By NUMUS editorial team
To calculate a merchant chargeback ratio, divide the chargebacks included in the metric by its defined transaction base, then multiply by 100. The arithmetic is straightforward. The important work is confirming what is counted, which dates apply, and whether the figures represent transactions or money.
A percentage without those definitions is difficult to compare. Before interpreting a dashboard number or explaining processing history to a new provider, write down the calculation behind it. Our merchant account guide explains the broader account relationship; this worksheet focuses on the numbers used in that conversation.
Define the numerator, denominator and period
The numerator is the figure being divided. The denominator is the base you divide it by. For a count-based metric, both should be counts. For an amount-based metric, both should be monetary amounts measured on the specified currency basis.
Complete these fields from the provider's documentation or reporting explanation:
| Definition | What to record |
|---|---|
| Metric and purpose | Exact dashboard or report name; operational analysis or a specified monitoring measure |
| Numerator | Included events or payments; count or amount; exclusions and treatment of repeated records |
| Denominator | Eligible transaction count or sales amount; included payment methods and transaction statuses |
| Date basis | Dispute received date, original payment date or another defined date |
| Reporting window | Start and end dates for each side of the calculation, plus time zone |
| Scope and version | Merchant accounts, channels, card brands, report filters and export date |
Keep unresolved fields marked “confirm with provider.” A sales dashboard and a dispute dashboard should not be assumed to use matching filters just because both display “August.”
Stripe provides a concrete example of date differences: its dispute activity calculation groups disputes by dispute date, while its dispute rate groups them by original charge date. These are Stripe's dashboard definitions, not interchangeable names for every provider's metric. Stripe: Measuring disputes.
A fictional chargeback-ratio worksheet
Every number below is invented for arithmetic practice. These percentages are not monitoring thresholds, acceptable targets or NUMUS approval criteria.
Imagine a business with these records, all in one currency and one account:
- July: 1,600 eligible successful card transactions.
- August: 2,000 eligible successful card transactions, totaling $100,000.
- August: 12 newly received chargebacks on 12 distinct payments, totaling $900 in disputed amounts.
- Of those 12 payments, four were originally processed in August; eight were processed earlier.
Assume these are the complete records available at the export date and there are no additional exclusions or duplicate events. To isolate the effect of each definition, calculate four different measures:
| Fictional measure | Calculation | Result |
|---|---|---|
| Received count divided by same-month sales count | 12 August chargebacks ÷ 2,000 August transactions × 100 | 0.60% |
| Received count divided by previous-month sales count | 12 August chargebacks ÷ 1,600 July transactions × 100 | 0.75% |
| Received disputed amount divided by same-month sales amount | $900 ÷ $100,000 × 100 | 0.90% |
| August payments disputed as of the export date | 4 August payments with chargebacks ÷ 2,000 August transactions × 100 | 0.20% |
The first two rows change only the denominator's month. The third measures dollars rather than transaction frequency. The fourth asks about a particular group of original payments. These are four different questions, so their answers need different labels. The previous-month row is a hypothetical reporting definition; it is not presented as a particular network's rule.
The $900 is disputed money, not necessarily a final loss after responses or recoveries. A large disputed transaction can influence an amount-based result more than a small one, while each contributes one payment to this example's count. Stripe's dispute reference also notes that a disputed amount can differ from the original charge, including when only part of an order is disputed. Stripe: Dispute amount.
Bring your own metric definition and dated figures to NUMUS. Discuss your processing history.
Preserve the dates behind the percentage
Save the export date alongside the reporting period. Stripe explains that a rate associated with original payment dates can change as later disputes arrive. A recent period should therefore be identified as a snapshot rather than silently treated as complete. Stripe: Measuring disputes.
For your worksheet, keep an original-payment reference and date beside the dispute reference and relevant event date. Adyen's reporting documentation separates payment, record and dispute dates, and identifies their time zones. It also distinguishes chargebacks from other dispute record types. That illustrates why counting every row in an export can produce the wrong numerator. Adyen: Disputed transactions.
Ask the provider which event and identifier its metric uses before removing duplicate-looking rows. Keep the original export intact and record any calculation changes separately.
Separate measurement from outcomes
A won case and a removed metric entry are different questions. Mastercard's merchant FAQ says winning a chargeback does not improve the chargeback ratio it describes. Do not subtract won cases from your worksheet without checking the applicable definition. Mastercard: Chargeback questions, question 8.
Likewise, a percentage alone does not explain an account's reserve provisions. If that is the issue in a proposed offer, use our rolling reserve guide to examine the actual withholding and release terms.
Avoid classifying a number as “safe” using a threshold copied from a different report. Ask which current definition and account-specific requirements apply to the decision being discussed.
Questions to take to the provider
Use the worksheet to obtain answers you can reproduce:
- Which exact report and metric should we use for this review?
- What is included on each side, and which dates govern each figure?
- Are we counting transactions, dispute events or monetary amounts?
- How do exclusions, reversals and multiple records for a payment affect this metric?
- Which current documentation applies to this account, and when did it take effect?
- What explains any difference between our calculation and your reported result?
If the denominator is zero, mark the calculation undefined and ask how the provider reports that situation. Do not enter 0% simply because the spreadsheet cannot divide.
Use our application checklist to organize the accompanying records. For a business needing a more specialized review, our high-risk merchant account page explains the initial discussion. NUMUS can help you prepare that conversation; the applicable provider determines eligibility and account terms.
Discuss your processing history.
Sources are linked beside the relevant explanations. Read our editorial approach.