The Signal You Repay From Is the Last One to Move

A revenue share collects from the slowest-moving number a merchant produces, and by the time it moves the decisions that caused it are months old.

Retail and food services sales fell 0.6 percent in July against a consensus expecting a 0.1 percent rise. The control group, the narrower measure that feeds GDP, fell 0.48 percent against an expected gain of 0.3 percent, its largest decline since January 2025.

Consensus had the number going up. It went down by most of a point. Whatever produced that had already happened inside thousands of businesses before it showed up in the aggregate, and the aggregate is the layer at which most lenders think about merchant health.

Receipts are downstream of every decision that matters

Consider the order in which a struggling merchant actually does things. Hours get trimmed first, because labour is the fastest lever and nobody has to be told. A supplier order gets delayed or split. A reorder gets skipped. The owner covers a gap on a personal card. Only after all of that does the till reflect anything, and even then it reflects it partially, because a merchant protecting revenue will discount to hold volume.

A revenue-share advance collects a percentage of receipts. That means the instrument is bolted to the last variable in the sequence. Everything informative happened upstream, weeks or months earlier, and none of it was visible in the number the lender is paid from.

This is not a criticism of the product. Repaying from receipts is what makes the product work for merchants with no fixed capacity to service a term loan. It does mean the repayment stream is a poor early-warning system, and it is routinely used as one.

The aggregate is not slow, it is also not about your borrower

The July release makes a second point worth more than the first. Within the same month, non-store retail fell 2.2 percent, autos and parts fell 1.8, gasoline stations fell 0.9 and electronics and appliances fell 0.5. Over exactly the same weeks, clothing and accessories rose 1.9 percent, health and personal care rose 0.7, miscellaneous retailers and bars and restaurants each rose 0.5, and building materials rose 0.3.

That is a four-point spread across categories in a single month. A lender reading “retail sales fell 0.6 percent” and inferring something about a specific merchant is reading a number that contains merchants moving hard in both directions.

It is also worth noting what the revisions did, because it cuts against the easy version of this argument. June was revised slightly up, not down: the headline from a 0.22 percent gain to 0.25, and the control group from 0.4 to 0.5. The prior month did not look like a warning even in hindsight. The turn arrived anyway.

What is actually observable earlier

Here is the part that matters operationally, and it is more encouraging than the above suggests. A lender funding from receipts is already receiving the upstream data. It arrives as bank transaction history and payment processor detail, and it is usually read for one purpose, sizing the advance at underwriting, and then not read again in the same depth.

Inside that same data sit the things that move before totals do. Transaction counts separate from transaction value, which is how a lender distinguishes a merchant discounting to hold volume from one genuinely growing. Deposit timing and frequency, which change when a merchant starts batching. Outbound payment patterns, which show a supplier being stretched. Payroll debits, which show hours being cut. None of these require new data collection. They require the existing data to be read continuously rather than once.

Our own view, formed from operations rather than from theory, is that most lenders in this segment have better information than they use, and the constraint is not analytics sophistication. It is that reading a live book at this depth is manual work that scales with the number of positions, so it gets done at underwriting when the file is in front of someone and never again while the position is open.

How We Approach It

The method is to move the read from a moment to a cadence.

Data Intelligence and Analytics builds the merchant-level view from the transaction data the lender already holds, separating volume from value and flagging changes in payment behaviour rather than changes in totals. The point is not a better score. It is a shorter distance between a merchant’s decision and the lender knowing about it.

Reporting and Intelligence Automation then runs that read across the live book on a defined schedule and delivers what moved, so a position deteriorating in the way described above surfaces while a conversation is still useful rather than after a remittance has already been missed.

Neither of these predicts a downturn. They shorten a lag, which is a more modest claim and a more achievable one. In a segment where the repayment stream is structurally the last thing to move, a few weeks of earlier warning across a whole book is worth more than any improvement in the accuracy of the original credit decision.

In most operations, far more work can be automated than leadership realizes. One discovery call is enough to size what automating it would return to your bottom line. Book it at https://cxocorporation.com/contact.