Running Meta ads for an ecommerce store, in the order that works

Work out five numbers before you open Ads Manager, because they decide your budget, your bid and whether a campaign is viable at all.

• Gross margin per sale. What you keep after the cost of the product. This, not revenue, is what pays for advertising.
• Allowable CPA. Gross profit per customer multiplied by your payback tolerance. If you make 60 euro of profit on a first order and you are willing to break even to acquire, your allowable cost per acquisition is around 60 euro. Strong repeat purchase lets you pay more up front.
• Conversion rate. Of the people who click, how many buy. This is a landing page and offer problem as much as an ad problem.
• Average order value. Which sets what a conversion is worth.
• Payback period. How long until a customer repays what you spent acquiring them. This decides how aggressive you can be.

If the arithmetic does not work at your allowable CPA, no campaign structure will save it. Fix the offer or the margin first. That sounds like a platitude until you have watched a store spend a quarter's budget discovering it.

The break-even calculator turns these into a target CPA and ROAS, and the budget guide turns that target into the daily floor your account needs to learn at all.

Four things have to be in place, and the order matters because each depends on the one before it. Getting the foundation right determines whether the rest scales.

1. Purchase signal, before anything else. Meta optimises toward the conversions you report. If purchases are under-reported, duplicated, or unmatched to a person, every campaign you build inherits that distortion. This is the one to fix first because it cannot be compensated for later, and it is covered in pixel, CAPI and match quality.

2. Catalogue, so the feed can do the work. A connected product catalogue is what allows dynamic product ads, which is usually the lowest-cost format available to a store because it shows people the specific things they looked at. Product IDs in your events have to match the IDs in the catalogue exactly, or delivery stays near zero.

3. Campaign type, once there is history to choose with. Automated shopping campaigns need enough conversion history to learn from; below that, manual with broad targeting gives the algorithm a simpler problem. The threshold and the trade are in Advantage+ versus manual.

4. Scaling, last. Adding budget to an account whose signal is broken breaks it faster. See scaling without breaking performance.

Treat a new store's first quarter on Meta as a data-gathering period rather than a performance one. Budgeting as though it were the latter is what leads to abandoning the channel before the numbers mean anything.

Expect costs to be worst at the start. A fresh account has no conversion history, so the system is pricing an unknown and exploring to find out who converts. That exploration is what you are paying for, and it resolves as volume accumulates. New account CPM covers why.

Expect the arithmetic to be tight before it is good. Your early CPA reflects a model with little to work from, so judging the channel on week two measures the learning phase rather than the business. Plan the budget so you can afford to reach the point where the numbers mean something, which is the calculation in the budget guide.

Expect to need more creative than feels reasonable. Conversion rates fall as the same people see the same ad repeatedly, so a store running one proven ad will watch it decay. Three to six genuinely different angles per ad set, refreshed as a pipeline rather than in a panic, is the working pattern.

What to watch in the meantime, rather than daily CPA: whether purchase events are firing cleanly, whether creative is producing different results from each other, and whether blended revenue over total spend is moving in the right direction. The MER calculator gives you that last one without depending on attribution.

What ROAS should an ecommerce store aim for?

The one that clears your own break-even, which depends entirely on your margin. Two stores on very different margins need very different ROAS to make the same profit, so a published target is meaningless without the margin beside it. Work out yours with the break-even calculator and use that as the line.

Do I need a catalogue if I only sell a few products?

It matters less with a small range, since dynamic product ads earn their advantage by picking from many items. With a handful of products, well-made single-product creative can do the same job. Connect the catalogue anyway when it is easy, because it costs little and it opens the format later.

Should I run retargeting separately?

Not at the start. Before there is a flow of people considering your products, a retargeting campaign has very few people to reach, so it spends its budget on high frequency against a small pool. Build prospecting first. Whether you need separate retargeting covers the decision once you do have volume.

My ROAS in Ads Manager looks good but the bank account disagrees.

Platform ROAS counts conversions Meta believes it caused, inside a window you chose, and can include purchases that would have happened anyway. Compare blended revenue over total spend, and new-customer acquisition cost, before deciding the account is working. Attribution windows explains the gap.

How much creative do I need to start?

Enough for genuinely different angles rather than variations, and three to six live per ad set is the working range. More than that and most never get meaningful delivery, because spend concentrates on whichever the system predicts will win.