When ad spend stalls, the first instinct is to blame the money. The card has expired, the limit was hit, the account was flagged. It is a reasonable first guess, and it is usually the wrong one.
On Brala's Best, a premium small-batch garlic-spread brand, spend had slowed and decisions were being made on bad attribution. We started where the evidence was cheapest to check.
Rule out the boring causes first
We looked at billing and it was clean. Ruling that out took very little time, and it mattered, because it moved the question from the account to the data.
Then follow the purchase
Meta's delivery system learns from the purchases you send back. If those events arrive late, arrive twice, or do not arrive with enough customer detail to be matched, the system is steering with a damaged map. It spends cautiously, or it spends on the wrong people.
The browser pixel alone was no longer enough. Ad blockers, private browsing and iOS all cut it off. The fix was to rebuild the purchase signal on the server and send it through the Conversions API, alongside the pixel, with a shared event id so each purchase counts once.
The same work moved the Meta Opportunity Score from 79 to 89, and external id coverage on the Conversions API from nothing to every event.
What to check on your own account
- Open Events Manager and look at the purchase event. Is it coming from the browser, the server, or both?
- Look at the match quality score for purchases. A low score means Meta cannot tie the sale to a person.
- Check whether the same purchase is being counted twice. Duplicates inflate results and poison learning.
- Compare the purchases in Meta with the orders in your shop for the same days. A big gap is a tracking problem, not a creative problem.
If your ads have gone quiet and nobody can say why, we start with the diagnosis. It is usually a few days of reading before anything is rebuilt.
Photograph from Unsplash.





