Fake COD Orders Are Costing You Money: How to Filter Them

Owners usually notice fake orders at the end of the month, when the courier statement arrives and the returned parcels outnumber what anyone expected. By then the money is gone. Every fix worth doing sits earlier, in the hours between the order landing in the store and the parcel leaving the warehouse.

Fake COD orders are costing me money. How do I filter them?

Catch them between the order and the courier, and do it within 2 hours of the order. That window is the only place a fake order costs nothing to cancel. Once it ships you have paid the outbound leg, and if your courier bills returns, you pay for the trip back too. On every store I build, the confirmation flow is designed on purpose for this: how and when an order is confirmed, so the cancelled ones are caught before they ship and cost delivery twice.

Filtering at the ad level helps less than people think. You can't target away someone who orders for fun. You can make sure their order never leaves the building.

What actually counts as a fake order?

Fewer orders are truly fake than owners believe. Cancellation rates in Egypt run between 15% and 40% depending on category and price, and most of that is real people who changed their mind, misread the price, or didn't expect the delivery time. The genuinely fake slice is smaller and easier to spot.

• Duplicates: the same phone number or address placing the same order more than once in a day.
• Unreachable numbers: wrong digits, switched off, or never answered.
• Incomplete addresses: no area, no building, a single word in the address field.
• Joke details: obviously invented names, or quantities nobody buys at retail.

The distinction matters because the fixes are different. A real buyer who changes their mind is an expectations problem on the ad or the page. A fake order is a filtering problem at confirmation.

How fast should I confirm a COD order?

Within 2 hours. Interest fades fast after an impulse order, and a call the next day reaches someone who has already forgotten what they ordered. This is the sequence I set up:

1. An automatic WhatsApp message the moment the order lands, with the product, the total including delivery, and the expected delivery time.
2. A call inside the 2-hour window if the message isn't answered or confirmed.
3. A clear status in the store: confirmed, waiting, or cancelled, so nobody ships a waiting order by mistake.
4. Only confirmed orders go to the courier.

The message restates the total with delivery on purpose. A surprise at the door becomes a refusal, so I'd rather lose the order at confirmation than pay to discover the same thing at the customer's door.

Which orders should I hold before shipping?

Any order that fails the 2-hour confirmation waits; it doesn't ship on hope. Beyond that, hold orders that match the patterns above until someone has spoken to the buyer: duplicates from the same number, incomplete addresses, and unusually large quantities on cash on delivery.

Keep a simple internal list of phone numbers that refused at the door before, and treat a new order from them as one that needs a real conversation, or prepayment, before it ships. I wouldn't block them outright on the first refusal. People refuse for reasons that aren't malicious, and a blanket ban costs you repeat buyers.

Should Meta count unconfirmed orders as purchases?

Not if you can avoid it. Meta learns from the purchase event, and if every fake and impulse order counts as a buyer, the algorithm goes looking for more people like them. In Egypt the gap between Meta's order count and the store's regularly reaches 40%, and part of that gap is orders that were never real to begin with.

Keying the purchase event to the confirmed order, rather than the completed checkout, gives Meta fewer events but cleaner ones. On accounts with heavy cancellations, that trade is worth making. See why orders show in Shopify but not in Ads Manager for the tracking side.

How much are fake orders really costing me?

More than the shipping line suggests, because they change the ROAS you need to break even. On a 618-product catalogue, break-even was 1.77 ROAS on paper and 2.11 once cancellations were counted. Campaigns clearing 1.9 had been treated as winners and were losing money on every delivered order (the case).

That is the real cost of letting bad orders through: you scale campaigns that look profitable and aren't. A confirmation flow that catches them before shipping brings the real break-even back toward the paper one, and that difference goes straight to margin.

If you want to see where your own store loses money first, send the link to the free audit. You get the most you can pay for an order, your ROAS after delivery, and your biggest leak within 48 hours.