When your business has a fraud problem, you’re not just working to stop the bad guys – you also have to ensure that good users aren’t blocked and can access your site. We understand your pain – your team has to juggle multiple decisions and lots of moving pieces in the fraud review process, all under a time crunch.

We built Formulas to help you accurately stop fraud in a way that’s easy, flexible, and customizable to your business needs.

What is Formulas?

Formulas is a feature that allows you to link Sift Science with your business’ systems so you can automate decisions and actions using insights from our powerful machine-learning platform.

Now you can easily customize and automate your business’ fraud workflows to reduce manual reviews, save time, and give your users a smoother online experience (all without having to involve your engineering teams!).

See Formulas in action:

What can Formulas do for me?

Formulas are incredibly flexible and powerful, enabling you to easily customize and automate business decisions and processes relevant to your needs.

Automated fraud management

  • Spend less time reviewing fraud and focus your efforts on what’s actually valuable. Implement Formulas to make automatic decisions based on a Sift Score or any other important signals so your team isn’t stuck reviewing the obvious stuff.
  • Examples:
    • Hold User for Review: If a user has been found to share a device with another user on your site, create a Formula to automatically flag him for manually review before their action is allowed to process.
    • Block Order: If your business policies don’t allow you to ship to Vietnam, set up a formula to block any order that is shipping to Vietnam from being processed.

Dynamic alterations based on fraud riskiness

  • Change the user experience for your customers based on how likely they are to be fraudulent.
  • Examples:
    • Preventing spam listings: You can easily introduce friction for suspicious users on your site. If a user with a high Sift Score attempts to post a listing on your site, you can add steps that require them to verify their identity through e-mail verification.
    • Making checkout faster (and easier) for return customers: If a user has a low Sift Score and has had legitimate past orders on your site, you could reduce friction in their checkout flow by not asking for the CVV number of their credit card.

Formulas are only available to Premium and Enterprise customers on their admin accounts..

  1. Do I get it right that ‘Formulas’ is just a good old rule-based antifraud approach implemented along with AI antifraud?

    1. Hi Vladimir,

      Yes, ‘Formulas’ is a rules-based system, but it’s also much more than a ‘good old rule-based antifraud approach’.

      Many of our customers need to enforce business policies like ‘Never ship to North Korea’. However, these black-and-white policies are generally not well suited for a machine learning solution. A business can now use Formulas to automatically block order fulfillment and shipping for any orders going to North Korea much like a traditional rules-system.

      The strength of ‘Formulas” lies in the fact that it allows businesses to easily customize and automate their fraud decisions while also leveraging Sift Science’s machine learning insights. A Formula can use our ML-driven Sift Score or hundreds of other attributes to automate business decisions in real-time. (e.g. A store can set up a Formula to automatically ask for the CVV number during the checkout flow for a customer with a high Sift Score, a forum can remove an e-mail verification modal for a new signup that’s low risk, etc.)

      With Formulas, there’s now immense flexibility for you to automate your business’ decisions in Sift Science. Your can easily connect Sift Science to your systems, and you don’t need to involve your developers to customize or manage your Formulas.

      If you have any questions, feel free to contact us at support@siftscience.com