3 ms·

>Several commentators and researchers have noted that imposing a financial cost on the sender may be the most effective way to discourage unwanted messages Sur

by rm999 14y ago

>Several commentators and researchers have noted that imposing a financial cost on the sender may be the most effective way to discourage unwanted messages Sure, if you measure your performance on true positives. But what about false positives? What if I and the recipient will gain utility from me sending him a message, but I refuse to pay a dollar to ensure he will see it? I think google proved algorithmic spam protection works well. It may not work perfectly (and IMO has gotten slightly worse over time), but it gets you pretty darn close without adding a tax. Automatic protection feels far more 'efficient' than the proposed one. I hope facebook's trial proves this.

Periodic 14y ago
I think the reasoning is that if you aren't already connected and it isn't worth spending $1 to make sure the other person sees the message, then how much utility does it really have? I suppose there could be a case where the utility to you is low, but the utility to the user is very high, but in this case they are probably seeking you out and you should invest in cost-per-click/conversion advertising instead of cost-per-impression. Facebook's proposed model has a very high cost-per-impression, which is what weeds out much of the spam. You have to get a very high conversion rate before it becomes economical, which implies that the message was indeed relevant. In fact, being an inbox message probably increases the conversion rate over other communication methods (such as ads and spam)
- rm999 14y ago
  The messages I get from non-friends are usually from old friends asking if I'm the correct rm999 (well, my real name). I think these messages do have a lot of utility, but I would never pay to send them. They also provide value to facebook by increasing the size of their social network. Uptown also brought up the semi-contrived (but still excellent) example of a found wallet, which is another case where the value to the sender is low but the global utility is high. Anyway, I get virtually zero e-mail spam and my e-mail address has been public for eight years. I don't see why facebook wants to shake things up. I consider spam a mostly solved problem.

by rm999 14y ago

Periodic 14y ago

I think the reasoning is that if you aren't already connected and it isn't worth spending $1 to make sure the other person sees the message, then how much utility does it really have? I suppose there could be a case where the utility to you is low, but the utility to the user is very high, but in this case they are probably seeking you out and you should invest in cost-per-click/conversion advertising instead of cost-per-impression. Facebook's proposed model has a very high cost-per-impression, which is what weeds out much of the spam. You have to get a very high conversion rate before it becomes economical, which implies that the message was indeed relevant. In fact, being an inbox message probably increases the conversion rate over other communication methods (such as ads and spam)

rm999 14y ago
The messages I get from non-friends are usually from old friends asking if I'm the correct rm999 (well, my real name). I think these messages do have a lot of utility, but I would never pay to send them. They also provide value to facebook by increasing the size of their social network. Uptown also brought up the semi-contrived (but still excellent) example of a found wallet, which is another case where the value to the sender is low but the global utility is high. Anyway, I get virtually zero e-mail spam and my e-mail address has been public for eight years. I don't see why facebook wants to shake things up. I consider spam a mostly solved problem.