3 ms·

> My personal experience with Amazon is that it's the shop to go to for stuff that is ok if it is broken. If I buy a cable for £10, and there's a 10% chance it'

by gwd 6y ago

> My personal experience with Amazon is that it's the shop to go to for stuff that is ok if it is broken. If I buy a cable for £10, and there's a 10% chance it's faulty, it's not a huge problem for me. I'm rather surprised at that, from someone who claims to be an economic grad. First of all, you must realize that that 10% isn't stable: the "Market for Lemons" effect will guarantee that if nothing is done to stop it, that 10% will become 90%. Secondly, you're not factoring in either your time or the opportunity cost of having to buy your second cable. If you order your cable and it arrives the next day but broken, you spend at least an hour figuring out that it's broken, and then have to order another one, and then wait at least another day for it to come.

lordnacho 6y ago
But a cable is such a simple item to find out if it's broken, it takes 5 seconds, plus I can just buy two and be 99% sure I have one that works. Akerlöf doesn't seem to apply here, there's no hidden information that comes out later. With a car there's all sorts of dimensions to it that might annoy you. Perhaps the idea is that all cables will go to crap eventually, but my detection cost is also fairly low on that. Additionally, and this is probably not a great dynamic, I can go to the website of a box-store, which will carry some of the same brands, presumeably vetted more than the junk shop.
- PicassoCTs 6y ago
  Its even cheaper to just send it to known problem detectors, and outsource the work of flaw detection to the customer.
- gwd 6y ago
  > Akerlöf doesn't seem to apply here, there's no hidden information that comes out later. The hidden information is how much quality control has gone into the manufacturing process. Seller A and seller B are both selling a cable for $10. Seller A bought thier cables for $8 because they spent a lot of time honing their manufacturing process and weeding out bad cables; the chance of getting a bad cable from them is 0.1%. Seller B bought their cables for $2, because they just went for the absolute cheapest option they could. The chance of getting a bad cable from them is basically 75%. You as someone buying it from Amazon can't tell the difference; so sellers like B make loads more money. Unscrupulous people hear about this and many more sellers like B show up. Sellers like A get frustrated and leave.
  - ballenf 6y ago
    Seller B got their hands on the rejected cables from seller A. They were in fact paid to scrap them and instead are selling them after putting a sticker over a label.
- vidarh 6y ago
  > there's no hidden information that comes out later Sure there is: How well a cable handles being bent, and whether or not it catches fire... Cables and especially chargers are some of the products I'm most cautious about buying from sellers I'm not confident about exactly for that reason. (not that origin is sufficient - the one charger I have had that got closest to setting the house on fire was a genuine Apple charger [many many years ago, they had a design flaw with a too-short sleeve on the cable on the early Macbook chargers; it'd bend and short and melt])
- xxs 6y ago
  >But a cable is such a simple item to find out if it's broken, it takes 5 seconds This is so incredible false, cable quality varies greatly: - wires, gauge and material. Copper is expensive and some cheat using aluminum or Al plated copper. Gold plated terminals, tinned copper, etc. - wire material and single/multi strand determines resistance and current rating, along with loses in the cable. - insulation and sleeves can be made from different materials with a lot of properties, incl. max temperature, flame retardant and so on: pvc - cheap, rubber - expensive, nylon/polyamid - expensive, silicon - very expensive. And that just touches the basics...
  - Chris2048 6y ago
    "find out if it's broken" isn't the same thing as cable properties/quality. If the cable fails to carry a signal, or fails to do so without interference, it should be pretty easy to determine this. Failure to do this would make it "faulty".
    - xxs 6y ago
      Lets say the cable catches fire due to high resistance/poor isolation (used at 40-50C or close to oven exhaust, causing thermal runaway) - is it broken? In that regard you can use a lamp cable to connect an oven - it will totally work for 5seconds, it might burn down later, though.
      - Chris2048 6y ago
        SO a QA process that tests for more than 5 seconds will find that.
- bbarnett 6y ago
  But a cable is such a simple item to find out if it's broken 5 seconds?! Poor assumption. You're presuming this is "only the types of cables you buy". Many times I buy cables to do things with devices, for which I have no working cable already. Examples: * My device stops syncing with my cable. Is it the device, or the cable? I buy a replacement cable, just in case. Replacement cable doesn't work. Now I think it's probably the device. I spent time here, checking to see what's changed, maybe an update, maybe a config setting, maybe something else. Or, is the device just broken? Do I send it away for repair? * I want a cable to use with something I've had for years. Lots and lots and lots of examples here, all sorts of devices, like serial cables to debug with routers/switches, test cables, even just an ethernet cable, if you may not have had one before. Now you test, and you can't tell. Is the port on the device bad? The settings you are using to connect? The device firmware? A problem with the port on the computer you're using to connect to the device? Hours and hours wasted. Worse, you advocate buying two. Great. This means that no matter what, I'm returning that spare cable. Now it's not '5 seconds' to determine it is bad, it's 'no matter what, I'm spending 20 minutes returning something I never wanted to begin with". Frankly, I'd rather spend $10 extra on some types/classes of $20 cable, to have it pre-tested even.
jp555 6y ago
"First of all, you must realize that that 10% isn't stable: the "Market for Lemons" effect will guarantee that if nothing is done to stop it, that 10% will become 90%." Are you sure? Is this ergodic? Does the ensemble average predict the time-series average?
fmajid 6y ago
Amazon has a huge counterfeit goods problem, and buying from Amazon itself as opposed to a third-party merchant on the Amazon platform is no guarantee, because of their "commingled stickerless inventory" program. Apple accessories in particular are heavily counterfeited, I'd bought some USB-Ethernet adapters for a meeting room with too much WiFi interference. They seemed to work fine then I got a new Mac and it refused to recognize the adapters, and they appeared with weird Chinese characters in System Profiler. On careful inspection, they turned out to be fakes, but highly convincing ones. Amazon refunded me, but the distinct impression I got was they don't care as long as they get the money from a sale, and refunds for the few customers who do complain are just a cost of doing business. https://blog.majid.info/amazon-counterfeits/ https://blog.majid.info/amazon-counterfeits/

by gwd 6y ago

lordnacho 6y ago

But a cable is such a simple item to find out if it's broken, it takes 5 seconds, plus I can just buy two and be 99% sure I have one that works. Akerlöf doesn't seem to apply here, there's no hidden information that comes out later. With a car there's all sorts of dimensions to it that might annoy you. Perhaps the idea is that all cables will go to crap eventually, but my detection cost is also fairly low on that. Additionally, and this is probably not a great dynamic, I can go to the website of a box-store, which will carry some of the same brands, presumeably vetted more than the junk shop.

PicassoCTs 6y ago
Its even cheaper to just send it to known problem detectors, and outsource the work of flaw detection to the customer.
gwd 6y ago
> Akerlöf doesn't seem to apply here, there's no hidden information that comes out later. The hidden information is how much quality control has gone into the manufacturing process. Seller A and seller B are both selling a cable for $10. Seller A bought thier cables for $8 because they spent a lot of time honing their manufacturing process and weeding out bad cables; the chance of getting a bad cable from them is 0.1%. Seller B bought their cables for $2, because they just went for the absolute cheapest option they could. The chance of getting a bad cable from them is basically 75%. You as someone buying it from Amazon can't tell the difference; so sellers like B make loads more money. Unscrupulous people hear about this and many more sellers like B show up. Sellers like A get frustrated and leave.
- ballenf 6y ago
  Seller B got their hands on the rejected cables from seller A. They were in fact paid to scrap them and instead are selling them after putting a sticker over a label.
vidarh 6y ago
> there's no hidden information that comes out later Sure there is: How well a cable handles being bent, and whether or not it catches fire... Cables and especially chargers are some of the products I'm most cautious about buying from sellers I'm not confident about exactly for that reason. (not that origin is sufficient - the one charger I have had that got closest to setting the house on fire was a genuine Apple charger [many many years ago, they had a design flaw with a too-short sleeve on the cable on the early Macbook chargers; it'd bend and short and melt])
xxs 6y ago
>But a cable is such a simple item to find out if it's broken, it takes 5 seconds This is so incredible false, cable quality varies greatly: - wires, gauge and material. Copper is expensive and some cheat using aluminum or Al plated copper. Gold plated terminals, tinned copper, etc. - wire material and single/multi strand determines resistance and current rating, along with loses in the cable. - insulation and sleeves can be made from different materials with a lot of properties, incl. max temperature, flame retardant and so on: pvc - cheap, rubber - expensive, nylon/polyamid - expensive, silicon - very expensive. And that just touches the basics...
- Chris2048 6y ago
  "find out if it's broken" isn't the same thing as cable properties/quality. If the cable fails to carry a signal, or fails to do so without interference, it should be pretty easy to determine this. Failure to do this would make it "faulty".
  - xxs 6y ago
    Lets say the cable catches fire due to high resistance/poor isolation (used at 40-50C or close to oven exhaust, causing thermal runaway) - is it broken? In that regard you can use a lamp cable to connect an oven - it will totally work for 5seconds, it might burn down later, though.
    - Chris2048 6y ago
      SO a QA process that tests for more than 5 seconds will find that.
bbarnett 6y ago
But a cable is such a simple item to find out if it's broken 5 seconds?! Poor assumption. You're presuming this is "only the types of cables you buy". Many times I buy cables to do things with devices, for which I have no working cable already. Examples: * My device stops syncing with my cable. Is it the device, or the cable? I buy a replacement cable, just in case. Replacement cable doesn't work. Now I think it's probably the device. I spent time here, checking to see what's changed, maybe an update, maybe a config setting, maybe something else. Or, is the device just broken? Do I send it away for repair? * I want a cable to use with something I've had for years. Lots and lots and lots of examples here, all sorts of devices, like serial cables to debug with routers/switches, test cables, even just an ethernet cable, if you may not have had one before. Now you test, and you can't tell. Is the port on the device bad? The settings you are using to connect? The device firmware? A problem with the port on the computer you're using to connect to the device? Hours and hours wasted. Worse, you advocate buying two. Great. This means that no matter what, I'm returning that spare cable. Now it's not '5 seconds' to determine it is bad, it's 'no matter what, I'm spending 20 minutes returning something I never wanted to begin with". Frankly, I'd rather spend $10 extra on some types/classes of $20 cable, to have it pre-tested even.

jp555 6y ago

"First of all, you must realize that that 10% isn't stable: the "Market for Lemons" effect will guarantee that if nothing is done to stop it, that 10% will become 90%." Are you sure? Is this ergodic? Does the ensemble average predict the time-series average?

fmajid 6y ago

Amazon has a huge counterfeit goods problem, and buying from Amazon itself as opposed to a third-party merchant on the Amazon platform is no guarantee, because of their "commingled stickerless inventory" program. Apple accessories in particular are heavily counterfeited, I'd bought some USB-Ethernet adapters for a meeting room with too much WiFi interference. They seemed to work fine then I got a new Mac and it refused to recognize the adapters, and they appeared with weird Chinese characters in System Profiler. On careful inspection, they turned out to be fakes, but highly convincing ones. Amazon refunded me, but the distinct impression I got was they don't care as long as they get the money from a sale, and refunds for the few customers who do complain are just a cost of doing business. https://blog.majid.info/amazon-counterfeits/ https://blog.majid.info/amazon-counterfeits/