Welcome to Sift Science!

Protecting your business from online fraud doesn't have to be intimidating. At Sift Science, we're committed to helping your business harness the power of big data to make smarter decisions. Sift Science uses advanced machine learning technology to enable your team to fight fraud more accurately and efficiently – all while giving your legitimate users a great customer experience.

Whether you're a global online retailer or a small mom-and-pop shop, these guides will show you how to use Sift Science to prevent fraud and other behaviors, so you can focus on what's actually important – growing your business.

So...what is machine learning?

Machine learning is the science of training computers to learn and recognize patterns in data, allowing them to make accurate predictions. This is something that the human brain is able to do naturally, and computer scientists use machine learning to replicate it.

Chances are, you've already run into machine learning in your everyday life – possibly without realizing it. For example, Netflix's recommendation engine uses machine learning to predict which movies you might enjoy, while your email account uses machine learning to flag and prevent spammy emails from flooding your inbox.

Sift Science applies machine learning to identify fraud. In the same way that Netflix predicts what movies you'll enjoy, Sift Science predicts which of your business' users will be bad (or good) based on the data that you send us.

How does Sift Science fight fraud with machine learning?

There are three major steps that make up Sift Science's machine learning system – a cycle that helps Sift Science keep learning and improving.

  1. Train
    • Your business sends data to train your customized Sift Science model. The more data (and the more relevant the data), the better our machine learning models are able to extract patterns and learn about the unique types of fraud your business is experiencing.
  2. Predict
    • Our advanced machine learning infrastructure analyzes your data. Our system evaluates 5,000+ different fraud signals, detects connections and linkages between your users and known fraudsters, and runs complex analyses to predict whether someone is fraudulent or not.
  3. Act
    • Here's where you come in! You give us feedback to tell us whether our prediction was correct or incorrect, training the model and restarting the cycle. We'll automatically learn from your feedback, which keeps improving our fraud detection accuracy (and makes your life even easier)!

If you're interested in learning more about machine learning, check out our free ebook, The Future of Fraud Fighting.

Read Next