Where COD merchants lose money: the five leaks

Leak 1 — Orders that were never real

The COD order form is free to submit, so a slice of every ad campaign's "results" is noise: wrong numbers, impulse orders, duplicates, and the occasional serial refuser. Every one of those that ships becomes a return fee; every one that sits unconfirmed while agents improvise is capital and attention burned.

The leak is rarely the existence of bad orders — that is physics — but the absence of a system: orders with no owner, no recorded outcome, no reschedule discipline, and no shared memory of who refused parcels before.

What closes it: a confirmation workflow where every order becomes an owned task with recorded outcomes (confirmed / no answer / reschedule / cancel), automatic reschedule resurfacing, fair auto-assignment with daily limits, and a fraud screen with customer history that blocks known-bad orders before anyone dials. In COD Scale that is the call-center workflow (Silver+), SmartRouting in the OMS Pro cockpit (Gold+), and Fraud Shield on every plan.

Leak 2 — Deliveries that fail

A returned parcel carries the full cost chain of a delivered one — ad spend, confirmation minutes, packaging, outbound shipping, weeks of transit — minus the revenue, plus a return fee. Return rate is the multiplier that turns an operation profitable or not with everything else held equal.

What closes it: doorstep-surprise prevention during the call (total restated with fees), structured wilaya/commune addresses, deliberate stop-desk offers where home delivery is weak, courier-to-zone matching measured per wilaya, and a watch on stalled parcels so a new attempt happens while the customer still cares. Covered in depth in the returns guide and the couriers guide.

Leak 3 — Ad platforms flying blind

Meta and TikTok optimize toward the conversions they can see. Browser-pixel-only tracking has been degraded for years — iOS privacy changes, ad blockers, and short-lived cookies mean a meaningful share of real conversions never reaches the platform from the browser alone. The algorithm then optimizes on partial data, and budgets drift toward audiences that look good rather than audiences that pay.

COD adds a second distortion: a submitted order is not money. If the pixel fires Purchase at submission, the platform happily optimizes toward people who fill forms — including the ones who refuse at the door.

What closes it: server-side events alongside the browser pixel (Conversions API on Meta, Events API on TikTok, Measurement Protocol on GA4, plus Google Ads and Snapchat), sharing one event ID so nothing double-counts — and a COD-aware strategy that sends Purchase only when the order is actually delivered, so the platform optimizes toward collected cash. In COD Scale this is server-side tracking with the Delivered Revenue strategy, from Silver.

Leak 4 — Finance that stops at revenue

A dashboard that shows order totals is showing you hope, not money. Between an order and its profit sit: the return probability, product cost, courier fees both ways, confirmation and packaging labor, ad spend, and the weeks the cash spends in the courier's drawer. Operations that do not compute delivered profit per order and per campaign routinely scale losers and kill winners, because revenue and profit rank campaigns differently.

The same blindness hides receivables: cash delivered but not yet remitted is real money with no owner watching it, and unreconciled courier statements are where parcels and fees quietly go missing.

What closes it: per-order P&L with real costs, campaign verdicts computed from delivered orders (Scale / Hold / Stop), receivables aging per courier, remittance reconciliation, and a cash forecast. In COD Scale this is the finance suite, from Gold.

Leak 5 — Multi-store chaos

Growth in Algerian COD usually means more stores: a second brand, a new country, a partner's catalog. Run through one shared dashboard and a pile of spreadsheets, stores bleed into each other — orders mixed, stock confused, agents unsure which brand they are speaking for, numbers that cannot be trusted per store.

What closes it: real workspace isolation — every operational record scoped to its store, cross-store lookups failing loudly instead of leaking, API keys bound to one store — with a hub that still aggregates the whole operation, and a choice between shared staff and per-store teams. In COD Scale this is the multi-store hub on every plan, with workstation roles from Gold.

The compounding effect

The five leaks feed each other. Unfiltered orders inflate failed deliveries; failed deliveries poison the ad signal if Purchase fires at submission; a blind ad platform buys more bad orders; finance that stops at revenue cannot see any of it happening; and multi-store chaos multiplies every other leak by the number of stores. Closing them is not five separate projects — it is one operating discipline, which is why COD Scale ships them as one platform.

Frequently asked questions