4 ms·

I was talking to a friend today who said, fairly succinctly, that Facebook's single biggest problem is that they Totally and Royally Suck(TM) at PR. I wonder w

by nlh 8y ago

I was talking to a friend today who said, fairly succinctly, that Facebook's single biggest problem is that they Totally and Royally Suck(TM) at PR. I wonder what would happen if they were just straightforward about this. "Friends you May Know is really cool! We use your phone's location data and do cross-checks on your address book to make the most intelligent suggestions. We even use some unique features of your camera to figure out if you know people!" They don't have to give away the store, but they could certainly avoid the creepy "try and guess what ridiculously creepy stuff we're doing to suggest friends" game.

dgzl 8y ago
Being upfront would bring more press... Alternatively, how many of your friends are going to read this article?
- beatgammit 8y ago
  Would it? The press likes to jump on surprising news, and if everything is upfront, there's not much to report on. They could even make a way to opt out if people complain (and few would actually opt out), which would save face in the media. I don't use Facebook because I value my privacy, and I'd probably have less success convincing people to leave the platform if they were upfront about the stuff they do. People care a lot less about their privacy than PR firms seem to realize...
drb91 8y ago
Actually enumerating what data might be used for strikes me as incredibly unlikely. They could also come up with a more substantial business model.
drdaeman 8y ago
> Friends you May Know is really cool! I've got over 9000 suggestions and not a single accurate one. Maybe they have the best technology and all the data out there, but I'm skeptical they can really use all the data they supposedly have on me. Same goes for any targeted advertisement, not just Facebook. Some come close (based on simple topic preferences), but in all those years I've yet to see any large number of the actually useful ones. All those data mining and AI-driven targeting seem to utterly fail, suggesting me to buy a second TV after I've just already got a new one. Or subscribe to Grammarly (every second YouTube pre-roll, seriously?!), when I've already bought that.
- smelendez 8y ago
  Interesting! It works pretty well for me. At worst it's usually people I don't know but logically could: friends of friends, people who worked somewhere I worked before or after me, people who went to my high school I never really knew. Occasionally it's full of apparent shady accounts: usually fake looking accounts of attractive women I have no friends in common with, then they'll disappear. I'm guessing scammers upload fake contact lists of random phone numbers to get on people's suggestion lists
  - drdaeman 8y ago
    It's almost always people I've never ever heard of for me. Just some random strangers that I don't see any connections with. Sometimes those are probably friends-of-some-"friends," but I haven't bothered to check every unknown name and face. I've recently moved to a different country, and now I get suggestions of some random people supposedly from here - even though I don't know anyone here yet. Well, except a few coworkers and a landlord, who - if they have Facebook accounts - aren't in the suggestions, even though Facebook knows where I live and work (I don't advertise it, but that's not private either). Paired with the fact that Facebook intentionally spams daily with "you have 20 new notifications", they just train me to filter and ignore their notifications. If someone's going to message me there, I won't see them until the routine monthly check.
- JoeSmithson 8y ago
  I take it a kind of weird 21st century compliment that my machine learnt ad suggestions are so poor. I think it indicates Google/Facebook don't have that much data for me. Sometimes it's fun game trying to work out how the algorithm possibly thought I would know people from other countries with no mutual friends. > subscribe to Grammarly Oh my god. I considered installing it just to stop the ads, but I see that doesn't help.
  - TeMPOraL 8y ago
    > Oh my god. I considered installing it just to stop the ads, but I see that doesn't help. I have mixed feelings about Grammarly. It's a perfect example of a thing that should not be an on-line service, but a fully off-line product. In its current form, it's essentially a keylogger.

by nlh 8y ago

dgzl 8y ago

Being upfront would bring more press... Alternatively, how many of your friends are going to read this article?

beatgammit 8y ago
Would it? The press likes to jump on surprising news, and if everything is upfront, there's not much to report on. They could even make a way to opt out if people complain (and few would actually opt out), which would save face in the media. I don't use Facebook because I value my privacy, and I'd probably have less success convincing people to leave the platform if they were upfront about the stuff they do. People care a lot less about their privacy than PR firms seem to realize...

drb91 8y ago

Actually enumerating what data might be used for strikes me as incredibly unlikely. They could also come up with a more substantial business model.

drdaeman 8y ago

> Friends you May Know is really cool! I've got over 9000 suggestions and not a single accurate one. Maybe they have the best technology and all the data out there, but I'm skeptical they can really use all the data they supposedly have on me. Same goes for any targeted advertisement, not just Facebook. Some come close (based on simple topic preferences), but in all those years I've yet to see any large number of the actually useful ones. All those data mining and AI-driven targeting seem to utterly fail, suggesting me to buy a second TV after I've just already got a new one. Or subscribe to Grammarly (every second YouTube pre-roll, seriously?!), when I've already bought that.

smelendez 8y ago
Interesting! It works pretty well for me. At worst it's usually people I don't know but logically could: friends of friends, people who worked somewhere I worked before or after me, people who went to my high school I never really knew. Occasionally it's full of apparent shady accounts: usually fake looking accounts of attractive women I have no friends in common with, then they'll disappear. I'm guessing scammers upload fake contact lists of random phone numbers to get on people's suggestion lists
- drdaeman 8y ago
  It's almost always people I've never ever heard of for me. Just some random strangers that I don't see any connections with. Sometimes those are probably friends-of-some-"friends," but I haven't bothered to check every unknown name and face. I've recently moved to a different country, and now I get suggestions of some random people supposedly from here - even though I don't know anyone here yet. Well, except a few coworkers and a landlord, who - if they have Facebook accounts - aren't in the suggestions, even though Facebook knows where I live and work (I don't advertise it, but that's not private either). Paired with the fact that Facebook intentionally spams daily with "you have 20 new notifications", they just train me to filter and ignore their notifications. If someone's going to message me there, I won't see them until the routine monthly check.
JoeSmithson 8y ago
I take it a kind of weird 21st century compliment that my machine learnt ad suggestions are so poor. I think it indicates Google/Facebook don't have that much data for me. Sometimes it's fun game trying to work out how the algorithm possibly thought I would know people from other countries with no mutual friends. > subscribe to Grammarly Oh my god. I considered installing it just to stop the ads, but I see that doesn't help.
- TeMPOraL 8y ago
  > Oh my god. I considered installing it just to stop the ads, but I see that doesn't help. I have mixed feelings about Grammarly. It's a perfect example of a thing that should not be an on-line service, but a fully off-line product. In its current form, it's essentially a keylogger.