The effective rate moved from 2.31 percent to 2.44 percent. It is on the board pack, it is thirteen basis points, and on the volume you run it is real money.
Someone asks what changed. The payments lead pulls the provider statements, the finance analyst pulls the ledger, and after two days the answer that comes back is that the rate went up.
That is not a failure of diligence. It is the number refusing to answer a question it was never able to answer.
TL;DR
- An effective rate is total cost divided by total volume. It is an average standing in for millions of separately priced events.
- Averages cannot be attributed. A rate can rise with nothing wrong in your pricing, and stay flat while something is very wrong.
- Federal Reserve data proves how wide the natural variance is: average interchange on one network's dual-message debit was 1.50 percent from exempt issuers and 0.44 percent from covered issuers in 2024.
- That is roughly 3.4 times, on the same network, decided by which bank issued your customer's card. You do not choose it and cannot see it in advance.
- So a rate movement is a question, never an answer. Attribution needs per-transaction comparison against the contract.
- Of the three layers in the rate, exactly one is governed by your agreement. That is the only one a dispute can reach.
Short answer
An effective rate is total processing cost divided by total processed volume. It is a blend, and a blend cannot tell you what moved it. Federal Reserve data for 2024 shows average interchange on one network's dual-message debit running at 1.50 percent of transaction value from exempt issuers and 0.44 percent from covered issuers, roughly a 3.4 times spread on identical rails, determined solely by which bank issued the card. A shift in which banks your customers happen to use will move your effective rate with no change to your pricing whatsoever. The same arithmetic works in reverse: a genuine increase in provider markup can sit inside a favourable mix shift and never surface. Attribution requires comparing each transaction against what your contract says it should have cost, then assigning the variance to a layer. Of the three layers, only provider markup is governed by your agreement, and only that layer is disputable.
What an effective rate is, and what it is for
Total cost, divided by total volume, expressed as a percentage. Nothing more.
It is a genuinely useful number. It belongs in a budget, a board pack and a unit-economics model, and it is the right figure for answering how much processing costs the business this quarter. Every payment business should know it.
It is the wrong instrument for the question people actually ask of it. "Why did it move" is a question about composition, and composition is exactly what an average destroys. Each transaction in that blend carried its own interchange, set by a schedule you do not control, its own network assessment, and its own markup under your contract. The average is one number standing in for millions of separately priced events, and once it is computed the information that would explain it is gone.
Averaging is not a reporting choice here. It is the step that removes the evidence.
The number that shows how wide the variance already is
Interchange is the largest layer in most card rates, and it is not one price. How much it varies is measurable, because the regulator publishes its own collection.
Under Regulation II, the Federal Reserve reports that a covered issuer "may not receive, for any electronic debit transaction, an interchange fee that exceeds $0.21 plus 0.05 percent multiplied by the value of the transaction, plus a $0.01 fraud-prevention adjustment, if eligible." The regulation itself sets the ceiling as the sum of "21 cents" and "5 basis points multiplied by the value of the transaction."
Two things about that cap matter more than the cap. It applies to covered issuers, meaning larger banks, and it applies to US debit. Issuers outside it are exempt, and exempt interchange is uncapped.
The Federal Reserve's 2024 figures show what that produces in practice:
| Network | Exempt issuers | Covered issuers |
|---|---|---|
| Visa dual-message debit | $0.62 average, 1.50% of value | $0.22 average, 0.44% of value |
| Mastercard debit | $0.63 average, 1.30% of value | $0.24 average, 0.49% of value |
| Discover debit | $0.83 average, 1.22% of value | $0.23 average, 0.72% of value |
On Visa dual-message debit, average interchange was 1.50 percent from exempt issuers and 0.44 percent from covered issuers. Roughly 3.4 times, same network, same card type, same rails. The only variable is which bank your customer banks with.
You do not choose that. You cannot see it before the transaction. And it is legitimate: nobody has overcharged anyone when a customer pays with a card from a smaller bank.
Two ways the rate misleads you
Once the variance is that wide, the effective rate fails in both directions, and the second failure is the expensive one.
It rises when nothing is wrong. Your customer base shifts a few points toward cards from exempt issuers. Interchange is pass-through, so your cost rises and your effective rate rises with it. Your contract has not changed, your provider has done nothing, and there is nothing to dispute. Two days of investigation confirm that the rate went up.
It stays flat when something is wrong. Your provider's markup drifts upward, through a tier that stopped applying or a rate that no longer matches the current contract version. In the same period your issuer mix moves toward covered issuers, and interchange falls. The two movements cancel. Your effective rate is unchanged, your reporting is calm, and the overcharge accrues every day inside a number that looks correct.
The first case costs you an investigation. The second costs you money for as long as it goes unnoticed, which, in an unattributed blend, is indefinitely.
The one layer your contract can reach
Not every layer is arguable, and knowing which is which decides where effort is worth spending.
| Layer | Set by | Paid to | Governed by your contract | Can you dispute it |
|---|---|---|---|---|
| Interchange | The card network's published schedule | The card-issuing bank | No | No. But you can verify it was applied correctly. |
| Network assessment | The card network | The card network | No | No. Uniform by design. |
| Provider markup | Your agreement with your provider | Your provider | Yes | Yes. This is the layer a dispute reaches. |
Interchange and network assessments are not negotiations. Raising them with your provider is raising something it also does not set. What you can do is check that the interchange applied to a transaction matches the schedule for that card, that corridor and that transaction type, which is verification rather than dispute, and it is where a great deal of quiet leakage lives.
Markup is different in kind. It is the layer your agreement prices, which makes it the layer your agreement can be held against. Everything a dispute can recover sits here, which is why the contract is the reference point rather than last month's invoice. If your reporting cannot separate this layer from the two above it, you cannot see the only number you are entitled to argue about.
Who pays and who sets each layer is covered in more depth in what revenue leakage is in payment operations and in why billing is harder for fintechs and banks.
What attribution actually takes
The requirement follows from the problem. If the average destroyed the composition, the composition has to be rebuilt from the level where it still exists.
That means per transaction, not per statement. For each one: reconstruct what the contract says it should have cost, layer by layer. Compare that against what was charged. Where the two differ, attribute the variance to a specific layer, and to a specific clause. Aggregate afterwards, and only afterwards, because an aggregate built from attributed variances answers the question that an aggregate built from totals cannot.
Done that way the answer to "why did the rate move" stops being a narrative and becomes a decomposition: this much from issuer mix, which is not actionable; this much from card-type mix, which is not actionable; this much from markup applied against the wrong tier, which is. The method is set out in how to calculate your true provider cost per transaction, and the statement lines it has to be reconciled against are covered in how to read a PSP invoice.
The calculation has to be deterministic, because its output has to survive being shown to the counterparty. A variance you can trace to a clause and a transaction is evidence. A variance produced by a model nobody can re-run is an opinion.
The bottom line
Your effective rate is a real number and it is not lying to you. It is answering the only question an average can answer, which is how much, and being asked a question it structurally cannot answer, which is why.
The Federal Reserve's own data puts the natural variance in one layer alone at roughly 3.4 times. Any signal smaller than that is invisible inside the blend, and provider markup errors are almost always smaller than that. So the number moves when nothing is wrong, holds steady when something is, and in both cases the investigation ends where it started.
Bluefyn reconstructs contract pricing and checks fees transaction by transaction, so that a change in cost resolves into named layers with evidence attached. Bluefyn never moves, holds, or custodies funds; it only analyses transaction and provider data.
A number you cannot attribute is not a control. It is a headline.
Frequently asked questions
What is an effective rate in payment processing?
Total processing cost divided by total processed volume, expressed as a percentage. It measures what processing costs you across a period. It does not identify what any individual transaction should have cost, and it cannot separate the layers inside that cost, which is why it answers how much but never why.
Why did my effective rate go up when nothing in my contract changed?
Most often because your interchange mix changed rather than your pricing. Interchange is set by the card network's schedule and varies by issuing bank, card type and transaction type, all of which are attributes of your customers rather than of your agreement. Federal Reserve data for 2024 shows average interchange on one network's dual-message debit at 1.50 percent of value from exempt issuers and 0.44 percent from covered issuers, so a modest shift in which banks your customers use moves your blended rate on its own.
Is interchange capped?
Partly, and narrowly. Regulation II caps what a covered issuer may receive on a US electronic debit transaction at $0.21 plus 0.05 percent of the transaction value, plus a $0.01 fraud-prevention adjustment where eligible. It does not apply to exempt issuers, whose debit interchange is uncapped, it does not cap credit interchange, and it is a US rule. Certain reloadable general-use prepaid cards and cards issued under government-administered payment programs are exempt from the standard even when the issuer is covered.
Can you dispute interchange?
No, and it is worth being precise about why. Interchange is set by the card network and paid to the card-issuing bank. Your provider passes it through rather than setting it, so there is no counterparty to argue with. What you can do is verify that the interchange applied to each transaction matches the published schedule for that card and transaction type. That is a verification question, not a dispute, and misapplied interchange is a common and recoverable finding.
Which payment fees can I actually dispute?
The provider markup layer, because it is the only one your contract prices. Tier thresholds, minimums, negotiated rates, contract versions and any charge the agreement defines are all in scope. Interchange and network assessments are not disputable, though both should be verified for correct application.
Does a flat effective rate mean my pricing is correct?
No, and this is the more expensive failure. Markup drift and a favourable movement in interchange mix can offset each other inside the same period, leaving the blended rate unchanged while an overcharge accrues underneath it. A stable effective rate is evidence that two numbers summed to the same total. It is not evidence that either number was right.
How do you attribute a change in the effective rate?
Rebuild the comparison at transaction level. Reconstruct the expected cost of each transaction from the contract, layer by layer, compare it against what was actually charged, and assign each variance to a layer and a clause. Then aggregate the attributed variances. The output separates the part of the move that came from customer mix, which nothing can be done about, from the part that came from a charge that did not match the agreement, which is recoverable.
