Show HN: I filter Amazon products for above-average ratings/below-average prices

I look for fake reviews by measuring the relationship between 1, 2, 3, 4 and 5 star reviews. There is a normal distribution and when I find outliers, it usually indicates that something suspicious is happening.

If I understand what you mean by inventory co-mingling, there is actually a drill-down that's built into Amazon itself. For example, the reviews of Samsung's HW-T450 look awesome at first blush: https://smile.amazon.com/dp/B085WTFCQ7/ - but, if you filter the ratings to only display that particular model, you'll see that 22% of the reviews are 1-star: https://smile.amazon.com/product-reviews/B085WTFCQ7/ref=cm_c...

This an automated check that I perform, but I also investigate the reasons for these disparities. (In this case, it's because the soundbar lacks a digital decoder to support the surround sound speakers that are sold as an additional accessory). On rarer occasions, the negative reviews can be discounted, even if they are legit. One example of this is when lots of customers misread the dimensions of a product (e.g. I assumed it would be larger...).

That's a good question. The original version of the site was published anonymously, and for whatever reason, the mystery of it attracted attention. The novelty faded pretty quickly though and the site hit the skids. I asked family and friends for their input and they suggested that I personalize it.

I guess deal sites are a dime a dozen, and this does go some way to personifying it, like it's "great finds shared by a friend".

I'd say there's merit to both approaches.