How it works
Ding’s fraud prevention algorithm is an adaptive machine learning system that assesses the risk level of each authentication in real time. It uses dozens of signals relating to each OTP and draws on data from our vast database to predict whether a request is likely to be fraudulent.
Our system is flexible and responsive, continually learning from new fraud patterns and adapting to new attack vectors while granularly filtering the attacker from real users without resorting to blocking an entire country or network block.
When you make an Authentication request, we retrieve the data we already have on the current user and add the signals you provide. The system then uses this information to correlate the authentication with ongoing spam attacks, bot farms, and other fraudulent activity.
Why we built it
Who hasn’t heard of a company affected by SMS fraud? Even the biggest names with huge engineering budgets are not immune to it.
Twitter is getting scammed by phone companies for $60M/year of fake 2FA SMS messages
Why is this happening? Because fraud is a cat and mouse game and keeping away from it is a constant investment. We’ve designed this feature to take the pressure off your engineering team, so you can concentrate on what’s important: your core business.
To make sure that you catch all spammers during your onboarding, make sure you send all the appropriate Signals to Ding when sending authentication requests. You can find the full list of signals in the API reference.