If you've taken three calls from three processors this year, you've probably noticed something odd: the numbers all land within a whisker of each other, and each caller is certain theirs is the best.

They're not all lying. The explanation is structural, and once you see it, most of what's strange about this industry stops being strange.

Your processor doesn't set the biggest number on your bill

A card fee splits three ways.

Interchange is the largest slice, and it is identical for everyone. Every processor in the country pays the same published rates. They cannot discount it, do not keep it, and have no more influence over it than you do.

Assessments are the same story on a smaller scale.

What's left — the markup — is the only genuinely competitive part of the price. And because it's the smallest part, competing on it means competing over a thin slice of a number the customer can't easily isolate.

That's the whole situation. Everything else follows from it.

Why interchange isn't one number

It's worth understanding the layer nobody controls, because it explains most of the variation you see month to month — and most of the arguments merchants have with the wrong party.

Interchange isn't a single rate. It's a large published table with hundreds of categories, and the rate applied to any given transaction depends on things that have nothing to do with your processor:

• What kind of card it was. A basic debit card and a premium rewards credit card carry very different interchange. Someone has to pay for the airline miles, and it isn't the airline.
• How it was presented. Tapped, dipped, swiped, or keyed. Card-not-present costs more because it carries more fraud risk.
• What kind of business you are. Merchant category codes carry different rates. A supermarket and a jeweller are not on the same table.
• How complete the data was. Transactions submitted with more information — address data, order detail — can qualify for better rates. Incomplete ones "downgrade."
• How quickly you settled. Batch late and a transaction can fall to a worse category.

Two consequences worth holding onto.

First, your rate can rise without anyone raising it. A month with more rewards cards costs more. That's your customers' wallets changing, not your processor's pricing.

Second — and this is the part processors rarely volunteer — some of that is genuinely fixable. Downgrades caused by late settlement, missing data, or keyed transactions that could have been tapped are real money, and they're operational rather than contractual. A processor who looks at your downgrades is doing something more useful than shaving a basis point off their markup.

What actually differs between processors

Strip out the parts nobody controls and here's what's left to choose between:

• How much markup they take, and whether you can see it as its own line
• Whether the price holds — or drifts upward at renewal
• How long they can hold you — the term, and the fee for leaving
• How fast they approve you, and whether you hear anything while you wait
• When your money lands — next day, or later
• Whether a person answers, and whether that person can decide anything
• What else they can sell you that isn't a rate at all

Notice that only the first is a number on a quote. The rest are the things you'll actually experience, and none of them show up in the comparison you're being invited to make.

Why the unsustainable quote exists

Here's the uncomfortable mechanic.

If the only visible variable is a number, and every competitor's costs are effectively identical, then the way to win the meeting is to quote a lower number than the last person through the door.

Do that honestly and you eventually run out of margin. Do it dishonestly and you have a business — provided you can make the difference back somewhere the customer isn't looking.

The places to make it back are well established:

1. The rate drifts at renewal. Legal, common, and typically announced by a line on a statement the merchant stopped reading months ago.
2. A fee appears under a name nobody can parse. Small, monthly, and unrelated to sales volume.
3. The tier definitions move. On tiered pricing, more transactions quietly get reclassified into the expensive bucket.
4. The term makes leaving expensive. If the exit costs more than the annoyance, most merchants stay.

None of that is fraud, exactly. It's what happens when the only variable is a number and everyone has the same number. The dishonesty isn't a character flaw in the industry; it's the shape of the industry.

And it's a bad business to be in even for the people running it. You spend your life defending a basis point against whoever called your merchant last Tuesday, and the merchant learns to distrust you by the second year.

Two quotes that look different and aren't

Here's how a comparison goes wrong in practice.

Two processors quote the same business. One leads with a headline rate that's clearly lower. The merchant picks it, and a year later is paying more.

(Illustrative comparison — the shape is what matters, not these particular figures.)

Processor A didn't necessarily lie. The 1.79% is real, for the transactions that qualify. The merchant compared the one number they were shown, which is exactly what the structure invites.

This is why the effective rate matters so much. It's the only figure that survives the categorisation, the naming, and the tiering — because it's just the money that left your account divided by the money you took.

Who's actually selling you processing

One more piece of the picture, because it explains a lot of the phone calls.

Most processing isn't sold by processors. It's sold by agents and ISOs — independent salespeople and small firms who write merchant accounts and earn a share of the revenue on the accounts they bring, month after month, for as long as the merchant stays.

That model has two effects worth knowing about as a buyer.

The good one: your agent has a long-term interest in you staying. An account that leaves in six months is worth very little to them. A good agent is genuinely the most useful person in this industry to have on your side, because they know the businesses on their own street and they lose money when you're unhappy.

The bad one: agents are often given the same commodity to sell as everyone else, and paid more for signing you at a higher rate. When the only tool is the quote, the incentive points the wrong way.

This is why what a processor gives its agents matters to you. An agent armed with nothing but a rate sheet has to win on the rate, and will be back a year later defending it. An agent with things that aren't a rate doesn't.

The three questions that actually separate processors

Since the quote can't tell you much, ask things the quote can't dress up.

1. "What would my effective rate have been on this statement?"

Hand over a real statement and ask for one number: total fees divided by total volume, as it would have been under their pricing. This is the only apples-to-apples comparison that exists. A processor who deflects to "it depends on your mix" is avoiding the question — your mix is right there on the statement they're holding.

2. "Is the rate fixed, and what happens at renewal?"

Then ask for it in writing. The verbal answer is always reassuring. The written one is the one that binds.

3. "What does it cost me to leave, and when?"

A company that needs a term to keep you has told you what it thinks of its own service. This is the single most informative question on the list, and it's the one people are most reluctant to ask.

What we did instead

We'll be straightforward about our own position here, because the argument above cuts against us as much as anyone.

We didn't want to run the business where you win by quoting a number you can't sustain. Not out of virtue — because it's a bad business. So we went and built the things that aren't a rate:

• Software that runs the invoicing and the chasing, included with processing rather than billed separately — because a processor with something else to sell doesn't have to win on the rate. See the software →
• A Square program for merchants already running Square who want to keep the counter and change what it costs. See the Square program →
• Dual pricing for businesses it genuinely suits, set up to the disclosure rules properly. How dual pricing works →

None of those is a number on a sheet. Each is something a competitor either has or doesn't.

And then the processing gets to be honest, because it isn't carrying the whole pitch. That's the entire idea, and everything else on this site is a consequence of it.

What to do with this

You don't need to switch processors to benefit from any of this. The useful part is smaller and more portable:

1. Work out your effective rate from your last three statements. Total fees ÷ total volume.
2. Find your term and your exit fee in your agreement, so you know what leverage you have.
3. Ask the three questions of whoever calls next — including, fairly, of us.

If the processor you already have answers them well, stay. That's a genuinely good outcome and it costs us nothing worth keeping to say so.

If you'd like us to look, send a recent statement. We'll tell you what your effective rate is, what we'd do differently, and when the honest answer is that you're fine where you are.