0.7% now, $480,000 in five years

Say we run $2M a year through a payment processor at 2.9% plus thirty cents per transaction. We can get 2.2% from a competitor with the same feature set and a higher dispute reserve. Our gap is 0.7%.

Year one, our gap costs $14,000. We priced the switch, decided $14,000 didn't clear the migration, and renewed the contract.

Year two, our volume grows to $3.4M and the gap widens to $24,000. Same math, same answer. Our cumulative cost is $38,000, and the switch is quoted at $25,000 of engineering, so it still loses on any single year.

Year five, our volume hits $9.8M. Our annual premium is $69,000. Our cumulative bill across the five years, with growth, is $480,000. When we finally migrate, it takes two engineers two days. We had booked $25,000 for it. Fiction.

This is how percentage fees compound, and most companies never run the multi-year version. The fee is small in any one moment. It grows with revenue. Over the horizon a company operates, the total dwarfs what any single-year review shows, which is why it stays buried on a spreadsheet nobody re-opens. The renewal gets decided annually, against a number that always looks like noise.

The vendors that behave this way share a shape: percentage pricing, bills tied to revenue or volume, switching costs that feel high and rarely are. Payment processors. Cloud infrastructure. Per-seat support tools. Revenue-based insurance premiums. Each is small at the unit and heavy in aggregate.

What breaks the pattern is an annual benchmark memo: current cost, alternative cost, switching cost, five-year cumulative gap. One person, two days per vendor. The memo drags the decision onto the five-year horizon, where it looks like a different decision.

The annual saving was never the point. Renew a percentage fee eleven times without re-pricing it and you have quietly written a check the size of an engineer's salary, or a chunk of a funding round's runway, to skip two days of work you already know how to do.