False decline
A false decline is a genuine customer’s application or payment declined because it is suspected of fraud. It is a false positive: a good customer is flagged as fraudulent. The cost can include lost margin on the sale or loan, and the customer, who may not return.
Example
For illustration, suppose a lender declines a genuine loan application that would have earned $120 in margin. The false decline costs $120 in margin, plus the possibility that the customer does not return.
Measuring false declines and their cost
The false decline rate is the number of false positives divided by the number of genuine applications or payments assessed. For illustration, 12 false positives among 600 genuine cases is 12 ÷ 600 = 2%. A separate ratio compares false positives with fraud caught. If those checks catch 6 fraud cases, 12 false positives per 6 caught is a ratio of 2 to 1.
A rate does not show the dollar cost on its own. Estimate lost margin from the declined sales or loans, then consider the customer who may not return.
Weighing false declines against fraud stopped
Compare the fraud loss a rule stops with the margin it gives up on good customers it declines. The comparison favours the rule when the fraud loss stopped is greater than the lost margin. The lender sets the rule using its policy and data.
A lender can send flags to manual review instead of declining them automatically. That trades analyst time for fewer false declines, so the lender weighs the review time against the reduction in declines.
Questions we get asked
Can a decline for a reason other than suspected fraud be a false decline?
No. The term refers to a genuine customer declined because they were suspected of fraud. A decline for another reason is a different outcome.
Does a flag sent for review count as a false decline?
A flag sent for manual review is not itself a false decline. If the reviewer declines a genuine customer as suspected fraud, that outcome is a false decline.
What does the break-even false-positive rate mean?
It is the false-positive rate at which the monthly cost of false declines equals the fraud stopped. It marks the balance point in that cost comparison.
Decisimo’s false-decline cost calculator compares the monthly cost of false declines with fraud stopped and shows the break-even false-positive rate.
