Ideas on fraud and fraud prevention
What we learned running fraud operations in the real world — detection, rules, metrics, and the total cost of fraud.
The patient one: the attacker who holds back so as not to draw attention and the regular customer leave the same trace
Small, boring purchases on the same card, none that stands out on its own. What actually separates the patient one from the regular mid-range customer.
Read → Fraud types · 5 of 9Social engineering: the one buying with details the victim handed over and the customer travelling leave the same trace
A card with every detail correct, used from a new device. What actually separates the one who got the details over the phone from the customer who is travelling.
Read → Fraud types · 4 of 9Identity rings: the person running five accounts and the family sharing a phone leave the same trace
Separate accounts sharing a device, an address or a phone line. What actually separates an identity ring from the family sharing the household card.
Read → Fraud types · 3 of 9First big purchase: the attacker with a single shot and the customer who just moved leave the same trace
A thin account and a large purchase right away. What actually separates the single shot attack from the customer who spent months saving.
Read → Fraud types · 2 of 9Velocity: the attacker racing the clock and the shop owner leave the same trace
Many purchases in a few hours on the same card. What actually separates velocity abuse from the customer who just got paid.
Read → Fraud types · 1 of 9Card testing: the attacker and the customer out of money leave the same trace
Minimum amounts, repeated declines and retries. What actually separates card testing from the student and the customer who ran out of money.
Read → Fraud types · IndexNine fraud types in payments and their legitimate twin
Every fraud type has an honest customer who leaves the same trace. Nine of them, with what actually separates them.
Read → Fraud detection · RulesRules That Learn: The Shadow Mode That Kills Fraud's Blind Spot
A rule that only blocks never tells you whether it was right. How intelligent shadow mode measures the real outcome of every rule — and teaches the model along the way.
Read → The Real Cost of Fraud · 8/8For Every $1 of Fraud, You Lose Almost $4 (and You Only Measure the First)
The number you report as the cost of fraud is 25%. The real bill is 3 to 4 times larger, and almost nobody adds it up.
Read → The Real Cost of Fraud · 7/8Rules, Lists, Velocities, Models, Graphs: The Problem Isn't Choosing, It's Combining Them
There's no Swiss army knife in fraud prevention. Combining the techniques well produces the quality data the engine needs to decide.
Read → The Real Cost of Fraud · 6/8Chargebacks Are the Thermometer, Not the Disease
Chargeback rate always reaches you late. What changes when you start thinking about fraud from the attacker's side.
Read → The Real Cost of Fraud · 5/8False Positives: The Damage Your Fraud System Does to Your Best Customers
False positives are the invisible cost of poorly calibrated fraud prevention. How to measure them without sacrificing real fraud detection.
Read → The Real Cost of Fraud · 4/8Scaling the Analyst Team Doesn't Scale
75% of the fraud analyst's day happens outside the fraud product. What has to change for every team hour to land where it matters.
Read → The Real Cost of Fraud · 3/8Your Fraud Stack Has 5 Vendors and None of Them Talk to Each Other
Five fraud vendors that don't talk to each other cost you more than the fraud itself. What a truly integrated stack actually looks like.
Read → The Real Cost of Fraud · 2/8If the Attacker Uses AI, Your Static Engine Has Already Lost
Having a trained model isn't the same as having an engine that learns from the adversary at the pace the adversary learns from you.
Read → The Real Cost of Fraud · 1/8The Invisible Cost of Slow Response
The lag between detecting a fraud pattern and blocking it in production is the metric that defines how much fraud actually costs you.
Read → Open Source · DataspotBuilding Dataspot: Lessons from Real-World Fraud Detection
How obsessing over fraud patterns led to an open-source tool that finds data concentrations — and why we released it to the community.
Read → Fraud detection · GraphsGraph Contamination in Fraud Detection: The Problem Nobody Talks About
Why graph-based fraud detection ends up hurting legitimate customers, and what the industry is missing about contamination by association.
Read →No posts found for your search.