Playbook: Handle a Refund Request
- Check payment, refund and dispute status before promising money back
- Use your margins, support time and refund policy to set approval limits
- Give the customer a response deadline your team can meet
- Handle refund requests before customers turn to their bank. Compare your refund and chargeback costs using your processor's fees.
- Empower your support team. Compare response times before and after giving agents refund authority, using similar requests.
Step-by-step guide for when a customer asks for their money back. The goal: resolve it before they call their bank. First check the payment state. If a formal card chargeback is open, use the dispute case and don't send a separate refund. If RDR already resolved it or a refund exists, don't credit it again.
Workflow Overview
| Phase | Key Tasks |
|---|---|
| Receive | Log the request, acknowledge to customer by the response time you publish |
| Assess | Pull order details, check for fraud signals, review history |
| Decide | Refund, partial refund, deny with explanation, or escalate |
| Resolve | Process refund, confirm to customer, update records |
When to Use This Playbook
- Customer emails, calls, or chats asking for a refund
- Customer posts a complaint on social media
- You receive a manual Ethoca or CDRN alert; check for an existing credit and follow its deadline
- Customer's bank contacts you before filing a formal dispute; use the provider's inquiry workflow
Step 1: Receive and Acknowledge
□ Log the request with timestamp, channel, and customer ID
□ Acknowledge receipt by your published response time; tell the customer when you'll follow up
□ Use this template:
"Hi [Name], I received your request about order [#].
I'm looking into this now and will follow up within
[24 hours / 1 business day]."
□ Do NOT ask the customer to "wait while we investigate"
without giving a timeline
Reply promptly and give a next-update time. Track how often unanswered requests become disputes; a fixed 24-hour cutoff won't tell you that for your business.
Step 2: Assess the Request
Pull the order details and check these signals:
□ Order details:
□ Transaction amount
□ Order date and delivery date
□ Product/service delivered?
□ Digital or physical goods?
□ Customer history:
□ First-time or repeat customer?
□ Previous refund requests? How many?
□ Previous chargebacks?
□ Fraud signals:
□ Does the request match the order? (e.g., "never received" but tracking shows delivered)
□ Multiple refund requests in short period?
□ Pattern matches known friendly fraud?
□ Timing:
□ Within your posted refund window?
□ If outside window, how far outside?
Quick Decision Matrix
| Request | What to check | Action |
|---|---|---|
| Clear delivery or billing error | Order, payment and refund status | Correct it through the supported workflow |
| Policy-covered return | Policy at sale, condition and any return requirements | Explain the return and refund steps |
| Conflicting records or repeat requests | Delivery, usage, prior credits and customer explanation | Investigate and promise a next update |
| Large or ambiguous request | Your approval limit and evidence | Escalate to the person authorized to decide |
Set approval limits from your own ticket size, margins and support time. This table applies only after the payment-state check above.
Step 3: Decide
Option A: Full Refund
Use when:
- The request is within your own refund approval limit
- Customer has a legitimate complaint
- Product was genuinely not delivered or defective
- You're outside the return window but the request is reasonable
□ Process refund through your processor dashboard
□ Use your provider's supported refund route for the original payment
□ Send confirmation email with refund amount and expected timing
□ Check your provider's estimate for this payment method; separate submission time from bank posting time
Option B: Partial Refund
Use when:
- Customer used part of the service before requesting refund
- Product was partially delivered or partially defective
- Shipping costs were incurred that can't be recovered
- You want to retain some value while satisfying the customer
□ Calculate fair partial amount
□ Explain the breakdown to the customer before processing
□ Get customer agreement before processing (email confirmation)
□ Process and confirm
Option C: Deny with Explanation
Use when:
- Strong evidence of friendly fraud (confirmed delivery, usage logs)
- Request is clearly outside policy and unreasonable
- Customer has a pattern of abuse
□ Prepare evidence (tracking, delivery confirmation, usage logs)
□ Respond with specific reasons, not just "policy says no"
□ Offer alternatives allowed by your policy and applicable law: exchange, store credit, partial refund. An open chargeback stays in the dispute workflow
□ Document everything in case this becomes a chargeback
□ Save all communication for potential representment
For an eligible direct request, offer an alternative allowed by your policy and applicable law. An open formal chargeback stays in the dispute workflow; don't send a separate partial refund. Track how often similar requests become chargebacks before and after the change, allowing the same follow-up time.
Option D: Escalate
Use when:
- Amount is large and circumstances are ambiguous
- Customer is threatening legal action
- Request involves a product safety issue
- You need manager approval per your internal policy
□ Inform customer of escalation and revised timeline
□ Escalate with full context (don't make the next person start over)
□ Follow up within the promised timeline
Step 4: Resolve and Follow Up
□ Process the decided action (refund, partial, denial)
□ Send confirmation to customer with:
□ What was decided
□ If refund: amount and your provider's timing estimate for this payment method
□ If denied: clear reasoning and alternatives offered
□ Update customer record with outcome
□ If this was a manual alert: report the outcome by its deadline; if RDR already resolved it, don't issue another credit
□ Flag patterns for review (is this product generating lots of requests?)
After Resolution: Prevent Repeats
After every 10 refund requests, review for patterns:
□ Are refunds clustering around a specific product?
□ Are refunds clustering around a specific timeframe (e.g., 30 days post-purchase)?
□ Is your refund policy clearly visible before checkout?
□ Are your billing descriptors recognizable? (Check: would a customer
recognize the charge on their statement?)
□ Are confirmation and shipping emails being sent and received?
Measuring Success
| Metric | Target | What It Tells You |
|---|---|---|
| Refund response time | Your published response commitment | Speed of customer resolution |
| Refund-to-chargeback ratio | Compare before and after changes on similar orders | Are requests resolving before customers dispute? |
| Repeat refund rate | Compare repeat requests within the same customer and product segments | Which requests need a closer look? |
| Refund rate | Compare similar products and sales periods, allowing the same time for refunds | Where are refund reasons changing? |
See Also
- Refund Strategy - The math behind refund vs. fight decisions
- Refund Policy Design - Building a policy that prevents chargebacks
- Chargeback Prevention - Stop disputes before they start
- Billing Statement Names - Fix unrecognized charges
- Chargeback Alerts - RDR, Ethoca, CDRN pre-dispute alerts
- Representment - When the customer disputes anyway