Integrating with Checkout.com's Risk SDKs is crucial for maximizing the efficiency of the Fraud Detection solution, especially for customer-initiated payment flows.
The SDKs capture advanced fraud signals that are leveraged in our machine learning (ML) model, significantly improving risk scores and reducing false declines.
Key enhancements include: Device data
The SDKs capture device identification, precise geolocation, spoofing attempts, and fingerprinting data.
This includes details like device fingerprint, IP address, IP country/city, model, OS, browser, timezone, and whether the browser is in incognito mode.
ML model improvements
Payments enriched with this device data are scored against an ML model that performs twice as well as models without device data.
This data can be used in your risk strategy via an ML risk profile or the :score: threshold rule property
How do Checkout.com's Risk SDKs enhance fraud detection?
Was this article helpful?
Share
Copy Link
Share via email
Articles in this section
- How to view which user created a risk rule?
- Risk score terminology
- How to reduce fraud from a specific issuer
- Can liability shift if the issuer doesn't support CVV checks?
- How to reduce false positive risk declines
- Why can't I start backtesting?
- Reconciling backtesting with reported fraud
- If I change my live risk strategy after starting backtesting, which strategy is assessed?
- Fraud risk scores on Pro only
- Are there any Fraud Detection webhooks?