There's no way he could make that guarantee. How can you prove a negative? The best Amazon can do is have policies, logs and access controls and punish employees if they break those policies. And Bezos committed to doing that.
This is exactly what Wells Fargo to get away with opening fraudulent accounts in their customers names.
As with most conspiracies the facts don’t add up super well in this case for Amazon - their whole company is based on trust. The cost of losing it would be huge, so they’d be super driven to stamp it out.
For example, look at The Verge Tech Survey 2020. 91% of those surveyed had a favorable opinion of the company, and 70% felt they had a positive impact on society. Most of the other polls have similar results.
I don’t trust Amazon to ship genuine to specification SD cards or original Apple power cables due to the commingling stock issue.
The original question here was not “does amazon provide a bad service to consumers”, but rather “does amazon abuse their data to compete unfairly”. Only companies would know that, end consumers won’t know why the amazon basics product exits, only if it arrives on time.
Realpolitik of sorts often matters far more than any trust at the company level - does going on your own or letting someone else take a cut pay off better? Would they get massively punished if they failed to carry through and burned your company?
Also, I don't think as a percentage of purchases that they are as common as people think - the stories get lots of attention, but the regular shopping is ignored.
Love the completely unfounded assertion. Based on what? "Top voted comments on HN say they've lost trust in Amazon, so everyone in the world must be the same"?
Let's look at actual data shall we? Turns out surveys indicate that "Amazon is the second most trusted institution in the United States, behind the military" (https://www.vox.com/the-goods/2018/10/25/18022956/amazon-tru...). Many other surveys indicate something similar.
If you have data that indicates otherwise, please do share.
That said, I would really like it if it could be a choice given to the buyer as well. Or at least make it clear when you're buying and getting something comingled.
Firing in this way when someone’s done nothing wrong would be constructive dismissal and prohibited by law.
The fact that Amazon treats their resigning employees in this way doesn’t increase my trust in them.
It actually just makes them look viscous and petty.
Who would want to be placed in the position of saying yes I was fired because I was suspected of being a security risk?
And if you're already on a back channel, then you don't care about playing fair.
But what's the rational the other way around? When an employee resigns, they obviously know that they are resigning before actually handling in their resignation. If they wanted to leave with something, surely they would have already stolen it? Or it's just an expectation of lazyness, in case nothing has been stolen yet?
If the employee should later indeed be found to have taken IP or other assets during their last weeks at work, it would reflect poorly on whomever decided they should work during the resignation period.
Hence, you let them go immediately, so you can shrug and say 'well, we did what we could' in case of problems down the road.
Feels a bit silly to me.
If you go to a competitor and aren't open about it? You might find yourself terminated with no additional pay.
And it makes sense. At Amazon, most people have varying levels of customer data and/or confidential company data. Amazon has to protect itself from data exfiltration/data theft.
This might catch extraordinarily stupid people from conducting industrial espionage, but really I see it as the company encouraging people to not give any notice at all (which, remember, is just a courtesy) when they quit. It’s shortsighted on the company’s part.
This was the exact argument put forth by defenders of Goldman Sachs when it was alleged they were screwing over their own customers in the financial crisis.
If a person keeps stealing from other people, do we just slap them with a fine each time they get caught and let them continue operating in society? No! They go to jail! We make sure they can't continue the crime! Because otherwise, the crime can become highly profitable; allowing any entity to profit from crime is simply incompatible with being a participant in a civilized society - If you let one entity get away with it repeatedly, soon enough, it becomes a viable business model and everyone will be doing it and society will implode.
If he and his shareholders are unwilling or unable to personally take responsibility for every single transgression committed by Amazon, then this company is unfit to exist within our society.
I think that corporate personhood is the biggest problem of our time and is a misnomer because not only do corporations have all the rights that people have, they actually have more rights! I don't have the right to influence politicians in Washington! I don't have the right to consistently get away with crimes with only a fine! I don't have the right to evade taxes by claiming that I live in a different country than where I actually live!
That's a rather different category of response to "Well I just can't guarantee this doesn't happen."
This is significantly different than a good faith organization having some members engage in criminal behavior.
It's amazing to me that a fraud perpetrated upon Wells Fargo by its own employees is spun as an act of malice by WF itself.
On one hand, I can see an employee who maybe has performance metrics that incentivize this behavior. Without sufficient oversight, any policy against using this data would become pointless.
I can also see that Bezos wouldn’t want to make a strong claim guaranteeing it doesn’t happen. He would perjure himself if it’s occurred even once.
I’m more interested to learn if it’s systemic or if it’s just outliers. One is a company incompetently benefiting from not enforcing its own policies whereas another is more egregious.
Amazon admitted to using aggregated seller data for competitive purposes. That's not illegal, but it should be, and congress should fix the gaps in the laws. Since Bezos doesn't want this, he dragged the conversation into the weeds (the aggregated/non-aggregated distinction and the proving-a-negative discussion).
That's fine. Everyone acted in their self-interest, as expected. What confuses me is how ridiculously effective the diversions were. This thread is full of people chasing the chaff. People who ought to know better.
It's a game where the politicians are in charge, subpoenaing the businessmen to come answer their questions, under their rules, at penalty of perjury.
Of course the businessmen do their best to present their side, but let's not pretend it's an even playing field.
>Of course the businessmen do their best to present their side, but let's not pretend it's an even playing field.
Your response is odd. The fact perjury is a possible outcome if a knowingly misleading answer is given doesn't change the equation of the fundamental social interaction.
Congresspeople are not fishing for soundbites. They have a legitimate interest in knowing what are the current practices being engaged in by major economic players. They need information pertinent to building consensus amongst their numbers to fulfill legislative mandate.
The capacity to give a concise number, or the propensity to avoid answering the question speaks volumes more to the necessity of regulation than any candid response would.
I mean, if you watch the questioning, it's very clear that they are.
Look at e.g. Alexandria Ocasio-Cortez' questioning of Mark Zuckerberg.
As I recall AOC had particular difficulty getting a straight answer out of him back then too, as did several other Congresspeople, including a few who were far more reasonable, accommodating, and less incentivized to go down the route of theatrics than AOC, and asked fairly standard, straightforward questions. Those questions were met with answers which were evasive and vacuous on average; downright equivocations and deceptions at worst.
As I only spend 1 minute a day on political news, there's a lot of scope for politicians to do non-soundbite things and me not to hear about it.
Yes. This has been discussed before. (See e.g. the 3 comments here: https://news.ycombinator.com/item?id=16807245 )
In particular:
> By the time you hear a pointed question in the committee hearing room or a speech in a floor "debate" from a Senator, they've been over it with staff, decided on a position, and calculated the optimal move to present to the cameras. That you enjoy or agree with the thrust of this particular show has no bearing on the theatricality of committee hearings as a tactic. They wouldn't do it if no one enjoyed it.
> It's not that politicians don't debate things in good faith. It's that they do it with and through staff (who are more like the people you find in high school debate than televised election debate), behind closed doors, and not on the floor.
The official questioning time is entirely staged. If a Congressman wants to know something, there are many ways for them to learn, but none of those ways involve televised oral questioning under tight time limits. That just isn't the purpose the formal questioning serves.
I don't completely disagree; political subcommittees do tend to be more practical than politics at large. But the committee a hostile environment for the people bought before it; if the politicians just wanted information they'd do it in writing or chat to Zuckerberg at a party or something. The committee process is a legitimate one for that sort of thing, but there are much easier ways to to figure out what is going on. If they actually care about matters of fact politicians tend ask domain experts to write a report - that is a much more comprehensive and nuanced approach.
Literally everything in a committee is about building a narrative to underpin political action. And the people appearing in front of the committee will generally not be the people who were advocating for that action.
Generally speaking, committee's don't happen if there is no reason to believe that any of the alternate means you brought up as examples would be sufficient to resolve the problem. In fact, many of those may have already been addressed or attempted to unsatisfactory result. The legislature has organized and tasked the committee with this investigation as an expression of the "Will of the People" to get to the bottom of business of redressing the grievances of the Constituency, and building policy on the basis of the facts collected. In Facebook's case, largely prompted by not adhering to a previously unnecessary to enumerate social norm, which many people have begun to become concerned with Facebook's primary business model being reliant, and the consequent leveraging of advantaged position via network effect exploitation to prevent competition.
As a corporate citizen of the United States, there is no adversity here. You're in this country, granted legitimacy by the authority of our institutions. If the political apparatus taps you on the shoulder and deems it necessary to look into the facts of what you're doing, nothing at all should change about what you're doing or the answers you give. That behavior of manipulating the response you give in response to interrogation by Congress is analogous to anti-debugger code; obfuscation being an act of perjury. This is fact finding and policy building by duly elected representatives. There should be no adversity. No legitimacy or illusion of legitimate competition acknowledged.
This is humanscale debugging by an authority mechanism implemented by a distributed decision making and consensus engine in the form of a Congress of elected representatives. I don't know about anyone else, but I don't accept the right of a Corporation or an Executive to evade or deceive official legislative apparata. The called already has going in their favor that a great deal of the population are highly resistant to meaningless change, and that there needs to be a non-trivial consensus for change.
Now I fully acknowledge we may have gotten away from the ideal of how things should work, and that when you're dealing with human beings, the level of nuance and communications overloading associated with a state bearing comment skyrockets; but that still does not to me justify lying by omission or outright to your legislature or it's representatives, or intentionally wasting their time for your gain or to intentionally frustrate policy making. If you can't be honest to the collective proxy of $the_rest_of_the_nation, you've got your priorities all sorts of misaligned, and if you're feeling that threatened by the experience, maybe, just maybe, it's finally coming back in unequivocal terms that something needs to fundamentally change about your operational model. I also acknowledge that making the determination of what is "frustrating policy making" and what is just best effort is left fuzzy, and in that fuzziness creates a dangerous policy tool to be abused. I make no claims or suggestion to what should or shouldn't be done in the event thereof, or what the determination that obstruction of policy making has occurred looks like. I'm only pointing out that if you accept as legitimate "defensive legislative perception management" you've undermined the status of the legislature from "wielders of the power of the collective social apparatus, and the authority and access to privileged information that comes with it" to "For sale: invest in a full time employee to haunt enough representatives, and you too can have a law."
There is no illegitimacy in any question being asked. Axiomatically, this is basically people with as close to societal root access as you can get trying to build a realistic representation of the socio-governmental system in order to plot a reasonable course to satisfy $rest_of_the_country.
TL;DR: So long as everything is being done by the book; there should be no self-preservation or overt attempts at manipulation being acceptable. Only as accurate of a rendering of facts as possible, and where a subjective question is asked, a candid answer provided.
For it to be any other way reeks to me of dysfunction, and an overt admission of a lack of integrity, and standing in opposition to the very thing that grants you the legitimacy and capability to live in the manner you desire.
From a social and civic point of view, it isn't and shouldn't be a contest. Just a mechanism at work.
The parallel between debugging and political committee hearings is pretty weak. Code tends to be repeatable and fairly easy to build consensus on (eg, a group of developers will have a fairly clear and reasonable band of opinions on whether networking code is succeeding or failing). In politics, even basic questions like what the question is are subject to uncertainty and there is no consensus on even basic points.
> I don't accept the right of a Corporation or an Executive to evade or deceive official legislative apparata
So if the official legislative apparata asked you to "explain, in detail, how your activity has broken the law?" how would you answer that if you don't believe you've broken the law? That sort of question is a little uncommon but entirely plausible in a committee hearing. It is impossible to answer directly; the person in the hot seat has to push back on the premise and answer a different question.
What you are saying there is implicitly supporting the idea that the people in front of the committee should agree with the agenda of the committee. That isn't reasonable.
Rep. Jayapal was pressing on this, and Bezos mentioned policy but never access control.
This is easier said than done. Data exists in multiple places. It transforms in the form of reports and presentations. Policing this is expensive, which is why it almost only happens under regulatory pressure.
I completely agree. Before GDPR, many companies did not knew who was accessing the data, what was the reason to store it and if there was any value to that data. Now, I have seen an increase in understanding of which data the companies have.
I have worked adapting services to the GDPR. The reality is that it is very cheap to design systems with 100% data accountability. The main problem is to adapt old systems full of technical debt where nobody knows if the data is used or not, or even how up to date it is.
Of course hard to see how a Congressperson would know about any of this. Congress should be bringing in domain experts to ask these kinds of questions.
I think in science you cannot "prove a negative" because you cannot observe all the events in the system (the universe).
But if you can observe all events in a system (Amazons data access logs) I think you can "prove a negative" by proving that the data was not accessed.
Breached and now you lied to Congress. You can’t control everything that happens in your company.
Why on earth would we believe him?
It could be a widespread practice he's happy to overlook.
I don't believe anyone in AMZN retail is looking at S3 bucket data, but I do believe they are looking at sales data from the online store, no doubt.
It's not like Walmart or Walgreens isn't doing this.
"We can't guarantee that policy has never been violated."
Perhaps there is a way to make this guarantee: Do not collect the data to begin with.
That would mean no reviews or Q/A section because engagement there is a good proxy for sales. Search would also be completely broken because Amazon wouldn't know what products were good.
Not to mention payment processing. How is Amazon going to pay sellers their money if Amazon has no data on what is bought and sold? Dealing with payments cannot be stateless for a number of technical and legal reasons (audits, chargebacks, accounting, etc).
Amazon collecting this data is completely legitimate. Are you concerned that Best Buy knows how many third party HDMI cables they sell while also selling their own?
If Best Buy, a brick and mortar, were anything like Amazon, they would be under investigation as well. Nonsense comparison.
You don't think stores should exist? How do you justify that?
>If Best Buy, a brick and mortar, were anything like Amazon, they would be under investigation as well.
I don't see how they are so different. Both are stores that sell third and first party items. BB even does significant online business. Care to explain?
There are less obvious differences such as the contract terms agreed between brands and Best Buy versus the ones between sellers and Amazon. Those details often are only discovered through litigation.
To suggest Best Buy was already doing what Amazon is doing is nonsense.
> This is not true of _all_ the brands carried by Best Buy.
Now you're just playing word games. Both Amazon and Best Buy sell some amount of stuff that could be sold direct to consumer, and some amount of stuff that could not. Even if the proportions are different, I'm not seeing how they differ in principle.
How would they acquire it? Amazon is normally their answer to "I'm trying to sell, but don't have the web traffic". Seems like if they could get the traffic without Amazon, they'd leave the platform, like Nike did.
The facts are that Amazon does not sell only payment processing, order fulfillment or customer services.
I might agree with you if this sentence used a generic term instead of "Amazon". However it specifically refers to Amazon.
I think people forget that Amazon did not operate as a middleman when they first appeared in the early 1990's.
Amazon was originally a bookseller, not a middleman for other sellers.
Not only does the sentence above from GP assume the existence of Amazon, it assumes that Amazon must act as a middleman for other sellers.
What if he said “we care about our sellers. To prevent abuse by our Amazon basics team we have isolated independent customer data from our amazon basics team and firewalled the teams with access to the data from the rest of the companies. Independent audits and training are used to minimize the risk of inappropriate access to this sensitive data. “
If Bezos said that wouldn’t you feel better?
There should be rules in place to stop them from effectively defrauding sellers (e.g. terminating accounts / promotions / offers without contract-backed cause, canceling wholesale orders outside of contract).
Even outside of the ethics of it, anyone else that wants to can offer a competing product on Amazon even if Amazon doesn't (and this happens like a dozen times a day).
this is not the problem that people are grappling with, though! the problem is that amazon can see that your product is doing well by looking at "their own" platform data and then competing with you based on your own sales data. I would hope that anyone entering in to a market that Amazon, or any big player, is in would already be aware of their competition.
Amazon can choose at any time to peek at the data and simply flood a market that they know they can dominate on their own marketplace -- does that not seem unfair to people? It seems to me that as long as Amazon is able and willing to profit off of sellers' and services' data on their platforms, they will incentive their employees to take advantage of that in the name of ~Consumer Value~, all else be damned.
Who isn't aware of it though? Every major retailer has their own store brand, and it's been like that since before any of us were born.
At least Amazon clearly marks their brand, unlike some companies create brands like "Open Nature" which is clearly meant to disguise their origin.
Even if the controls were somewhat weak in practice, his inability to commit to that even on paper is damning.
Just because your safeguards will not be perfect doesn’t mean you shouldn’t have any. It is akin to saying as there are many zero days out there so I won’t secure my infrastructure at all .
I've no idea if anything at all will come of it, or whether anything that does will be good or bad, but this hearing wasn't just about getting verbal commitments of being good from the CEOs (despite there being several questions did ask for verbal commitments in various areas).
How is this different than every other brick and mortar store that offers store brand items? This is a practice society has been completely happy with for decades.
Amazon also has 50% of e-commerce marketshare and nearly double Walmart's share of all retail sales.
There are very strong ethical and economic grounds for breaking up these corporations. It merely comes down to the law. It really doesn't make sense that corporations are allowed to keep breaking the law and only get fined whereas if people keep breaking the law, they go to jail.
If people were treated the same as corporations and we abolished prisons, people would form mafias and go on rampages, hunt down billionaires, seize all their assets, then use the assets that they stole to pay the fines... Clearly that doesn't make any sense, it would be total anarchy; and yet this is exactly what we allow corporations to do.
Besides while they are pissy about accessing the data it isn't /against the law/ per se. A mom and pop scale website doing the same thing wouldn't even be cared about as anti-competive. It isn't like price fixing where it is a crime in itself.
Amazon hasn't been doing it (why not?), and isn't doing it (why not?), but Bezos has committed to "look into it very carefully". Great (sarcasm).
It is another matter that law should be strong enough to prevent this in the first place and AMZ or any other retailer shouldn't be needed to put in the contract explicitly.
E.G., not just anyone can look at seller data, have to have a ticket with some IT group to get limited access to a specific seller's data with a specific reason stated for the access etc.
My point being - regardless of access control the people that are allowed to actually have the data could still leak it in the end.
I don't buy it, but I imagine that'd be the excuse.
Given the mountains of anecdata on this subject all pointing to Amazon openly and intentionally mining third party sales data, it seems overwhelmingly more likely that Bezos dodged the question because the answer would make Amazon look bad, not because Amazon is mostly innocent and just can't prove it 100%.
Just as companies earn the monetary rewards of their actions, they should also earn responsibility for their actions.
Corporations are people, my friend - Mitt Romney.
Quite the opposite, the complexity is enormous there's even an entire area of study for it "Organizational Theory".
Have you looked at ELon Musk, Steve Jobs and many others?
Isn't there a saying for this? "the plural of anecdote is not data"?
This reminds me of the meme that facebook/google is listening in on you without you knowing it, the evidence for which is mountains of anecdotes of people claiming that [product] showed up on their ads shortly after them talking about [product], despite them never expressing an interest in [product] to facebook/google. Despite the mountains of evidence, no one was able to provide a shred of solid evidence of the fact (eg. decompiled binary, network capture, runtime analysis).
A self-serving non-answer is even less reliable than a pile of anecdotes. It would be utterly ridiculous to elevate this one to the dignity of "data."
Amazon has a policy, as described in this article, of not allowing individual seller's data to be used in non-aggregated form by internal parties. This is purportedly to make third parties feel less directly competed with, and is a guarantee not provided by my above examples (Trader Joe's, etc.).
There is not a mountain of evidence that they look at unaggregated data, and it is absolutely difficult to prove a negative on. For example, let's say there's a single brand breaking through in a new electronics category. Let's say some enterprising internal brand manager is looking for up and comers to pull into Amazon Basics. Why not just search the site for that type of electronic? Or look at short term best seller lists?
I'm on the fence about what should or shouldn't be legal in this situation. But if Amazon can't do it, it seems clear that Trader Joe's, Costo, etc. shouldn't be able to do it, and that's quite a can of potentially anti-consumer worms.
Amazon has the policy, and the journalists have found sources saying they break them, so at the very least an internal investigation is warranted, or some non-criminal external investigation for transparency's sake. But as far as I understand (I am so ready to be corrected), Amazon does not even need to have such a data aggregation policy, much less have an iron-tight way of policing it. What this really looks like to me is Amazon shifting from being a third party seller platform to a more highly curated internal/external hybrid platform, just as maker supermarkets did in the 90's. Believe me there was a lot of handwringing then about the practice at supermarkets, but now it is commonplace and one can argue that it is ultimately a benefit to the ecosystem.
> that's quite a can of potentially anti-consumer worms.
No kidding.
> an internal investigation is warranted
We agree.
Costco at best knows what I bought at their store, Amazon knows a lot more than that. Just from the data they collect on the store they will know how many people compared what products with what else, what are they searching for, what is in who's shopping list, they will likely know demographics and lot more information about the buyers as well.
It is not remotely the same problem between online and offline.
It's not a binary offline / online business. Retailers are a mix of both, with some, like Walmart, quickly expanding their online presence as well.
How do we make companies follow it is different issue altogether . First step is the make liability exposure high for violating (like COPA ?) and provide consumes right to access what they have on you.
There is no technical way privacy of this or any other kind can be guaranteed, companies internal controls are always going to be opaque and data collection looking from the interfaces we can audit may not visible .
We can only make it higher risk legally and give consumers frameworks to make it easier to expose violations
Aggregate data on what consumers bought from the retailer. This is the retailer's data.
Individual sales data from sellers of the retailer's platform. This should be PII and treated like any other sensitive data.
Aggregate data on what consumers bought from all sellers on the retailer's platform. This is the retailer's data.
Individual data on consumer actions on the retailer's platform prior to purchasing an item. This can / should be PII. This includes search history, shopping cart, listings visited, etc.
Aggregate data on consumer actions on the retailer's platform prior to purchasing an item. This is the retailer's data.
---
My position is it's unreasonable to ask retailers to limit their use of aggregated, anonymized data. And unless you limit that use, it doesn't materially change the current situation. The major difference between Walmart, Costco and Amazon is Amazon has a lot more aggregate data.
A grocer can get away with whitelabeling, because the manufacturer of the product is bought from as a supplier. Therefore every brand/product in the store, regardless of the manufacturer or "brand" is essentially the grocer's product. In that sense what grocer's are doing is more inventory management. None of their suppliers are "harmed" by that in house data processing.
Amazon, on the other hand, is fundamentally harming the members of it's marketplace because they never make a purchase to prop up other members of the marketplace. Instead, they use those marketplace actors in order to target high return corners of the market in which to compete without taking any of the risk; and crucially, without also making aggregate data available to all sellers.
In effect, every Amazon seller is a risk sink. Everyone else suffers the downsides of poor market fit, but if there's a sector with good profitability, you can bet that Amazon is on it as soon as possible with a brand new Amazon Basics offering. Essentially leveraging their scale to extract as much retail volume out of the market for themselves at everyone else's expense.
Dismissing this looking at seller data due to it being aggregated is like marketing one's tobacco as toasted. You're drawing attention to something irrelevant in the hopes that no one keeps going down the path that leads them to realize you're abusing your advantaged position to the detriment of everyone else involved.
But you can put various things in place. E.g. limited access, (external) audits, etc. Such things are entirely standard and expected for any big company.
Saying "cannot guarantee" is too much nitpicking. Such needs (limit the access, etc) happens all the time. Say you want to takeover another (competitor) company. During that process you'll need to figure out a lot of information. However, you're competitors and you aren't allowed to do that. The common solution is to "ring fence" that team. Such "ring fencing" is a normal occurrence.
Similarly, you can also educate what is allowed and what not on a regular basis.
The company I work for has annual repeating courses for various basis things. Meaning, GDPR, competition law, fraud, etc. They're highly annoying and enough people try to ignore the course and go straight to the questions. It'll be difficult to pass if you didn't do it. Not doing those course will get you an email
After doing all that you'll still have enough cases where people were found to have committed fraud, etc. But that's entirely different than 'cannot guarantee'.
I know you went into things I said above. However, immediately saying "cannot guarantee" is too much of a technical answer. Yeah, there might be cases. But your whole response is a lot of proactive things. You also need to check after the fact, plus do more than just a policy. There is more than enough possible. Further, the question that was asked is entirely normal.
Answering “no” while under oath would be a good start.
I don't see how you can guarantee this without also scrutinizing and monitoring employees to a level that would feel incredibly invasive.
Other CEOs gave truthful answers, which is ... out of 10s of thousands of data points, I can't give you a blanket yes or no without commiting perjury.
don't make third party sales data available to Amazon at all? Can't exploit data you don't keep. The same way Signal prevents data problems: https://twitter.com/signalapp/status/1280166087577997312?s=2...
The method is so good, it would even save Amazon storage space!
Customer disputes? Returns? Presumably you'll need some sort of history. At the very least saying Person X bought Y product
Alternatively a proper secrecy management system that allows customers/vendors to expose particular data to Amazon for those purposes without handing everything and the kitchen sink over, with fine-grained controls.
As a consumer, would you really trust a merchant that opted out of Amazon’s return policy? It’s already to the point where I filter on Amazon Prime eligible and if possible sold and shipped by Amazon.
On some purchases I would, I mean this is basically like an ebay purchase with no return policy.
Point being, is there a technical argument against giving the vendor/consumer control over this, as long as the nature of the sale is transparent? I don't think so.
And even if you're saying that's too free-wheeling there's point two. Keeping only a track record of the data for the return period and only making it accessible if a customer demands a refund is technically possible.
The anti-competitive /privacy breaching data-processing that is the topic of the thread can be completely avoided.
> On some purchases I would, I mean this is basically like an ebay purchase with no return policy.
Not at all. eBay’s money back guarantee overrides any seller policies. You can definitely return eBay items that are sold as “no returns”, if they are defective, counterfeit, etc.
Why do you need aggregated sales data for these use cases? Presumably bad internal actors aren't going through order history records one at a time and pasting into Excel! This problem could only exist if you could query large amounts of sales records. Cynically I suspect Amazon has tooling specifically to identify lucrative products.
Even if you didn't keep purchase history for customers, Amazon would need to keep purchasing records for accounting purposes. In a multinational corporation you simply can't have money going into a black hole with no audit trail.
For accounting purposes you can still have line items with less specific category labels.
To be honest you probably don't even need to do much other than gate the data between teams (CSR vs product) and use permissions that prevent bulk aggregation across records for anyone not on a specific non-product team. Lots of businesses have internal policies like this (banks for example).
The reason they don't isn't due to the implementation challenges but because they don't have to (yet).
Should Bezos / Amazon be fined or penalized? Absolutely not. His statements just prove what should already be obvious. Businesses won't police themselves.
This applies to all marketplace platforms where an incentive exists to abuse your access to third party sales data (like app stores).