A spike today. A story from last month.
Align the denominator, time window and dispute reason.
A fictional merchant sees 36 new disputes this week, up from 12 last week. A colleague divides 36 by this week’s 1,000 payments and calls it the fraud rate. You need to explain what changed before recommending a policy change.
Fictional scenario · Synthetic data · No real customer records- new disputes
- 36
- previous week
- 12
- payments this week
- 1,000
What would
you look at?
Open at least two files before making a decision. You can read every file.
0 of 3 files opened
A useful definition01 / Cohort viewTransaction dates+
Thirty of this week’s disputes relate to purchases made four to six weeks ago. Six relate to newer purchases. This week’s payment volume is lower than last month’s. A same-week ratio is not a measure of confirmed fraud in this week’s sales.
Events can arrive after the underlying transaction. A useful comparison defines which transactions and outcomes belong together.
02 / Operations noteReasons & fulfilment+
Twenty-four disputes cite goods not received, eight cite unrecognised payments, and four cite duplicate processing. A fulfilment delay affected an older order cohort. These are reported reasons, not completed investigations.
A dispute is not automatically fraud. The reason mix points to different owners and evidence.
03 / Review planDeadlines & comparison+
Response deadlines vary across the open cases. The team has transaction dates, shipment records and customer contacts. Compare older purchase cohorts at the same elapsed time and use the payment provider’s definitions for any scheme-monitoring metric.
Operational deadlines need attention while analysis continues. An internal learning metric must not be presented as a network-program calculation.
Make the call.
Leave your reasoning.
Choose what you would do under this scenario’s assumptions. Your notes stay in this browser. Do not enter real customer information.
The trade-off
Still unknown
Which claims are valid, what evidence each case needs, and how comparable cohorts develop as more outcomes arrive.
A useful handoff
Separate confirmed findings from reported reasons. Record cohort definitions, response deadlines and the operational owner for each cause.
Completion records participation, not competence or certification. You can change your answer and compare again.
Further reading
Visa / Dispute resolutionReference checked 5 November 2026. The scenario, runbook and debrief are original Antifrog teaching material; the source does not endorse them.