November 13, 2026
A new merchant goes live on Monday. By the following week, it may already have processed its first fraudulent payment. That is not a worst case. Fraudio's latest data study found that one in 10 newly observed merchants experiences fraud within its first 90 days of trading, and for those affected, the median time to first fraud is just 33 days.
The timing is the problem. Fraud arrives at exactly the point when a merchant has the least transaction history, so there is no settled picture of what normal looks like. This guide explains what the data shows, why judging each payment on its own misses most fraud, and how Fraudio helps acquirers, payment service providers and payment facilitators protect new merchants from their first transaction.
Fraudio's fraud detection scores transactions in real time and profiles merchants from day one, using networked data from across issuers, acquirers and payment facilitators. New merchants are benchmarked against comparable merchants in that network, so you get useful context before a merchant has built its own history. Payment companies can test it on their own historical data before committing.
1. New merchant fraud at a glance
2. What Fraudio's data found
3. Why new merchants are so exposed
4. Why a single transaction rarely gives fraud away
5. How Fraudio protects new merchants from day one
6. Five ways to reduce fraud in a merchant's first 90 days
7. How to evaluate a merchant fraud detection tool
8. Everything you need to know about new merchant fraud
9. FAQs
| Question | Short answer |
|---|---|
| How common is fraud at new merchants? | One in 10. In Fraudio's cohort, 10.5% of newly observed merchants had at least one fraud-labelled transaction within 90 days. |
| How soon does it start? | Fast. The median time to first fraud was 33 days, and a quarter of affected merchants saw it within nine days. |
| Are new merchants riskier than established ones? | Yes. The fraud rate per transaction was 2.9 times higher in the first 90 days than in programmes more than a year old. |
| Can you spot fraud from the payment alone? | Rarely. 82% of fraudulent transactions looked unremarkable on amount, currency and channel, close to the 84% of legitimate ones. |
| What does work? | Context: links between cards, merchants and earlier fraud, plus benchmarks from comparable merchants. That is what Fraudio's networked data provides. |
Fraudio analysed a global sample of more than 127.5 million transactions processed between January 2025 and April 2026. Within that sample, it identified 3,923 merchants whose first recorded transaction took place during 2025 and which had not appeared in the data before. It then tracked how many of them saw fraud, and how quickly. The full study is available on the Fraudio data-led insights page.
Of the 3,923 new merchants, 10.5% experienced at least one fraud-labelled transaction within their first 90 days. For an acquirer onboarding hundreds of merchants a month, that is dozens of new accounts facing fraud before their first quarter is out.
For merchants hit during their first 90 days, the median time from first transaction to first fraud was 33 days. A quarter encountered fraud within nine days, and three quarters within 61 days. Almost one in five saw their first fraud in their very first week, and 47% within the first 30 days.
The first week stands out. Once the figures are adjusted for the different lengths of each time band, first frauds occurred at almost three times the daily rate during week one compared with days 31 to 90. Risk does not build slowly as a merchant grows. It peaks almost immediately.
New merchants do not just pick up more incidents because they process more payments over time. Per transaction, the fraud rate during a merchant's first 90 days was 2.9 times higher than in payment programmes more than a year old. Early-stage merchants carry a disproportionate share of the risk.
Most fraud controls learn from history. They compare a payment with what a merchant usually sees: typical ticket sizes, usual currencies, normal channels, regular customers. A new merchant has none of that. In its first days, every payment is unusual and none of it is, because there is no baseline yet.
This is the cold-start problem. Fraudsters can exploit it from either side. Stolen cards can be tested on a merchant whose controls have not yet learned what normal looks like. Bad actors can also open merchant accounts to process fraud quickly, then disappear before patterns become obvious. Either way, waiting for history to build means paying for the lesson in losses.
As Fraudio's CEO and co-founder, João Moura, puts it: "Fraudsters don't wait for a business to build six months of transaction history before they begin attacking it." He adds that payment companies should not have to learn what fraud looks like by experiencing losses first. Fraud prevention needs to start with context from day one.
Fraudio also tested how much a payment reveals on its own. Each transaction was compared with that merchant's typical activity on three attributes: amount, currency and channel. On that basis, 82% of fraudulent transactions looked unremarkable. So did 84% of legitimate ones. Those attributes alone do almost nothing to separate fraud from genuine payments.
Connections tell a very different story. Across the wider dataset, 30% of fraudulent transactions involved a card already connected to an earlier fraud. For legitimate transactions, the figure was 0.075%. That is a gap of roughly 400 times.
Fraudio is clear that this is a retrospective measure. Fraud labels can be confirmed by chargebacks well after the original payment, so the figure does not show exactly what could have been known in real time. The direction is still plain. The amount may be normal, the currency may be normal and the channel may be normal. What matters is what happened before, how entities connect and whether ordinary-looking behaviour fits a wider pattern.
That is also why a new merchant's own data is not enough. A card linked to fraud at another merchant, or in another portfolio, is invisible to a system that only sees one merchant's short history.
Fraudio is a unified transaction risk platform for the payments industry. It scores transactions in real time through an API and continuously profiles merchant behaviour, so acquirers, PSPs and payment facilitators can act from a merchant's first payment rather than after its first loss.
Viva Wallet used Fraudio while growing 7x in transactions, increased its risk team's efficiency by 600% and reported an 8x return on investment. You can test the approach on your own historical data first, through a free Proof of Results engagement with no commercial obligation.
Learn more about monitoring merchants from their first transaction: www.fraudio.com/merchant-fraud-detection
The data shows first fraud clusters in week one and peaks well before day 90. Monitoring that ramps up only once a merchant reaches a certain volume, or after its first review, starts too late. Apply your closest scrutiny at go-live and keep it through the first quarter. Fraudio monitors merchants from their first transaction, with no warm-up period.
A new merchant has no history, but its category, its peers and the wider network do. Benchmarking a new merchant against comparable merchants turns an empty baseline into a useful one. Fraudio compares each merchant with its peers in a networked dataset spanning issuers, acquirers and payment facilitators.
Amount, currency and channel left 82% of fraud looking ordinary. Links between cards, merchants and earlier fraud did far better. Your scoring should see whether a card, device or counterparty has appeared in fraud elsewhere. Fraudio analyses transactions, merchants, cards and related entities together, in real time.
Slowing onboarding to cut risk costs you good merchants. A better trade is to onboard quickly and watch closely, with clear thresholds for review, reserves or holds when signals appear. Fraudio's traffic-light scoring lets risk teams set their own thresholds, so friction lands only where the risk is.
Blended portfolio fraud rates hide early-life risk. Track fraud rate, time to first fraud and chargebacks for new merchants as a separate cohort, so you can see whether controls work where the risk is highest. Fraudio's reports give risk teams consistent data on each merchant's behaviour over time.
1. Day-one coverage. Ask how the tool scores a merchant with no history. If the answer is "after a learning period", the riskiest weeks are uncovered.
2. Network data. Check whether models learn from data beyond your own portfolio, so fraud seen elsewhere can be recognised in yours.
3. Entity links. Confirm the tool connects cards, merchants and related entities, rather than judging each payment's fields alone.
4. Speed. Check response times under load, and confirm scores arrive within your authorisation window.
5. Alerts and investigation support. Look for alerts on merchant behaviour changes and reports your team can act on.
6. Test on your data. Ask the vendor to score your historical transactions, including recently onboarded merchants, and report catch rate and false positives.
| Topic | Key point |
|---|---|
| Prevalence | 10.5% of new merchants saw fraud within their first 90 days |
| Speed | Median time to first fraud of 33 days. A quarter within nine days |
| Peak | Week one, at almost three times the daily rate of days 31 to 90 |
| Relative risk | Fraud rate per transaction 2.9 times higher than in programmes over a year old |
| Isolated checks | 82% of fraud looked unremarkable on amount, currency and channel |
| Context | 30% of fraud involved a card linked to earlier fraud, against 0.075% of legitimate payments |
| Best control | Network context and merchant monitoring from the first transaction |
Fraudio gives new merchants the context they cannot have on their own. It scores every transaction in real time, profiles merchant behaviour from day one and draws on networked data to recognise fraud already seen elsewhere. It is built for acquirers, PSPs, payment facilitators and processors, and goes live in days to weeks with usage-based pricing and no setup fees.
Talk to our team about protecting your newest merchants: www.fraudio.com/contact-us
In Fraudio's analysis of 3,923 merchants first observed in 2025, 10.5% experienced at least one fraud-labelled transaction within their first 90 days. That is roughly one in 10 new merchants.
Among merchants hit in their first 90 days, the median time to first fraud was 33 days. A quarter saw fraud within nine days, almost one in five in the first week, and 47% within 30 days.
New merchants have no transaction history, so there is no baseline for normal behaviour. Fraudsters exploit that gap. Fraudio found the fraud rate per transaction was 2.9 times higher in a merchant's first 90 days than in programmes more than a year old.
Rarely, if you only look at the payment itself. Judged by amount, currency and channel against the merchant's usual activity, 82% of fraudulent transactions looked unremarkable, close to the 84% of legitimate ones. Context, such as links between cards and earlier fraud, is far more informative.
It is the gap between a merchant going live and building enough history for models to learn what normal looks like. Controls that depend on a merchant's own history are weakest in this window, which is also when fraud risk peaks. Network data and peer benchmarks close the gap.
Fraudio scores each transaction in real time and monitors merchant behaviour from the first payment. Its models use networked data from issuers, acquirers and payment facilitators, and compare each merchant with comparable peers, so a new merchant benefits from context it has not built yet. Integration takes 3 to 14 days.
Not necessarily. Slower onboarding costs good merchants. Fast onboarding paired with close monitoring through the first 90 days, and clear thresholds for review or holds, protects revenue and limits losses.
Fraudio offers a free Proof of Results engagement with no commercial obligation. It scores your historical transactions, including recently onboarded merchants, so you can compare results with your current setup.
Fraudio analysed a global sample of more than 127.5 million transactions between January 2025 and April 2026. The new merchant cohort comprised 3,923 businesses first observed during 2025 that had not appeared in the data before. The analysis uses fraud confirmed in Fraudio's transaction data.
How about trying our solution and experiencing the next generation for yourself?