An ISO sent me its top-10 most profitable merchant list last week. After we layered in fully loaded card network fees, half of them no longer belonged in the top 10.
Most acquirer and ISO teams have a solid handle on interchange. Far fewer model the rest of the card-brand cost stack at the merchant level. That gap is where profitability analysis goes wrong.
Interchange is the big visible line item, but its effective rate moves materially with debit-regulation status, card product, ticket size, channel, and qualification. Regulated debit at $0.21 plus 5 basis points behaves nothing like a rewards, commercial, keyed, or downgraded credit transaction. It's also not the full cost stack.
What else is in the network fee stack beyond interchange?
Here's what else lands against the book.
Core volume assessments. The Visa assessment (Acquirer Service Fee) and Mastercard's Acquirer Brand Volume Fee. Roughly 13 to 14 basis points on settled volume, and the first thing interchange-only models drop.
Per-authorization and retry fees. Auth submission, network access, excessive reattempt, and integrity fees tied to decline-and-retry behavior. Scale with attempts, not approved volume.
Per-clearing and per-transaction fees. Network base, clearing, settlement, and force-post or unmatched-clearing fees. Flat per item, so low-ticket merchants get hit hardest.
Digital commerce and ad valorem fees. Mastercard DEF and Visa DCSF. Visa's DCSF moved April 1, 2026 and now bundles separate token, updater, credential, and enrichment fees. Both carry minimums, and DEF caps at $0.50 above $2,000. Low-ticket CNP hits the floors. DCSF is uncapped, so that's where high-ticket exposure lives.
Fixed and registration fees. FANF, acquirer license fees, and specialty merchant registration. Mastercard's Specialty Merchant Registration Program doubled the annual fee from $500 to $1,000 effective May 1, with transaction-based charges starting June 3.
Cross-border and settlement-currency fees. ISA, IAF, Mastercard cross-border assessments, and APSF. Material on any book with meaningful international volume.
Behavior-based and compliance fees. Fallback Avoidance, force-post events, dispute fees, and Transaction Processing Excellence fees. Small per event, large in aggregate when the behavior isn't fixed.
In CNP-heavy, cross-border exposed, high-ticket, retry-prone, or specialty portfolios, the non-interchange stack, base assessments included, can add roughly 30 to 80 basis points. On a $500M acquirer book with that profile, that's $1.5M to $4M annually invisible to interchange-only models.
The build is straightforward: tag each fee category on the network invoice, map it to its driver, allocate by merchant, then rerank on network-cost-adjusted contribution. That isn't full profitability, you still need revenue and loss data, but it's the layer most models are missing. The merchants hitting unfavorably on multiple drivers are the ones quietly eating your residual.
Happy to share what a full network-cost profitability model looks like.
Related reading: the Mastercard specialty merchant registration fee increase, and why card network fee pass-through is harder than it seems for BIN sponsors, acquirers, and ISOs.