Why a 7-day COD settlement cycle is the right number

The most consequential number in COD finance is not the delivery rate. It is the settlement cycle — the days between a confirmed delivery and the USD wire landing in the merchant's bank account.

Fufills publishes 7 days. Most LATAM operators publish 14, 21, or 30. A few quietly run on 45.

This piece is the math behind why 7 is the right number, why faster is structurally fragile, and what 30-day operators are actually doing with the float.

A COD order looks linear from the outside: the carrier delivers, the buyer pays cash, the merchant gets the money. Inside the operation it is a six-step finance pipeline:

1. Cash collection at the door. Carrier driver receives local-currency cash from buyer.
2. Driver-to-hub remittance. Driver deposits cash at carrier hub, typically same-day or next-day.
3. Hub-to-carrier-bank deposit. Carrier deposits aggregated cash into their operating bank account. Usually T+1 or T+2.
4. Carrier-to-platform remittance. Carrier wires aggregated collected funds to the COD platform's local-currency account, on an agreed schedule (weekly, biweekly).
5. Currency conversion. Platform converts local currency to USD via FX desk.
6. Platform-to-merchant wire. Platform wires USD to merchant bank of record.

Six steps. Each step has its own SLA, its own reconciliation, its own failure modes. The settlement cycle is the sum of those steps plus reconciliation buffer.

If the operator is running clean SOPs at every step, the natural floor is roughly:

That is 5–10 days, end-to-end, with no idle time. 7 days is the operating-pessimist median of that band. A platform that publishes 7 and hits 7 is running each step at the operational floor with a small reconciliation buffer.

Some operators advertise T+1 or T+2 COD settlement. Mechanically, this is only achievable in one of two ways:

Option A — the operator fronts the cash. They settle the merchant before the carrier remits the underlying cash. This requires a sizeable working-capital line, and it works fine right up until a chargeback wave, a fraud event, or a carrier-remittance delay. Then the operator is short, and the merchant is exposed to the operator's balance sheet.

Option B — the settlement is partial. The operator wires "expected" funds based on a delivery rate they have not actually reconciled. When reconciliation completes, the merchant receives a clawback for over-paid balances. The merchant sees fast money on day one and an awkward debit on day fifteen.

Neither model is wrong by itself. Both are wrong when sold as a feature without disclosure. Fufills runs neither. We wire after reconciliation, on the 7-day cycle, in full. The merchant never receives a clawback for a settled invoice.

The 30-day end of the market is generally one of two patterns:

Pattern A — the operator is using merchant funds as working capital. The operational floor is 7 days. Extending settlement to 30 days creates a 23-day float per dollar collected — across an aggregated book, that float is the operator's working-capital line. The merchant is, in effect, financing the operator's growth at 0% interest.

Pattern B — the operator's reconciliation is broken. Disputes, short-pays, and returns are not closing inside a clean window because the SOPs are not running. Settlement extends because the platform has no confidence it can wire the right number until the dispute backlog clears.

Both patterns are bad for the merchant. Pattern A is bad because it puts merchant cash on the operator's balance sheet. Pattern B is bad because it signals operational drift — and the next missed week is usually a 45-day cycle, then 60.

Three reasons:

It is the operating floor with discipline. Not the floor of fronting cash, not the floor of cutting reconciliation. The floor of running every step inside SLA and posting the reconciliation buffer publicly.

It is auditable. A merchant can compare invoice date to wire-received date and check whether the published number matches the operating number. We expect to be checked. If a Fufills cycle slips past 7 days, the published target was overstated; we publish the actual median quarterly in the dashboard.

It compounds with the rest of the stack. Hard-gated confirmation drives low RTO, which drives clean reconciliation, which drives stable 7-day cycles. The pieces lock together. Operators who run ungated COD cannot hit 7 even if they want to, because reconciliation is permanently chasing returns.

Three questions to ask:

1. What is your published settlement cycle, and what is your actual median last quarter? (If the published number and the actual number differ by more than a day, the published number is marketing.)

2. Do you front cash, or do you wire after reconciliation? (Both can be acceptable. Only one of them is consistent with a stable operation over five years.)

3. What was your worst single-cycle delay in the last 12 months, and what caused it? (Every operator has had at least one. Operators who claim they have had none are not telling you about one.)

The platform that answers all three openly is the platform that has thought about COD finance ops the way it deserves to be thought about: as the actual money pipeline, not the marketing claim.

That is the number Fufills designed itself around. Seven days, wired in full, after reconciliation. We expect merchants to audit us against it.