1600 reviews, almost a perfect rating. The reviews and questions are about other products like a bread machine. I reviewed the product as 1-star but my review doesn’t show up. I reported this listing to amazon several days ago but it remains.
Raise your hand if you couldn't knock out a prototype algorithm over a weekend when handed that set of data. FFS, with something that obvious, your "machine learning" could be a bevy of if statements.
1. This goes a long way to show that there isn't probably zero human interaction with the "Amazon's Choice" listings and that it's completely algorithmic.
2. Making fraudulent listings on Amazon is extremely effective. So effective that it can be sent to the top of the charts.
My question is this: even if the listing gets removed, I'm sure it's been up long enough to make an insane amount of money, probably far more than they paid for the review-bots or the stock (if they even sell the correct thing). If they get the listing removed and "banned," what's to stop them from making another shell corp and doing it all again? What's the barrier for entry for fraudulent listings?
Tangentially related, if "Amazon's Choice" is based on reviews and popularity, what stops it from becoming an echo chamber? Age of product doesn't seem like a good factor, nor time spent in the amazon choice spot.
I've known an amazon worker who worked on the amazon's choice functionality at one point- it's all algorithmic with little to no human interaction. It's also skewed towards the products amazon stands to profit the most off of, obviously.
Amazon desperately needs to fix their completely broken third-party marketplace. There are entire categories of products I simply won't buy from Amazon as its near impossible to find anything that even approximates an authentic listing with genuine reviews. Yet Amazon does nothing. Amazon genuinely doesn't care that their customers are regularly getting ripped off, given they get their cut regardless.
The depressing thing is that there's probably some Amazon employee sitting at a desk in Seattle or somewhere, who keeps a meticulous model that tracks and compares their losses from returns on fraudulent listings vs. what it would cost them in effort and what they'd lose in commissions from actually doing something about the issue, and Amazon execs have basically decided that since they're making more money from running a marketplace that enables fraud, its somehow justified to continue enabling fraud.
Its basically the same situation as that Ford Pinto case from the 80s where Ford used a cost-benefit analysis to decide that it would be cheaper to settle cases where people were injured or died as a result of a faulty fuel-system, rather than redesign said system.
Hopefully at some point another marketplace with much higher standards and far greater customer protections will emerge and eat into Amazon's market share forcing them to take the issue seriously.
17 comments
[ 2.8 ms ] story [ 60.5 ms ] threadEdit:
This is actually more fun: https://www.amazon.com/stores/node/13968870011?_encoding=UTF...
All of the product have either 5 stars with hundreds of reviews or 0-1 starts with tens of reviews...
1. This goes a long way to show that there isn't probably zero human interaction with the "Amazon's Choice" listings and that it's completely algorithmic.
2. Making fraudulent listings on Amazon is extremely effective. So effective that it can be sent to the top of the charts.
My question is this: even if the listing gets removed, I'm sure it's been up long enough to make an insane amount of money, probably far more than they paid for the review-bots or the stock (if they even sell the correct thing). If they get the listing removed and "banned," what's to stop them from making another shell corp and doing it all again? What's the barrier for entry for fraudulent listings?
Tangentially related, if "Amazon's Choice" is based on reviews and popularity, what stops it from becoming an echo chamber? Age of product doesn't seem like a good factor, nor time spent in the amazon choice spot.
The depressing thing is that there's probably some Amazon employee sitting at a desk in Seattle or somewhere, who keeps a meticulous model that tracks and compares their losses from returns on fraudulent listings vs. what it would cost them in effort and what they'd lose in commissions from actually doing something about the issue, and Amazon execs have basically decided that since they're making more money from running a marketplace that enables fraud, its somehow justified to continue enabling fraud.
Its basically the same situation as that Ford Pinto case from the 80s where Ford used a cost-benefit analysis to decide that it would be cheaper to settle cases where people were injured or died as a result of a faulty fuel-system, rather than redesign said system.
Hopefully at some point another marketplace with much higher standards and far greater customer protections will emerge and eat into Amazon's market share forcing them to take the issue seriously.
https://webcache.googleusercontent.com/search?q=cache:HeZP55...