A payment processing agent closes a merchant once and gets paid every month for as long as that merchant keeps taking cards. It is the most durable compensation model in sales, and the most awkward to calculate.

Most commission guides assume a deal closes, a rep gets paid, and everyone moves on. Residual compensation breaks that assumption on day one. The deal never really closes, the payment never really stops, and the amount changes every month based on what the merchant did. Here is how the model works, written for someone who has to actually run it.
An ISO buys processing from a sponsor bank or a larger processor at a wholesale price. That price is the buy rate. The ISO then sells to a merchant at a higher price, the sell rate. Everything the agent earns comes out of the gap between the two.
Say the buy rate is 2.10% plus 10 cents a transaction and the merchant is boarded at 2.60% plus 15 cents. The spread is 0.50% plus 5 cents. On a merchant processing $80,000 a month across 1,600 transactions, that spread produces $400 from volume and $80 from transaction count. The agent is then paid a share of that $480, commonly somewhere between 40% and 70% depending on whether they brought the deal, service it, or both.
This is why residual statements are so hard to check. The number a rep is owed depends on a merchant's behaviour last month, a wholesale rate the rep never sees, and a split that may differ per merchant. Three moving parts, none of them under the rep's control.
A rep on a residual plan is not building a quarter. They are building an annuity. A merchant boarded in year one is still paying in year four, so a book of two hundred small merchants can out-earn a book of twenty large ones, and it survives the loss of any single account.
That changes what good management looks like. Attrition is the number that decides an agent's income, and attrition is invisible on a commission statement that only shows what was earned. An agent who added eight merchants and lost eleven had a bad month that looks like a fine one.
It also changes what a rep needs to see. A quota-carrying rep wants to know whether they hit. A residual agent wants to know which merchants moved, which ones dropped off, and whether the split they were promised is the split they got.
Larger ISOs recruit sub-agents, who recruit their own. Each layer takes a cut of the layer below, and the cuts are usually expressed as a share of residual, which is a different base from the spread. That distinction sounds academic until you compute it.
Take a merchant generating $480 of monthly spread. The office keeps 30%. The closing agent takes 50% of what remains. The sub-agent who sourced the lead takes 20% of the closer's share. Run it in the wrong order and the sub-agent is paid on the gross instead of the closer's net, and every payment that month is wrong by a small amount that nobody notices for a year.
The order of operations is the part that gets lost when a plan moves from a contract into a spreadsheet. The contract says what each party gets. It rarely says what each party gets it from.
Override on the closer's share
What the contract says
Override on the gross
What the spreadsheet does
Same contract, same percentages, one order of operations apart. The closer is $62.40 short on a single merchant, every month, for as long as that merchant processes.
Residual income arrives from the processor as a file, usually monthly, usually late, and usually in a format that changed slightly since the last one. Somebody then has to match every merchant ID to an agent, apply the right split, handle the merchants that boarded mid-month, and produce a statement each agent can check.
In most ISOs that person is one finance analyst with a spreadsheet that has grown for six years. It works until they take a holiday. The spreadsheet has no audit trail, so when an agent disputes a figure the only way to answer is to rebuild the calculation by hand, which takes a day and produces a number that is close but not identical.
The failure is rarely a wrong formula. It is that nobody can prove the formula was right.
One merchant at $80,000 monthly volume and 1,600 transactions, boarded at 2.60% plus 15 cents against a 2.10% plus 10 cent buy rate. This is the arithmetic a residual statement should show and usually does not.
| Component | Calculation | Amount |
|---|---|---|
| Volume spread | $80,000 × 0.50% | $400.00 |
| Transaction spread | 1,600 × $0.05 | $80.00 |
| Monthly statement fee | Retained by office | $0.00 |
| Gross residual | Spread available to split | $480.00 |
| Office retention | 30% of gross | −$144.00 |
| Agent share | 50% of the $336 remaining | $168.00 |
| Sub-agent override | 20% of the agent share | −$33.60 |
| Agent net for this merchant | $134.40 |
Illustrative rates chosen to show the order of operations. Real buy rates vary by processor, card mix and merchant risk profile.
Sales compensation tools are built around a deal that closes once. Residual models need a calculation that reruns every month against changing volume, splits that vary per merchant rather than per person, and a waterfall where the order of operations changes what everyone is paid. Commish was built for a payments customer first, so this is the ordinary case here rather than the exception. If a vendor cannot show you a residual run with the workings visible, they are describing a roadmap.
Bring a processor file and an agent agreement. We will show you the same book calculated in Commish, with every merchant traceable to the rule that paid it.