If you run a calendar-year P&L, 2027 planning likely starts about now. Before the templates start being filled, look at what the first half of 2026 actually did to card economics, because that trajectory is your planning input, not last year's run rate.
Here's what changed, and what I cover below is only a slice of it.
The pace picked up. More changes, landing more often, more of them off-cycle. Earlier this year I wrote about four mechanics in Mastercard's Spring 2026 release alone, and that was a fraction of the full release. Pick any month this half and the same gap holds: what made the headlines was a small share of what actually hit the schedules.
More of them attach to behavior, not volume. Force posts, chip fallback, excessive cross-border retries, stale preauths. You pay for what your operation does, not just what it sells, and much of it bills whether or not the sale clears.
More of them ride in on compliance. TLID, specialty merchant registration, BIN sponsorship rule shifts. The obligation and the fee arrive together, with a hard effective date attached.
And the quiet ones cost the most. Repricing of existing line items, rebundling, scope creep. None of it gets a headline, which is exactly why it lands in the variance report unexplained, and why the changes you read about are never the whole bill.
Over the top of all of it, the US interchange settlement received preliminary approval, and regulators in the UK and Australia moved the same direction: more disclosure, more pressure on caps, less room.
So when you build the 2027 number, three things:
Don't budget network fees flat to volume. H1 showed per-unit cost rising through scope, price increase, and behavior charges, most of it never announced loudly. Plan fee growth above volume growth, and write the assumption down so it survives the review.
Plan a scenario, not a number. If you issue in the US, the interchange settlement is a long term compression risk to model, not footnote. In other markets it's the regulators setting the direction, Australia just lowered its caps, the UK is moving toward it. Wherever you sit, the fee base is moving faster than a flat assumption can hold.
And budget the capability, not just the cost. Keeping up with this manually is a headcount problem now. The 2027 question is whether you staff it, tool it, or keep absorbing the misses.
If your network (scheme) fee line still surprises you at month end, you don't have a fee model. You have an invoice. Budget season is when that gets expensive, or gets fixed.
If reducing network fee cost, optimizing card profitability, and better managing network compliance is on your 2027 list and you want tools and a second set of eyes on it, let's talk.