4 ms·

Speaking from experience, porn sites send the worst traffic imaginable (just a bunch of horny guys who don't convert). Then again, every site on his network see

by jarin 16y ago

Speaking from experience, porn sites send the worst traffic imaginable (just a bunch of horny guys who don't convert). Then again, every site on his network seems to target a fetish of some kind. I don't think we'll run into any problems advertising SetForMarriage.com (since it serves basically the exact opposite market as Ashley Madison), but it is a little concerning that Google/Bing/Facebook might start to crack down on the dating industry as a whole. Granted, a lot of sites deserve it due to their ultra-shady business tactics, but we're trying to at least try to build a legit dating site that actually cares about its customers. For example, we try to answer all support emails within an hour, we try our best to block and delete fake profiles (most dating sites don't), we let non-premium users reply to messages (we got inspired by the deleted OK Cupid post about it), and users can cancel their own subscriptions via a big red button.

jrockway 16y ago
Interesting. How successful are you in blocking fake profiles? How do you even measure success?
- jarin 16y ago
  It's difficult to measure, but we kind of eyeball it by looking at the decrease in fake profiles with each thing we add. CloudFlare was hands down the most effective thing we've done so far (by blocking users by country and/or by Project Honeypot info). It virtually eliminated all of the automated bot accounts and most of the human spammers. A couple of days ago I changed the signup process to use the Facebook Registration tool (under the assumption that Facebook is way better at weeding out fake profiles than we are), and that seems to be working pretty well so far. It's too soon to tell how it will affect our signup rates, but we'd rather err on the side of quality profiles. I'm also looking at running profiles through Mollom and some image-checking service I can't remember the name of if we still have fake profiles slipping through the cracks after all that, but we'll continue to go through the profiles every day and try to find the bad guys. We also gave a nice lady a lifetime premium account for reporting a ton of suspicious profiles when we were first starting out. :)
  - underwater 16y ago
    The image checking service is probably http://www.tineye.com/ http://www.tineye.com/
    - jarin 16y ago
      It's something like Oculus, or something like that. I've tried to use the Tineye browser plugin to check profile images, but it doesn't even pick up Raven Riley images (for some reason, like 5-10% of spam profiles have used Raven Riley as their image).

by jarin 16y ago

jrockway 16y ago

Interesting. How successful are you in blocking fake profiles? How do you even measure success?

jarin 16y ago
It's difficult to measure, but we kind of eyeball it by looking at the decrease in fake profiles with each thing we add. CloudFlare was hands down the most effective thing we've done so far (by blocking users by country and/or by Project Honeypot info). It virtually eliminated all of the automated bot accounts and most of the human spammers. A couple of days ago I changed the signup process to use the Facebook Registration tool (under the assumption that Facebook is way better at weeding out fake profiles than we are), and that seems to be working pretty well so far. It's too soon to tell how it will affect our signup rates, but we'd rather err on the side of quality profiles. I'm also looking at running profiles through Mollom and some image-checking service I can't remember the name of if we still have fake profiles slipping through the cracks after all that, but we'll continue to go through the profiles every day and try to find the bad guys. We also gave a nice lady a lifetime premium account for reporting a ton of suspicious profiles when we were first starting out. :)
- underwater 16y ago
  The image checking service is probably http://www.tineye.com/ http://www.tineye.com/
  - jarin 16y ago
    It's something like Oculus, or something like that. I've tried to use the Tineye browser plugin to check profile images, but it doesn't even pick up Raven Riley images (for some reason, like 5-10% of spam profiles have used Raven Riley as their image).