopped offering "great" support... resulting in a growing number of slowly (if at all) answered requests for help. Only recent reviews (last 12 months?) should be considered... that's what I'm interested in knowing, i.e. is this vendor currently active in properly supporting his modules?
I totally understand your point and do not disagree with calculating in the number of sales somehow. We can go back and forth all day long probably, but in order for this to work buyers NEED TO POST REVIEWS (good or bad). I can't tell you how many market items that I know for a fact have been sold, but yet do not have a single review on them. So until people start posting reviews, none of this will work properly.
One of the suggestions i made a while back on that was to do something like netflix does after you have watched a movie.
A few days after someone makes a purchase, the site should present a section at the top of the page showing the module they purchased and ask them to review it. I believe that will increase the number of reviews, and the more reviews there are, the more accurate any automated system will be.
A few days after someone makes a purchase, the site should present a section at the top of the page showing the module they purchased and ask them to review it. I believe that will increase the number of reviews, and the more reviews there are, the more accurate any automated system will be.