The Fraud Losses Hiding in Plain Sight: What "Customer Error" Really Costs
A customer mistake or a scam? Getting that distinction wrong doesn’t just skew your fraud data—it distorts risk, sends budgets in the wrong direction, and undermines your ROI calculations.
Many financial institutions have a fraud problem they don't count as fraud. Instead, the losses often end up in a different bucket: customer error.
It's not hard to see why. In many of these cases, the customer logged in themselves, passed authentication, and approved the payment. From a transaction perspective, everything appears legitimate. So when the customer challenges the payment, the “error” classification can seem reasonable.
But that's just on paper. In reality, there is a world of difference between a customer accidentally sending money to the wrong account and a fraudster manipulating them into sending it to the right one.
When a Scam Looks Like a Mistake
A genuine customer error is exactly that: a mistake. A mistyped account number, an accidental transfer, or a misunderstanding that leads to money being sent where it shouldn't. Nobody influenced the decision. Nobody engineered the outcome. The customer simply got it wrong.
An authorized push payment (APP) scam is something entirely different. The payment may be authorized, but the decision behind it has been engineered. The customer is pressured or manipulated into sending money to a criminal under false pretenses. The transaction looks legitimate. The circumstances that led to it are anything but.
The problem is that authorization tells us how the payment was made, not why. Yet when both losses end up in the same reporting bucket, that crucial distinction disappears—and the fraud case quietly slips from view.
Small Classification Error, Big Consequences
On the surface, misclassifying a scam as customer error looks like a paperwork problem. The money is gone either way, right? But in fraud prevention, correct labels mean correct understanding—and correct decision-making. When scam losses are recorded as customer errors, banks may be systematically understating their true fraud exposure.
1. Fraud Disappears from the Numbers
Having a clear picture of the fraud targeting your institution is essential. By misclassifying scams as customer error, fraud cases vanish from reporting and distort the institution's view of its risk landscape. Fraud losses may appear stable or even declining, while the reality looks very different. That only compounds the visibility problem already created by underreporting.
The result is a blind spot. The institution still suffers the loss, but the loss is no longer reflected in its fraud metrics. Over time, that makes it harder to understand the true scale of scam-related fraud, as well as the tactics and techniques driving it.
2. Prevention Becomes Harder to Justify
You can't manage, budget for, or report on fraud you've reclassified out of existence. When scam losses are hidden behind a “customer error” label, the business case for additional controls, analysts, and fraud prevention technology appears weaker than it really is.
The irony is that these are often the very investments needed to address the problem. By making fraud appear smaller on paper, misclassification can quietly deprive prevention teams of the resources required to stop growing losses.
3. Institutions Learn the Wrong Lessons
When a scam is treated as customer error, the root cause becomes a customer mistake rather than criminal manipulation. The response then shifts accordingly: more payment warnings and better user experience design.
While those measures have their place, they don't address the heart of the problem. If the loss was caused by social engineering, the priority should be understanding how that manipulation happened. Otherwise, institutions risk treating a criminal problem as a customer problem.
The Stakes Are Rising
This classification challenge is becoming more serious because fraud itself is changing. Historically, many fraud cases involved unauthorized activity: stolen cards, compromised credentials, or account takeover. Today, the customer is the one pressing “send.” And the losses are climbing fast: in the UK, APP fraud losses jumped by almost 20% year over year in 2025.
That creates a problem for classification frameworks built around authorization. The customer logged in, passed authentication, and approved the transaction. Yet the decision behind the payment may have been manipulated. As scams become more sophisticated, the gap between what happened and what traditional fraud signals can explain continues to grow.
At the same time, regulators are paying closer attention to APP fraud and how institutions manage it. As reimbursement requirements and scam-related obligations expand, accurately distinguishing customer mistakes from customer manipulation becomes more important than ever.
Finding the Fraud Behind the “Error”
Making hidden fraud visible starts with understanding what traditional investigations often miss. A transaction record can reveal who sent the money, when it happened, and where it went. But those details alone rarely explain what led the customer to make the payment in the first place.
The missing piece is context.
If authorization alone can't distinguish a genuine payment from a manipulated one, fraud teams need additional evidence. Was the customer on a live call while making the transaction? Was their behavior consistent with how they normally bank? Were there other unusual patterns that suggest external influence? These are exactly the kinds of signals that can help separate a genuine customer error from a scam in progress.
ThreatMark provides investigators with evidence that helps explain the circumstances behind a transaction. This evidence can include signs that the customer was engaged in a phone conversation during the payment process, indicators of remote access or coaching, unusual behavioral patterns, and transaction context that differs from the customer's normal activity.
Together, these signals help turn a seemingly legitimate transaction into a more complete picture of what actually happened.
Why Accurate Classification Matters
Banks don't need to treat every authorized payment as suspicious. They need better evidence to distinguish a customer's mistake from a scammer's success.
That distinction matters for far more than classification. It shapes how institutions measure fraud, where they invest their prevention budgets, how they demonstrate ROI, and how prepared they are for a regulatory landscape that increasingly treats APP scams as a distinct and measurable risk.
In this landscape, accurate classification isn't just cleaner reporting. It's the foundation for understanding the true cost of scams.
So how much of your “customer error” is actually fraud? Find out.
