Hacker News Reader

Comment by crooked-v | Hacker News Reader

My experience with targeted ads pretty much comes down to this classic tweet: https://twitter.com/kibblesmith/status/724817086309142529

> Amazon is a $250 billion dollar company that reacts to you buying a vacuum by going THIS GUY LOVES BUYING VACUUMS HERE ARE SOME MORE VACUUMS

For a personal example, some months back I bought a pair of identical first aid kits (one for home, one for car), and Amazon still thinks that endlessly recommending me tourniquets, splints, medical scissors, and tactical trauma kits is somehow going to result in more sales.

View on HN

> THIS GUY LOVES BUYING VACUUMS HERE ARE SOME MORE VACUUMS

Everyone thinks this problem is really annoying, but to me it's a good sign - it indicates the current technical capabilities on using and abusing personal information is still limited, the predicted machine learning apocalypse is not here yet. In other words: despite the large quantity of personal information collected by advertisers, they still cannot exploit it efficiently enough, so far all they can do mostly is showing you the same things over and over - advertisers are still unable to develop a perfectly manipulative algorithm that makes you buy things (edit: or they have it, but it's still not found in wider deployment).

Effectiveness of targeted ads directly translates to effectiveness of mass surveillance. Perhaps we should enjoy the useless shopping recommendations from Amazon for now. When every single shopping recommendations made by the algorithm is the thing you always wanted (coming soon!), you'll know the cyberpunk dystopia is officially here.

This point of view also reconciles the two different opinions on machine learning in the tech community. The first person tells you that it's nothing but overhyped marketing, while it has some specialized uses, but everything is exaggerated as "AI" to get money from VCs, and in many areas, it's effectiveness is just too limited. The second person warns you that collecting personal information at an industrial scale is extremely dangerous, with the recent progress on "AI", they can use your data against you in ways never possible before. These two are not a contradiction: The first view is where we were previously, and the next view is where we are going to.

> it indicates the current technical capabilities on using and abusing personal information is still limited

Does it really though? I would expect a successfully manipulative algorithm to be much more subtle than this; just because we don't notice it, doesn't mean it's not there.

I don't deny its existence (in conference papers, in labs, in black projects by government and corporations, or even in a tiny startup company). I'm just saying so far it's not mature enough to be widely deployed on large platforms, like Amazon's banner ads.

Since events like this famous case, where Walmart outed a teenage girl pregnancy, it's well-known they mix in some "weird" or irrelevant things to give a sense of plausible deniability. "oh, pregger testers, haha, how weird they can be <nervous laughter>".

https://www.forbes.com/sites/kashmirhill/2012/02/16/how-targ...

This was Target not Walmart and I say that not to nitpick but rather highlight that arguably Walmart is somewhat an upper tier tech company.

If this is what Target does in 2012, Amazon in 2020 not only inserts plausible deniability to optimize for lack-of-creepiness factor. They're probably running AI bots on forums and social media injecting seamless product placement in commentary you see in your browser.

If we observe that the targeted advertising doesn't seem all that good, it's a bit of a leap to suspect there's actually a great algorithm that's so subtly manipulative we don't notice it. By Occam's razor, the algorithm is probably just showing too many vacuums because it's not that good.

What if they are not aiming for "efficiency" in the same sense as you are describing it. What looks grossly inefficient to us may actually meet the minimum level of "efficiency" these "entrepreneurs" of online advertising need to generate a return and declare their "business" a success.

How effective is spam? From the perspective of the spammer, it's effective enough.

In both cases, these forms of "business" are certainly effective in annoying a large number of people, a tradeoff the operators apparently deem worthy of making.

Possible. It's why I said,

> advertisers are still unable to develop a perfectly manipulative algorithm that makes you buy things (edit: or they have it, but it's still not found in wider deployment).

Perhaps the expensive processor time or R&D just doesn't worth it...

Yet (But it's coming soon?).

> but to me it's a good sign - it indicates the current technical capabilities on using and abusing personal information is still limited

I don't think it is a good sign.

I think it shows that, even if your personal information is quietly being abused, individual businesses will still prioritise spamming you.

So, even if you willingly give up your personal information in exchange for personalised ads, you will never see the claimed benefit to you.

One side of the contract is broken.

"you will never see the claimed benefit to you"

Targetted ads are considerably more effective, even if they are annoying.

When you 'buy' something, generally you get the 'benefit' of the surplus.

The ads very effectiveness is a rough measure of their economic benefit, even to you.

How we 'emotional feel' about ads that seem to track is is not really a measure of that. It's just annoying.

> Effectiveness of targeted ads directly translates to effectiveness of mass surveillance.

Assuming you're correct about the technical points of efficacy, we can still turn this around as a social problem and hypothesize that the willingness to use unproven, annoying targeted ads directly correlates to the willingness to use intrusive mass surveillance.

You bought a vacuum here are more vacuums might directly translate to you attended a protest so you're a permanent malcontent. Worse yet, the "metrics" of social monitoring aren't as black-and-white as future purchases, so badly-used surveillance might falsely think itself useful simply through confirmation biases.

> but to me it's a good sign

Or ... perhaps they do this selectively to quiet the informed few :)

Well, advertisers don't really want to spy on you and control your mind. What they want are stable, well-defined target audiences with well-known statistical properties and control groups.

Basically, applied sociology.

They can't get that directly with the current state of the Internet, so they have to rely on hacks.

> Well, advertisers don't really want to spy on you and control your mind. What they want are stable, well-defined target audiences with well-known statistical properties and control groups.

What we're evolving is predators of human state of mind. For every well-understood target audience there is an optimal revenue extraction method. With more data new sub-audiences appear, allowing experiments to find even more effective advertising for maximizing revenue extraction in those subgroups.

A huge fraction of advertising is nothing more than gaslighting us into believing we have a need or desire for a product or service. The other fraction is exploiting biases in our brains to influence our purchasing decisions toward particular brands or markets.

No. We don't live in an Isaac Asimov novel. Hari Seldon doesn't exists.

Advertisers want reproducible results backed by science and numbers. And the only way to get that today is with old-fashioned statistics - normal distrubutions, laws of large numbers and carefully controlled sample selection processes.

Nowhere here does the sci-fi of "mind control" factor.

Most the ads I get on Facebook are for dumb looking mobile games. Looking at "Why are you seeing this ad" it's basically "Is this person female and aged 25-35"

It's annoying because: The ads are extremely disctracting: I can't help but notice them. They look like they for the same game the last 500 ads were for with a reskin. And I haven't played a mobile game since 2012 and facebook is supposed to be able to track all of this right?

The other ads are for womans clothes or shoes which is reasonably accurate. But you need almost zero tracking to determine that I might be receptive to online clothes store. (I'm not, I buy clothes in persons because sizing is hard)

I also once got ads from Jordan Pertsons thing... WOW HOW WRONG CAN IT GET

> The ads are extremely disctracting: I can't help but notice them

Sadly this sounds like it's perfectly WAI.

> I also once got ads from Jordan Pertsons thing... WOW HOW WRONG CAN IT GET

I mean, what's wrong with that from a marketing perspective? He has a fair few female fans so the targeting can't be that wrong.

What's stopping you from installing an ad-blocker like uBlock Origin? You can't help but notice them (no one really can), and it doesn't sound like they add value to your life.

Sounds like she's on mobile considering the ads

I am. But still get them running Facebook from Firefox with Ublock installed. I also get facebook ads on Firefox and Edge on Desktop. As an aside I switched to the mobile app recently considering the Ad Block on firefox wasn't working.

Be sure to use UBlock Origin, not UBlock.

https://github.com/gorhill/uBlock/

Facebook is a pretty high profile website, so it is unlikely that UBlock Origin won't block all adds there (I don't use Facebook myself, so I can't verify). In a normal desktop browser that is.

Of course. Facebook is a serious cat & mouse with adblocking. I once tried to look at an ad with the devtools inspector and the ad the disappeared. Somehow FB was detecting the inspector and remove the ad so it couldn't be inspected. (Well I'm sure that could be used against them)

https://old.reddit.com/r/uBlockOrigin/comments/fnkwa4/facebo...

There is new-ish one that might work.

EDIT: No, now it's only the text that identifies it as "Sponsored" the rest of the ad does not disappear.

EDIT2: For an example here is how the word "Sponsored" is represented in the DOM:

  <span aria-labelledby="jsc_c_cm" aria-label="Sponsored" class="j1lvzwm4 stjgntxs ni8dbmo4 q9uorilb gpro0wi8"><span
        aria-label="Sponsored" aria-labelledby="jsc_c_cn"></span><span
        class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">S<span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">n</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">t</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">S</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">n</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">p</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">p</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">o</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">n</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">i</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">s</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">i</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">o</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">S</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">r</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">o</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">e</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">l</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">o</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">e</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">c</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">n</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">s</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">c</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">s</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">o</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">r</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">e</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4">d</span><span
            class="b6zbclly myohyog2 l9j0dhe7 aenfhxwr l94mrbxd ihxqhq3m nc684nl6 t5a262vz sdhka5h4"
            style="position: absolute; top: 3em;">d</span></span></span>

In some cases, targeted advertising is still quite dumb. Advertisers themselves choose who to target. In your example, someone may have said “We need more women interested in Jordan Peterson! Target white women 25-35.”

College Humor--may it rest in peace :( :( :(--did a great bit about this a few years ago. https://www.youtube.com/watch?v=KbKdKcGJ4tM It is absolutely ridiculous how useless targeted ads actually are... and seeing a real person try to sell you this stuff really underscores the insanity.

The usual explanation is that someone who just bought a vacuum cleaner to decide they don't like the vacuum cleaner they just bought and need a new one. Probably more likely than it is for someone who hasn't thought about buying a vacuum cleaner for a long time to suddenly have a need for one, even.

If I bought a vacuum I didn’t like from Amazon I’d return it thanks to their generous return policy. Targeting ads based on returns could make some sense I guess if they wanted to roll the dice on recouping losses.

That actually would be very smart targeting - especially given Amazon actually asks you for a reason of return. So they could target people that said this vacuum doesn't work or they need a different model, etc. by offering them similar one but from other manufacturer, etc. But I don't think this happens, and it's require developing much more sophisticated models than dumb "bought together" ones they seem to be using now.

But does that help Amazon?

Buying a vacuum, returning it, and buying a different one is less profit than just keeping the first vacuum, so I don't see why they'd want to encourage this.

I would say that the more parsimonious explanation is that:

a) they've got a model that has learned that people who buy X tend to buy X again, and

b) they don't have a model that can properly distinguish between classes of X for which this is true and classes of X for which it isn't. They are aware of the problem and have made several attempts to fix it over the years but the results have never been sufficiently compelling to incorporate into the actual recommendation model in prod

I think another factor that may come into play here is distinguishing between "products that serve the same purpose" and "related products" e.g. between "vacuum cleaners" and "vacuum cleaner accessories" (filters, bags, additional attachements...), the later likely being purchased in addition to the former. In this case it seems like there's probably a clear line but could certainly get blurry in some cases. If two products have some overlapping features, but also some unique ones, how different do they have to be for people to be likely to buy both?

I always loved when it would say "You bought No. 26 in a comic series, do you want No. 19 and 13?"

Yes, they're commonly bought together, but there's mostly, if not completely, a one-directional relationship-- someone buying 13 might want 14, 15, 16, but not the other way around.

This feels like it's just a matter of adding one more signed value to the "people who buy A buy B" record table... a weight indicating if the take is mostly A->B, B->A, or neutral.

Or, instead of the usual recommendation widget, just show a link to "All volumes in this series". Or even incorporate it into the website structure like YouTube playlists do.

Some years ago, I saw a presentation on Neo4J by a developer for a major British supermarket. (I think it was Sainsbury's.)

They wanted to print an appropriate coupon after someone paid for their shopping. "£1 off hamburgers next week" or similar, to get the person to return to the shop the following week.

They said a very important feature was to avoid giving a discount for a product someone has just bought or a closely related product. This would annoy customers who felt they'd missed out. Therefore, they were updating Neo4J with the just-made purchases and calculating the recommendation in the time it took to process the payment.

Maybe Amazon's chaotic mess of similar products makes this much more difficult than for a large supermarket, or maybe Amazon doesn't think the annoyance is a big deal.

» Maybe Amazon's chaotic mess of similar products makes this much more difficult than for a large supermarket, or maybe Amazon doesn't think the annoyance is a big deal.

I think amazon.com is no longer day one if it ever really was. I've seen at least one instance where they lowered price and refused to price match sale that hasn't even been delivered yet. I think they have data shows that not all people will go through the hassle to return and rebut and/or there's enough lock-in that amazon.com does not care about frustrating users.

Maybe the people who would be frustrated would go through the hassle to return and buy again. I mean if you have prime and you haven't opened the box, it is just two more days.

I always just figured that was because it’s much easier to figure out that I was looking for a vacuum cleaner than it is to know that I’ve already purchased one and don’t need to be targeted any more. Especially if the “looking for a vacuum” signal gets spread further, to more systems/parties, than the “doesn’t need one anymore” signal does.

Maybe it's because the purpose is to make money selling ads, rather than to necessarily sell you the consume what you need?

ie these systems have multiple customers, you, and the ad or marketplace sellers, and the behaviour of serving up ads for things youve already bought is happening to milk the advertisers/marketplace sellers?

You are not the customer of these ads. The customer is the company selling vacuums.

Your profile has "interested in vacuum cleaners" and then the ad system monetizes this fact by selling ads to vacuum cleaner companies. The ad system makes money every time you see one.

Then the vacuum sellers are getting ripped off just as much as I am.

More so, since they are paying their ad dollars- bidding in a real-time auction for the privilege of filling the ad slot on the website you are looking at.

Someone who has just bought a vacuum cleaner is probably one of the people least interested in vacuum cleaners.

Suppose you try it out, find you don't like it, and decide to return it within a week, after which you want to buy another one. How often does that happen? 5%? That might still make you significantly more likely to buy the next vacuum than a random member of the population.

Of course, if you bought a vacuum and returned one, that would be the ultimate targeting point. I don't know if "they just returned a vacuum cleaner" is a datum provided to advertisers.

It also might be that people usually don't return it until they've chosen one that they think would be better (possibly ordering the new one, even waiting until they receive it and trying it out before starting the return process).

That's the prevailing common-sense wisdom, but it's completely wrong.

These algorithms work on statistical foundations, and the probability of you buying a vacuum cleaner if you already bought one is always higher than the probability for a randomly sampled person from the general population.

Not all ad / search systems know you've made a purchase. Google probably does though. Of course they should probably shut up about it.

I bought a laptop this year - naturally the ads changed accordingly.

Out of curiosity I clicked - "Out of Stock". No surprise here, because there were literally less than 10 units in my country sent in one batch, but why advertise it then?

Number of visits to the product page is probably used as a signal for how much to restock.

Vacuums is not the worst of it. I made a mistake of looking for a junk removal company with shields down. Every marketing software on the planet got the idea "oh, this guy is really into junk removal, let's help him!" and all ads in every place where I couldn't disable them were about junk removal for half a year at least. Needless to say the junk I need removed was long gone after the first week. Now if I ever need junk removed again in a year or two... well, actually I already forgot all those ads, so it was just a waste of money for them.

Ahh, well next time you need junk removed, the company that spent the most on advertising to you will seem the most "familiar" and trustworthy! It's the same reason you see Coca Cola ads everywhere despite the fact that nobody needs to be informed about its existence.

> It's the same reason you see Coca Cola ads everywhere

This is one thing I don't understand, even though I know it must obviously work (to be clear, I'm not disputing it works. I just don't understand why/how). I know their purpose is to "remind you" Coca Cola exists, but unlike with junk removal services, you will never "need" Coca Cola and either you like it or you don't. People who like Coca Cola tend to drink it regardless of ads (some are addicts), and don't need the reminder. People who dislike Coca Cola won't ever drink it and the "need" will never exist in the future, unlike with junk removal.

So again, what's the rationale for these ads? Who needs a reminder that Coca Cola exists? I bet they are data-driven and their studies show that people do consume more soft drinks after seeing these ads everywhere -- I'm not disputing this -- but I really don't understand why it works.

Maybe the target is young kids exposed to these ads for the first time? But in general, the people in Coca Cola ads look like teenagers or in the 20-30 age segment, not kids.

It's not to remind you Coca-Cola exists. It's to make an impression that everybody drinks it and not doing it makes you weird. OTOH, drinking it makes you in sync with what cool people do. "Liking" it is largely a formed habit - I used to drink a lot of Cola, until I realized what it's doing to my body. Now I don't touch the stuff, hasn't been drinking it for years, and I find it absolutely disgusting if I taste it (I did a few of times over the years just to see if I'd find it tasty or not after not drinking it for so long). So it's not like there's some "innate" liking, it's a formed habit, and the ads help you to maintain the habit (or make it hard to quit it).

Oh, I fully agree it's a formed habit. Unfortunately I'm already addicted to it. However, the ads do nothing to me: I'm not a 20 or 30 something anymore, and I find ads of "cool people doing cool things" both fake and embarrassing.

(I know Coca Cola is terrible for your health and I keep telling myself I'll just stop drinking it. Maybe some day I will...)

I myself just decided not to drink it for a month. I had to think it through because it turned out I has a variety of habits that led me to having another glass of Cola, and I had to consciously modify them and keep the goal in mind for the whole time. It is a bit tiring, but you can do it for a month if you really concentrate on it. And then I discovered by the end of the month the craving is practically gone. My habits changed. And then a bit later I discovered I actually don't even like the stuff, so it takes me no effort to keep away from it. I was additionally motivated by rather scary blood test I've got which hinted my health could be in danger very soon if I don't stop, so keeping it in mind was easier.

Probably not - I literally don't remember a single one.

Also, I think Coca-Cola is completely disgusting and you'd have to pay me quite a lot of money for me to agree to ingest the stuff, so maybe I'm not a very typical consumer. The fact that I know it's literally a poison to my body, in most plain sense (and probably yours too) doesn't help either :)

I bought a glowforge about 1.5 years ago for use in a side business. Youtube still shows me ads for glowforge about 2-3 times a week. I kind of feel bad that GF is wasting their ad spend on this type of crap.

This kind of stuff feels extra stupid given it's 1-to-1 personalised targeting.

I blame Bret Victor. He wrote an influential paper that was longish and people read up to the "last value predictor" section, had an a-ha moment, went to implement it, and ignored the rest. He hasn't update the paper since so no one knows what else to do. The marginal improvement was to take the last value, relate it to a bunch of other people whose last value was the same, and show me their next values.

It's a little bit different, but theres actually some reason behind this.

Buyers Stockholm Syndrome or post-purchase rationalization.

https://fifthperson.com/avoid-stockholm-syndrome-buy/

https://cxl.com/blog/post-purchase-rationalization/

After reading both of those articles... what's the reason behind it? The first article gives us nothing but definitions (it's oriented toward the consumer, so it doesn't spend any time prescribing vendor behavior). The second article has a section titled "How to Help Your Customers Rationalize Post-Purchase", which lists a bunch of things you can do to help your customers avoid buyer's remorse, but I don't see how any of their recommendations support the behavior being discussed.

I wonder (and really have no proof of this one way or another) whether this actually is successful a good bit of the time. HN readers are generally pretty thoughtful with their purchases, but maybe a lot of Amazon customers really will buy two vacuums? Maybe they see there’s a nicer one, and just can’t help themselves?

From these examples we should conclude it works sometimes for emotional or artistic items subject to impulse buys, because Amazon sells billions and knows their business, but maybe not for utility items.

The problem is, it's hard to distinguish for an algorithm. Vacuums no, but tools or kitchen gadgets, maybe

Exactly. Moreover, the ad about the second vacuum cleaner only has to be more effective than whatever random ad distribution was used previously, which would be a rather low bar.

We all have the ad that follows us. For years, mine is the forever spin top. I thought about having a nice top as a desk toy after seeing Inception (yes, that long ago). I checked out the tops, saw they came with a $70 price tag and noped out. I see these ads weekly if not daily.

Not all of us mate, I don't actually recall any ads because I block them extensively .. through software online, through habits of avoiding tv etc, changing channels or muting radio ads, and then mentally tuning out anything that gets past that .. I love the peace of mind.

I'm an avid consumer but I believe my interests are best served by a pull model where I research the things I want to get best in class. Then anything that spends an excessive amount of effort on advertising is wasting their budget which could go toward a better product and lower prices! So that helps me avoid advertising driven products, as well as advertising. #winning :D

They are right though.

Those Vacuum and First Aid ads are probably 10x more effective than random ads.

Randomly showing ads for people is actually not economical in most ways.

Showing ads for stuff that people have randomly bought, probably is.

It 'feels' annoying because totally random ads don't jump out at you, while the targeted ads do - you notice them.

Ironically, it's the very fact that you notice them gives hints as to why they are powerful.

I think the point is that there is a middle ground between showing completely random ads and showing what someone already bought.

If I bought a vaccum maybe I want vacuum accessories, or vacuum bags (if it's not a reusable), or maybe I want more appliances for the house, and so on.

Yes, but the 'accessories' is actually not a 'middle ground' - that's actually very advanced thing to do. We may get there.

The 'middle ground' is literally 'bought vacuum -> show vacuum' - and it works.

People are emotional about this stuff and aren't thinking about it in terms of behaviour, response etc..

Some things are negative, toxic, this one is just innocuous: you only notice the vacuum ad because you bought one. Otherwise, it would be a completely random ad for something you truly don't care about and you wouldn't notice it.

It's actually technically more efficient to show you stuff 'you might want to buy' than stuff you 'don't even care about a tiny bit'.

Edit: I should add - you see ads for generic products all day long - in fact probably the least value-creating ads are those for really well established brands. Crest toothpaste, laundry detergent etc it has not improved materially in decades (small things maybe) and yet they spend billions on packaging, positioning, huge ads etc.. And yet for some reason it doesn't quite bother us.

> Amazon still thinks that endlessly recommending me tourniquets, splints, medical scissors, and tactical trauma kits is somehow going to result in more sales.

And that tactic does increase sales for Amazon, on average. Their endless split testing would have long proven that.

This doesn't sound like an advertising problem, it sounds like a data problem.

At some point Amazon ran tests to see whether this stuff worked and realised that it did - they'd probably see the same results sending emails and ads that are just a big Amazon logo.

And yet, since that tweet was posted Amazon has become 6x larger, crossing $1.5T in market cap. So maybe their advertising team knows something we don't?

Yeah, they know exactly how stupid the average company’s marketing team is that they’re willing to pay for this kind of stuff.

Maybe the Dutch website is different but I think it is one of the worst webshops I've been on. When you search for a product you end up in a flood of pages with half related products. No filters, nothing. So I never find what I am looking for and just leave.

Now I don't think Amazon is stupid so there must be a reason for this experience.

I also think they know something we don't.

When I was 16 about, I read a biography of Howard Hughes. The only thing that really stuck with me was a line regarding the Hughes empire in the 70s when he was at his craziest but the money still kept pouring in, the idea was that basically once you got to a certain size there is a long time before your mismanagement can hope to decrease the money flow.

Also, how much of that 1.5T is because of AWS?

About 63% https://zdnet4.cbsistatic.com/hub/i/2020/10/29/c612e8ee-3387...

It's probably because people just shop from them, not because I bought the gay pride shirt, hat, wallet, car seat cover, pencils, earrings, and bracelet that Amazon thought I wanted. I still cannot figure out why Amazon almost exclusively goes on these advertising tangents of junk I have a rather explicit disinterest in. If targeted advertising showed me things I would find interesting I would be all for it to some extent, but when it shows me automotive stuff for vehicles I don't own after I buy a digital micrometer I can't begin to think it's a good system at all. (Examples mentioned are real experiences of Amazon's suggestions of things I may like)

As a reader, I must say that at least in the books section the recommendations work. If I buy a book I like, I look for automated recommendations, because the chance I might like those books is rather high.

I guess with some categories it may work, like books, movies, video games etc

Yep, if I liked one book by G. R. R. Martin, it makes sense to suggest I would like another one. If I bought one Samsung tablet, it's much less likely I'd need another one a week later, unless that one was defective. They should introduce more information about goods - but it's probably too expensive to model that, and they get paid handsomely anyway, so why bother.

Even if that one was defective, you're likely to exchange it (especially if you're buying from Amazon). Not buy a whole new one.

How do you know it's related to their advertising team? I buy a lot of stuff on Amazon (mainly because their delivery is - usually = reliable, their selection is good and their return policy is generous) but I can't remember the last time I bought anything by clicking on an ad. That didn't happen for many years at least. In fact, I'm not sure I ever bought anything on Amazon by clicking on an ad, if you exclude affiliate links and people just reviewing stuff and then giving Amazon link.

If vacuum sellers are willing to bid top dollar to advertise to people that just ordered a vac, who is Amazon to say they are wrong? Maybe they are betting on a dissatisfied return by people that clearly were in the market for a vacuum. Maybe the transaction isn't data that is available to the bidders.

Advertizing works. even on those thar believe themselves immune.

Far more smart systems already exist and are already used to sway people on a large scale.

The smartest system might not be the most profitsble for the intermediate add platform owner, so in those cases you might be on the receiving end of 'stupid' adds because it means more profit for the platform owner.

I'm geniunly curious why this got down-voted. Would you mind posting why?

Reply on news.ycombinator.com