Probably several signals including clustering via IP address.
A couple of years ago I discovered I'd been banned from AirBNB. They didn't notify me this happened and it was months after I'd last used the service (which left a good review), so I only discovered once I tried to log in much later. I know it wasn't related to the prior stay because they'd been emailing me to thank me for being a part of the "AirBNB community" for months afterwards.
As far as I could tell the only possible reason was that my girlfriend had a dumb moment and booked a place for a friend, and that friend then invited other people round in the apartment. It wasn't trashed or anything but the neighbours complained about the noise (they were playing loud music or something). So by the transitive rule of AirBNB: random person I never met makes too much noise once -> girlfriend banned -> me banned.
We weren't married and the accounts weren't obviously linkable in any way, except that we use the same internet connection. So presumably they're clustering based on that.
AirBNB claims in the article that it "employs an appeals process for people who feel they have been unfairly banned". This isn't true. The only way to appeal was to click a button, and because they don't tell you why you were banned and back then I hadn't figured out they were clustering users, all I could write was something like "I can't log in and don't understand why". Then you get a stock email back a few hours later saying the ban was upheld. That's not an appeals process worthy of the name.
I understand the chain of reasoning that leads to this sort of thing, but ultimately hotels don't pull this kind of stunt and have lots of other advantages so AirBNB is dead to me even if one day the ban is reversed. If the model AirBNB is using requires this sort of hyper-aggressive banning of users who have not actually done anything wrong then it's a broken and unsustainable model. Competitors like booking.com clearly don't need anything like this either, so that's where my business goes these days.
For convenience, Airbnb lets you send/share the itinerary with other guests via their platform. That's not literally every customer's relationship status, but it would be trivial to use that to ban known associates of bad actors.
Probably several signals including clustering via IP address.
A couple of years ago I discovered I'd been banned from AirBNB. They didn't notify me this happened and it was months after I'd last used the service (which left a good review), so I only discovered once I tried to log in much later. I know it wasn't related to the prior stay because they'd been emailing me to thank me for being a part of the "AirBNB community" for months afterwards.
As far as I could tell the only possible reason was that my girlfriend had a dumb moment and booked a place for a friend, and that friend then invited other people round in the apartment. It wasn't trashed or anything but the neighbours complained about the noise (they were playing loud music or something). So by the transitive rule of AirBNB: random person I never met makes too much noise once -> girlfriend banned -> me banned.
We weren't married and the accounts weren't obviously linkable in any way, except that we use the same internet connection. So presumably they're clustering based on that.
AirBNB claims in the article that it "employs an appeals process for people who feel they have been unfairly banned". This isn't true. The only way to appeal was to click a button, and because they don't tell you why you were banned and back then I hadn't figured out they were clustering users, all I could write was something like "I can't log in and don't understand why". Then you get a stock email back a few hours later saying the ban was upheld. That's not an appeals process worthy of the name.
I understand the chain of reasoning that leads to this sort of thing, but ultimately hotels don't pull this kind of stunt and have lots of other advantages so AirBNB is dead to me even if one day the ban is reversed. If the model AirBNB is using requires this sort of hyper-aggressive banning of users who have not actually done anything wrong then it's a broken and unsustainable model. Competitors like booking.com clearly don't need anything like this either, so that's where my business goes these days.
For convenience, Airbnb lets you send/share the itinerary with other guests via their platform. That's not literally every customer's relationship status, but it would be trivial to use that to ban known associates of bad actors.