Chargeback management tools are often bought on a headline win rate that nobody can reproduce, under a pricing model that rewards the vendor for disputes the merchant would have won anyway. The market now mixes three different jobs: stopping disputes before they become chargebacks through network pre-dispute services, fighting the ones that arrive with well-built evidence, and reporting the numbers acquirers and card networks watch. This vendor-neutral buyer's guide is for payments, risk and finance leads at merchants, subscription businesses and marketplaces choosing dispute and representment software or a managed service. It covers requirements, capability areas, evaluation criteria, vendor questions, red flags and a proof of concept. Fraud screening and chargeback guarantees are covered in the Ecommerce Fraud and Chargeback Protection buyer's guide, and the dispute process itself in the Chargeback Management and Dispute Guide.
Key takeaways
- Prevention and representment are different products. Pre-dispute alerts and resolution services keep a claim from becoming a chargeback; representment wins money back after one. Most merchants need both, priced differently.
- Resolved pre-disputes usually do not count against you. Processor documentation states that disputes resolved through Visa Rapid Dispute Resolution or Ethoca Alerts do not count towards dispute rates, which matters for network monitoring programmes.
- Evidence quality is decided at checkout. Visa Compelling Evidence 3.0 needs prior transactions matching on IP address or device ID plus another element, so the tool must receive that data from your order systems.
- Define win rate before comparing vendors. By count or by value, with or without cases the vendor chose not to fight, and including or excluding pre-arbitration, the same results can look very different.
- Pricing models change behaviour. Contingency fees on recovered value, per-case fees, per-alert fees and subscriptions each create different incentives; model them on your own dispute mix.
Define requirements
- Baseline: 6 to 12 months of disputes by network, reason code, amount, product and channel; current win rate by count and value; refund-before-dispute volume; dispute rate as each acquirer reports it; staff hours per case.
- Mix: the share of true fraud, friendly fraud (cardholder made the purchase but disputes it), and service disputes such as item not received or cancelled subscription. Each needs a different response. See the First-Party Fraud Guide.
- Payment stack: processors, acquirers and payment service providers in use, and whether each supports the pre-dispute programmes and API access the tool needs.
- Data sources: order management, shipping and delivery confirmation, customer account logs, device and IP data, subscription and cancellation records, customer service tickets.
- Operating model: software your team runs, a managed service that writes and submits responses, or a hybrid. Decide who approves auto-refunds and auto-accepts.
- Constraints: data residency, card data scope, procurement rules on contingency contracts, and the monitoring programme thresholds your acquirer has warned you about.
Capability areas
| Capability | What it does | What to check |
|---|---|---|
| Pre-dispute alerts | Receives issuer notifications of confirmed fraud or a customer dispute before a chargeback is filed so the merchant can refund, stop shipment or cancel | Network coverage, alert latency, automated refund rules, how quickly outcomes are returned (Mastercard's Ethoca documentation recommends a final outcome within 24 hours) |
| Rules-based pre-dispute resolution | Automatically resolves qualifying disputes under merchant rules, such as Visa Rapid Dispute Resolution | Rule granularity by amount, reason and product; reporting of credits issued; per-case fees; exclusions such as partial amounts |
| Order data sharing | Returns order details to the issuer when a cardholder does not recognise a charge (Visa Order Insight, Mastercard Consumer Clarity) | Which fields are sent, merchant descriptor quality, data minimisation, enrolment through your processor |
| Compelling Evidence 3.0 handling | Identifies and submits qualifying prior transactions for Visa reason code 10.4 card-absent fraud disputes | Automatic matching of the 120 to 365 day history, IP and device data capture, one-attempt handling |
| Representment automation | Collects evidence, builds reason-code-specific responses and submits them through processor or acquirer portals and APIs | Templates per network and reason code, evidence pulled automatically from your systems, deadline tracking, human review queue |
| Analytics and reporting | Measures win rate, recovered value, dispute rate trends and root causes | Definitions published, raw case export, reason-code and product drill-down, monitoring programme forecasts |
| Root-cause feedback | Feeds dispute causes back to fraud rules, descriptors, fulfilment and customer service | Links to your fraud tool, alerts on emerging reasons, recommendations you can act on |
Evaluation criteria
| Criterion | What good looks like | Weight guide |
|---|---|---|
| Prevention coverage | Supports the pre-dispute and order data programmes available through your processors for your main networks | High for high-volume digital goods and subscriptions |
| Evidence automation | Pulls delivery, login, device, IP and usage evidence automatically; CE3.0 matching built in | High |
| Measurable outcomes | Win rate and recovered value reported by count and value, with every case exportable | High |
| Control over decisions | Merchant-owned rules for auto-refund, auto-accept and fight, with audit history | Medium to high |
| Integration and security | Documented APIs; scoped, rotatable credentials for processor connections; SSO and role-based access; no full card numbers needed | High |
| Commercial model | Fees modelled on your mix; no fees on cases you would refund anyway; clear exit and data return | High |
| Service | Named analysts for managed service, quality checks on responses, reason-code expertise in your regions | Medium |
Pricing models compared
| Model | Incentive it creates | Works best when |
|---|---|---|
| Contingency (share of recovered value) | Vendor fights cases likely to win; may skip low-value or hard cases | You lack staff and want no upfront cost; check whether easy wins you already achieve are charged |
| Per case submitted | Vendor paid whether or not you win | You control which cases are fought and have reliable evidence |
| Per alert or per resolved pre-dispute | More alerts mean more fees, including refunds you would have given | Dispute rate is near a monitoring threshold and avoiding chargebacks is worth the fee |
| Subscription or platform fee | Predictable cost; vendor less tied to outcomes | Your team operates the tool and volumes are stable |
Many contracts combine models. Rebuild every quote against the same 12 months of your own disputes before comparing.
Questions to ask vendors
- How do you define win rate, and will you report it by count and by value, including cases you chose not to fight?
- Which pre-dispute and order data programmes do you support through our specific processors and acquirers, and which require separate enrolment?
- How do you identify and submit Compelling Evidence 3.0 cases, and what data do you need from our checkout and account systems?
- Which of our dispute reason codes do you not handle, or handle only manually?
- What fees apply to a pre-dispute alert that ends in a refund we would have issued anyway?
- How do you track network and processor response deadlines, and what happens if one is missed?
- How do processor credentials work: scoped API keys or OAuth, who holds them, how are they rotated and revoked? See the API Key Management Guide.
- Do you need full card numbers or any sensitive authentication data, and what is your PCI DSS responsibility position?
- Can we export all case data, evidence and outcomes in a usable format at any time and at exit?
- What root-cause reporting do you provide, and how does it feed our fraud rules and customer service?
Red flags
- A win rate quoted without a definition, a time period or the denominator.
- Contingency fees applied to cases resolved by refunds, reversals or issuer withdrawals the vendor did nothing to win.
- Alert services that auto-refund everything by default, with no merchant rules on amount, reason or customer history.
- Generic response templates that ignore reason code and network requirements.
- No raw case export, or historical data held back at contract end.
- Requests for shared logins to processor portals instead of scoped, auditable credentials.
- Promises to remove you from network monitoring programmes without showing how resolved disputes are counted.
Proof of concept
- Freeze a baseline: 90 days of disputes, outcomes and dispute rate by network and reason code, with your own win rate definition written down.
- Connect read-only first: give the vendor scoped, revocable credentials and limited data; confirm no card numbers flow.
- Run alerts in shadow mode: receive pre-dispute alerts and log the recommended action without auto-refunding, then compare with what your team would have done.
- Split representment: send a random half of new disputes to the vendor and keep half in house for a like-for-like comparison over at least one full dispute cycle.
- Test CE3.0 matching: check how many reason code 10.4 disputes the tool finds eligible and whether its matched data is correct.
- Price the result: apply each vendor's fee model to the trial cases and calculate net recovery after fees.
- Plan rollback: confirm you can revoke credentials, stop alert enrolment through the processor and export all case data before signing.
Track the outcome with the measures in the Fraud Metrics and KPIs Guide: dispute rate by network, net recovered value, cost per dispute and time to respond.
How NHI Mgmt Group can help
We provide independent market analysis, RFP strategy and evaluation support for fraud, identity and security programmes. Browse vendors in our NHI and AI products directory or contact us about independent evaluation support.
Related NHI Mgmt Group resources: Chargeback Management and Dispute Guide · Ecommerce Fraud and Chargeback Protection · First-Party Fraud Guide · Payment Fraud Guide