Automating Compelling Evidence for Chargebacks with Dynamic Arguments

As online transaction volumes soar, so do disputes. With global chargeback numbers projected to grow 42% by 2026, and fraudulent chargeback rates rising annually, merchants have no choice but to learn to fight chargeback fraud or risk potentially devastating annual losses – sometimes of up to 25% net income. To dispute illegitimate chargebacks, merchants are required to submit compelling evidence. But what exactly constitutes compelling evidence, and how can merchants leverage automation technology to streamline this complex process?

Compelling evidence in a chargeback dispute refers to any documentation or information that, based on the card scheme rules, proves the legitimacy of a transaction and contradicts the cardholder’s dispute claim. This evidence serves as the merchant’s primary defense against unwarranted chargebacks, especially those stemming from first-party, or friendly fraud. Merchants collect this evidence before sending it to their acquirer/PSP, who sends it to the card scheme, through which it flows to the issuer, who weighs the evidence before making a decision. 

Specific cases require specific kinds of evidence. For instance, for a physical goods online store that received a “Goods or Services Not Provided” dispute, a signed delivery receipt would be considered compelling evidence, and could help strengthen the evidence to win the dispute for the merchant. For an online merchant selling digital goods, the same “Goods or Services Not Provided” dispute would need a detailed service description and usage logs to help strengthen the evidence. In cases of suspected friendly fraud, evidence of the customer’s past purchasing behavior or additional information about their IP address might be particularly compelling. Some common forms of compelling evidence include:

• Transaction receipts may prove compelling in disputes where the customer claims they didn’t make the purchase.
• Delivery confirmations (signatures, photographic evidence) are essential for “goods not received” claims, these prove that the item was successfully delivered to the customer.
• Records of consent to terms of service are critical for subscription-based services or when disputing claims about unclear policies, demonstrating that the customer agreed to specific terms.
• Customer account activity history may compellingly demonstrate patterns of legitimate purchases, especially when a long-time customer suddenly claims fraud.

While the concept of compelling evidence might appear straightforward, the process of gathering and presenting it is far from simple. This is partly because of the huge diversity in how particular PSPs, acquirers, card schemes, and issuers demand that disputes are handled. Though each PSP and acquiring bank works according to the rules of the card schemes, they each have different processes and different ways of handling disputes. Each reason code requires different kinds of evidence, each card scheme has different rules, and each Issuer reviews the evidence in a different way. 

It’s easy to see how the number of combinations can quickly become astronomical. This amounts to a large degree of flux relating to how each chargeback must be handled – and a huge amount of work for payments teams if they are handled manually. If time is not taken to ensure all of these processes are followed according to the whims of each representative part of the payment ecosystem, your chances of winning the cases greatly decreases.

The time constraints faced by issuers further complicate matters – typically, their employees have three minutes to review up to 20 pages of documentation. While rules around evidence requirements are set by card schemes, they are subject to the issuers’ interpretation, and preferences differ drastically across different institutions. This means that evidence must be presented concisely, with the most crucial information prominently displayed – ideally according to the specific preferences of the issuer. Other complicating factors include: 

• Business model diversity: Many merchants operate multiple business models (subscription, physical goods, digital goods, services), each requiring different end-user flows and, consequently, different evidence arguments. This creates exponentially more work for administrative staff, who must learn the nuances of several argument types. 
• Case-specific narratives: Each chargeback has its own unique story based on the case data, necessitating a tailored response. Generic templates typically fail to address the specific nuances of individual cases.
• Human decision-making: For most chargeback (non-automated) solutions, humans make decisions on how to build the evidence, adding an element of subjectivity and capacity for error that is largely absent in automated solutions. This human factor frequently leads to inconsistencies and omissions in the representment of evidence.

All of these factors create multiple permutations for each evidence document, raising several critical questions. For every chargeback, we need to ask: 

1. What data points should be used? Different chargebacks require different types of evidence. Certain data points may be compelling in one context, and irrelevant in another. Ensure that your arguments are founded on the most impactful points.
2. What arguments should be made? The arguments you choose to present can make or break your case. They need to directly address the reason for the chargeback while effectively utilizing the available evidence to counter the cardholder’s claims.
3. How should these arguments be designed? Design choices, such as syntax and font, directly affect comprehension. Given that issuers only have minutes to review each case, well-designed arguments can make your evidence more compelling and easier to understand.
4. In what order should these arguments be placed? The order of your arguments can influence their effectiveness. Place your strongest points first to capture the reviewer’s attention and set a persuasive tone for the rest of your evidence.
5. How much evidence should be submitted? Too little evidence may not sufficiently support your case, while too much could overwhelm the reviewer.
6. How should the evidence be presented? Well-organized, clearly labeled, and easily digestible evidence can make it easier for the reviewer to understand and accept your argument.

The challenge of scale

Due to the difficulty of accounting for these permutations at scale, internal teams and other chargeback solutions typically use generalized templates for common use cases based on industry norms and one-size-fits-all reason codes. This decreases the quality of the evidence and fails to provide tailored responses.

Moreover, when chargeback volumes fluctuate, manual teams—whether internal or through a service provider—struggle to handle all cases or provide top-quality evidence consistently. This challenge is particularly acute during peak seasons or unexpected surges in dispute volume. The result is often a trade-off between quantity and quality, where neither is optimal.

To address these challenges, Justt has developed Dynamic Arguments – a unique feature that harnesses machine learning to respond to each chargeback with a bespoke and highly optimized response within seconds – regardless of volume. This fully automated solution frees your employees from time-intensive chargeback-related work, allowing them to focus on core business activities. 

Justt’s chargeback mitigation tool not only saves money by automating the dispute process, but by – on average – more than doubling chargeback win rates within a short period of time. 

Our solution achieves these results through several key processes:

1. Ongoing data collection: Justt continuously collects and structures potential evidence sources from merchants (when available), PSPs, and third-party sources. This comprehensive pool of 500+ data points forms the foundation for compelling evidence documents.
2. Case-specific tailoring: Based on this rich dataset, arguments are tailored for each specific case. The system analyzes the unique aspects of each dispute and selects the most relevant and compelling evidence to present.
3. Expert oversight: Domain experts constantly review regulatory and industry changes to update the model’s assumptions, ensuring compliance and effectiveness. This human expertise is essential for keeping the system aligned with the latest best practices and requirements, and the preferences of card schemes and issuers.
4. Continuous optimization: Ongoing A/B testing of arguments and data points in order to tackle the subcategories of cases that receive weaker results. This allows the model to choose and design the arguments and data points that will most likely lead to the highest chargeback recovery rate. This data-driven approach ensures that the system is always improving and adapting to changing patterns in chargeback disputes.

The result is a machine learning system that improves over time as it processes more data, leading to more wins month over month. This approach requires minimal human intervention on Justt’s side and, importantly, no human intervention on the merchant’s side – Justt’s system handles the complexities of evidence compilation and presentation at scale, producing clear and nuanced representments with a success rate that human teams simply cannot match.