FACTORING,
AT SMALL SIZES.

The cost of assessing a counterparty is largely fixed. On a large invoice it is a rounding error and on a small one it exceeds the margin, so the market simply does not serve below a threshold rather than pricing it higher.

Most of the diligence is establishing that the invoice is real, that the goods were delivered and that the buyer has paid before. On a platform where the order, the receipt and the payment history are already records, that work is a query rather than an investigation.

Factoring rates are quoted per period and read as annual. 2% for 30 days is roughly 24% a year — a reasonable price for closing a cash gap and an expensive one for funding growth, and the difference should be stated rather than left to be discovered.

It is a query, not an investigation, and it runs on the four things GRAJ already holds: the goods were delivered, the invoice is unpaid, this buyer has settled with this seller before, and whether they paid by the date they were given. Nothing asks a model — a language model guessing whether a shop is good for forty dollars is both worse than counting and a bill. The advance is 80% of the invoice, or 90% where the buyer has settled three or more and never late.

The invoice is raised as a funding request on GRAJ and a funder offers against it, through the same request, offer, agreement, drawdown and repayment path as every other financing here. GRAJ advances no capital and carries no credit risk; it takes its 5% on the capital, once, as it does on everything. The price shown before an offer arrives is what the record suggests a funder would ask, quoted per period and annualised, and it says so.

A receivable already with a terms financier cannot also be raised to a funder, and a request pinned to an invoice is unique while it is live across every participant on GRAJ. Both are refused by the database rather than by a screen, because the one path with no application check in front of it is a direct write.