The Acquisition Trap is Real

This is the first in a series of articles derived from “Meeting the Growth Challenge Requires Precision Over Volume,” a white paper produced by Curinos on why the old playbook for retail banking is failing, what the data actually shows, and how the banks winning in 2026 are making decisions differently.

Banks have never spent more to grow. The cost to acquire a checking customer has doubled since 2018, to $559. Institutions have leaned into digital, stood up always-on acquisition engines, and flooded the market with rate offers. By almost every measure, the industry is trying harder than ever.

But the results tell a different story. In 2025, only one in ten institutions achieved 3% growth for both customers and deposits; half saw no growth at all to either or both. More money, more channels, more offers — and, for most, no more growth.

This isn’t a spending problem or an effort problem. The trap is more subtle: most banks optimize acquisition in isolation and grade it on the wrong number. Cost per acquisition is easy to calculate and easy to defend, which is why it has become the default scorecard. But it measures only the cost at the front door. It says nothing about whether the customer funds the account, holds a balance, or quietly leaves within the year.

And the front door has been moving. Digital channels reached 40% of new relationships in 2025 and will pass 55% by 2027. Yet after twelve months just 41% of digitally acquired relationships remain on the books, compared with three in four opened at the branch — and replacing a lost customer costs about four times more than keeping one. The cheaper channel is producing thinner relationships, and CPA metrics can’t see it.

None of this argues against acquiring, or against digital; both are essential. It argues against the scorecard. But if CPA is the wrong number, what’s the right one? The rest of the articles in this series gets at the answer.