Introduction

The purpose of this manual is to guide you through the process of sending a sample of your historical data to Fraudio.

Firstly - It is important for us to receive a sample with as many data fields as possible. The more data fields we get, the better your fraud scores will be! If at anytime though this process you have questions please contact us!

The manual is broken down into 3 areas:

1. Key Points for a Smooth Process. We have learned from experience that there are some key points to get right so we have created some easy steps below.
2. What Data to Gather. We have a comprehensive list of fields that make up the full Fraudio schema - the more of this data you can send - the better we can determine what data we need for sending you the best possible fraud scores.
3. How to Send the Data. The data samples can be sent to us by e-mail.

1. Key Points for a Smooth Process

How much data to send

• We want to receive a data sample containing at least 6 months (and up to 12 months) worth of all transactions from all of your merchants and/or cards.
• Chargeback data - we also need all chargebacks reported for the entire data sample - excluding chargebacks for the last month (which becomes your blind test month)

<aside> 💡 Because the majority of chargebacks do not get reported until 45-90 days from the date of the original transaction, it's important to consider this when sending chargeback data.

</aside>

• We prefer that you send us raw data. Don't waste time during this first iteration preparing/pre-processing your raw data.

<aside> 💡 In order to map your data to our AI brain schema we need to know what your data fields represent e.g. cardbin = Card Bin Number

</aside>

2. What Data to Gather

We need many different data columns to predict fraud. You can either send as much information as possible per transaction in one single file (after joining all of the necessary data tables on your side) or you can send us your raw data tables (with parent identifers) and we will do the joining on our side.

The data we need is:

2.1 All Payment Data

This is any data that gives information about payment transactions. Payment data can be acquiring data, issuing data, wallet transfers/payments, bank transfers, etc. To find out the details of the data we are expecting including datatypes, descriptions, & example values - head to our Fraudio Data Schema Index page and select the view: