Fraud filters are a standard control for card-not-present merchants. Most already run some mix of address checks, velocity limits, and risk scoring. For acquirers and processors, the useful question is no longer whether a merchant has filters. It is whether those filters are set correctly.
Calibration matters because the pressure runs in one direction. Fraud reports and disputes feed monitoring ratios such as Visa’s VAMP, so a merchant near a threshold has every reason to tighten its rules. Tighter rules decline more good customers. That costs the merchant revenue, and it costs the acquirer processing volume. Acquirers have a direct stake in how their merchants tune their filters.
What Fraud Filters Can & Cannot Do
Fraud filters screen transactions for signs of criminal fraud. They look for mismatched data, unusual velocity, and risky patterns, then approve, decline, or flag the order. They are built for one problem: a stolen card used by someone other than the cardholder.
They do not address other sources of disputes. Friendly fraud, affiliate fraud, and business email compromise scams can all pass through a filter, because the transaction often looks legitimate. In friendly fraud, the buyer is the real cardholder and may be a long-time customer.
Filters also differ from authentication. When 3D Secure 2 authenticates a transaction successfully, fraud liability shifts to the issuer, but consumer disputes such as items not received are not covered. Merchants should treat filters and authentication as separate controls with separate jobs.
The Hidden Cost of False Declines
A false decline is a legitimate order rejected as fraud. These losses are easy to miss. A declined transaction never settles, so it never produces a chargeback or a TC40 report. A merchant that tracks only fraud metrics can look healthy while turning away good customers.
The scale is significant. Signifyd, citing Datos Insights, projects that false declines will cost ecommerce brands more than $231 billion globally in 2026. The same source cites the Merchant Risk Council’s 2026 report, in which 38% of merchants put their false decline rate between 2% and 5% of orders.
Chargeback spikes make the problem worse. A spike in chargebacks often triggers an overcorrection, in the form of tighter rules that create more false declines. The merchant then loses sales without fixing the cause of the disputes.
Why Filter Performance Degrades
Three patterns account for most of the problem.
The first is ruleset bloat. Merchants add a rule for each new fraud pattern and rarely remove old ones. Legacy rules-based systems grow more restrictive as rules accumulate, which raises false positives over time. A rule written for one pattern can also become obsolete when that pattern shifts, yet continue to generate false positives.
The second is uniform rules. New customers, repeat buyers, subscriptions, and high-value orders carry different risk. A rule that suits one group can be wrong for another.
The third is untested enforcement. Switching on strict enforcement without a passive learning period typically causes a spike in false positives.
How to Calibrate & Maintain Filters
Test new rules in shadow mode. Let the rule score live transactions without acting on them, then compare its calls with real outcomes before enforcing it.
Review three outcome sets on a regular schedule: flagged transactions that proved to be fraud, declines that proved to be legitimate, and approved transactions that later became chargebacks. Together they show whether a rule catches fraud or only creates friction. Merchants can estimate false declines by tracking successful retry purchases and customer complaints after declines.
Segment rules by customer type and transaction profile. A returning customer with a clean history should not face the same screening as a first-time buyer placing a large order. Retire rules that no longer catch fraud.
These reviews fit naturally into the quarterly operational review. Filter performance is one of the findings that review is meant to surface.
What Acquirers & Processors Can Do
Ask merchants for their false decline rate alongside their fraud and chargeback ratios. A merchant that cannot provide it is not measuring half of its filter performance.
Share TC40 and chargeback data. Merchants often lack visibility into which approved transactions turned out to be fraud. That information shows where rules are failing and supports better tuning.
Advise against reactive tightening. When a merchant’s ratio approaches a threshold, the first step is root cause analysis by reason code, not blanket rules. Tightening everything treats a symptom and costs sales.
Escalate when the signals conflict. Rising decline rates alongside flat or rising chargebacks suggest the filters are miscalibrated, and the merchant needs a direct review.
Fraud filters need maintenance like any other control. Acquirers and processors who help merchants keep them calibrated protect fraud outcomes and approval rates together.
