Right idea, wrong moment: fraud validation in the purchase funnel

The strategic question

A business unit came in with a clear, reasonable-sounding request: run fraud validation every time a user selected a payment method. The instinct behind it was sound — protect the business from fraudulent orders. The question I had to answer wasn't whether fraud validation belonged in the funnel. It was: where, exactly, does it belong?

The constraint

Payment-method selection happens early and often. A large share of users who tap a payment method never complete a purchase — they're comparing options, checking eligibility, or just browsing. Running a fraud check at that moment meant two costs stacking on top of each other: added friction and drop-off for people who were never going to buy, and a paid fraud-check API call fired for every one of them, buyer or not. The request was solving for risk while quietly generating cost and friction against a mostly-non-buying population.

The decision I owned

I reframed the request. Instead of accepting "add a risk control here" at face value, I pushed the conversation to a more specific question: when does the customer actually signal intent to purchase? Payment-method selection isn't that moment — confirmed purchase intent is. I moved the fraud validation trigger to that later point in the funnel.

This meant saying no to the request as originally scoped, while saying yes to its underlying goal. The business unit kept full fraud protection — nothing about their risk coverage was reduced. What changed was precision: the check now fired for people who were actually buying, not for everyone who touched a payment option on the way to deciding.

The outcome

The original design would have triggered a paid fraud check on every payment-method selection; the final design triggered only at confirmed purchase intent, cutting call volume by construction.

What I'd tell another product leader

When a request arrives framed as "we need a control here," treat the "here" as the part still open for debate — the underlying need rarely is. Most of these conversations aren't about whether to say yes. They're about finding the point in the funnel where the control does its job without taxing everyone who was never going to convert.