E-commerce
October 28, 2026
Are you wondering what to really keep in mind about online payment fraud? The essence boils down to one sentence: in most cases, the merchant refunds the victim, loses the merchandise, and pays the chargeback fees related to the credit card.
The subject deserves an update, as fraud is shifting. Chip cards have reduced counterfeiting in stores, and part of the attacks has shifted to the web, where no physical card is needed. The classic trap: hardening prevention so much that you reject perfectly legitimate orders.
So what essential reminders to keep about online payment fraud? On the agenda:
How is online payment fraud precisely defined?
Why is fraud increasing despite protections?
What signals should alert without blocking your customers?
How does Qstomy help handle payment incidents?
What checklist to apply before each activity spike?
Let's go.
Summary
How is online payment fraud precisely defined?
Online payment fraud encompasses several realities: the use of stolen or falsified payment information to place an order, and the identity theft that often accompanies it. The fraudster is not trying to buy: they are trying to get the merchandise.
Once the order is shipped, they have it delivered to an intermediary, a freight forwarder, or pick it up directly at a doorstep. The package disappears, along with any simple trail to follow.
These crimes have multiple victims. First, the person whose card or information was stolen. They are reimbursed in most cases. Then the merchant who processed the fraudulent order, as they fund this reimbursement. They lose the merchandise, return the amount, and pay the associated chargeback fees.
Understanding this chain changes how we view prevention: your goal is not just to block, but to not pay for someone else.

Convert over 2,000 customers on average per month with Qstomy.
The world’s 1st Shopify AI dedicated to customer conversion



Empowering 200+ e-commerce merchants
Why is fraud increasing despite protections?
Two explanations stand out. The first relates to the chip card. Today, a large majority of cards are equipped with chips and almost all major retailers have changed their in-store payment systems. Good news: counterfeit card fraud has significantly declined since the adoption of these cards. Bad news: criminals have shifted their activity to the web, where no physical card is required.
The second explanation relates to the availability of stolen personal data. In the first quarter of 2018, reported data breaches remained numerous and represented more than one billion exposed records in just three months.
With so much information circulating on the black market, the raw material for fraud is not lacking. The volume of identity theft victims has reached a record level, and attacks on online commerce have increased by more than 30 percent compared to the previous year.
In other words, the problem is not a lack of vigilance on your part: the playing field has shifted.
How much does fraud actually cost a store?
The cost is not limited to the order amount. You must add three components: the shipped and lost merchandise, the reimbursement to the person whose information was used, and the chargeback fees levied by the card network.
At the store level, this translates to an expected annual loss expressed as a percentage of sales. Benchmarks published in the reference report on online fraud place this loss around 0.8 to 0.9 percent of e-commerce revenue each year.
This figure may seem small, but it adds to often tight margins. On an activity of several millions, the bill quickly becomes significant, not to mention the time spent handling disputes.
There are also indirect victims. A person responding to a work-from-home ad may find themselves questioned for having transmitted stolen goods. Your brand name, however, appears in the story.
Why does mobile fraud deserve separate attention?
Mobile commerce is growing. In surveys conducted with online merchants, a notable share already reported generating more than half of their e-commerce revenue from orders placed on a phone. Projections went further: within a few years, one in five online retailers could generate nearly a third of their sales on mobile.
The key point is not the volume, but the difference in nature. A mobile buyer is more likely to use a mobile wallet, an alternative payment method, or a cryptocurrency. The observable technical signals are therefore not the same as on a computer.
Hence a simple recommendation: track fraud by channel. Separate mobile indicators from desktop ones.
You will then know whether prevention must be identical everywhere or if mobile requires its own profiles. Perhaps nothing differs. Perhaps everything differs. Only measurement will tell.
What signals should alert without blocking your customers?
The natural reflex is to refuse any order with a different shipping or billing address. This is also the worst calculation: legitimate buyers regularly ship a gift elsewhere or pay with a card registered to a different address.
Instead, look for a combination of signals. An unusual cart, a rush on delivery, a shipping address resembling a transit point, a first purchase of a high amount: it is the accumulation that matters, not the isolated element.
The true indicator of health is the false positive rate. A store that massively rejects orders out of fear of fraud loses real sales. Industry surveys showed that on average a merchant rejected about 2.6 percent of incoming orders, and even more for orders over 100 dollars.
On a volume of 25 million, this represents more than 600,000 rejected orders each year. The issue is therefore not to block more, but to block just right.
How to track fraud by channel without overburdening your daily routine?
Start small. For each disputed order, note three pieces of information: the channel used, the payment method, and the reason for the dispute. After a few weeks, profiles emerge.
Next, distinguish the indicators. A mobile dispute rate and a desktop dispute rate do not have the same meaning, because payment methods differ. Mixing them is like drowning the signal in noise.
Then measure what you reject. A prevention tool that blocks a lot is not necessarily effective. Compare the amount of fraud prevented and the amount of legitimate orders lost.
Finally, keep a record of decisions. If a rejected order turns out to be legitimate, you must be able to understand why you blocked it. This is how you gradually adjust criteria, instead of stacking contradictory rules.
Should you automate prevention or keep control?
Automated prevention tools exist and are useful for volume. They compare, score, and flag. Their known limitation is that overly zealous prevention is costly, and this cost never appears on the subscription invoice.
The reasonable solution is hybrid. The tool processes the flow and flags suspicious cases. You retain the decision on ambiguous situations, those where a quick contact with the buyer is enough to clear the doubt.
This contact changes everything. A simple question about the order, asked before shipment, often makes the fraudster abandon the attempt and reassures the honest customer. You transform a blunt refusal into a relationship.
The right balance is therefore neither fully automatic nor fully manual, but a threshold you adjust. Your goal remains twofold: reduce fraudulent transactions and reduce false positives, without having to choose between the two.
What to do when a customer disputes after an unusual payment?
Some situations look like fraud but are actually simple technical glitches. The classic case: the payment is captured, but no order appears in your back-office. The customer, however, sees the debit and is worried.
Handle these cases separately, with a clear message. Confirm that the amount has been charged, explain that you are checking the system side, and give a specific timeframe. A quick explanation avoids a chargeback, which costs far more than a message exchange.
Same logic for an order pending payment, or for a payment method temporarily unavailable. The customer does not want technical jargon; they want to know if their order exists.
These responses should be prepared in advance. A standard, short, clear response saves time for the whole team and reassures the buyer at the moment they doubt most. To go deeper, see how to manage captured payments without an order created.
How to differentiate a fraudster from a clumsy customer?
Not all incidents are fraud. A customer may enter the wrong number, use a gift card combined with a card payment, or pay with a local method you are less familiar with. These cases deserve dedicated handling, not automatic suspicion.
The right reflex is temporal. A fraudster wants to go fast: they make multiple follow-ups, insist on immediate shipping, and refuse any verification. A clumsy customer, on the other hand, responds willingly and corrects mistakes.
Also look at the history. A buyer who has been ordering for months with reasonable variations is not a stranger. A first high-value order shipped to an atypical address requires verification.
Finally, document your exchanges. If you need to contest a chargeback, your messages are your best evidence. This traceability protects both your margin and your reputation.
Which mistakes harden prevention at the expense of sales?
The first mistake is confusing caution with systematic distrust. Adding a refusal rule for every incident eventually blocks loyal buyers, and these customers do not return.
The second mistake is failing to measure the impact. Prevention that rejects orders without quantifying the loss in revenue assumes it is free, when in fact it is very costly.
The third mistake is ignoring customer service. An unhandled chargeback becomes a banking dispute, with fees and delays. Conversely, a quick response defuses most situations.
The fourth mistake is reasoning in silos, forgetting the multichannel dimension. A customer who orders on mobile then asks for advice in-store does not fit into narrow categories.
Finally, assume everything is resolved after hardening. Fraud evolves, and your prevention must keep up. To learn more, consult our reminders on the commandes en attente de paiement.
How does Qstomy help manage payment incidents?
Qstomy is an AI agent for Shopify that guides the buyer toward purchase and supports them after the order. For payment incidents, it acts as a primary filter: reassure quickly, explain, and prevent doubt from becoming a chargeback.
Concretely, it answers questions on a moyen de paiement indisponible by proposing an alternative without losing the sale, and it handles questions about the moyens de paiement locaux that your customers are less familiar with.
It also manages awkward cases: gift cards combined with a card payment, orders pending payment, payments captured without an order created. The customer receives a consistent response without waiting for a human to be available.
For the merchant, this reduces repetitive exchanges and secures conversion at the most fragile point of the journey. Over a hundred merchants already use this approach. The e-commerce chatbot use cases cover both pre- and post-purchase.
What checklist to apply before each activity spike?
Before a busy period, check four points. First, review your refusal rules: an old rule still active may block legitimate orders today.
Next, prepare your standard responses for payment incidents: amount charged, missing order, payment method failure.
Then track your metrics by channel, separating mobile and desktop, including chargeback amounts and rejected order volumes. Finally, ensure customer service has a fast response channel.
In short
The merchant absorbs most of the fraud cost. Blocking too much is as costly as fraud. Measure, talk to your customers, and adjust.
Frequently Asked Questions
Is a different delivery address a sign of fraud? No, not on its own. Sending a gift elsewhere remains common.
Should all suspicious orders be rejected? No. Mass rejections destroy your legitimate sales. Verify via contact.
Does mobile change prevention? Not necessarily, but track channels separately to know. To go further, see the uses of an AI agent, integrated customer service in SEO and the platform choice.

Enzo
October 28, 2026


