Synthetic identity: the customer who never existed and the genuinely new customer leave the same trace

How it attacks

The person buying from you does not exist. The document is real in shape, the name sits well with it, the phone answers, the email has years of manufactured age. None of it points at a person.

They sign up like any new customer and do what a careful new customer does. Buy little. Pay on time. Come back and buy a bit more. Give them a credit line and they use it below the cap and clear it before the date.

What they are after is not today's purchase. They want every on-time payment to nudge you into giving them more room, and that room to grow on its own, because your own rules reward whoever pays well.

When the line reaches where they wanted it, they use all of it in one day. Large purchases, one after another, until nothing is left available.

And that is the end of it. There is nobody to collect from, because the holder was never there. The address leads to no one, the phone stops answering, and the collections file goes round and round against a person who was never born.

Who looks like it and is honest

The genuinely new customer. Signed up recently, started small because they do not know you yet, and as it works out for them they buy more. They pay on time because they want their limit raised.

The one rebuilding. They had a bad year, ended up with no usable history and are starting over. They are careful on purpose: they pay before the date because they cannot afford another stumble.

The young person with their first financial product. They have no past to look at, and everything they do is a first. Small purchases, slow growth, a thin file.

The same thing happens to all three: you judge them on very little past, and their good behaviour is exactly what asks you to trust them more.

What actually separates them

• Human mistakes. A real customer gets things wrong: mistypes their document number, writes the address two different ways, misses a payment and pays it the next day, calls to ask something obvious.
  • The new customer: leaves that noise in your file without noticing.
  • The synthetic identity: comes out spotless from the very first form. The absence of mistakes is itself a signal.
• How old the details are relative to each other.
  • The new customer: their phone, their email and their document have different, untidy ages, because each one was got when it was needed.
  • The synthetic identity: their details appear in the world at roughly the same time, shortly before signing up with you.
• Who they share things with.
  • The new customer: shares a device or an address with their family, and those people behave like family.
  • The synthetic identity: shares pieces of itself with other files that also started recently and also behave well.
• What they use the credit line for.
  • The new customer: uses it unevenly, depending on what they need that month, and sometimes leaves it untouched.
  • The synthetic identity: uses it evenly and always below the cap, like someone managing a score.
• How they pay.
  • The new customer: pays when they can, sometimes right on the day, sometimes the day before, and the amount varies.
  • The synthetic identity: pays with the same lead time every time, as if it were a scheduled job.
• What they do on the day they use it all.
  • The new customer: when they finally spend big, it is for something specific and in one go.
  • The synthetic identity: drains what is available across several purchases in a row, and their last move is the largest the line allowed.

How it mutates once you detect it

1. You demand more proof at sign-up. They bring it. The document survives the check because it was built to survive it, and now your file has it marked as verified.
2. You delay the limit increases. They stretch the good phase. They keep paying on time for as long as it takes, because waiting costs little next to what they take at the end.
3. You start to suspect the file without a single stumble. They add mistakes on purpose: pay a day late, abandon a cart, write their own address wrong once.
4. You detect that several files share details. They separate the identities: different devices, different addresses, and they grow them at different times so they do not look like one litter.
5. You watch for the sudden use of the whole line. They split the final day into two or three, and leave a remainder unused so the account never touches the ceiling.

At the end of that staircase you are left with a file your system built alongside you. Every limit you raised was a reasonable decision given the information you had. What that information was missing was a person.

Closing

The synthetic identity feeds on the part of your system that works: rewarding whoever behaves well. By the day it spends everything it is already too late. The moment to see it is in the quiet months before, when the only odd thing was that it behaved too well. That comes down to whether you can look at one file against the other files, or only against its own history.

So how do you solve it?

Tuning this by hand takes days, and every day costs chargebacks and good sales. That is what we are solving at Frauddi. We will show you on your own data.

Book a free demo

Next: the bad-faith dispute, the real customer who buys, receives and weeks later does not recognise the charge, and the parent disputing in good faith what their child bought (9 of 9).