Skip to content
Documentation
All docs

Analyze

Chargebacks

Every dispute, with RefundSentry's prediction accuracy overlaid.

The Chargebacks surface shows every Shopify Payments dispute filed against your store, with RefundSentry’s prediction accuracy overlaid. It’s the “did we catch it?” view — the empirical test of whether your risk scoring is doing its job.

What you see

  • Trend chart— Chargeback count and dollar volume over time.
  • Reason distribution— Pie chart of dispute reasons (fraudulent, unrecognised, product not received, etc.).
  • Financial impact— Total disputed dollars and average days-to-resolution.
  • Prediction accuracy— What fraction of chargebacks RefundSentry scored HIGH or MEDIUM before the dispute landed. The headline number that justifies the product.
  • Return correlation— How often a chargeback followed a return we already flagged. Catches the return-then-chargeback abuse pattern.
  • Dispute table— Per-dispute date, amount, status, reason, and link into the related order.

Date range

The selector covers the last 3 / 6 / 12 months. Chargebacks settle slowly, so “last 7 days” isn’t a useful unit here.

How chargebacks land

RefundSentry subscribes to shopify_payments/disputes/create and /update webhooks. On install, we also backfill the last ~12 months of disputes via the Shopify Payments GraphQL API so the accuracy chart has historical data to work with. Disputes are linked back to the originating order and (when available) to the return that preceded them.

What this surface is good for

  • Calibrating thresholds— If our prediction accuracy is < 60%, your HIGH zone is probably set too tight.
  • Identifying gaps— Chargebacks we missed entirely (scored LOW or never scored) are the most valuable post-mortem. Open one, read the signal breakdown, and ask whether a signal weight or new policy could have caught it.
  • Justifying spend— The Financial Impact card is the export-to-PDF artefact for your finance team.

Next steps