The Chargeback Ratio Is the Wrong Metric

Optimising for chargeback ratio rewards refunding winners and fighting losers.

The chargeback ratio, disputes divided by approved transactions, has become the default health metric for online merchants. For any business below 10,000 transactions a month it is also unreliable. The ratio is volatile at low volume: a single bad week can push a steady business past Visa's 1.5% Excessive threshold without anything having changed about the underlying portfolio, while a quiet week can mask drift that matters more than the headline number does.

Over-trusting the ratio creates two losing patterns. A finance team that treats the ratio as the metric will refund defensible cases to keep the headline number down, surrendering recovery the rebuttal would have produced. The same team will burn expert hours fighting unwinnable cases on thin records, because losing them would worsen the ratio further. Both patterns sacrifice margin without improving the underlying business.

The better metric is per-case expected value, calculated case by case: the probability of winning the dispute, multiplied by the recovery if the merchant does win, minus the expert hours required to fight. Cases with negative expected value get refunded, cases with positive expected value get fought, and the merchant tracks the realized win rate by reason code, by issuer, and by disputed-amount band so the next case is priced more accurately than the last.

Visa's monitoring programs do care about the ratio, and crossing the 1.5% Excessive threshold has consequences: $8 per-event fees, an acquirer reaching for rolling reserves, and a remediation plan that surfaces operational gaps the merchant had been quietly absorbing. Below the threshold, the ratio is a leading indicator at best and a distortion at worst. The realized win rate per reason code is the metric that decides whether the dispute portfolio runs at a profit.