Meta says your campaigns sell. Your sales figures disagree

Monday budget review. Ads Manager shows last month's purchases, and you write the number down. Then you pull the same month from your store or your CRM, and it's lower. Nobody is quite lying: Meta also counts people who saw an ad and bought days later, and it can count one purchase twice if the signal reaches it by two routes.

Next month's budget gets split on Meta's number. Before you move another euro, you want to know which of those sales really came from the ads, and what it costs you to keep trusting the rest.

What we check in your Meta account, and why in this order

Measurement comes first. If what Meta learns from is miscounted, everything else in the analysis rests on bad data.

1. Pixel and Conversions API.
The pixel is the code that tells Meta, from the browser, when someone buys or submits a form. The Conversions API sends the same signal from your server, so it still arrives when the browser blocks the pixel. If you send both, each event needs the same ID so Meta recognises the pair and counts it once (Meta calls this deduplication). We check this event by event and match it against your real orders or leads.

2. Campaign structure.
How many campaigns and ad sets (the groups inside a campaign that hold the audience and the budget) compete for the same people, and how much money each one gets. Meta needs a minimum of results per ad set before it stops testing and settles; split a large budget into too many small pieces and each one keeps testing on your money.

3. Creative and fatigue.
With broad targeting, the ad itself does a lot of the selecting: Meta shows it to people who resemble those who reacted to it. We look at how much spend sits on a handful of ads, how often the same person sees the same ad, and when results start to slide. That slide, with the same audience, is what we call fatigue. More in Meta Ads frequency.

4. Audiences.
Whether the campaigns meant to find new customers exclude the ones you already have, and how much of what an automated campaign like Advantage+ claims comes from people who already bought from you. Advantage+ is the campaign type where Meta decides on its own who sees the ad; we explain it in what is Meta Advantage+.

5. Attribution.
Attribution is the rule that decides which ad gets credit for a sale. By default, Meta adds up people who clicked in the last 7 days and people who only saw the ad the day before buying. We split those two and set them against your sales.

What you get and what we need from you

A report with every finding, where to see it in your account and our estimate of what it costs you each month. Then the fix order, including what isn't worth touching. And a working session to walk through it with whoever owns the budget.

Access is read-only: the ad account and Events Manager through your Business Manager, plus an export of orders, or of leads with the ones that turned into sales. We change nothing during the audit, and you can remove our access when it's done.

The report is yours, written so that whoever runs the account can act on it. If you'd rather we did, that's a separate conversation; this is how we work as a Meta Ads agency.

What you can do today, before talking to anyone

Go to Events Manager, open your purchase or lead event and see where it comes from: browser, server or both. If it's both, check that Meta says it is matching the two copies of each event. If it isn't, a single sale may be counting twice.

Then export your ad report for the last 90 days and run it through the dead spend auditor. It shows which part of your budget brought in no results at all. The file never leaves your browser.

If either of those makes you uneasy, request an audit of your Meta account.

Frequently asked questions