Systems and methods for predicting chargebacks

A computer-implemented method for determining the likelihood of a transaction resulting in a chargeback using a computing device coupled to a payment processing device, the method comprising:

• extracting transaction information about a plurality of transactions authorized by an account, the transaction information comprising at least one of authorization data and clearing data of each of the plurality of transactions;
• extracting account profile information about the account used to authorize the plurality of transactions, the account profile information comprising historical chargeback information about the plurality of transactions authorized using the account;
• generating a chargeback prediction model for the account based on the extracted transaction information and the account profile information of the plurality of transactions;
• receiving a candidate transaction paid for using the account and extracting transaction information from the candidate transaction;
• analyzing merchant prior performance information associated with a merchant identified in the candidate transaction to determine a chargeback risk associated with the merchant;
• generating a chargeback prediction score indicating a likelihood that the candidate transaction will result in a chargeback based on the analyzed merchant prior performance information, the extracted transaction information for the candidate transaction and the generated chargeback prediction model, wherein the chargeback prediction score indicates a likelihood that a cardholder for the candidate transaction will request a return of funds provided during the candidate transaction;
• determining, based on the chargeback prediction score, that the candidate transaction is likely to result in a chargeback;
• and
• based on the determination that the candidate transaction is likely to result in a chargeback, causing the transaction to be denied in real time at the time of authorization, thereby conserving computational resources and reducing processing time that would otherwise be used for processing the chargeback.

One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon for determining a likelihood of a transaction resulting in a chargeback, wherein the computing device includes a memory device and a processor in communication with the memory device, and wherein when executed by said processor, the computer-executable instructions cause said processor to:

• extract transaction information about a plurality of transactions authorized using an account, the transaction information comprising at least one of authorization data and clearing data of each of the plurality of transactions;
• extract account profile information about the account used to authorize the plurality of transactions, the account profile information comprising historical chargeback information about the plurality of transactions authorized using the account;
• generate a chargeback prediction model for the account based on the extracted transaction information and the account profile information of the plurality of transactions;
• receive a candidate transaction paid for using the account and extracting transaction information from the candidate transaction;
• analyze merchant prior performance information associated with a merchant identified in the candidate transaction to determine a chargeback risk associated with the merchant;
• generate a chargeback prediction score indicating a likelihood that the candidate transaction will result in a chargeback based on the extracted transaction information for the candidate transaction and the generated chargeback prediction model, wherein the chargeback prediction score indicates a likelihood that a cardholder for the candidate transaction will request a return of funds provided during the candidate transaction;
• determine, based on the chargeback prediction score, that the candidate transaction is likely to result in a chargeback;
• and
• based on the determination that the candidate transaction is likely to result in a chargeback, cause the transaction to be denied in real time at the time of authorization, thereby conserving computational resources and reducing processing time that would otherwise be used for processing the chargeback.