Nobody builds a control process around a thirteen-cent charge.
Plenty of banks would disagree with that, and they'd be half right.
For 2023 the US Federal Reserve reported $12.95 billion in debit network (scheme) fees, averaging 12.9 cents per transaction. Two figures from the same report, and you meet them in opposite ways. The per-transaction number lands on an invoice or in a schedule and looks immaterial by any reasonable threshold. The aggregate never lands anywhere, because summing it isn't anyone's job.
Where the disagreement is fair. Most large issuers and acquirers do have network fee controls. Invoice validation, variance review, network relationship management, someone who owns the line. That's real, and it's more than most smaller institutions have.
Where I'd push back. Those controls are almost always calibrated to the line items that look material on their own. Interchange, the headline assessments, anything that moves a basis point. There are tens of thousands of potential fees that wouldn't clear that threshold.
The process may exist. The granularity doesn't.
Which is a harder problem to see, because everything looks controlled.
Two more things about that Fed figure. The split: acquirers and merchants paid 64.9%, issuers paid 35.1%. That second number is the one I'd sit with if you're on the issuing side, because almost every public conversation about card fees is framed around merchant acceptance cost, and the regulatory attention has followed. Issuer-side network fees get a fraction of the scrutiny while carrying more than a third of the burden.
And scope. That figure covers debit card transactions, so even the most useful aggregate reporting available describes one part of the card business.
Here's the test I'd run as an issuer or acquirer. Take your ten largest network fee line items by annual value. For each one, can someone in the building name the service it pays for, the rule that triggers it, and whether changing your own behavior would reduce it?
In my experience the first three are easy and the answers thin out quickly after that. Which is fine, right up until the ones nobody owns start growing.
This sits inside a briefing I've put together on what regulators across seven markets have concluded about network fees. Every figure sourced.