Hacker News Reader

Comment by Havoc | Hacker News Reader

>Until you have a listing for your headphones that comes with 5,000 pre-made positive reviews!

Fk. That's possible? How has Amazon not done anything about that?

Like I can get keeping fake stuff out is a losing battle, but that seems like a rather significant design flaw if you can "transfer" reviews like that.

View on HN

I've sat in on calls with Amazon Employees where they tell you to do exactly this to boost sales.......and then deny that they ever did it when you are contacted by the IP rights holder.

https://brandservices.amazon.com/projectzero

>I've sat in on calls with Amazon Employees where they tell you to do exactly this to boost sales.

Well that's next level depressing.

It's weird. All these amazon complaints - they ring true to me...yet I've thus far escaped the negatives thus far despite heavy amazon use (say package every 48hrs).

What team did those people work for?

The one that contacts you and wants to “help you sell more”.

The point is that each of these types of teams has no insight into the underlying structure of how Amazon actually works. As a business you are forced to buy into the racket in the hopes that you are not accused of being an IP infringer...try hiring a decent law firm to help you get out of this debacle....oh wait they all can’t help because they work for Amazon already.

States attorney or class action? Going after Amazon for structuring a system enabling fraud on purchases they get a commission on sounds like a good position for anyone able to represent consumers in bulk.

Think about all the AI/Machine learning work going on. I really think it's not a losing battle, that if Amazon really cared, they could train and maintain a "product listing integrity" AI fairly easily.

That they don't already have such a thing to dissuade the gaming going on to dissuade the sketchy practices essentially means to me Amazon condones the behavior.

I'm going to be kind to them and assume they are using the AI to augment the obviously simple "rating delta over time" I'd do as the first pass solution.

Like "Over a 3 month period this product reviewed at 3.7 and the next 2.1" maybe look into why.

During a recent Amazon all-hands meeting, the company leadership talked about how there are something like 5,000 employees working on various marketplace frauds, including using ML on some of this.

I have no idea why it's so hard, but I don't think Amazon is being as complacent as it seems from the outside.

Frauds where Amazon loses money, or frauds where Amazons customers do not get the product of its purported reputation, or maybe not even the product they thought they ordered.

They were specifically talking about marketplace fraud, where Amazon customers are misled but no money is lost. Customer fraud (against Amazon) is a different thing, though there's a bit of overlap when it comes to fraudulent returns.

This is actually rather easy to solve without ML/AI. They need to track the change delta (ex: with jsondiffpatch) and weight each part, presumably using levenshtein or text mining with weighted dictionary, and limit changes to 2 or 4 in 24 hours period. Then weekly compare those changes and treat the highest one as need to be reviewed by customer.

Additionally, each customer review should tied to specific change version, that they can easily compare with current one.

Easier, no editing. Create a new listing for any changes. Allow create listing from.. or linking to old listing for repprting. Show new version available text instead of buy again. Solved.

Amazon thinks it can automate itself out of these problems

Well here you probably can - that kind of substitute a product can't be that hard to detect.

It could, by saying that the category for a product in the store is immutable and that if you want to change it, you must create a new item. If that doesn't work, extend it to the item's title and removing any of the pictures of the product (you can add more, but it's append only). Should work fine to prevent this.

That requires programming it... I mean amazon literally wants to automate these issues without actually programming it.

No, it thinks it can get away with turning a blind eye to these problems.

There are literally several thousand people who work on marketplace fraud and quality issues. It's a huge concern.

Just like how there are literally thousands of people who work on ensuring customers' privacy at Google?

So easy for armchair quarterbacks to make snippy remarks. Truth is that fixing this kind of issue is probably way harder than you think.

Ebay doesn't have this problem and the solution is extremely obvious. Amazon has no third party seller rating. There is no way to confirm the trustworthiness of a third party seller. People cannot review sellers, they can only review products so sellers can easily get away with reputation loss without having to create new accounts by simply moving on to the next product because Amazon as a whole absorbs the reputation loss. I personally loathe buying on Amazon because of this.

> Amazon has no third party seller rating ... People cannot review sellers

Sellers have reviews and ratings just like products. You can leave public feedback and a star rating for sellers you've purchased from [1]. On the 'Your Orders' page click on 'Leave Seller Feedback'. This page also describes how to do it [2].

You can see these ratings while selecting which seller's offer to buy from. For example, many 3Ps are selling '5 Hour Energy Shot Berry - 24 count' [3]. Alongside each seller's offer, you can see their star rating and positive feedback percentage. For the first seller I see a 5-star rating and "96% positive over the past 12 months (1,122 total ratings)". You can click on that to see individual feedback.

It's my understanding that these ratings influence whether a seller's offer will be chosen for the 'Buy Box', which is the default offer that's shown on the product detail page. When an offer appears in the BuyBox, underneath it will say something like: "Sold by 7-Eleven, Inc and Fulfilled by Amazon". The seller name is a hyperlink to the seller's profile page where you can see their ratings and reviews [4].

Amazon also has a slightly different problem than Ebay, as I understand it, because Amazon is trying to have a single product detail page for each "product". There should be in theory exactly one page for "24-pack of 5 Hour Energy Berry". I'm not experienced with Ebay, but it looks like every Ebay listing is completely separate and independent. Doesn't that make it harder to comparison-shop for the best offer for a specific product among multiple sellers? Is there an easy way to see all offers for one specific product? I couldn't see an easy way to do this while experimenting with Ebay search right now. Amazon is trying to make this comparison shopping easy by consolidating all offers for the same product into one page, and then selecting a recommended offer based on various factors (including seller rating, shipping speed/reliability, etc.) that matter to customers.

(However, because there is a single page for each product, and sellers have some ability to create and modify pages, that occasionally leads to abuse, like the kind being discussed elsewhere in this thread.)

[1] https://www.amazon.com/feedback [2] https://www.amazon.com/gp/help/customer/display.html?nodeId=... [3] https://www.amazon.com/gp/offer-listing/B0021L9XK4 [4] https://www.amazon.com/sp?_encoding=UTF8&asin=&isAmazonFulfi...

pssht. I emailed jeff@amazon.com to complain about the general problem of counterfeit BeyBlades and provided links to scam items. Rather than take down the scam links, or even acknowledge the problem which was causing my inability to trust amazon for these purchases, the response was that they could refund any specific items I bought that were scams.

Welcome to Day Two

I'm not seeing enough here about how Jeff Bezos is actively supporting and creating a system that enables and promotes fraud.

Sounds like a great idea for a deep learning neural network...or just lock all the fields down after a product is created.

Amazon hasn't done much about anything

Reply on news.ycombinator.com