The challenge
Six Ad accounts across three portfolio divisions running siloed campaigns with inconsistent scaling cadence.
A real-estate developer portfolio in Egypt.
The scale first: six Meta ad accounts, three portfolios each covering a group of projects, 120 campaigns, five and a half months from mid-January to late June 2026, producing 4,776 leads.
The blended average cost per lead is what the business owner was looking at, and it lies.
The six accounts ran side by side without a single governing logic. One campaign delivered leads at a fraction of the cost of another campaign in the same account, an eightfold gap that nobody could explain, because everyone was reading the average.
We opened the history of all 120 campaigns one at a time: every budget change, every pause, every scale. Each campaign received one of three verdicts: proven winner, mistake, or bleeder.
What emerged: 16 winning campaigns and 24 losing ones, with the losers costing more than three times as much per lead. The blended average in the report is those two mixed together. It describes nobody's performance.
The two largest accounts produced 1,999 and 1,391 leads at CPLs of 220 and 228, meaning volume was performing acceptably and the losses sat in small scattered campaigns nobody was watching.
Best campaign: a residential project in New Cairo, 496 leads in 17 days on a fixed daily budget with not one jump. Worst campaign: a Gulf market-entry test, 66 leads, Egyptian creative against a cold Gulf audience, closed on day 18.
The pattern repeated on both sides. Across the 16 winners: fixed budget, refreshed creative, frequency under 3. Across the 24 losers: budgets moving up and down every few days, and the wrong audience.
The practical output was a written set of scaling rules for the whole portfolio, covering what gets closed, what gets budget, and when.
An average hides the disaster and the brilliance at the same time. As long as you judge an account by one number, you aren't judging anything.
