Nobody is fighting hardest at the front door

Socure president Matt Thompson, writing on Liminal’s Friday Five, makes a plain argument that lands harder than its bullet-point form: new account fraud is the origin point of nearly every downstream loss, and the industry still underdefends it. “Every fraud loss has an origin story, and most start at the same place: the moment an account is opened.” His disappointment is specific — he expected buyers to fight for the last basis point of fraud capture at onboarding, and mostly they don’t. Understaffed teams and competing priorities leave the threshold underguarded while the expensive cleanup happens months later.

The economics he lays out are the whole case. An account caught at creation costs almost nothing; the same account caught six months on carries a transaction history, a credit line, maybe a role in a mule network, and a remediation bill many multiples larger. Thompson cites Socure research finding fraud rings that pushed nearly 25,000 synthetic identities into more than 35,000 applications, often within 48 hours of an identity being created, against a backdrop of roughly $16 billion in reported U.S. consumer fraud losses in 2025 — up about 25% in a year. The vendor incentive is worth naming: Socure sells identity decisioning at exactly this front door, and “buy better onboarding” is the house recommendation. But the underlying asymmetry is real and independent of who’s making the pitch.

The useful prescription is the buying discipline, not the alarm. Thompson’s push for a proof of value — score each vendor against a retrospective sample of your own account-opening decisions, including the fraud you missed and the good applicants you declined — is the rare vendor essay that tells buyers to test rather than believe the marketing. A basis point of capture with a dollar figure attached, he notes, stops being a technical metric and becomes a budget argument. That’s the sentence worth keeping.