But they are hard to reconcile with why Amazon keeps showing me a rice cooker ad, after I buy a rice cooker from them.
I've done similar things a lot, in this case it was because the first cooker had too large a minimum cooking amount. I'm currently looking at buying my mother google WiFi - which I bought a few weeks ago. And a few months back I rebought a tool I had just purchased because I lost it.
I currently work for a (responsible) own site adtech (personalization) company. Its surprising what behaviours our clients find profitable for their business.
It seems like an effective ad would be this: Say you bought a rice cooker on Amazon for $75. Now Walmart shows you an ad for the same rice cooker for $60. That would certainly stick in my head- not for the rice cooker but for the price difference that Walmart can provide.
With the Google wifi I had forgotten, and it caused me tk reconsider and write it on my whiteboard.
The tool I wpuld purchase anyway.
One thing to keep in mind is that the purchase data is per-user, so it's different from the segment approach I was mentioning above. It would require a whole new schema* that is queried whenever bids are processed for an ad event.
* Sorry for using the word schema here. It was used a lot on a team I was on in a similar context, and it bothered me at first but I got used to it. What I mean by schema is a "blob" with purchase data that is periodically processed by the bidding subsystem to add bid exclusions per user.
Is it possible the ratio is so large to where while it obviously could be optimized to exclude people who have since purchased said item - the optimization is only a rounding error, and may not turn out to be worth it?
But maybe they should. The recipients of those 9,000 ads are being trained to ignore the recommendations because they’re so useless, and may continue to ignore its suggestions even after it improves. Likewise, a few clueless recommendations can “spoil” good ones shown at the same time.
None of this shows up in the test set, of course, but people tend to turn off their brains when ‘evaluating’ ML stuff.
How can you be sure that "random" untargeted ads don't have just as bad of an effect? Ultimately Amazon can see the numbers and we can't, and they made their decision based on that. It's pointless to argue over theories without data.
It’s quite possible that those are still the best items to recommend to us, to mop up any “unhappy with the one they got, so save them with a return/rebuy” sales.
For 99% of people that probably doesn't happen, but the people with the money to do that probably spend a TON of money, so they are worth targeting with ads.
You might not be happy with the purchase and on the lookout for a better model.
The law article cited in ths news story seems to be saying that even though Amazon's practices haven't visibly hurt consumers in terms of retail prices, this is due to the ability to play legal games with corporate accounting that conceal the present and future anticompetitive effects of Amazon's business model, and that its status as the dominant platform for online retailers presents some degree of threat to competition.
Lina Khan published a very informative and influential review on Amazon and antitrust in 2017. https://www.yalelawjournal.org/note/amazons-antitrust-parado...
What Sussman actually wrote was "In addition, [this paper] calls for the establishment of a viable framework under which a case may be brought against such firms [as Amazon] by injured competitors."
The journalist says no antitrust laws need be changed. What he fails to mention is the small matter of this framework Sussman says is needed before any case can be brought. It does not exist.
Sussman did not describe the framework in any detail or how it would work. All one can find in the paper is this single sentence: "It is, therefore, necessary to develop some means of oversight that will allow the public to determine whether firms such as Amazon are engaging in such activities."
"Linda Khan published a very informative and influential review on Amazon and antitrust in 2017."
I can remember when a story on the Khan paper was posted to HN. I thought at the time the journalist seemed just a tad overly optimistic. It felt odd to see a law review article being so "hyped" by a journalist. It almost seemed like he/she did not want the reader to actually read the paper. There was so much paraphrasing that the risk of the journalist getting something wrong would be very high.
In 2018, there was another paper in Yale LJ by Baker and Morton that discussed Amazon's anticompetitive activities. This paper was actually influential.1 Unlike Sussman's single sentence calling for "some means of oversight" and a framework under which cases can be filed, the authors included an entire section on "US Enforcement Challenges", about five pages devoted to how enforcement might proceed. The paper was cited by US Sen. Blumenthal in a letter to DOJ and FTC urging them to open investigations of Amazon's predatory pricing.2 (He did not cite Khan's paper.)
Not sure if the Baker and Morton paper received any hype from journalists, but it is indisputable the paper and the letters to DOJ and FTC had an impact. Amazon finally removed the MFN clauses from its agreements with US sellers, as it had done for sellers in Europe in 2013.3
1 https://www.yalelawjournal.org/feature/antitrust-enforcement...
2 https://www.blumenthal.senate.gov/imo/media/doc/12.19.18%20-... https://www.blumenthal.senate.gov/imo/media/doc/12.19.18%20-...
3 https://www.bbc.com/news/business-23881202 https://www.axios.com/amazon-price-practice-antitrust-elizab...
...so, not predatory pricing then?
If predatory pricing is effective, an algorithm can be taught, or learn to, exploit it if fed with the right data.
This, quite obviously, shouldn't change anything regarding the potential legality, as it's entirely self evident that a sufficiently powerful machine learning algorithm will exploit any patterns not expressly forbidden.
As always, the responsibility for following the laws, and making sure these extensions of our own agency we call machines also follow the same laws fall on whomever is making the machine act in their stead.
One can however be assured that several companies will argue otherwise, argue as if they were simply a victim of some sort. Which might be true in a sense. Victims of their own devices, both figuratively and almost literally at the same time.
[1] https://www.russellmcveagh.com/insights/march-2019/ai-and-pr...
[2] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3209781
How does it come about that that is that a concept?
Does the thought go something like: "We neglected to instruct the algorithm not to collude, and failed to monitor outcomes for collusion, therefore we're not responsible?"
> and each gives its software total autonomy to set prices
Ah, yep.
see Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209.
essentially, all you have to do is prove that there is "dangerous probability" that a monopolist can recoup their losses after all competitors have been priced out.
Predatory pricing is either legal or not (in the eyes of the court, or legislations).
Not to this shareholder.
If Amazon is playing a dangerous game, then it's risking my money, which is not OK.
If you're so sure of that, then why did you buy AMZN? Why have you not sold your AMZN?
Amazon always ships packages ahead of time, at my leisure, and often cheaper than retail. They always process returns and complaints in my favour, without contest. They reach out to local logistics companies in order to improve delivery, and I can tell by the fact that those companies now have branded vehicles (and much happier and more attentive personnel than UPS!), and better verification processes (they found a logistics provider that actually emails the signature given at the door to you when the delivery is processed) that they've done well out of that.
I feel they're serving me very well as a customer, and every ordinary person in my life who isn't competing with Amazon is richer and happier because of Amazon's services.
> definitely not ... the businesses on the platform
The businesses fulfilling through Amazon would absolutely not have received my business were it not for Amazon; in many cases they do not have the expertise to handle the orders correctly, and I do not trust all of them to process the payments securely to the degree that I trust Amazon to.
They have every opportunity to improve their ordering and fulfillment systems, but they choose Amazon because it's where the customers are comfortable already.
I dreamt of a logistics integration system much like Amazon Marketplace when I was 16, and every year they make it closer to that dream, I think it's great. I'm willing to accept an argument from concrete evidence that they have somehow deliberately and systematically broken the law in so doing, but I've yet to see it.
I don't trust Amazon not to send me fake items or cheap Chinese knockoffs. There is absolutely no quality control on Amazon's end. Amazon has the resources to fight fake reviews, yet fake reviews are still rampant. As a result, I cannot trust Amazon's rating system at all.
While this is true, at least the part about fake reviews being pervasive, the fact is that I don't trust any reviews on a retailer's website anymore, especially those that aren't verified purchases, and the only retail site where I trust verified purchase reviews at all (and not all of those either) is Amazon.
> I get free shipping from Walmart.com
Yes! walmart.ca is great, very prompt and simple to work with them.
> AliExpress, which 90% of electronics on Amazon are sourced from.
More like 90% of obscure commodity electronic bits and bobs. But yes, some of these rinky-dink sellers and unrecognized brands do mess up the experience a bit, but imagine ordering directly from their individual weird off the shelf fulfillment sites and hoping your credit cards don't get burned. It's not any better of an experience, and at least with Amazon you generally have recourse outside the seller.
> I don't trust Amazon not to send me fake items or cheap Chinese knockoffs
I don't trust AliExpress not to send me fake items or cheap knockoffs either, but that doesn't stop me from trying sometimes. At least with Amazon you have recourse in my experience.
I feel like this is a dramatic market development. In the era of brick-and-mortar shops, a lot of those products were only available as branded versions. Think when you had to go in to a store and get a printer cable in 2001, and you were paying $25 for a Belkin cable in a fancy box because it's all they sell.
Branding from a consumer perspective is about building a message of "we're the quality one you know and trust" or "we're the cool/desirable one you recognize." It isn't necessary for a lot of these products: they're not sexy and the bar of competence is so low that you're unlikely to be burned.
AliExpress and friends have completely reinvented that market by tapping into that sentiment. Their products are unbranded (or as-good-as-unbranded, with a name with no meaning to the buyer) but trade on cheap and good enough. I'm sure a lot of the bits-and-bobs industry is up all night sweating.
Unfortunately as a non-Prime customer, Amazon's shipping has been getting so slow that the gap is closing and it's getting harder to see their value.
Only on Amazon do you find listings for headphones with 400 positive reviews for some BluRay version of Harry Potter, where those 400 positive reviews are displayed as if they're for the headphones themselves.
> I don't trust AliExpress not to send me fake items or cheap knockoffs either, but that doesn't stop me from trying sometimes. At least with Amazon you have recourse in my experience.
AliExpress has never given me a reason to distrust them, while Amazon has. I once received a defective item from a seller on AliExpress and I was refunded my money without having to ship the item back. I'm okay with that recourse.
The 2003 documentary "The Corporation" covers this subject in detail: "Documentary that looks at the concept of the corporation throughout recent history up to its present-day dominance." [1]
The general consensus around these parts seems to be that Amazon is a dog of a company.
In fact almost all large businesses use these types of data to maximize profits, I would assume. Is there a functional difference between changing a digital page to optimize return and changing a store shelf configuration at Wal Mart to do the same? Because of course Wal Mart uses analytics to determine their store layouts and deal placements etc.
The fact that the US only cares about how action like this effects the price an end-consumer pays in a good
The purpose of the team is to calculate the proper price to sell items for, the "discount" to vendors, and the items to put in the "buy box" in order to make the most profit for Amazon. I think the what is described here in the article is only the tip of the iceberg. Amazon isn't doing something as simplistic as cutting out the competition with low prices... they are feeding all the numbers into neural networks and analyzing and predicting competition and profits, and then doing what the systems tell them is the most profitable.
That is, anything that Amazon can predict will make them more money or gain more market share... they are doing. I have full confidence that they are doing all of the terrible things outlined and more. The bigger question is whether government and law allow them to continue doing this. I think that it will continue to be allowed, despite being super shady and destroying both small and medium size businesses.
It cannot easily be proven that there is anyone orchestrating the monopolistic processes, because Amazon has essentially turned the process of dominating the market into an equation and they just do what the equation tells them to.
Amazon will claim "we are just being smarter; we mean no harm", and without providing open minutes of their internal discussions it will be impossible to prove wrongdoing.
What is needed is open accountability for internal plans, but that is against the "American way" so I doubt it will happen anytime soon.