In the rapidly developing and competitive banking industry, chargeback cases are a growing concern. Chargebacks can strain the bank’s system and may lead to an increase in customer reports, causing dissatisfaction.
THE OBJECTIVE
The objective was to, first of all, ensure compliance with relevant regulations and keep customer service at the highest level. We offer a reliable solution to decrease erroneous chargeback reports.
EXPECTATIONS
Neontri’s goal was to provide a solution to decrease the number of erroneous chargeback reports that could stem from users’ mistakes. With too many chargebacks, banks must also comply with agreements with payment providers.
To decrease the chargebacks, there were two issues that need to be addressed: decreasing chargeback reports to fit in the agreement limits, and finding the reason why so many erroneous reports were made.
According to the Midigator report on 2022 chargeback statistics, the high number of fraud-based chargebacks is due to a poor classification process. Because of limited knowledge and insight, banks assign incorrect codes to chargebacks that are “friendly” fraud.
This kind of chargeback means that a customer files an illegitimate chargeback report. It may happen when, for example, the customer doesn’t recognize a purchase and thinks they’re a fraud victim. Good categorization and proper transaction labels can reduce this.
Such categorization is possible through data enrichment technology. With this solution, transactions are enriched with merchant logos, addresses or locations, and contact information to facilitate payment identification for clients.
OUTCOMES
Chargeback costs are higher today than in 2019, so reducing their number is important. Moreover, according to Clearly Payments, the most common chargeback reasons are unknown or fraudulent purchases (35%). The unknown purchases most likely stem from the lack of context in transactions, so customers report chargebacks on payments they may not remember or identify by mistake.
To decrease erroneous chargebacks reported by customers, we would implement a solution that could provide clients with transaction information that would make it easier to identify where and when the transaction took place.
The solution was expected to decrease the number of reported chargebacks because customers could easily recall their transactions thanks to data enrichment.
THE PROCESS
Our job was first to identify the problem and find a fitting solution. Once we knew the erroneous chargebacks were caused by a lack of transaction information, we proposed data enrichment to fill in the missing information that caused clients to report chargebacks.
Snowdrop was the solution we implemented for data enrichment. We used it to provide the client with a proper data enrichment system for banking transactions.
Why this technology? The quality of data introduced through the data enrichment process must be the highest. Another reason was the cost optimization. Moreover, the client needed a real-time data enrichment solution for transactions to boost user experience.
CHALLENGES
One of the challenges was to find a solution that could handle big data in retail and provide real-time data enrichment for a massive user base. Snowdrop Solutions was chosen in this case as it could handle that.
The next challenge was to fulfill the technical requirements, such as caching transaction data within the permissible period in terms of legal requirements that was different for various purchases. Then, implementing real-time data enrichment was another issue.
We had to address the legal requirements for information storage, which depended on the market and transaction types. This is why we came up with a valid solution for all transactions, which required in-depth preparation and customization.
The process didn’t impact the project timeline. The implementation of the data management system took about three months.
COOPERATION
As the bank has been our partner for a long time, we knew how to handle cooperation effectively. To do so, we utilized the agile methodology with two-week sprints and iterated requirements depending on the performance testing covered by the client.
Our team included three programmers and a project manager. The bank was responsible for acceptance testing.
TECHNOLOGY
The key technologies used in the project were: Python 3, fastAPI, Apache Cassandra (DataStax), C++ 11, Apache Kafka, RESTful API, OpenShift (environment), and Terraform.
C++ was used for performance reasons to process and save data.
Python and FastAPI were used for API integrations as they enabled easy changes to fit requirements.
The two main components were handled using different technologies to address specific needs and requirements. In this project, we implemented the data enrichment technology offered by Snowdrop Solutions.
RESULTS
The expected results were to significantly decrease erroneous chargebacks regarding friendly fraud cases. Looking at the customer service reports, this is a reasonable expectation based on the data provided by the bank. Many of the cases of friendly fraud could be a result of a lack of transaction classification and proper labels.
The business impact of this project included a lower cost of handling chargeback reports with payment providers and customer service. With the data enrichment of transactions, users are better informed about their transactions and don’t have to identify the purchases that are easily labeled in real-time. As a result, the number of reports concerning unidentified transactions to customer support should also be decreased.
Many banks don’t offer such technologies for their customers. The implementation of data enrichment gives them leverage and increases their competitiveness in the banking market.
Paweł Scheffler
