Meta Ads in Algeria for e-commerce: a method to measure your own numbers, with no invented benchmarks

This guide does not give you a “good CPM in Algeria”: no reliable source publishes one, and the ranges on competitor blogs do not say where they come from. It gives you a method: what to measure, how to compute what your ads really earn when part of the orders are never delivered, and when to decide. The figures in the examples are invented, to illustrate.

Before you spend

Three subjects are settled before the first campaign: how to pay Meta from Algeria (the text is unclear, read it), the framework of your activity if you sell goods (auto-entrepreneur and selling goods), and how you contact your customers afterwards (outreach and law 18-07).

1. Measure: pixel and Conversions API

Install the Meta pixel on your site. If you also add the Conversions API, send the same event with the same name and the same ID on both sides: the order number, for example. That is what lets Meta count the order only once. According to Meta’s documentation, the two sends must arrive within 48 hours of each other (a point to re-read on the page, which our tools could not open).

A limit specific to cash on delivery: the “purchase” event fires when the customer fills in the form, not when they collect the parcel. Meta therefore counts orders that will be cancelled or refused at the door. That is why you must recompute with your own rates.

2. The learning phase

According to Meta’s help, an ad set leaves the learning phase after about 50 optimisation events in 7 days, and a significant edit (targeting, creative, optimisation event, bid) restarts it. In practice: a budget too small to reach that volume keeps the ad set in an unstable zone, and changing three things a day stops it settling.

A simple calculation: minimum weekly budget = target cost per result × 50. An invented example: if you aim for 500 DA per order, that is 25,000 DA a week per ad set, about 3,570 DA a day. If that figure exceeds your budget, test fewer ad sets or choose a more frequent event (add to cart), knowing it is further from the sale.

3. What you really earn: cash-collected ROAS

The ROAS shown by Meta assumes every order is paid. With cash on delivery:

cash-collected ROAS = displayed ROAS × confirmation rate × delivery rate

An invented example: displayed ROAS 3, 60% of orders confirmed by phone, 75% of the confirmed ones actually delivered. Cash-collected ROAS is 3 × 0.60 × 0.75 = 1.35. If your gross margin before ads is 40% of the selling price, you need a cash-collected ROAS above 2.5 (1 / 0.40) to cover the advertising: in this example the operation loses money while the dashboard showed 3.

Your confirmation and delivery rates are not another seller’s: they depend on your product, your calls, your carrier. Write them down every week. Our COD profit calculator does the maths with your own variables, and the article on the real cost of a delivered parcel details the method.

4. Testing without getting lost

One test at a time: one creative, one audience or one offer. Before launching, set the amount you accept to lose on this test, for example a multiple of your margin per delivered parcel, and stick to it. Let it run long enough for the ad set to leave the learning phase before judging. Keep what worked in a table: date, change, cost per order, confirmation and delivery rates.

5. What not to believe

Screenshots of ROAS posted in groups with no delivery rate. “Magic budgets” and benchmarks from competitor blogs. Carrier rates copied from a third-party site: see the method to choose a carrier. If a figure has no dated source, measure your own.