Split funding and the holdback the statement never shows
Most advances announce themselves. A fixed debit leaves the account every business day, in the same amount, under the same name, and any careful reader finds it. Split funding does the opposite: the funder takes its share out of card settlement before the money reaches the bank at all, so the repayment is never a line on the statement — only a smaller deposit is. Here is how to recognise one in a file, how to size what was kept, and why that size belongs in a range.
- Repaid before the money lands — no debit to count
- Revenue and burden both understated at once
- An estimate with its assumption attached, not a guess
What split funding is
Merchant cash advances are collected in two broadly different ways. The familiar one is a debit: the funder pulls a fixed amount from the operating account every business day or every week until the purchased amount is delivered. The other is split funding — also sold as a lockbox or a split withholding — where the merchant’s card settlement is routed so that the funder keeps an agreed share of every batch and passes the remainder on.
The share it keeps is the holdback: a percentage of takings rather than a flat daily figure. That is the feature merchants are sold on, and it is a real one. A slow week costs the merchant proportionally less, because the holdback moves with the receipts instead of ignoring them. Nothing is ever taken out of the account, so nothing bounces when the account is thin.
For an underwriter reading that merchant’s statements three months later, the same mechanic is the problem. An obligation that is settled before the money arrives leaves no trace on the side of the ledger where obligations are normally found.
Why the statement cannot show it
A bank statement is a record of an account, not a record of a business. It reports money that arrived and money that left. A split holdback does neither: it is deducted upstream, at the processor, and the account simply receives less. There is no debit to find because no debit was ever made.
That produces two distortions at once, and they point the same way. Card revenue reads low, because what lands is net of the holdback. Financing burden reads low, because the usual measure — what the funders debited, against what the business earned — is missing its numerator entirely. A file repaid at source therefore looks cleaner than a file repaid by debit carrying exactly the same obligation, and it looks cleanest precisely where the obligation is heaviest.
In the worst case the position disappears altogether. If the only trace of the funder is an inbound credit in its name, a quick read files it as money coming in — a recent advance, perhaps — rather than as the visible remainder of money already collected.
Four signals that a position is being repaid at source
No single signal is conclusive on its own. Together they separate a split-funded position from an ordinary deposit relationship.
The funder appears only on the credit side
A recognisable funder name arrives as an inbound payment, repeatedly, with no matching debit stream anywhere in the period. A funding event is one credit. A stream of credits from a funder is a remittance, and a remittance implies a deduction before it.
Card deposits step down while the business does not
Settlement deposits fall by a consistent proportion from a particular date, while deposit count, non-card revenue, and the general rhythm of the account hold steady. A quieter month moves every number. A holdback moves only the card side, and by a stable share.
The cadence belongs to the processor, not to an ACH schedule
Remittances follow batch settlement: every business day, weekend takings bunched into Monday, amounts that vary with trade. A daily ACH debit is flat and indifferent to what the business actually took.
The descriptor names a routing arrangement
Funds that pass through a settlement or lockbox account before reaching the operating account usually say so in the descriptor. It is the one place the arrangement is written down on the statement itself.
What the remittance tells you about what was kept
The visible remittance is not silent about the invisible holdback. If the funder keeps a share r of gross takings and passes on the rest, then the deposit you can see is the takings less that share. Turn it around and the amount withheld before a remittance of R arrived is R × r / (1 − r).
Work it on a real shape of file. Suppose $40,000 of remittances across a statement month. At a 10% holdback the funder kept about $4,400 on top of what you can see. At 20% it kept $10,000. Identical evidence, the same account, the same period — and an answer that more than doubles depending on a term the bank never printed.
So the rate is the whole question, and the statement does not carry it. That leaves two honest responses. Ask the merchant for the agreement and use the real number. Or state a range, say what it assumes, and let the reader decide how much weight it carries. What is not honest is a single confident figure derived from an assumption nobody declared.
How we report what the statement cannot show
When MetrikData finds a position repaid at source, the position row stops pretending to a total debited. It shows an estimated range for what was withheld, with the assumption printed beside it — the typical 10–20% retention seen in split arrangements — so nobody reads the figure without also reading what it rests on.
The monthly figure is normalised over the stream’s own span rather than over the calendar, with a floor of three weeks. A fortnight of remittances near the end of a statement is not stretched into a full month of holdback on the strength of two weeks’ evidence.
The headline burden stays the figure the statement supports. The estimate travels next to it, as a second reading of the same ratio with the holdback included, and again in the footnote under the obligations table. It is never quietly folded into the headline, because a number the statement cannot vouch for should not be able to hide inside one it can.
If the merchant produces the agreement, the band collapses to the rate it states and the range becomes a figure. The arithmetic does not change; only the confidence does.
Common questions
What is split funding in a merchant cash advance?
A collection method where the funder is repaid out of card settlement instead of by debiting the bank account. The processor routes each batch so the funder keeps an agreed percentage — the holdback — and the merchant receives the rest.
Does a split-funded advance show up on bank statements?
Not as a repayment. The holdback is deducted before the money reaches the account, so there is no debit to find. What remains visible is a smaller deposit, or a remittance arriving in the funder’s name.
How do you work out the holdback from a net deposit?
At a retention rate r of gross takings, the amount withheld before a remittance of R arrived is R × r / (1 − r). A $40,000 month of remittances implies roughly $4,400 withheld at a 10% holdback and $10,000 at 20%.
Why publish a range instead of a single number?
Because the holdback rate is a contract term and the statement does not carry it. The remittances bound the answer but do not settle it. A range with its assumption stated is what the evidence actually supports; a single figure would be a guess wearing a decimal point.
Is split funding worse for the merchant than a daily debit?
Not inherently — a holdback flexes with receipts, so a slow week costs less and nothing bounces. The risk is what it hides. Because both revenue and burden read low, a split-funded merchant can look like they have room for another advance when they do not.
Related
See what a statement is holding back
Upload one file and read the positions, the cadence, and the burden from the transactions themselves.