Bifalgorithm post

Let's say you offer a free trial and then you notice the same people keep creating new accounts to get it again. How do you stop them? The first idea is obvious. One free trial per IP address. Sounds reasonable, but it breaks pretty quickly. Imagine four people living in the same house. They all use the same Wi-Fi. From your service point of view, they may all appear behind the same public IP address. So if one person uses the trial first, everyone else could get blocked. The same problem happens in offices, universities, hotels or mobile networks where lots of legitimate users can share an IP. So IP alone is not enough. What about device fingerprinting? that account history sent out velocity behavior. And maybe challenge suspicious users instead of blocking them immediately. Some people will still get through. And occasionally a legitimate user might look suspicious. That's why good fraud systems are based on probability, not certainty. You're not asking, is this definitely the same person? You're asking, given everything we know, how likely is it that this person is abusing the free trial?

Let's say you offer a free trial and then you notice the same people keep creating new accounts to get it again. How do you stop them? The first idea is obviou…