The challenge
468 campaigns across product categories with fragmented messaging and optimization flags.
ROAS 2.34.
In electronics that number is "fine." Not a loss, not a win worth celebrating. It's exactly the kind of number that keeps a business owner in the same place for a year without moving.
An electronics store in Egypt, three months, 468 campaigns.
We audited all 468 campaigns, gave each a verdict, then mapped spend against revenue by type. The picture inverted.
Proven winners: 16 campaigns. They returned ROAS 21 on their spend. That is 7.5% of spend producing 68% of revenue.
Message-cheap-deep: 53 campaigns at ROAS 1.37, cheap messaging campaigns generating many conversations and few sales.
Photo-view inflated: 62 campaigns whose spend was tied to image views rather than purchases. The numbers look attractive because image engagement is high, but that spend isn't tied to purchase at all.
Value contamination: 43 campaigns with purchase values corrupted, events recording wrong amounts so the visible ROAS isn't real.
Killed bleeders: 24 campaigns at ROAS 0.1. Closed.
Add it up: roughly 70% of the spend sat in campaigns that were either bleeding or reporting untrustworthy numbers, while 68% of revenue came from just 16 campaigns.
The decision was straightforward: shift budget from the four lower categories into the 16 at the top and anything resembling them. But tracking had to be fixed first, so value contamination wouldn't recur, because without that any new ROAS is another lie.
That 2.34 was never an average. It was 21 blended with 0.1. The work isn't improving the average. The work is separating the two.
An account that looks unremarkable usually has one brilliant campaign buried under four hundred ordinary ones.
