3 ms·
Weeding out fake products is not a software or machine learning issue. It's a process issue and much simpler to solve. You require all your merchants to provide
by sytelus 8y ago
Weeding out fake products is not a software or machine learning issue. It's a process issue and much simpler to solve. You require all your merchants to provide their legally traceable identity. Your agreement should require merchants to put payments on hold when fake product complaints are received with arbitration fees. All big brick-and-mortar shops do this. It would be very hard to introduce fake product in Fred Meyers or Costco by any merchant without financial and potentially legal consequences. Amazon just needs electronic version of that process.
- twblalock 8y agoThe trick is to implement such a process without the overhead the process has at a place like Costco. Keep in mind that Amazon operates in many countries. There is also significantly more incentive to fake a legal identity for sellers at Amazon than at brick and mortar stores. Even if it costs a million dollars it might still be worth doing -- you might make a lot more than that by the time you get caught.
- fmajid 8y agoThe overhead is exactly what keeps the bad guys out. Without it, the spam and counterfeit problem will continue. Amazon can't have its cake and eat it too.
- sytelus 8y agoOverhead should be paid by merchants as one time fee for coming on to the platform. Amazon doesn't need to subsume this.