Skip to main content

Refund Strategy (Operator Field Manual)

Cold reality: refund fast when it is cheaper than a chargeback, fight when evidence is strong, and route edge cases by ticket size and model.

What Matters (5 bullets)​

  • Refund vs fight math: dispute fees + ops time often exceed small tickets.
  • Ticket-size grid: under $25 almost always refund; above $500 fight with evidence.
  • Model matters: digital goods are harder to prove; physical goods rely on delivery proof; subscriptions hinge on cancellation proof.
  • Speed beats perfection: fast refunds reduce “I don’t recognize this” disputes.
  • Keep receipts organized: evidence templates matter less than transaction-specific proof.

Test to Run (2 weeks)​

  • Implement a triage grid by ticket size and dispute type.
  • Track: refunds issued within 24 hours, chargeback submissions, win rate, repeat dispute rate.
  • Success if dispute count drops and win rate improves without revenue loss beyond grid targets.

Scale Callout​

  • Under $100k/mo: Default to refund under $100 unless high-risk pattern. Document cancellations.
  • $100k–$1M/mo: Add reason-code routing and renewal reminder emails for subscriptions. Capture delivery proof on physical goods.
  • Over $1M/mo: Dedicated dispute owner, evidence packs by reason code, alerting for ratio spikes.

Where This Breaks​

  • No cancellation proof for subscriptions; add renewal reminders and confirmation logs.
  • Partial shipments or backorders without customer comms.
  • Refund promises without tracking (ops says yes, finance never issues).
  • International shipping with weak delivery confirmation.

Refund vs Fight Grid (Operator)​

  • Fraud claim + ticket < $100: refund.
  • Fraud claim + ticket $100–$500: refund unless strong device/IP/login evidence.
  • Fraud claim + ticket > $500: fight with device + login + delivery proof.
  • Service/quality complaint: refund unless clear misuse; aim for save via support first.
  • Recurring billing: if cancellation proof exists, fight; else refund and fix comms.

Support as Risk Control​

  • SLAs: under 1-hour first response during spikes; under 24h normal.
  • Empower refunds up to a threshold without manager approval.
  • Provide agents a script for “not recognized” calls: descriptor, date, item, renewal terms.

Next Steps​

Setting up triage grid?

  1. Apply refund vs fight grid - By ticket size and type
  2. Implement by scale - Right-sized for volume
  3. Run 2-week test - Track refunds, CBs, win rate

Empowering support?

  1. Set SLAs - Under 1hr during spikes
  2. Give refund authority - Threshold without approval
  3. Provide recognition scripts - "Not recognized" handling

Fighting specific disputes?

  1. Check reason codes - What evidence defeats each
  2. Build evidence packs - By reason code
  3. Follow representment guide - Full workflow