Mitigate Fraud with Machine Learning Technology – Sift Science Integration

Mitigate Fraud with Machine Learning Technology – Sift Science Integration

Sandra Wróbel-Konior

Content Marketing Specialist at SecurionPay

Content Marketing Specialist with a tech-savvy personality, experience in writing and passion for reading. Staying up to date with the latest social media trends, in love with GIFs.

Views 812

Mitigate Fraud with Machine Learning Technology – Sift Science Integration

26.10.2016 08:15 am

Payment fraud is growing and is now one of the biggest problems for online business owners. As a merchant, you need to constantly monitor payments to predict the possibility of risk. Why don’t you use an automatic solution to limit the fraud attempts?

According to a LexisNexis® RiskSolutions report, merchants lost an average of 1.32% of revenue to fraud and its related costs. Fraudsters are getting smarter with the new technology so you need to use highly effective anti-fraud tools. One such tool is Sift Science with real-time machine learning.

Now you can easily integrate Sift Science with SecurionPay (you’ll find the instructions at the bottom of this blog post).

What is a Sift Science?

Sift Science is a tool which lets you mitigate the risk of online fraud. It is based on advanced machine learning technology so you can fight fraud more efficiently. It doesn’t matter if you’re a global online retailer or run a small, local e-shop. Machine learning makes it easier to recognize patterns in data. In short, computers are trained to learn from data to make accurate predictions, so it works kind of like the human brain.

Need an example?

You get into machine learning when you use your e-mail inbox. When you flag a specific message as a spam mail, it would be added to your spam folder automatically next time. That’s how machine learning works – it learns and remembers your behavior and preferences.

Machine learning system in Sift Science comes with three major steps which are: Train, Predict, and Act. First, you’re sending the data to train your customised Sift Science model to extract patterns. Then, the data is analysed and you evaluate the effectiveness of Sift Science prediction. In all, your feedback helps improve the fraud detection accuracy.

What is fraud scoring and how do you use a Sift console?

Let’s dig a little deeper into the Sift Science features.

To make it easier to evaluate whether a charge is fraudulent, you can see a risk score between 0 and 100 (the higher the score, the more suspicious the transaction) which is assigned to each of your customers. It’s updated every time the tool identifies new information about a customer, so it predicts which users are safe and which are fraudsters.

Example of a Sift Score
Example of a Sift Score

In short, when the charge is processed, the advanced algorithms rate the risk level. The scoring solution gives you the option to determine your business’ score thresholds based on your tolerance for risk. Every decision you make (block order or ban a user) gives Sift information which helps better identify fraudsters.

For detailed instructions, see a Sift Science guide.

To receive the most accurate scores, you should focus on a manual review to train the model during the first 30 days. The Sift team recommends reviewing every order over 60 and make a Decision in the Sift Science console.

After a month, Sift examines what score threshold makes sense for automation. Typically, it starts with blocking when a score is higher or equal to 95 and accept when a score is lower or equal to 20. Then they test it for 2-3 weeks, and if the results look good, they increase thresholds, e.g. block when score>=85 and accept score<=30. The process continues until the merchant reaches a comfortable, steady state level.

To learn more about how to enable Decisions and how to optimize the process for efficiency, watch this webinar: An Overview of Sift Science’s New Workflows and Review Queues.

You perform all the actions (identify and investigate suspicious users, take business actions, etc.) in the Sift Science Console. Here are the best practices for how to use the console:

 

How do you integrate SecurionPay with Sift Science?

The integration process is simple and takes a few clicks and a copy and paste action. Here’s how to connect your accounts:

  1. First, go to Sift Science and register your account (or log in if you have one).
  2. Then, go to your SecurionPay account settings and click on the Integration tab. You will see the Sift Science integration. The next step is to connect your account.
Sift Science integration
After pushing the ‘Connect my account’ button, you will see the fields to enter in the keys from Sift Science account.

Note that there are fields to paste Sandbox API Keys or Production API Keys. Double check which mode is activated in your Sift Science account settings.

3. Go to Sift Science console, choose the ‘Developers’ tab on the left side toolbar and then click on ‘API Keys’ tab.

Sift Science API Keys

 

Copy the keys and paste them in your SecurionPay dashboard. To complete integration, click the ‘Connect’ button.

Sift Science integration via SecurionPay

Now, you can give Sift Science a try with your data included.

Automatically blocking fraudsters will save you time, let you prevent chargebacks and fraudulent transactions before they happen. Are you ready to strengthen the fraud protection on your website?

 

This article originally appeared on SecurionPay.com

Latest blogs

Colin Neil Adyen UK

P2PE – Silver Bullet or Snake Oil?

Fraud is an ever-present problem for merchants, especially with the increasing number of payment providers, start-up challenger banks, and online shopping sites providing different levels of fraud protection. To ensure the best customer and business Read more »

Thomas Pintelon Capilever

Credit servicing - Much more than just a back-end process

While credit origination is considered as a very customer-centric process, the credit servicing part that comes afterwards is usually considered as a purely operational, back-end process. However, substantial added value and competitive advantage Read more »

Simon Cureton Funding Options

Due Diligence Vs Speed: Why It’s Not Either-or When It Comes to SME Lending

The coronavirus pandemic has put the SME lending market under the microscope with extensive debate on how to deliver the right financial support to SMEs at speed. Traditional retail banks usurped fintechs and were gifted the golden CBILs and BBLs Read more »

Keith McGill Equifax UK

Fraud Continues to Rise, but Faster Digitisation Will Bring Benefits

The results of the Cifas report show the changing face of fraud and identity theft across the UK. With a 13% rise in reports to the National Fraud Database from 2018, it’s clear that even before the pandemic struck there were a number of challenges Read more »

Darren Capehorn Icon Solutions

Unbanked and Unconnected: Supporting Financial Inclusion Beyond Digital

Many of us take it for granted, but accessing basic financial services is fundamental to our economic and social development. It is hard to ‘get on’ if you are forced to hide life savings under the mattress, or rely on predatory loan sharks for Read more »

Magazine
ALL
Free Newsletter Sign-up
+44 (0) 208 819 32 53 +44 (0) 173 261 71 47
Download Our Mobile App
Financial It Youtube channel