Hacker News new | past | comments | ask | show | jobs | submit login

I don't know about anyone else, but Amazon.com's recommendations for me are laughably bad. "You bought a 48-pack of AA batteries? Your recommendations for the next few weeks will only be batteries, since you are clearly a battery collector!"

I don't remember ever looking at their recommendations and finding them useful. Why would anyone want to use the same algorithms?




Came here to say this. I've been an Amazon customer for a very long time (2003). I don't think I have ever bought anything from one of their recommendations, that are, as you say, only based on your most recent purchases and that assume, for some reason, that you need another version of what you just bought.

If I just bought a big pack of batteries I won't be needing batteries for a long time; and they should know this because they know that I buy batteries every 18 months or so.

Also, I don't have a dog, I have never bought anything pet related on Amazon (or any other website for that matter) and yet they keep suggesting pet food.

Amazon recommendation engine is abysmally bad. Picking a random item from my orders history would be just as good and so much simpler to implement.


Off topic, but if you need to buy batteries each 18 months, I hihgly recommend you rechargeables.


When making this switch its best to commit. Spend $40 on a good recharger then buy 64 rechargeable AA batteries. I did the same about eight years ago and couldn’t believe how quickly my cache of batteries got eaten up by random devices and accessories and toys. I actually had to buy more, BUT! the change was complete and I didn’t need to buy alkalines anymore (till my kiddo came down with a case of type 1 diabetes and we needed them for her insulin pump). Everything else gets eneloops and the ones I bought back then still seem to be going strong.


> You bought a 48-pack of AA batteries? Your recommendations for the next few weeks will only be batteries, since you are clearly a battery collector!"

Patrick explained it why it might be like so: https://mobile.twitter.com/patio11/status/982208307057246209 (thread: https://threadreaderapp.com/thread/982208307057246209.html )

> I don't remember ever looking at their recommendations and finding them useful.

Amazon does various forms of recommendations (no way limited to just recent purchases) based on

1. Purchases

2. Searches

3. Browsing history (via trackers, ads, affiliate links)

4. Activities (on their digital devices, tagging things on their websites in wishlist, adding preferences in your account etc)

5. Global and Local trends (people in your area/country buy..., people with similar buying pattern buy..., It's Eid and you're Muslim, so you buy...)

...And probably many other signals I might be missing.

They must have got pretty good people working with this and enough data to get their models right.

Also, Amazon makes decisions on new retail businesses to start depending on those signals. https://techcrunch.com/2016/11/03/amazons-private-label-bran...

That, and considering the fact that Amazon has an unshakeable culture of making data-driven decisions for everything it does...

I guess what I'm trying to say is, Amazon and its subsidiaries may not be as dumb at AI/ML as it might seem to be by gauging against one datapoint.

Also see: https://www.cnbc.com/2014/04/09/big-data-knows-youre-pregnan...


Patrick's explanation is that because of buyer's remorse, and because if a thing is going to break, it often breaks early in its expected lifetime, people are more likely to buy an item (e.g. a refrigerator) soon after buying another one than they are to buy the item in general.

Okay, that makes sense to me. They go from 0.02% likelihood of wanting to buy a fridge to, let's say, 2% likelihood. A marvelous leap!

But even though they are now much more likely to buy another fridge than the average person ... I would think that they would be more likely still to buy other types of products than to buy another refrigerator.

Even if their likelihood of needing to buy more consumable products (e.g. deodorant) is only 5%, that's still double the likelihood of their likelihood to buy a fridge again.

So I would think that there would be some bias toward more frequently purchased items, even if your likelihood of buying a less frequently purchased item does indeed increase after purchase.

I'm sure the math checks out somehow, and that the creators of the recommendations algorithm wouldn't be pursuing strategies that don't work, but I don't think Patrick's explanation fully captures it.


The revenue on the refrigerator to Amazon is 10x or more higher than the deodorant.


That seems to miss the fact that if you buy a fridge on Amazon and decide to return it, you are most likely going to look for another fridge on Amazon regardless of whether Amazon puts fridges in the recommendation section or not, since you already know you need a fridge and that Amazon sells fridges.

My only explanations are that that kind of thing happens when they can't figure out anything good to recommend.


> They must have got pretty good people working with this and enough data to get their models right.

You seem to be taking this an an axiom and running with it. My impression too, is that the recommendations are pretty shoddy.


I worked previously at a recommendation systems company where we also generated this sort of recommendation. It looked incredibly shoddy to us too, but what we found out was:

- yes, people who just bought a lot of batteries are actually more likely to buy more batteries soon. same for people who just bought a fridge.

- these recommendations may cause some sort of "burn out", where people stop looking, responding, or decide to unsubscribe. But this wasn't much worse than other recommendations. including the awesome ones.

- the loss from burnout was greatly outpaced by the wins in actually targeting the right users

Yes, these are bottom-of-the-barrel recommendations. But they work. And people don't care that much even if they disagree. Maths said so.

I think this is a similar story to the complains on non-literal searches in search engines. People think they want one thing, but they want something else.


They should.

Perhaps they are not.

Perhaps the best they can do, is to recommend to all of us the same/similiar product again because thats the only thing they got working: Catching those x percent of people who send the tv back and buy the second recommended one.


I tend to agree regarding the home page content, but I’ve found the “customers also bought” sections quite helpful lately, especially when looking at industrial/technical products.


Unfortunately the customers also bought section is not really "recommendations" as much as a table join.


That is a rather simplistic view of it


Its not my view, that's what it is. No machine learning is necessary to generate that section. Just a lot of intersected purchases from a huge amount of customers, sorted by frequency.


Do you work at Amazon, so you know this is how it works? Or are you projecting how you imagine it works?

Because if I'm an Amazon dev I'm 100% not just showing a 45 year old woman who buys a TV the same "other people bought" as a 25 year old man. I'd definitely run the same ML on each section, because the whole point is to understand the closeness of connections between personal traits and purchases, and it is impossible for a human to outperform in that task.


I used to work at Amazon. Demographics aren't needed when you have a ton of customers for which to take the intersection, and sort by weighted (frequency, time of purchase.) Any non related purchases won't show up nearly as often or close in purchase date as related ones on a large sample of customers. This is just a consequence of having big data, which makes the noise floor insignificant. ML shouldn't be thrown at everything indiscriminately.


This was the collaborative filtering of 15 years ago. It's likely they have made improvements since then.


I also find it a much more helpful section than 'recommendations'.

Something simpler is more effective in this case, in that it gives me much more relevant results. Why is that unfortunate?


I'm really surprised to see this sentiment echo'd so much on here. I find Amazon's suggestions for me to be incredibly good, though mostly with respect to books. Do other people feel their book recommendations are excellent, but not so much for other products? Am I alone in this?


The book recommendations are good.

Everything else sucks though.


I suspect that's because it's fueled more by people than pure algo, as in take into account what other people with this book have in their library etc.


Perhaps you're a dev working for an ecommerce company, and your CEO has just sent an email down the line saying "hey, Amazon has recommendations and they're a successful shop, so we should have recommendations too!" - and you want the simplest way to get the job done and move onto something more useful.

Or perhaps you think they might work well for your particular sales domain. They clearly don't work very well for product categories with lots of different items which fulfil the same purpose (e.g. one thousand kettles, of which you only need one, or one thousand TVs) but they might work great for funny slogan T-shirts or for board games.


  "hey, Amazon has recommendations and they're a successful shop, so we should have recommendations too!" 
Sounds like Groupon, who sends me 10 emails a week pushing Smog Check vouchers... which I need once every 24 months at most.


Several e-commerce and logistics companies avoid using AWS/anything amazon for the simple reason of not giving more dollars to their competitor. I have seen this in my current and last company with this as stated reason.


I have a Kindle. I read mostly hard science fiction and popular science. However, Amazon’s recommendation thing always shows me women’s romance novels, Danish crime and period dramas.


I think this is because people who buy batteries are likely to buy them again soon, at least more likely than any other product. Maybe you bought the wrong type, maybe you need more than you thought?

I read a good post about this using washing machines as an example but I can't find it.


They have all the data. They should be able to do better than this.


My theory regarding this it's the following: I helped for a while in a family related retail business and there was a lot of need for consumable products like pens/papers etc or just barcode scanners/toners and so on...

Recommending for similar products at the end of a purchase raised awareness towards those items for the next inevitable batches. I believe amazon wants its own brands to win or otherwise a strong competition among others producers.

I can't believe it's just math.


My personal favourite was when my partner bought a book on breastfeeding through my account, and for weeks afterwards my recommendations were full of erotic novels.


I think the main problem comes when you already bought them...

I mean, I understand Amazon shows me a washing machine while I'm searching for one and I clicked on several without buying, but not when I already bought one! It's not something I would buy every other week...


I am constantly recommended MAGA Hats, it started out as silly but now it's very annoying.


Amazon indeed seem to be very proud of their "more of the same" algorithm. It occasionally recommends me useful books but they seem to mix in a lot of stuff that they are just looking to sell opportunistically. They also inexplicably include products that I've bought already; from them. Also they seem confused about things like book series. Why would I buy part 8 of 30 books where I have never shown any interest in either the author or indeed parts 1-7. I regularly use them to check reviews for things I'm interested in. I quite often end up buying them elsewhere; or not at all.

Regardless, this could be a useful service for some but I do think you would need to worry about things like GDPR if you are planning to offer this to amazon. I'm assuming they've thought this through and are providing ways for companies to use this without getting in trouble. For the same reason, I'm pretty sure that Amazon is going to be very careful not exposing themselves to legal trouble here as that could become very expensive for them. So, I'm not so concerned with them trying to grab the data for their own purposes.


Other customers who bought 'AA Batteries one third gross' also bought: "Long Dong Silver vibrator" "Jewel Cannabis Delivery Inhaler" ...


Showing you them is indeed more profitable. That's why they show it.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: