Had lunch with Glenn from iovation today. Beautiful fall day here in Boston, perfect for sitting outside and chatting for a few hours. I am seriously getting interested in fraud detection services, lots of interesting solutions, partnerships and deals going on. iovation is clearly on a roll these days and it’s nice to see dating sites acknowledge that their services are capable of reducing fraud, which results in less spammy profiles and makes for a better user experience. Anyone who is starting a dating site should be checking out anti-fraud services, which come in many flavors.
I’ll be posting soon about Sobayli, a client which is doing tremendous work reducing spammy profiles and enabling dating sites to efficiently managing member-created content.
Related posts:
{ 3 comments… read them below or add one }
Have you heard of any creative uses of services like Akismet, Defensio, or Mollom for stopping profile and messaging spam?
Each of these offers an api that will accommodate most custom applications wanting to reduce spam.
I’ve integrated with Defensio for a few projects and think it works well. Akismet and Mollom are almost identical and work quite well too.
Akismet would be interesting, although it’s catching a lot of valid comments as spam.
iovation is using a shared database of fradulent activity and there are other companies out there providing the same for images.
I would not rely on Akismet and I don’t know enough about the others to judge them.
A common black-hole list for all dating and social sites in in order.
Defensio & Mollom try to be better akismets. The think I like about akismet is that you can help “train” their algorithm by reporting false positives and false negatives.
Also if you provide it with sufficient information (such as ip address of the client and email address of the writer) it does a pretty good job.
I had built some moderation tools around our use of akismet that allowed us to review what Akismet flagged as spam and help aid the system. I’ve minimally tested out Defensio and Mollom. We will most likely try a run with Defensio and see how it compares.